How to Read an HPLC Chromatogram

An HPLC chromatogram is a graph. The x-axis records time. The y-axis records how strongly the detector is responding at each moment. The peaks are where compounds emerged from the column and passed the detector, and the area under each peak is the number used to calculate purity. That is the full mechanical description. What the graph means, and what it does not mean, takes considerably longer to explain, and that explanation is what a reader of a peptide certificate actually needs.

The axes and what the detector is recording

The y-axis in most peptide HPLC work is the output of an ultraviolet detector, expressed in milli-absorbance units. The detector shines light at a chosen wavelength through the column eluent and records how much is absorbed. More compound present means more absorption means a higher signal. Peptides are typically detected near 214 nanometres, a wavelength at which the peptide bond itself absorbs strongly. This makes the method broadly responsive across peptides rather than dependent on the particular side chains in the sequence.

The x-axis is elapsed time in minutes from the moment of injection. The trace is flat near zero, rises into a peak wherever a compound is present in sufficient concentration, and returns to flat. The flat parts are the baseline. The baseline is not silence exactly, because detector electronics produce low-level random noise, but it is the signal when nothing meaningful is eluting.

One important consequence of ultraviolet detection: anything in the sample that does not absorb at the chosen wavelength produces no response. It is not registered as an impurity. It is simply absent from the measurement. Water, many counterions, and certain excipients fall into this category. The detector is not measuring everything that is in the vial. It is measuring everything that absorbs.

The void time and the solvent front

At the very beginning of any reversed-phase HPLC run, before any retained compounds start to elute, there is usually a disturbance in the baseline caused by the sample solvent and any completely unretained material passing through the column. This is the solvent front, sometimes called the void peak or the unretained peak, and it corresponds to the dead time of the system.

Dead time is the time it takes for a compound that does not interact with the stationary phase at all to travel from the injector through the column to the detector. It reflects the volume of mobile phase inside the column and in the connecting tubing. Any compound eluting at or near the dead time has essentially zero retention and is indistinguishable from the solvent front. Peaks that appear very early in the run, before retained compounds start arriving, are typically not the target peptide, and a chromatogram should be read with that in mind.

The void time also provides a practical lower bound. A target peptide with a retention time only slightly above the dead time is poorly retained, which often means a method that provides inadequate separation from early-eluting impurities. Retention time is therefore informative not just about identity but also about whether the method is doing meaningful separating work at all.

Retention time and what it tells you

Retention time is the elapsed time between injection and the apex of a peak. It is the primary basis for identifying which peak is which compound in a chromatogram.

How that identification works in practice is important to understand. A laboratory runs a certified reference standard of the target peptide under the same gradient, the same column and the same flow rate on the same day. The standard produces a peak at a particular retention time. If the sample produces a dominant peak at the same time, that is supporting evidence that the sample contains the target peptide.

It is supporting evidence, not proof. Retention time is a property of the compound under a specific set of conditions, and a different compound can elute at the same time under those conditions. This is called coelution, and it is the central limitation of reversed-phase HPLC used alone for identity purposes. A retention time match combined with a molecular weight match from mass spectrometry is substantially stronger evidence than either alone.

Retention time is also sensitive to small changes in column condition, mobile phase composition and temperature. Results from different laboratories, or from the same laboratory at different times with different columns, are not directly comparable without the method details being identical. This is why a purity figure or a retention time reported without the method that produced it has limited meaning.

Peak area, peak height, and which one purity uses

A peak has two obvious numerical properties: its height above the baseline, and the area enclosed between the peak and the baseline. Both can in principle be used for quantitation, but peak area is the standard choice in modern HPLC for a clear reason.

Peak height is sensitive to peak width. If column conditions change slightly between runs and peaks broaden, the height drops even if the same amount of compound is present. The area, by contrast, is conserved. A broader peak at the same mass is shorter but wider, and the area stays the same. Area is therefore a more robust measurement of how much material is present.

For purity by area normalisation, the calculation is straightforward: the area under the main compound peak divided by the sum of all peak areas in the run, multiplied by one hundred. The result is a percentage representing the share of the detectable material that is the main compound.

Three things can distort this calculation silently. First, anything that does not absorb is absent from all areas, not just the denominator. Second, compounds with very different absorption properties at the detection wavelength will have different detector responses per unit mass, so area percent is not the same as mass percent. Third, the integration threshold setting determines what counts as a peak and what is dismissed as noise. Raise the threshold and small peaks disappear from the total, which increases the apparent purity of the main peak without any change in the sample. Integration decisions are analyst decisions, and different choices on the same data produce different numbers.

Baseline drift during a gradient run

Reversed-phase HPLC for peptides almost always uses a gradient rather than isocratic conditions. The mobile phase starts predominantly aqueous and becomes progressively more organic, typically acetonitrile, over the run. This gradient washes more hydrophobic compounds off the column as the organic content rises.

The gradient itself affects the baseline. Water and acetonitrile have slightly different ultraviolet transparencies, and as the ratio changes over the run the detector records a slow rise in background signal. This is baseline drift, and it is a normal feature of gradient chromatography rather than evidence of something wrong.

The practical consequence is that the baseline under each peak is not flat. It is sloped, and the integration software has to estimate where the true baseline lies beneath the peak in order to calculate the area correctly. Where the software draws that line matters. A tilted integration baseline set too high underestimates the peak area. A line drawn too low inflates it. For a large, well-resolved main peak this decision has modest impact. For small peaks near the rising baseline late in a run, the integration choice becomes the dominant source of uncertainty in the result.

Tailing, fronting, and resolution between adjacent peaks

A well-behaved peak is roughly symmetrical, with the leading edge rising at a similar rate to the trailing edge falling. Real peaks are often not symmetrical.

Tailing occurs when the trailing edge decays more slowly than the leading edge rises. The peak has an elongated right-hand tail. The most common cause in reversed-phase peptide work is secondary interactions between the analyte and unreacted silanol groups on the silica surface. Tailing is quantified as the tailing factor as defined in USP General Chapter 621: the total peak width at five percent of maximum height divided by twice the front half-width. European Pharmacopoeia methods instead report an asymmetry factor measured at ten percent of peak height, so the two numbers are not directly comparable. A tailing factor above about 1.5 indicates a peak that will be difficult to integrate cleanly, and that may partially overlap adjacent peaks in a way that affects their measured areas.

Fronting is the opposite shape, with a gradual leading edge and a sharp trailing edge. It is less common in peptide work and often indicates sample overloading or a mismatch between the sample solvent and the mobile phase at injection.

Resolution describes how well two adjacent peaks are separated. A resolution value of 1.5 or above indicates that the two peaks are essentially baseline-separated, meaning they can be integrated independently with acceptable accuracy. Below 1.0 the overlap makes the integration assumptions progressively less reliable, and below about 0.8 the peaks are substantially merged, and the integration software is making decisions about where one ends and the other begins that cannot be verified from the trace alone. When two peaks of similar area overlap substantially, their individual areas are unreliable, and so are the purity figures derived from them.

Coelution and photodiode array peak purity

The most consequential limitation of a single-wavelength ultraviolet chromatogram is coelution. If two compounds elute at the same retention time they appear as a single peak. The impurity is absorbed into the main peak area and reported as part of the main compound. The chromatogram looks clean because it is one peak. The sample is not pure because it is two compounds.

One tool for detecting coelution is photodiode array (PDA) detection. Instead of measuring absorbance at a single wavelength, a PDA detector records the full ultraviolet spectrum across the entire peak as it elutes. If the peak is a single pure compound, the spectrum is consistent from the leading edge to the trailing edge. If a second compound is coeluting, it is likely to have a different spectrum, and the spectral profile changes across the peak. Software calculates a purity angle from this variation and compares it against a purity threshold to flag spectral heterogeneity.

PDA peak purity is a useful test. It is not definitive. Two compounds with very similar ultraviolet spectra can coelute and pass a PDA purity check. The absence of a PDA flag indicates no detected spectral variation. It does not confirm a pure compound. The more reliable tool for identifying what is inside a peak is coupling the chromatograph to a mass spectrometer, so that each eluting compound can be assigned a molecular weight as it emerges. The companion article on HPLC versus mass spectrometry covers why the two methods answer different questions and why both are needed.

Limit of detection, limit of quantitation, and what is below the threshold

No analytical method can detect an infinitely small amount of anything. Every HPLC method has a limit of detection, the lowest concentration at which a peak can be reliably distinguished from baseline noise, and a limit of quantitation, the lowest concentration at which the area can be measured with acceptable accuracy and precision.

The conventional signal-to-noise criterion for the limit of detection is approximately 3:1. A peak whose apex is three times the baseline noise level can be identified with reasonable confidence as a real peak. The limit of quantitation is set higher, typically at a signal-to-noise ratio of around 10:1, reflecting the additional requirement that the measurement be numerically reliable rather than simply present.

What this means for a purity result is precise. A compound present in the sample at a concentration below the limit of detection produces no peak. It is not reported as an impurity. It is not counted in the total area. The purity figure is calculated as though it does not exist. This does not mean it is absent. It means the method cannot see it at that level.

The implication is that a purity figure is a statement about what the method could detect, at its sensitivity threshold, on that day. Trace-level impurities that matter for reasons unrelated to purity, such as bacterial endotoxin or residual synthesis reagents, exist below the detection floor of a standard ultraviolet HPLC method regardless of what the purity figure says.

What a clean trace still does not prove

A chromatogram with one large peak, a flat baseline, no tailing of consequence, and nothing visible in the minor-peak region is a good result. It is worth stating carefully what it establishes and what it does not.

It establishes that under the conditions of that method, at the sensitivity of that detector, the sample produced one dominant peak. If a reference standard was run and the retention time matched, there is supporting evidence that the dominant compound is the one expected. If mass spectrometry confirmed the molecular weight, that evidence is considerably stronger.

It does not establish that no other compound is present. Coeluting impurities are invisible. Compounds below the detection limit are invisible. Compounds that do not absorb at the detection wavelength are invisible. Counterions and bound water are invisible and are real mass in the vial.

It does not establish that the sample is safe for any use. Chromatographic purity is a measurement of chemical composition. It does not assess microbial load, endotoxin, sterility, residual solvents below the detection threshold, or the appropriateness of the compound for any purpose. A Certificate of Analysis reports what was measured in the sample submitted to the laboratory, and nothing beyond that.

Novagen is a third-party analytical laboratory. We report what the instruments measured on the sample that arrived. Every certificate we issue carries a verification key that resolves against our published register at novagenlaboratories.com, so the result can be confirmed at the source rather than taken on the document alone. Current turnaround from receipt of samples is 21 days.

Read next: HPLC versus mass spectrometry, and why purity and identity are different questions, or a line-by-line guide to reading a peptide Certificate of Analysis.