An HPLC chromatogram is a two-dimensional plot where the x-axis shows retention time (in minutes) and the y-axis shows detector response, typically in absorbance units. Each peak represents a compound in your sample, and the position, height, shape, and area of those peaks tell you what’s in the sample and how much. Once you understand what each element of the plot means, you can extract a surprising amount of information from what initially looks like a series of bumps on a line.
What the Axes Tell You
The x-axis represents retention time: how long it took a compound to travel through the column and reach the detector after injection. Compounds that interact weakly with the column’s packing material pass through quickly and appear on the left side of the chromatogram. Compounds with stronger interactions take longer, so their peaks appear further to the right.
The y-axis represents the detector’s response to whatever is passing through at that moment. With UV detectors (the most common type), this is measured in absorbance units at a specific wavelength, such as 210 nm or 254 nm. A taller signal means more of that compound is hitting the detector at that instant. When nothing is eluting, you see the baseline, a relatively flat line near zero.
Identifying Compounds by Retention Time
The primary way to identify a compound on a chromatogram is by its retention time. If you inject a known standard of caffeine and it produces a peak at 4.2 minutes, then a peak at 4.2 minutes in your unknown sample is likely caffeine, assuming you used the same column, mobile phase, flow rate, and temperature.
Two related time values help refine this identification. The void time (sometimes called dead time) is the retention time of a compound that doesn’t interact with the column at all. It represents the minimum time any molecule needs to simply travel through the system. The adjusted retention time subtracts this void time from the total retention time, giving you only the time a compound actually spent interacting with the stationary phase. This adjusted value is more useful when comparing results across slightly different system configurations.
Keep in mind that retention time alone isn’t definitive proof of identity. Two different compounds can share similar retention times. For confirmation, analysts often use a second detector (like a mass spectrometer) or run the sample under different chromatographic conditions to see if the peak still matches the standard.
Measuring How Much Is Present
To go from “that peak is caffeine” to “there are 12 micrograms of caffeine per milliliter,” you need quantification. Modern software does this by integrating the area under each peak, which is more reliable than measuring peak height alone. Peak area stays proportional to the total amount of compound even if the peak broadens or changes shape slightly between runs. Peak height is only proportional to concentration when peak width and shape remain constant, which is hard to guarantee.
That said, peak height has one advantage: it’s less affected by neighboring peaks that overlap. When two compounds elute close together and their peaks merge at the base, measuring height at the apex of each peak can sometimes give a cleaner result than trying to divide the shared area between them.
Building a Calibration Curve
To convert peak area into concentration, you first run a series of standards at known concentrations. For each standard, you record the peak area, then plot area against concentration. If the relationship is linear (and it usually is over a working range), you can fit a straight line through the points using least-squares regression, typically with at least three standards of varying concentration. The slope of that line is your calibration factor: the ratio of detector response to analyte concentration.
Once you have this calibration curve, you inject your unknown sample, measure the peak area of the compound of interest, and read the corresponding concentration off the curve. If the relative standard deviation of your calibration factors across standards is 20% or less, the EPA considers the calibration linear enough to use a simple average response factor for calculations rather than the full regression line.
Evaluating Peak Separation
A chromatogram is only useful if the peaks are adequately separated. Resolution is the metric for this, calculated from the difference in retention times between two adjacent peaks divided by the average of their peak widths. A resolution value of 1.5 means the peaks are baseline-resolved, with a clear return to baseline between them. In practice, analysts aim for a resolution above 2.0 to build in a safety margin, since real-world conditions can shift slightly between runs.
If your peaks overlap, three factors control resolution. Efficiency (related to the number of theoretical plates in the column) determines how narrow the peaks are. Selectivity describes how differently the column retains two compounds. And the retention factor reflects how long compounds spend interacting with the stationary phase versus just flowing through. Adjusting the mobile phase composition, column temperature, or column type changes these factors and can pull overlapping peaks apart.
What Peak Shape Reveals
An ideal peak is symmetric and Gaussian: it rises and falls evenly around its center. Deviations from this shape point to specific problems.
- Tailing: The back half of the peak stretches out longer than the front. This often results from secondary interactions between the analyte and active sites on the column packing, particularly with basic compounds on older or lower-purity silica columns.
- Fronting: The front half of the peak is broader than the back, creating a shark-fin shape leaning forward. The most common cause today is physical collapse of the column bed, which affects all peaks equally for both standards and samples. When only sample peaks show fronting but standards look normal, the problem is more likely related to the injection: too large a volume, a solvent mismatch between the sample and mobile phase, or pH differences in the injection solvent.
- Split peaks: A single compound produces what looks like two peaks or a peak with a notch. This can result from injecting too much volume or using a sample solvent that’s much stronger than the mobile phase.
A useful rule of thumb for injection volume: you can inject up to about 15% of the volume of the first peak of interest without distorting peak shape, as long as the sample is dissolved in mobile phase or a weaker solvent.
Reading the Baseline
The baseline carries information too. A clean, flat baseline means the system is stable and free of contamination. Two common baseline problems look quite different and have different causes.
Baseline drift is a slow, gradual rise or fall in the signal over time. Temperature fluctuations are a frequent culprit, especially with refractive index detectors, which are notoriously sensitive to even small temperature changes. UV detectors are affected too, but less dramatically. During gradient elution (where the mobile phase composition changes over the run), drift can also come from trace impurities in one of the solvents. As the proportion of that solvent increases, the impurities accumulate and the baseline creeps upward.
Baseline noise looks like a fuzzy or jagged signal, with rapid high-frequency fluctuations. A data acquisition rate set higher than necessary can reveal short-term signal variations that would otherwise be smoothed out. With UV detectors specifically, anything that reduces the amount of light reaching the detector (a failing lamp, a dirty flow cell, or monitoring at a wavelength where the mobile phase absorbs strongly) will increase noise.
Spotting Ghost Peaks
Ghost peaks are peaks that don’t belong to your sample. They appear consistently or sporadically and can mislead your analysis if you mistake them for real analytes. Common sources include carryover from a previous injection (when the autosampler or injection needle wasn’t cleaned thoroughly enough), contaminants in the mobile phase or sample vials (plasticizers leaching from caps are a classic offender), degradation of the column’s stationary phase at high temperatures or extreme pH, and contamination from system hardware like pump seals or tubing.
The simplest way to identify ghost peaks is to run a blank injection with pure solvent and no sample. Any peaks that appear in the blank are coming from the system or solvents, not your sample. Compare the blank chromatogram to your sample chromatogram, and any peaks present in both are suspect. Ghost peaks tend to be small and may appear at inconsistent retention times, though carryover peaks will show up at the same retention time as the analyte from the previous injection.
Signal-to-Noise Ratio and Detection Limits
When you’re working with very low concentrations, the question becomes whether a tiny peak is a real signal or just noise. The signal-to-noise ratio (S/N) quantifies this. You measure the height of the peak and divide it by the amplitude of the baseline noise in a nearby region where no peaks elute.
A signal-to-noise ratio of 3:1 is the generally accepted threshold for the limit of detection: you can say the compound is present, but you can’t reliably measure how much. For the limit of quantitation, where you need a number you can trust, the accepted S/N ratio is between 6:1 and 10:1. Below these thresholds, the peak is too close to the noise floor to distinguish with confidence, and any concentration you calculate would carry too much uncertainty to be meaningful.