A clean outline is not a component count
An HPLC trace records detector response against time. The detector adds the signals present in its flow cell; it does not draw a separate peak for every molecule. Two components with nearly identical elution profiles can produce a smooth combined peak. A shoulder is a reason to investigate, but the absence of a shoulder is not evidence that only one component is present.
A diode-array detector (DAD, also called PDA) compares UV spectra across a peak. Spectral variation can expose interference. Similar spectra, a weakly absorbing minor component or a large concentration imbalance can conceal it. A software result called ‘peak purity passed’ describes that comparison under its settings, not chemical purity of the entire sample.
Agilent 5991-7931EN, p. 6: UV and mass-selective detection of coeluting compounds
Change selectivity, not just peak width
Selectivity describes how differently a chromatographic system retains two components. In a simple isocratic comparison, the separation factor is α = k₂/k₁, with k₂ ≥ k₁ and retention factor k = (tᵣ − tₘ)/tₘ. Here tₘ is the hold-up time. Efficiency describes band spreading: a more efficient system produces narrower peaks under comparable conditions.
Synthetic example: tₘ = 1.00 min and retention times 4.00 and 4.10 min give k₁ = 3.00, k₂ = 3.10 and α ≈ 1.033. Narrower peaks may help this pair. If both retention factors are exactly equal, narrowing both peaks alone cannot separate their centers. Changing stationary-phase chemistry, solvent type or suitable pH can change relative retention; none guarantees improvement for every pair.
Choose the experiment for the suspected interference
Start with the question: is the issue the target peak's identity, its integrated area, or an impurity hidden beneath it? A detector change can distinguish signals without improving the physical separation. Conversely, a chromatographic change can separate components whose UV spectra look alike. Keep the original run and integration settings so the comparison is traceable.
Coelution checks and what they leave unresolved
On narrow screens, scroll the table sideways. You can focus the table and use the arrow keys.
| Check | What it may reveal | Remaining limitation |
|---|---|---|
| DAD spectra across the peak | Spectrally different contributions | Similar spectra can escape detection |
| Extracted-ion traces in LC–MS | Different detectable ions at the same time | Isomers, adducts and ion suppression need interpretation |
| Different separation selectivity | A formerly combined peak splits | Reassign identity and calibration in the changed method |
| Blank and matrix controls | Background or preparation interference | They do not cover every unknown compound |
Agilent 5991-7931EN, p. 6: UV and mass-selective detection of coeluting compounds · ICH Q2(R2), §3.1–3.3: validation of analytical procedures
Define what would change the conclusion
For a quantitative result, investigate whether interference changes the reported amount at the relevant level. A visibly separated standard pair does not automatically challenge the sample's closest interferent. Use representative material, the intended matrix and the concentration balance that matters to the result; record the critical pair and the decision criterion before comparing runs.
In the Peak Explorer, bring the two centers closer while keeping width fixed, then increase width while keeping the centers fixed. Both worsen separation, but by different mechanisms. The model has two known Gaussian components; a real unknown trace does not disclose its hidden component count. Numerical resolution between visible neighbors cannot certify that no third component is underneath either one.