What is the halo of minuscule dots seen around some LEDs?

[RoGeorge]

Some models of LEDs light up with a visible halo. 

That is visible with the naked eye, especially against a black background.  The halo can be captured with a camera, too, though to the eye the halo is not that thick.

To the eye it looks like the halo is made out of minuscules dots of light sprinkled around the LED.  Those sprinkled dots seem "fixed" in the air when changing the viewing angle.  More visible from red SMD LEDs, or from clear 5mm red LEDs, especially for punctiform light sources.  LASER pointers always have that halo when pointed to a wall, much stronger than the red LEDs, and visible no matter the pointer's color.

What causes that visible halo?

[AndyC_772]

I think with a laser pointer, it's an interference pattern. The light from the laser is coherent, and where there are multiple paths to reach the same point on the wall (eg. because of dirt, imperfect lens coatings etc), those different paths represent a different number of wavelengths. Hence in some cases they add constructively, and others, destructively - hence the pattern.

I'm not really sure what I'm looking at in that photo - it just looks overexposed, I don't see any structure or pattern.

[AndyBeez]

Agree with @AndyC_772, it's the interference pattern created by the light waves inside the laser beam.

Point the laser beam into a glass of water and you might see the pattern split apart. Do not point laser's into your eyes to see the effect. 

[RoGeorge]

If so, do ordinary LEDs emit coherent light, too, just like a LASER diode would emit?

[tom66]

If so, do ordinary LEDs emit coherent light, too, just like a LASER diode would emit?

No.  Ordinary LEDs emit incoherent, diffuse light.  If you check the datasheet for an LED it will show the typical emission pattern.

[SiliconWizard]

As Beyonce once said,

Baby, I can feel your halo,
Pray it won't fade away

[shapirus]

those different paths represent a different number of wavelengths

did you mean phase shifts? different wavelengths would mean different colors.

[BILLPOD]

CATARACTS

[Benta]

It has a lot to do with your eyes, one technical term is "spatial resolution".
The "sharpness" with which a human eye sees the world is dependent on brightness (of course), but also on wavelength.
Its best resolution is in the green area, 2nd best in red and worst in blue. This is one of the reasons why blue LED annunciators tend to look a bit "smeared".

A lot of work in this field was done in the 1930s and was the basis of analog color TV broadcasting, first seen in NTSC in the US.
The largest color bandwidth in the signal was allocated to green and red, the least to blue.
If you stand close to an old NTSC/PAL analog TV showing analog video, it's very easy to see. The green contours are sharp, so are the red, but blue areas are not sharp or well-defined.
But your eye (and brain) doesn't notice it when viewing. Pure cheating, in fact. 

[Someone]

If so, do ordinary LEDs emit coherent light, too, just like a LASER diode would emit?

No.  Ordinary LEDs emit incoherent, diffuse light.  If you check the datasheet for an LED it will show the typical emission pattern.

That would be incorrect, LEDs are producing light in roughly the same mechanism and as the emitters get smaller interference effects get stronger:
https://www.nature.com/articles/s41598-017-06215-x

Like polarization it is not a binary yes/no but a question of how much. Zero or 100% are not the answers in real life engineering.

[RoGeorge]

https://www.nature.com/articles/s41598-017-06215-x

That link revealed the proper name, thanks! 
It's 'speckle', Wikipedia has more details, 'subjective speckle pattern' and better pics:


Photo source:  https://en.wikipedia.org/wiki/Speckle_(interference) 

[Benta]

Yes, well, that's about lasers. The speckle effect is well known there.
Your OP was about "some models of LEDs".
What exactly do you mean? Is your question about laser LEDs or plain LEDs?

[tom66]

No.  Ordinary LEDs emit incoherent, diffuse light.  If you check the datasheet for an LED it will show the typical emission pattern.

That would be incorrect, LEDs are producing light in roughly the same mechanism and as the emitters get smaller interference effects get stronger:
https://www.nature.com/articles/s41598-017-06215-x

Like polarization it is not a binary yes/no but a question of how much. Zero or 100% are not the answers in real life engineering.

I am not sure if we are reading the same thing.  That paper states "For a light source with a broad spectral bandwidth such as LED, it should be regarded as an incoherent source."  They talk about sLEDs being coherent; these are *not* LEDs, though they are an interesting device that I only found out about from that paper, so thanks for that.

LEDs are not coherent light sources.  To be coherent the emission of light from the device must be parallel, or as close to that as practically measurable, and there is effectively no coherence in the output of a typical LED.  LED emission is almost exclusively from stimulated emission, the electron directly dropping through the bandgap and emitting light, whereas laser diodes experience gain through stimulated emission.  The gain process means the emission of light from a laser diode is coherent like other laser sources.   

[RJSV]

Noticed similar effect, with old school incandescent colored lights:
   Had noticed, with a little blue incandesent bulb, it had a RED halo!  Wondering how that could be, the halo consisted of a sparse pattern of thin (red) lines, emanating, seemingly from that blue lamp.
But, a bit of head bobbling and I realized that there WAS a red bulb, but you just couldn't see any of it, being barely blocked,.by the blue bulb.
   Was easy to duplicate, when I got home later.

[Someone]

No.  Ordinary LEDs emit incoherent, diffuse light.  If you check the datasheet for an LED it will show the typical emission pattern.

That would be incorrect, LEDs are producing light in roughly the same mechanism and as the emitters get smaller interference effects get stronger:
https://www.nature.com/articles/s41598-017-06215-x

Like polarization it is not a binary yes/no but a question of how much. Zero or 100% are not the answers in real life engineering.

I am not sure if we are reading the same thing.  That paper states "For a light source with a broad spectral bandwidth such as LED, it should be regarded as an incoherent source."  They talk about sLEDs being coherent; these are *not* LEDs, though they are an interesting device that I only found out about from that paper, so thanks for that.

LEDs are not coherent light sources.  To be coherent the emission of light from the device must be parallel, or as close to that as practically measurable, and there is effectively no coherence in the output of a typical LED.  LED emission is almost exclusively from stimulated emission, the electron directly dropping through the bandgap and emitting light, whereas laser diodes experience gain through stimulated emission.  The gain process means the emission of light from a laser diode is coherent like other laser sources.

You again jump to the binary yes/no quantification, which is plainly incorrect. Observing coherence in non-laser sources is easy:
https://iopscience.iop.org/article/10.1088/1361-6404/ab3035
Showing coherence from LED pumped phosphor, or for a more in depth look:
https://opg.optica.org/prj/fulltext.cfm?uri=prj-10-11-2460&id=509788
"The radiation from white LEDs is generally understood to have low spatial coherence across the spectrum and be more or less unpolarized"
see how the language is softened to acknowledge there is some coherence and polarisation. Keep saying its none/zero and I'll keep replying with the same point, this is not a binary all/none one/zero yes/no characteristic.

Coherent light does not need to be parallel, it can be point/spherical source. They are independent characteristics that you lump together (which could be simultaneously necessary for some applications) as some generalisation which is incorrect.

People really need to avoid making claims about fields which they don't understand.

[tom66]

"The radiation from white LEDs is generally understood to have low spatial coherence across the spectrum and be more or less unpolarized"
see how the language is softened to acknowledge there is some coherence and polarisation. Keep saying its none/zero and I'll keep replying with the same point, this is not a binary all/none one/zero yes/no characteristic.

I didn't say the light had no coherence, I said it was incoherent. Here is a definition of incoherent light:

Incoherent light is light that contains waves whose wavelengths are not in phase with each other and that do not oscillate at the same frequency. Not only do the photons in incoherent light oscillate at different frequencies, but they also oscillate in different directions.

That matches what an LED emits.  Yes, there may happen to be an coherent point of emission on the LED by chance arrangement or, possibly on some device that could be constructed in a lab environment.  The point is, the emission of light from an LED in a normal application is not, on the whole, coherent.  It therefore is not a coherent source of light.   As far as any normal use of an LED is concerned, the light exhibits little to no coherence (it is "incoherent", like ABS is, "inflexible" -- you can flex it, but not really that much.)

Coherent light does not need to be parallel, it can be point/spherical source. They are independent characteristics that you lump together (which could be simultaneously necessary for some applications) as some generalisation which is incorrect.

Ok, fair enough, a simplification on my part applicable to the most notable difference between LEDs and laser diodes.  But still, LED light is not coherent.  Not for any normal LED in any normal application.  I await the datasheet for a coherent LED that I can buy.

[David Hess]

Coherent sources, and LEDs are somewhat coherent, have speckle:

https://en.wikipedia.org/wiki/Speckle_(interference)

[Someone]

"The radiation from white LEDs is generally understood to have low spatial coherence across the spectrum and be more or less unpolarized"
see how the language is softened to acknowledge there is some coherence and polarisation. Keep saying its none/zero and I'll keep replying with the same point, this is not a binary all/none one/zero yes/no characteristic.

I didn't say the light had no coherence, I said it was incoherent. Here is a definition of incoherent light:

Incoherent light is light that contains waves whose wavelengths are not in phase with each other and that do not oscillate at the same frequency. Not only do the photons in incoherent light oscillate at different frequencies, but they also oscillate in different directions.

Except the LED does have coherence in its emissions
in prefix = negation, in-coherent is without coherence. Except this entire thread is someone asking why they observe coherence effects from an LED.... perhaps because the LED emits (some/significant/enough) coherent light?

Light as a practical use is not incoherent or coherent as binary choices, please just stop trying to label things with primary school level science. Light from most practical sources is somewhere between and quantifiable by various metrics, there most certainly is coherent light coming from an LED, significant amounts of it which cause all sorts of interesting optical effects. It seems everyone else here gets it as a parameter of qualities of light that has some continuum of more/less rather than yes/no.

If you want to get silly, please present the incoherent light source (an imaginary construct which does not exist) and we can illuminate all the spherical cows with it. Even thermal sources of light still contain coherence!

[T3sl4co1l]

Indeed it is a speckle pattern, and despite LEDs not being coherent sources, the spacial coherence can still be enough to observe such effects -- thanks to their intensity and small size, they make good point sources.

I've observed this with my strings of, uh, whatever they were, 7x7mm (4-pin THT) red superbright LEDs.  Presumably the speckle is characteristic of the component lens, the air between (mostly irrelevant at indoor conditions), and your eyes.  So it varies as you move around, and seems to track your eyes say as you rotate your head.

You can see similar patterns in the bokeh when defocused, or viewing through a magnifier or such.  (I feel a loupe/microscope shows these fairly often, as the depth-of-field is quite poor, and we often work with strongly specular materials e.g. "bright" solder joints and flux, glossy plastic parts, solder mask, etc.).

Also contributes (or contributed by) normal eye wear/aging; I've noticed a little bit of, astigmatism I suppose, in the last I don't know five years or so; bright sources (particularly at night, both because of contrast and wide iris) having more of a halo, plus some streaky shapes which can be shadowed in certain regions (e.g. hold my finger near my eye, partially shadowing the light source from one side or another).  Well, I was told a long time ago that I probably wouldn't need glasses until I was 40, and there's a few years left before that, so that seems on track.

Tim

[SL4P]

BTW, consider that many lasers, even visible beams, can emit IR in those ‘side lobes’, so even not looking into the beam, or observing lasers kicking off glass - might expose your eyeballs to harmful doses of invisible IR laser emissions.

[David Hess]

Also most lasers, especially laser diodes and gas lasers, have very short coherence lengths making them useless for holography or Interferometry, yet they all produce speckles.

[Siwastaja]

What exactly do you mean? Is your question about laser LEDs or plain LEDs?

"Laser LED" is a misnomer and as such, so is "plain LED".

Laser diode = LD. OP is clearly asking about LEDs, not laser diodes.

[RoGeorge]

Yes, normal LEDs.  The one that reminded me to ask was a SMD red LED from an Arduino Nano devboard.

I wonder how much would it be possible to get the projection of a diffraction pattern from such a red LED with visible speckle.  From the laser pointers is trivial to project the diffraction pattern of a narrow slit.  The LED is not strong enough.  I've tried with a closer paper screen but it's hard to say if that's diffraction or just a shadow. 

[SiliconWizard]

https://www.sciencedirect.com/science/article/abs/pii/S0022231309003871