UV and IR Wavelength Selection
A Component Engineering White Paper — wavelength, bandwidth and binning as yield-critical parameters in curing, disinfection and sensing systems
Queendom LEDs · Component Engineering Group
1. Why Wavelength Is a Yield Parameter, Not a Comfort Parameter
In visible lighting, a few nanometres of wavelength drift is largely a colour question. In UV curing, UV disinfection and IR sensing, it is a yield question. A photoinitiator absorbs in a narrow band; a microbial action spectrum peaks at a specific wavelength; a sensor’s filter passband is a few tens of nanometres wide. Selecting the emitter peak and spectral width is therefore a process decision with a measurable cost when it is wrong. This paper explains how to select UV and IR wavelength and bandwidth so that the system, not the component alone, meets its target.
It complements three application notes in our resource center: the UVC 275 nm disinfection dosage note, the UVA curing and sensing design note, and the SWIR wavelength selection for food sorting note.
2. How Wavelength and Bandwidth Are Specified
Three quantities govern whether an emitter will work in the target process: the peak wavelength, the full width at half maximum (FWHM), and the binning tolerance on both. Buyers who specify only the peak wavelength routinely discover that the delivered parts span a band wide enough to split the process.
It is useful to picture the emission spectrum as a shape rather than a number. The peak tells you where the shape is centred; the FWHM tells you how broad it is; the bin tolerance tells you how far the shape may move between production batches. A process that depends on a narrow chemical absorption band cares about all three, because energy emitted outside the absorption band is paid for but not used. This is why two emitters with identical peak wavelength and identical radiant flux can produce measurably different process results.
| Parameter | Definition | Why it decides yield |
|---|---|---|
| Peak wavelength (λp) | Wavelength of maximum radiant flux | Must overlap the absorption or action spectrum peak |
| FWHM | Spectral width at half maximum | Wide FWHM spends energy outside the useful band |
| Bin tolerance | Allowed spread of λp in production | Wide bins create non-uniform process results |
| Shift with temperature | Change of λp and FWHM per °C | Predicted results drift from measured results |
| Shift with drive current | Change of λp with forward current | Over-driving moves the peak off the process window |
3. Selecting in the UVC Band
Disinfection applications are anchored near 265–280 nm. Within that band the choice is narrower than it appears: 275 nm is the common engineering compromise between germicidal efficacy and optical extraction, and it is the wavelength at which Queendom UVC components are binned. Note that 275 nm (J-15 series) and 254 nm (Z-14 low-pressure lamp reference) are distinct references and must not be treated as interchangeable when calculating dose.
| Band | Typical λp | Primary use | Bandwidth concern |
|---|---|---|---|
| UVC | 265–280 nm | Disinfection, dose delivery | Wide FWHM wastes wall-plug efficiency |
| UVB | ≈308 nm | Phototherapy, controlled dosing | Dose must track wavelength precisely |
| UVA | 365–405 nm | Curing, fluorescence, sensing | Must match photoinitiator absorption |
| NIR | 850–940 nm | Illumination, sensing, machine vision | Filter passband overlap |
| SWIR | 1050–1550 nm | Food sorting, moisture, material ID | Water/moisture absorption bands |
4. Selecting in the UVA Band for Curing
UVA curing is a matching problem between the emitter and the photoinitiator, not a power problem. A 395 nm emitter and a 365 nm-sensitive adhesive can both be called ‘UVA’ while producing very different cure depths. The correct method is to read the adhesive’s absorption spectrum, select the emitter peak to fall within it, and then size the irradiance from the required dose.
| Photoinitiator class | Absorption anchor | Preferred emitter band | Note |
|---|---|---|---|
| Tpo / Tpo-L | ≈ 380–405 nm | 395–405 nm | Deepest cure in thick pigmented layers |
| Irgacure 819 | ≈ 370–405 nm | 385–405 nm | Broad absorption, tolerant |
| Irgacure 184 | ≈ 365 nm | 365–375 nm | Shallow surface cure, often combined |
| Cationic systems | ≈ 365 nm | 365 nm | Sensitive to moisture and temperature |
5. Selecting in the IR and SWIR Bands
IR selection is dominated by the detector and the environment rather than by a chemical response. For illumination and machine vision, the emitter must sit inside the camera’s response and outside the dominant ambient noise. For SWIR material sensing, the emitter must sit on a characteristic absorption feature of the material being sorted.
5.1 IR illumination and machine vision
850 nm and 940 nm are the two workhorses. 850 nm offers more silicon camera response at the cost of a faint visible red glow; 940 nm is nearly invisible and is preferred for discreet illumination, but demands a camera with adequate sensitivity. The choice is made by the camera, not by the emitter.
5.2 SWIR for sorting and material identification
SWIR emitters are selected against the absorption spectrum of the target. Moisture absorbs near 1450 nm; many plastics have distinctive features between 1100 and 1700 nm. The emitter bandwidth must be narrow enough not to average across the feature, otherwise the contrast that makes sorting possible is washed out.
6. Temperature and Current: The Two Hidden Variables
Most wavelength selection is done at one temperature and one current, then expected to hold in the field. It does not. Both the peak wavelength and the FWHM move with junction temperature and with drive current, and the movements are large enough to matter in narrow-band processes.
| Variable | Direction of shift | Typical magnitude | Process consequence |
|---|---|---|---|
| Junction temperature up | Peak moves to longer wavelength | 0.1–0.3 nm/°C | Curing and disinfection dose drop as the emitter warms |
| Forward current up | Peak moves to shorter wavelength, FWHM widens | Several nm over the rated range | Over-driving leaves the matched band |
| Pulsed vs continuous | Pulsed operation reduces self-heating | Peak stays closer to nominal | Pulsed curing can be more consistent |
| Ageing | Peak drift over life is small but real | Part of the total budget | Include drift in the acceptance window |
7. Common Mistakes and How to Avoid Them
| Mistake | Why it happens | Consequence | Avoidance |
|---|---|---|---|
| Specifying peak wavelength only | FWHM treated as a detail | Process results scatter across the band | Specify FWHM and bin tolerance together |
| Treating 275 nm and 254 nm as equivalent | Both described as ‘germicidal UV’ | Dose calculation is wrong, disinfection incomplete | Keep the references separate in the dose model |
| Choosing UVA power over UVA match | Irradiance is easier to specify than spectrum | Under-cured adhesive despite high power | Match emitter peak to photoinitiator absorption |
| Ignoring temperature shift | Datasheet values taken as fixed | Field results drift from lab results | Include shift-per-degree in the design model |
| Choosing IR emitter before the camera | Emitter selected as a lighting part | Poor signal-to-noise at the chosen wavelength | Select wavelength from camera response first |
| Using wide-FWHM parts for SWIR sorting | Availability drives selection | Loss of contrast, sorting errors | Choose narrowband parts matched to the absorption feature |
| Validating at one current only | Design current assumed to be the operating point | Shift under over-drive is missed | Test at the extremes of the drive range |
8. How to Write the Specification
Wavelength requirements are frequently under-specified because the process is described in words rather than in numbers. The following fields make the requirement measurable and therefore testable at goods-in.
| Specification field | What to state | Why it removes ambiguity |
|---|---|---|
| Peak wavelength | Target λp with a window, not a single value | Allows binning to be meaningful |
| FWHM | Maximum permissible spectral width | Prevents wide-band substitution |
| Bin tolerance | Maximum spread within and between lots | Keeps the process inside its window |
| Radiant flux / irradiance | At the work plane, at stated distance | Separates optical from electrical specification |
| Operating conditions | Drive current and junction temperature at rating | Makes the shift allowance explicit |
| Measurement method | Spectrometer or radiometer, with calibration reference | Makes acceptance testing reproducible |
8. Worked Example: UV Curing Cell for an Adhesive Bond Line
A production cell must cure a 1 mm adhesive bead on an opaque black housing. The adhesive datasheet lists a peak absorption near 385 nm and a required dose of 1.5 J/cm² at the bond line. The selection proceeds as follows.
| Step | Value for this cell | Conclusion |
|---|---|---|
| Absorption anchor | 385 nm | Emitter band must be 385–405 nm, not 365 nm |
| Bandwidth | Feature is broad | Moderate FWHM acceptable; binning still matters |
| Required dose | 1.5 J/cm² at the bond line | Optical budget calculated from irradiance × time |
| Operating temperature | Emitter warms by 40 °C | Peak shift of roughly 4–12 nm accounted for |
| Verification | Spectrometer plus radiometer at the work plane | Both spectrum and dose measured, not assumed |
The instructive part of this example is that the drive current is not the primary variable. A higher current that shifts the peak away from 385 nm can reduce cure quality even while increasing measured power. Matching the spectrum and then sizing the irradiance is the reliable order of operations.
10. A Wavelength Selection Checklist
| Step | Key question | Output |
|---|---|---|
| 1. Target response | What absorbs, or what does the detector see? | Required peak band |
| 2. Bandwidth | How narrow must the emission be? | FWHM specification |
| 3. Binning | What spread can the process tolerate? | Bin tolerance |
| 4. Conditions | How do temperature and current shift the peak? | Shift allowance |
| 5. Delivery | What irradiance or dose is required at the work plane? | Optical and drive budget |
| 6. Verification | How will the delivered spectrum be measured? | Acceptance test method |
| 7. Specification | Can the requirement be read one way only? | Frozen specification text |
12. Conclusion
Wavelength selection in UV and IR is a matching discipline. The question is never ‘which wavelength is best’, but ‘which wavelength and bandwidth match this chemistry, this detector and this environment, within the tolerance the process can absorb’. Specifying peak, FWHM, bin tolerance and shift together is what separates a process that works in the lab from one that works in production.
Download the full white paper
UV-IR-Wavelength-Selection-White-Paper.pdf — complete edition with full test matrices, derating curves and reference data.
PDF: UV-IR-Wavelength-Selection-White-Paper.pdf
The full edition adds absorption-spectrum overlays, dose worksheets and a binning tolerance calculator. Queendom · Component Engineering Group.
This white paper is published as part of the Queendom LED technical library. For datasheets, test reports and application notes referenced above, see the LED Components Support and Resource Center sections. Engineering enquiries: sales@queendomlamp.com.















