This question already has an answer here:

I suspect the answer to this is going to be "no", however, just in case anyone knows something...

I need to broaden the emission peak of a near IR LED. Is there any generally applicable method of doing this in a reliable manner?


marked as duplicate by Marcus Müller, JRE, Warren Hill, Elliot Alderson, Finbarr Feb 21 at 17:53

This question has been asked before and already has an answer. If those answers do not fully address your question, please ask a new question.

  • \$\begingroup\$ I had a funny thought, theoretically if you wave if fast enough would you get some sort of Doppler effect, but with light? (theoretically) \$\endgroup\$ – Wesley Lee Feb 20 at 17:59
  • 1
    \$\begingroup\$ Use your LED to shine on light emitting material with other lambda... \$\endgroup\$ – smajli Feb 20 at 18:05
  • \$\begingroup\$ @MarcusMüller Problem with a filtered halogen light source is intensity, and more to the point I have to modulate it at kHz frequencies. Not sure that is doable \$\endgroup\$ – Dirk Bruere Feb 20 at 18:28
  • \$\begingroup\$ The broadband IR diode from one of the answers I've linked to might be a solution. Other than that: in this situation, you usually don't modulate the light source, but a reflecting or transmitting material. See: DLP \$\endgroup\$ – Marcus Müller Feb 20 at 19:03
  • 1
    \$\begingroup\$ @Dirk, you should include the requirement about modulation in the question post. \$\endgroup\$ – The Photon Feb 20 at 19:54

Run it at the highest current possible. It will get hot, and radiate IR just from the heat (like any hot object.) If you over do it, you will end up with a DED (dark emitting diode.)

Other than that, LED spectrums are pretty much set in stone.

The wavelength is determined by the materials the LED is made of. The wavelength will vary a (tiny) bit with temperature, but that's about it.


For a particular LED? No.

A quick web search for "broad spectrum IR LED" threw up the following interesting news release, Osram presents the world’s first broadband infrared LED from 2016.

enter image description here

Osram Opto Semiconductors is using converter technology for infrared emitters for the first time. The result is an LED that emits broadband infrared light in a wavelength range from 650 to 1,050 nanometers (nm). The main application is near-infrared spectroscopy, for example for analyzing food.

Further down the technology is explained thus:

The basis of the SFH 4735 is a blue 1 mm2 chip in UX:3 technology. Its light is converted into infrared radiation with the aid of a phosphor converter developed specifically for this application. A residual blue component in the light helps users target the area they want to investigate. The emission spectrum of the SFH 4735 has a homogeneous spectral distribution in the infrared range. The chip is mounted in the proven and compact Oslon Black Flat package which is characterized in particular by good thermal resistance.

Checking the SFH4735 datasheet, the Vf is 2.95 V at 350 mA which ties in with the blue LED as IR would typically be 1.4 V for the low power devices at least.

They're quoting 60 μW/nm @ 750 nm and 30 μW/nm @ 980 nm. (The article is missing the '/nm' in the units.)

€6 each on Digi-Key.

Prompted by @asdfex's comment I had a look further down the datasheet and the relative spectral emission is quite horrendous.

enter image description here

Figure 1. The relative spectral emission graph shows a blue peak at >> 100 times that of the IR.

  • 2
    \$\begingroup\$ That should read µW/nm, not µW in the second last line. "Residual blue component" is a nice wording for "90% of the emission is in the blue" \$\endgroup\$ – asdfex Feb 20 at 18:28
  • \$\begingroup\$ Thanks. Looks possible, but I need 10s of mW at around 2000nm \$\endgroup\$ – Dirk Bruere Feb 20 at 18:30
  • \$\begingroup\$ Your question doesn't say that, @Dick. d:^) \$\endgroup\$ – Transistor Feb 20 at 18:45
  • \$\begingroup\$ @asdfex: You're right. See the update. \$\endgroup\$ – Transistor Feb 20 at 18:55

This gets into mechanisms. The ones that come immediately to mind are:

  • QM uncertainty in the energy levels and finite lifetimes. This is a matter of the construction of the LED and I don't think you can do much about that.
  • Thermal (heat.) In effect, this is just Doppler broadening under another name. The added heat provides higher mean velocities and the random mix of these can lead to an effective broadening. You could consider the idea of raising the temperature under some kind of control.
  • Collisions also broaden by reducing the lifetime of some state (reducing the \$\tau\$, in effect.) The lower atmosphere, compared to the upper atmosphere, broadens the emission lines of, for example, CO2 because of the higher pressure and much shorter mean time between collisions.
  • Fluorescence and Phosphorescence mechanisms. (Look up Stokes and anti-Stokes, for example.) Fluorescence is a much faster mechanism because it uses allowed QM state transitions. Phosphorescence is much slower because it uses dis-allowed (illegal) QM state transitions (illegal doesn't mean it doesn't happen, it means it happens rarely, in these cases.) I don't think this will be of much use because it usually shifts over some distance and doesn't "broaden" as I think you want.

Of the above, I think you are stuck with trying out temperature. It's probably not going to get what you want. But you can at least easily try it and see.

It seems a little strange to me that you'd want to broaden an IR LED, though. They are inherently pretty broad, already -- FWHM of 60 to 100 nanometers? Something like that? If you need a very broad source, then you switch to black body radiation curves. (Heated filament, for example.)


You can always add filters to 'cut' spectrum, getting more of it from a semiconductor device without changing the material will be difficult.


Not the answer you're looking for? Browse other questions tagged or ask your own question.