A purity percentage is a summary of a chromatogram, and summaries lose information. The trace itself shows how many species the method resolved, how well they separated, where the software drew its integration boundaries, and whether the detector was behaving. Reading it takes a few minutes and answers a question the number cannot: is this figure supported by the data behind it. This guide covers the axes, peak shape, baseline, and integration marks, in the order worth checking them.
What the axes show
A chromatogram plots time on the x-axis, in minutes from injection, against detector response on the y-axis, usually UV absorbance in milli-absorbance units. Each peak is a resolved species reaching the detector, with anything the method failed to separate traveling inside the same peak, and its position on the time axis is its retention time, determined by how strongly it interacted with the column relative to the mobile phase gradient. For a reverse-phase peptide method, more hydrophobic species elute later. Retention time is method-specific rather than intrinsic, so it is comparable across runs only when the column, gradient, flow rate, and temperature match. That makes it useful in one particular way: the target peak should appear at a consistent retention time across batches run by the same method, and an unexplained shift signals either a method change or a different species. The y-axis magnitude matters less than the shape, since response depends on injection amount, though a trace where the main peak is driven off-scale is not usable for integration.
Peak shape
Shape is where most of the diagnostic information sits. A well-behaved peak is close to symmetric, roughly Gaussian, and returns cleanly to baseline on both sides. Departures are informative. Tailing, where the peak's descending side stretches out, is the most common and usually indicates secondary interactions with the column, though it can also reflect column degradation or overload. It matters for purity because a long tail makes the boundary between peak and baseline ambiguous, giving integration software latitude that changes the reported figure. Fronting, a stretched leading edge, typically means the column is overloaded, in which case resolution of nearby species is compromised. Shouldering, a visible bulge on one side, is the important one: it usually indicates a partially resolved second species that the method could not separate, and depending on where integration boundaries fall it may be counted as target. A peak noticeably broader than its neighbors, without a clear shoulder, can indicate the same thing at lower resolution. Symmetry is therefore not cosmetic; it is evidence about whether the main peak is one compound.
Baseline and noise
The baseline is the detector's signal when nothing is eluting, and its behavior bounds what the run could detect. It should be flat, low, and stable. Drift, a baseline sloping upward across the run, is common in gradient methods because mobile phase composition changes absorbance, and mild drift is normal, though pronounced drift complicates integration by making the peak boundary depend on where the software places the baseline. Elevated noise, visible as thickness or jitter in the trace, sets the detection floor: impurities producing signal below the noise level cannot be distinguished, so a noisy run is blind to small peaks and reports a flattering purity by omission. Sharp spikes unrelated to elution usually indicate bubbles or electrical interference rather than sample content. Sudden baseline steps can mean a wavelength or lamp event mid-run. The practical read is that a clean baseline supports the claim that minor peaks would have been visible had they existed, while a noisy one does not.
Resolution between peaks
Resolution describes how completely two adjacent peaks are separated, and it determines whether their areas can be assigned accurately. Baseline resolution, where the trace returns fully to baseline between peaks, allows unambiguous integration. Partial resolution, where peaks share a valley that never reaches baseline, requires the software to split the area at some chosen point, and that choice affects both peaks' reported values. Peaks eluting close to the target are the ones that matter, since they are the ones whose area can be misassigned into the target's. Reading resolution alongside the reported number is what reveals whether a 99% figure came from a well-separated main peak or from a merged cluster integrated as one. It also indicates whether the method was appropriate: a gradient that compresses everything into a narrow window produces high apparent purity by failing to separate, which is a method limitation rather than a sample property.
Integration marks
Most reported chromatograms show the integration boundaries, as tick marks, vertical drop lines, or a baseline segment drawn under each peak. These are worth examining because they show what the software decided. Check that the boundaries enclose the full peak rather than clipping a tail, that drop lines between partially resolved peaks fall at the valley rather than somewhere arbitrary, and that the constructed baseline under a peak follows the actual trace rather than cutting across a raised region. Also check whether small peaks were integrated at all: an integration threshold set high enough will ignore minor species entirely, which removes them from the total area and raises the reported purity. A peak table listing retention time, area, and area percent for each detected species is the most useful accompaniment, since it makes the arithmetic checkable. If the percentages do not sum sensibly, or if the table lists fewer peaks than the trace visibly shows, the reported figure and the data do not agree.
Reading a chromatogram against a claim
Put together, the workflow is short. Confirm the trace is for the lot in question and that the method parameters are stated. Find the main peak, check its retention time against the expected value for the method, and assess symmetry for shoulders or unusual breadth. Look at the baseline for noise and drift, which tells you what a small peak would have had to exceed to be seen. Scan for minor peaks, particularly those close to the target, and check their resolution. Read the integration marks and the peak table, and verify that the reported purity follows from the areas shown. Finally, compare against a previous lot's trace if one is on file, since consistent chromatograms across batches indicate a controlled process while a shifting minor-peak pattern at constant reported purity indicates variability the number is not capturing. A supplier that provides traces routinely is making claims that survive this process, which is the entire point of asking for them. Where the underlying question is whether a compound will perform, not just what it contains, that is a separate measurement covered in purity versus potency.
FAQ
What does a shoulder on the main peak mean?
Usually a second species that the method partially resolved. Depending on where integration boundaries were drawn, some or all of its area may have been counted as target.
Is peak tailing a quality problem with the compound?
Not necessarily. Tailing more often reflects column condition, secondary interactions, or overload than sample composition, but it does make the peak boundary ambiguous and therefore the integration less definitive.
Why does baseline noise affect a purity figure?
Impurities producing signal below the noise floor cannot be detected or integrated. A noisy run omits them from the calculation, which raises the reported purity without the sample being cleaner.
Can a chromatogram confirm compound identity?
Not on its own. Retention time consistent with a reference under the same method is supporting evidence, but identity confirmation requires mass spectrometry.
What if a supplier provides only the percentage?
The claim is unverified. The percentage is a derived value, and the trace is the evidence behind it, so a supplier declining to provide it is withholding the checkable part.
Research Use Only: All compounds sold by Prove It Performance are intended exclusively for laboratory research. Not for human or animal consumption. These products are not drugs, supplements, or food. Statements have not been evaluated by the FDA. Must be 21+ to purchase.
Stay in the loop
New compounds, research updates & exclusive subscriber deals.
