Signal-to-Noise, Baseline and Limit of Detection on a Chromatogram

A chromatogram from a purity test is read against its baseline and its noise, not in isolation. Three linked terms decide whether a small feature on the chart counts as a real peak or just background: the baseline, the signal-to-noise ratio (S/N), and the limit of detection (LOD). This page sets out what each one measures, how they relate, and why they sit behind any purity figure on a research-peptide Certificate of Analysis (COA). It is reference documentation about how an analytical chart is read, nothing more.

Research use only. Not for human or animal consumption. This page describes how a laboratory chart is interpreted. It does not describe any use, effect, or outcome of any compound.

What the baseline is

The baseline is the detector's reading when nothing is eluting from the column. On a chromatogram it is the roughly flat line that the trace sits on between peaks. A stable, level baseline near zero is the reference against which every peak is measured: a peak's size is its rise above the baseline, so where the baseline sits directly affects the area attributed to each peak.

A baseline is rarely perfectly flat. It can:

  • Carry small fluctuations (noise) from the detector and electronics.
  • Drift slowly upward or downward over a run, often with a changing mobile phase.
  • Shift where integration software places its start and end points for a peak.

Because area is measured from the baseline up, the way the baseline is drawn under a peak is part of integration. For how integration feeds the reported figure, see area percent vs weight percent hplc.

Signal-to-noise ratio (S/N)

Noise is the small, random variation in the baseline when no component is present. Signal is the height of a peak above that baseline. The signal-to-noise ratio compares the two: how large a peak is relative to the background it has to be seen against.

  • A peak many times taller than the baseline noise has a high S/N and is read with confidence.
  • A feature only slightly above the noise has a low S/N and may be hard to distinguish from the background itself.

S/N matters because it sets the floor for what can be measured reliably. A noisy baseline raises that floor; a quiet, stable baseline lowers it. This is one reason a clean, well-resolved trace is easier to interpret than a noisy one, and why the same sample can read differently on instruments with different baseline behaviour.

Limit of detection and limit of quantitation

Two thresholds are defined directly from S/N:

  • Limit of detection (LOD) is the smallest amount of a component that produces a peak distinguishable from baseline noise. It is commonly described around a signal roughly three times the noise (about S/N 3). Below the LOD, a component may be present but cannot be told apart from background.
  • Limit of quantitation (LOQ) is the smallest amount that can be measured with acceptable confidence, not merely detected. It sits higher than the LOD, often described around S/N 10.

The practical point for reading a chromatogram: absence of a peak is not proof of absence. A component present below the LOD simply does not register as a distinct peak. So a clean-looking trace shows that nothing above the method's detection floor was seen, not that nothing at all is there. This bounds how a purity figure should be read; see what 99 percent hplc purity means.

Why these terms sit behind a purity figure

A single purity percentage on a COA is the visible output of a process governed by baseline, noise and detection limits underneath it:

  • The baseline decides peak area. Where the software draws the baseline under a peak changes the integrated area, and area drives the area-percent figure.
  • Noise decides what counts. Tiny features near the noise may or may not be integrated, which affects both the main peak and any minor peaks counted against it.
  • The LOD bounds the claim. A figure derived from a method can only speak to components the method can detect; anything below the LOD is outside the result.

None of this is visible from the percentage alone. It becomes interpretable when the named method and conditions accompany the result. A chromatogram only confirms separation and relative quantity, not identity, which is a separate confirmation; see hplc purity vs ms identity.

Reading these features on a report

Treat baseline and noise as things you look for, not assume:

  • Is the baseline shown and stable? A flat, level baseline near zero is easier to trust than a drifting or noisy one.
  • Are minor peaks resolved from the noise? Small peaks should sit clearly above the baseline noise, not blur into it.
  • Is the method named? S/N and LOD are method-dependent, so the figure is only interpretable with the conditions stated beside it.
  • Does the lot match? A trace, however clean, is only evidence for the batch it was run on; lot-to-label matching is the final check. See how to read a peptide coa.

Where to go next

Research use only. Not for human or animal consumption. This page is reference documentation about an analytical chart and does not describe any use, effect, benefit, dose, or outcome of any compound.

FAQ

What is the baseline on a chromatogram?

The baseline is the detector's reading when nothing is eluting from the column, drawn as the roughly flat line the trace sits on between peaks. Every peak is measured as its rise above the baseline, so where the baseline is drawn directly affects the area attributed to each peak. A stable, level baseline near zero is the reference the rest of the chart is read against.

What does signal-to-noise ratio (S/N) mean on a purity result?

Signal-to-noise ratio compares a peak's height above the baseline (the signal) with the small random variation in the baseline when nothing is present (the noise). A high S/N means a peak is many times taller than the background and is read with confidence; a low S/N means a feature is only slightly above the noise and is hard to distinguish from it. S/N sets the floor for what can be measured reliably.

What is the limit of detection (LOD) on a chromatogram?

The limit of detection is the smallest amount of a component that produces a peak distinguishable from baseline noise, commonly described around a signal roughly three times the noise. Below the LOD, a component may be present but cannot be told apart from the background, so absence of a peak is not proof of absence. The LOD is method-dependent.

What is the difference between LOD and LOQ?

The limit of detection (LOD) is the smallest amount that can be detected as a distinct peak above noise, often around a signal-to-noise of about 3. The limit of quantitation (LOQ) is the smallest amount that can be measured with acceptable confidence, not merely detected, and sits higher, often around a signal-to-noise of about 10. LOD answers whether something is seen; LOQ answers whether it can be measured.

Does a clean baseline mean a sample contains nothing else?

No. A clean baseline shows that nothing above the method's detection floor was seen, not that nothing at all is present. A component below the limit of detection does not register as a distinct peak. This is why a purity figure is only interpretable with the named method and conditions beside it, and why the lot must match the certificate before the result is relied upon.

Research use only. Not for human or animal consumption.