Spectral resolution

The smallest wavelength difference at which an instrument can distinguish two neighboring spectral features.

A spectrum can display many wavelength points while neighboring absorption features still merge into one broad peak. Separation depends on the instrument's actual response to nearby spectral structures, not on the density of the graph's horizontal coordinates.

Evaluate resolution using spectral features

Spectral resolution describes the smallest spectral difference an instrument can distinguish. It is often assessed from the full width at half maximum of the response to a narrow line, or whether two neighboring narrow lines can be separated under a specified criterion. IUPAC's definition also notes the influence of instrument line shape and the natural width of the measured features.

Wavelength accuracy asks whether a peak is in the correct position; resolution asks whether nearby peaks can be separated; sampling interval describes how densely data are recorded. These concepts are related but not interchangeable.

The sample itself can limit separation

If two absorption bands are intrinsically broad and strongly overlapping, narrower instrument bandwidth may still not produce two distinct peaks. Conversely, a narrow-line reference source reveals the instrument line shape more directly and is useful for resolution verification.

At low signal-to-noise ratio, fluctuations may hide small structures that are theoretically separable. Excessive smoothing further merges local features, so resolution assessment must specify sampling, smoothing and fitting methods.

Resolution should serve the analytical question

Routine nucleic acid or protein concentration measurements mainly use broad absorption features near specified wavelengths, with different resolution requirements from fine spectral-line analysis. Pursuing narrower bandwidth may sacrifice light throughput without adding useful information for the target method.

To assess whether performance is sufficient, use references or samples suited to the target spectral shape and wavelength range, and specify peak separation, a peak-to-valley criterion or full width at half maximum. A nominal number alone cannot prove better discrimination for every sample.

The value of resolution lies in preserving the spectral information an experiment needs to distinguish, rather than in achieving the smallest number.