Why Purity Is Measured by Peak Area, Not Peak Height

A peak has two obvious numerical properties: its height, measured from the baseline to the apex, and its area, the region enclosed between the peak curve and the baseline. Both can be calculated from a chromatogram. Only one of them accurately represents how much compound is present. Understanding which one, and why, is the foundation for reading any purity figure on a peptide Certificate of Analysis.

The short answer is that area is used because it is proportional to the mass of compound injected regardless of how the chromatographic band behaves inside the column. Height is not. But that short answer contains several assumptions, and those assumptions define exactly where a purity figure is reliable and where it begins to mislead.

Why area and not height

As a compound travels through an HPLC column it does not move as a sharp plug. The band spreads. This is called band broadening, a fundamental consequence of the physics of separation. Diffusion, non-uniform flow paths through the packed stationary phase, and slow equilibration between the mobile and stationary phases all contribute as the band moves from injector to detector.

The consequence for peak shape is direct. A compound that broadens emerges spread over a longer time window, so its peak is wider. Because the total amount of compound is unchanged whether it arrives in a narrow band or a broad one, a wider peak is also shorter: the height falls as the width grows.

Area is conserved across this broadening. A peak that is twice as wide is also roughly half as tall, and the two changes cancel so that the enclosed area remains approximately constant for the same amount of injected material. Peak height is not conserved. It reflects width as much as it reflects amount, which makes it an unreliable basis for quantitation whenever peak widths vary across a chromatogram or across runs.

In practice, peak widths vary systematically. Column age, flow rate, temperature and sample loading all change peak width, and therefore height, without changing the mass of compound present. Area is robust to all of them in a way that height is not, which is why quantitative chromatography uses area as its default.

Area normalisation and the purity calculation

Once area is established as the measurement, the purity calculation is straightforward. The area under the main peak is divided by the sum of all peak areas in the run and multiplied by one hundred. The result is expressed as a percentage and reported on the certificate as area percent purity.

This approach is called area normalisation: the main peak area is normalised against the total detected area to express composition as a proportion rather than an absolute quantity.

Area normalisation is attractive because it needs no known concentration standard and no precisely known injection volume: the main peak and the impurity peaks are scaled by the same total, so minor run-to-run variation in sample loading cancels. For a sample whose components have similar detector responses, it gives a reasonable and reproducible estimate of composition.

But it rests on an assumption. The calculation treats every peak as though it represents the same amount of mass per unit of area. That is only correct if every compound in the mixture produces the same detector response per mole, or per gram. In UV detection that means equal molar absorptivity at the chosen wavelength. The assumption is often a reasonable approximation. It is not always true, and understanding where it fails is the most important thing to know about a peptide purity figure.

The response factor and why it is not always one

The quantity that describes how strongly a compound absorbs ultraviolet light at a given wavelength is its molar absorptivity, sometimes called its molar extinction coefficient. It is specific to the compound and specific to the wavelength. A compound with a high molar absorptivity at the detection wavelength produces a large peak relative to its concentration. A compound with low molar absorptivity produces a small peak for the same concentration.

The ratio between a compound's molar absorptivity and that of a reference compound at the same wavelength is called its relative response factor. Area normalisation assumes that the relative response factor of every compound in the mixture is one: that every compound absorbs equally. This is the central assumption behind every area percent purity figure.

For peptide HPLC run at 214 nanometres, the assumption works reasonably well for a specific class of impurities. The 214 nanometre wavelength is chosen because the peptide bond absorbs there, at the pi-to-pi-star transition of the amide group. Deletion sequences, truncated peptides, and most synthesis-related fragment impurities carry peptide bonds. They also absorb at 214 nanometres. Their response factors relative to the target peptide are often close to one, and area normalisation returns a reasonable approximation of their proportion in the mixture.

The assumption breaks down in three situations that matter in practice.

Aromatic side chains absorb much more strongly. Tryptophan, tyrosine and phenylalanine contribute substantially to UV absorbance at 214 nanometres, beyond the peptide bond contribution alone, and histidine adds a weaker contribution. Tryptophan in particular has a molar absorptivity at this wavelength that is considerably higher than that of a plain peptide bond. A peptide fragment that happens to contain an aromatic residue produces a disproportionately large peak relative to its actual mass, and a fragment without one produces a correspondingly smaller peak. Area normalisation will overestimate the contribution of aromatic-rich impurities and underestimate aromatic-poor ones. The error is not random: it follows the amino acid composition of the impurities present.

UV-transparent species are simply absent from the calculation. A compound that does not absorb at 214 nanometres produces no peak. No peak means no area. It contributes nothing to the denominator of the purity calculation and nothing to the numerator. It is invisible to the measurement, not penalised by it. The purity figure is then calculated against a total that excludes it entirely, and the result is a distribution among detectable compounds that may look clean while the vial contains material the method cannot see. Certain counterions, inorganic salts, water, and some organic solvents fall into this category.

Non-peptide contaminants have unpredictable response factors. A residual synthesis reagent, a solvent or a non-peptide contaminant may absorb strongly at 214 nanometres, weakly, or not at all, and area normalisation has no mechanism for correcting this: it treats the peak area as though it carries the same absorptivity as the main compound. A UV-absorbing contaminant is detected, but its true proportion is not reliably captured by its peak area without a known response factor for that compound.

Integration decisions and baseline effects

Even before the response-factor question arises, the area numbers that enter the purity calculation depend on decisions made during integration. Integration is the process of identifying where each peak begins, where it ends, and where the baseline lies beneath it.

For a well-resolved, symmetrical peak on a stable flat baseline these decisions are unambiguous: the software places the start and end points at the return to baseline and the enclosed area is calculated with minimal uncertainty.

In practice, peptide chromatograms are gradient runs, and gradient runs produce baseline drift. As the mobile phase composition shifts from aqueous to organic over the course of the run, the UV transparency of the eluent changes gradually, and the detector records a slow upward slope in background signal. Integration software estimates the true baseline beneath each peak by interpolating across this slope. Where the estimate is drawn determines the area assigned to the peak, and different software settings or analyst choices on the same raw data can produce different area values.

Tailing complicates integration further. A tailing peak extends a slowly decaying tail into the retention window of the next-eluting compound. The software must decide where the main peak ends and where the adjacent impurity begins. There is no unambiguous boundary in the raw trace. A generous integration of the main peak tail absorbs impurity area into the main peak and inflates apparent purity. A conservative one does the opposite. The reported purity number reflects this decision as much as it reflects the actual composition of the sample.

An integration threshold setting determines what counts as a peak at all. Small peaks near the noise floor are called as peaks or dismissed as artefacts depending on where the threshold is set, and raising it removes them from the total area, which increases the apparent purity of whatever remains. The sample has not changed. The calculation has.

What area percent actually states

Taken together, these properties define what an area percent purity figure is: a statement about the proportion of detectable UV-absorbing signal that was assigned to the main peak, under a specific method, with specific integration settings, subject to the assumption that all compounds have equal molar absorptivity at the detection wavelength.

That is a narrower claim than it might appear. The figure does not state the mass of peptide in the vial, does not count UV-transparent species, and does not correct for compounds whose response factors differ from the main compound. It also depends on the method: a shallower gradient over a longer run resolves co-eluting impurities that a steeper, shorter run merges into the main peak, so the same sample returns a lower purity number from the more discriminating method because it can see more of what is there.

None of this makes area percent the wrong measurement. It is the correct measurement for the question it answers, and when the method is demanding enough and the assumptions roughly hold, it gives a reproducible, meaningful picture of composition. It becomes misleading only when it is read as more than it is.

When response factors matter most

For routine purity assessment of a synthetic peptide, area normalisation with the equal-response-factor assumption is usually adequate. The major impurities are related peptide fragments, their response factors are close enough to the main compound that the approximation holds, and the figure gives a reproducible picture of material quality.

The assumption becomes consequential when the precise level of a specific impurity is needed rather than a general composition picture. If a degradant or process-related impurity has a response factor substantially different from one, area normalisation will misstate its proportion. Correcting for this requires a reference standard for that impurity and a calibration step, which replaces area normalisation with a calibrated method reporting a corrected concentration rather than a raw area percent. This is common in pharmaceutical development, where the absolute level of an impurity may need to be known accurately to support a safety assessment. A certificate reporting area percent is answering a different question from one reporting corrected impurity levels against characterised reference standards.

Where peak height remains useful

Peak height is not without analytical value. It simply serves a different purpose from area when quantitation is the goal.

Height is a useful diagnostic for column performance. Plate count and peak efficiency are derived from height and width together, and a drop in peak height across successive runs of the same sample signals a degrading column in a way area alone would not immediately reveal. Height also helps for trace-level impurities at the edge of detection: when the start and end of a very small peak are hard to define against baseline noise, height can be less sensitive to integration boundary uncertainty than area, since it does not require those boundaries to be set precisely.

These are specific, limited applications. For the quantitative purity calculation reported on a certificate, area is the standard and height is not the alternative.

Reading the number honestly

An area percent purity figure is best read as a description of the material the detector could see, expressed as proportions within the detectable population. The main compound accounted for that share of the detected signal, under that method, on the day the sample was run. The figure is useful and reproducible: a chromatogram showing one large, clean main peak on a flat baseline is meaningfully different from one where the main peak is surrounded by a forest of related impurities, and the number captures that difference.

What the figure does not state is the mass of peptide in the vial, the identity of that peptide, the purity a different method would have returned, the contribution of UV-transparent species, or any conclusion about safety or suitability for use. Each of those requires a different measurement or a different certificate field.

Novagen is a third-party analytical laboratory. The purity figures on our certificates are area percent values produced by reversed-phase HPLC with UV detection and reported alongside the method that produced them. Every certificate carries a verification key that can be confirmed against our published register at novagenlaboratories.com/verify. A key that resolves to a matching record confirms the result is the one we measured. A key that resolves to nothing was not issued by us.

For a full account of what area percent means relative to the other numbers on a certificate, the companion post on what 99 percent purity actually means covers that in detail. For the underlying chromatogram that produces the areas, how to read an HPLC chromatogram walks through each feature of the trace and what it tells a careful reader.