Guider 

Peak wavelength and bandwidth answer different questions

Wavelength, usually reported in nanometers (nm), locates light along the spectrum. Bandwidth describes the width of a spectral feature. Neither is an electrical wattage rating or a treatment duration.

Peak, center, and centroid coincide for a symmetric spectral line; an asymmetric distribution can separate them. “Center” should therefore be defined rather than assumed to mean “peak.” These distinctions follow the spectral-line definitions in the Gigahertz-Optik technical reference. [1]

A worked example—not a device measurement

Imagine a single peak at 660 nm, with half-maximum crossings at 650 and 674 nm. Its FWHM is 674 − 650 = 24 nm; the midpoint of those crossings is 662 nm. Its centroid cannot be calculated from those three numbers alone. This deliberately asymmetric example shows why “660 nm peak” and “662 nm center” need not contradict each other.

FWHM is not the full emission range. It marks half-height crossings, not points where the light disappears. It also does not mean that half of the total optical energy lies inside that interval: curve height and integrated area are different quantities. [3]

Read both axes before comparing curves

Start with the horizontal axis: which wavelengths are included? Then identify the vertical quantity. “Counts,” “relative intensity,” and calibrated spectral irradiance are not interchangeable. Ocean Optics distinguishes detector counts and instrument sensitivity from radiometric calibration; NIST describes spectral irradiance as a calibrated measurement with associated uncertainty. [3] [4]

• A normalized curve may set its own maximum to 1 or 100%. It helps compare shape, but removes the original amplitude scale. Two separately normalized peaks of equal height do not establish equal optical output.
• An absolute spectrum needs a defined quantity and units. Spectral irradiance describes incident power per area per wavelength interval; total irradiance over a stated band involves integrating that distribution. [4]
• A logarithmic axis is not a linear height scale. Do not identify “half maximum” by simply taking half the plotted vertical distance; use the report’s calculation method.

For a mixed-band device, ask for each band’s peak and width, plus its contribution on a shared calibrated scale where quantitative comparison is intended. A single global maximum identifies the tallest feature of that plotted quantity. It does not inventory every channel or describe the total power in each band.

The instrument has a bandwidth of its own

An LED’s FWHM characterizes the source. A spectrometer’s optical resolution characterizes how finely the instrument can distinguish spectral features. Both may be expressed in nm, but they are different specifications. Optical bandwidth can broaden measured features; the Gigahertz-Optik reference illustrates this effect and discusses bandwidth correction. [2]

A densely spaced data export is not proof of equally fine optical resolution. Ocean Optics distinguishes dispersion in nm per detector pixel from optical resolution. Ask for both the sampling interval and optical resolution, as well as wavelength accuracy. More decimal places in the peak result do not supply that missing information. [3]

The report should identify the instrument, relevant calibration, and measurement uncertainty. Wavelength calibration concerns the horizontal scale; radiometric calibration concerns the measured output. Neither can be inferred from a smooth-looking graph. These are measurement-quality questions, not a demand that readers build a laboratory setup at home. [2] [4]

Match the spectrum to the tested device and mode

A component datasheet, a finished-device test, and a screenshot without a sample identifier provide different information. Before calling a device “accurate,” establish what was tested and what the specification allows. A nominal label without a defined tolerance and matching measurement is insufficient for a pass/fail judgment.

As a bounded regulatory example, FDA’s January 2023 draft guidance for specified US photobiomodulation (PBM) medical devices recommends testing technical characteristics on the final device, reporting wavelength or wavelength range, and identifying output mode and irradiance. The official page still labels it a draft, not for implementation, at this article’s check. It is not an international operating standard, and some low-risk general-wellness products may be outside its scope. [5]

Where near-infrared and mixed-band reports become incomplete

Near-infrared (NIR) reporting needs a measurement range that includes the claimed NIR band. For example, a plot ending at 780 nm cannot verify an advertised 850 nm channel. That is a limit of the displayed evidence, not proof that the channel is absent.

Similarly, a red-only measurement does not verify a red-plus-NIR operating mode. If channels operate together or in sequence, request reports identifying those conditions. For separated peaks, ask how each FWHM was calculated; a single width spanning a valley between bands can obscure the individual distributions.

Record the collection position, distance, optics, output setting, and timing within operation. These make the measurement interpretable; they are not instructions to use the device at that test distance or for that test duration. Do not replace its instructions with laboratory conditions.

A practical checklist for an LED spectral report

Use these questions to separate a useful report from an unexplained chart. They are an editorial verification checklist, not a universal certification protocol.

If a field is missing, describe the limitation precisely: “NIR output is not verified by this visible-range plot” is more defensible than “the device has no NIR.” If only one sample was measured, do not treat its result as a guarantee for every unit.

Questions a wavelength label leaves unanswered

Does a narrower bandwidth mean a better device?

No general ranking follows from width alone. FWHM describes a spectral feature; choosing a clinically useful spectrum requires evidence for the intended application. A narrow curve is not an effectiveness or safety score.

Does “660 ± 10 nm” mean a 20 nm FWHM?

Not necessarily. The notation might describe a permitted peak range, a nominal specification, or another defined quantity. Only an explicit bandwidth definition establishes FWHM. Ask what the plus/minus interval represents before comparing it with a measured width.

Is dominant wavelength another name for peak wavelength?

No. Dominant wavelength is a colorimetric descriptor, whereas peak wavelength locates a spectral maximum. Do not substitute one for the other when checking a spectral report. [1]

What if two laboratories report slightly different peaks?

First compare the sample, mode, definitions, test conditions, wavelength calibration, and uncertainty. Without those details, the difference alone does not establish a defect or a meaningful change in performance. Calibration and uncertainty belong in the interpretation. [2] [4]

What a useful report lets you conclude

A useful spectrum replaces an isolated wavelength label with a defined distribution under documented conditions. Keep the conclusion that specific: what light was measured, in which mode, over which range, and with what measurement limits. Leave device accuracy, safety, and clinical equivalence unconfirmed wherever the necessary evidence is missing.