Optical Interference: LEDs & Coherence

[Zachariah]

Hi, i have a question about the optical interference, supposing we have two LEDs which spectra are partially overlapped, can we have interferences?? is there any chance to have a partial coherence between the two lights??

 

[Drakkith]

Hmmm. I believe LED's emit non-coherent light, so I don't think you will see interference with light from another LED no matter what the spectrum of each LED is. To see an interference pattern, I think you'd first need to pass the light from each one through something like a very small slit.

 

[Zachariah]

Thanks for your fast replay, actually i am not intending to visualize interference pattern as with Young slits, but just looking for an evidence for the existence of a constructive or destructive superposition of waves.

 

[jtbell]

The waves from two sources always superpose, but you don't get a visible interference pattern unless the sources are coherent.

If you could take an instantaneous "snapshot" of the light arriving at a screen from two incoherent sources, you would see an interference pattern. However, it changes very rapidly and randomly, on a time scale on the order of the period of oscillation of the waves, i.e. about 10^-15 second. So it simply "washes out" as far as the eye or any photographic device (that I know of, anyway) are concerned.

 

[Zachariah]

I agree, in this case we cannot visualize the pattern, and even if we take an image of it with the camera, we will get the mean intensity which is the sum of the individual intensities, but in my case, the result intensity is always below that sum, that pushed me to suppose the existence of destructive interference, even if i cannot visualize it. would that be true?

 

[Drakkith]

I agree, in this case we cannot visualize the pattern, and even if we take an image of it with the camera, we will get the mean intensity which is the sum of the individual intensities, but in my case, the result intensity is always below that sum, that pushed me to suppose the existence of destructive interference, even if i cannot visualize it. would that be true?

It's true that the light interferes both constructively and destructively, but you would not be able to see evidence of either of these just by looking at the incoherent light. How far off is the total measured intensity vs the sum of the individual intensities?

 

[Zachariah]

To be clear, i tried two LEDs with spectra totally separated, i took the image of the incoming light via an optical fiber, the mean intensity of a ROI is equal to the sum of individual intensities, then i used the two LEDs mentioned above, and the resulted intensity is less than the sum by more than 10% depending on their power, the more i increase their power the more the gap go larger.

 

[Tom.G]

Is you detector linear with respect to the incident power?
If you are measuring the sum of the LEDs with an optical fiber are their angles equally off-axis from the fiber entrance angle?

 

[Zachariah]

Is you detector linear with respect to the incident power?
If you are measuring the sum of the LEDs with an optical fiber are their angles equally off-axis from the fiber entrance angle?

Yes, the detector is linear with respect to the incident power.
I actually mix the two lights inside an integrant sphere coupled with the optical fiber, all measures are taken without changing any parameter, i only turn on and off LEDs.

 

[SlowThinker]

Still must be a measurement error. Either the power source is shared for both LEDs and cannot supply both at high power, or there's a problem measuring the light intensity.
You can't have interference from two independent sources of light.

 

[Zachariah]

Still must be a measurement error. Either the power source is shared for both LEDs and cannot supply both at high power, or there's a problem measuring the light intensity.
You can't have interference from two independent sources of light.

To be sure i don't have a measurement error i tried several configurations:
- Two LEDs with separated spectra --> good result
- Two LEDs with overlapped spectra, LEDs are similar (fabricated from same semiconductor materiel) --> intensity decreases
- Two LEDs with overlapped spectra, LEDs are not similar (fabricated from different semiconductor materiel) --> good result
I even made the output pass through a diffraction grid and plotted the substraction of the sum of the individual spectra and the mixture spectrum, i found that the region responsible for this phenomena is located within the overlapping region.

 

[SlowThinker]

To be sure i don't have a measurement error i tried several configurations

Still it has to be an error somewhere. It definitely is not interference. All optical phenomena are linear, at least at the intensities a mere mortal can access (I know that very strong lasers use nonlinear mirrors).

This sounds like the detector can't take light that is too intense.
What happens if you switch one light source on and off at say 1 Watt (or whatever typical value it has) and increase the other from 0 to 10 Watts? Is the step the same at first then flattens out? Or is it 10% less throughout the range? Also what is the accuracy of the detector, can you actually perform this experiment?

If you have enough equipment around, you can try to split the light in two using a piece of glass at around 45 degrees, measure the two images (straight and reflected) and add the intensities.

 

[Zachariah]

Still it has to be an error somewhere. It definitely is not interference. All optical phenomena are linear, at least at the intensities a mere mortal can access (I know that very strong lasers use nonlinear mirrors).

This sounds like the detector can't take light that is too intense.
What happens if you switch one light source on and off at say 1 Watt (or whatever typical value it has) and increase the other from 0 to 10 Watts? Is the step the same at first then flattens out? Or is it 10% less throughout the range? Also what is the accuracy of the detector, can you actually perform this experiment?

If you have enough equipment around, you can try to split the light in two using a piece of glass at around 45 degrees, measure the two images (straight and reflected) and add the intensities.

The imager was not saturated in any of those measurements, i worked with moderated powers.
Using low powers leads to good result, and as long as i increase the power, this phenomena becomes critical.
The detector accuracy is taken into consideration.

I think the experiment u suggested is similar to using 2 identical LEDs having same spectrum, i already tried it and i still have a slight lose in power.

 

[Tom.G]

Since the usual problems have been largely addressed, here is another track to follow.

The LEDs could be pumping each other when their spectrums overlap. There is a known similar problem with LASER diodes. A possible way to test this would be:

• Place an 90%/10% beamsplitter in front of each LED. 90% beam going to your present integrating sphere & detector
• Monitor each of the 10% beams with additional detectors and readouts
• If the individual, 10%, detectors track with the power anomaly in your main setup, then the LEDs are talkin' to each other... either optically, thru a common power supply, or thermally.

 

[Zachariah]

Since the usual problems have been largely addressed, here is another track to follow.

The LEDs could be pumping each other when their spectrums overlap. There is a known similar problem with LASER diodes. A possible way to test this would be:

• Place an 90%/10% beamsplitter in front of each LED. 90% beam going to your present integrating sphere & detector
• Monitor each of the 10% beams with additional detectors and readouts
• If the individual, 10%, detectors track with the power anomaly in your main setup, then the LEDs are talkin' to each other... either optically, thru a common power supply, or thermally.

Ok i will try this, i keep u tuned.
Is this protocol has an appellation so i can read more about it? Link?

 

[sophiecentaur]

IS it possible that the two LEDs are sharing a power supply and its output power is limited?
Edit - sorry, that's been thought of already.

 

[Zachariah]

IS it possible that the two LEDs are sharing a power supply and its output power is limited?
Edit - sorry, that's be thought of already.

LEDs need 3V and consume some milliamperes, the power supply output is 3V/1A

 

[sophiecentaur]

LEDs need 3V and consume some milliamperes, the power supply output is 3V/1A

OK but is there any common series resistance in the circuit? Power supplies and lack of decoupling are very common faults. Sometimes they are very subtle.
I don't see the cause as being due to interference because you would have to expect fringe patterns that would move about as the equipment is physically distorted. You would need a fantastic level of alignment to see partial cancellation all across the field of view. This is in addition to the other reasons from other people, above.
A diagram (complete) would be an advantage in chasing something as odd as this.

 

[Tom.G]

Is this protocol has an appellation so i can read more about it? Link?

Here is one that's along the same lines as, but not exactly identical to, the optical crosstalk conjecture. When the spectrums of the two LEDs overlap, that could be similar to a 3-mirror optical cavity when used with a LASER source.

Quote from: http://opticalengineering.spiedigitallibrary.org/article.aspx?articleid=2195791

"The power spectrum of the laser changes due to interaction between the lasing field and the small backscattered field, which re-enters the laser cavity with results described by Eq. (1)."

 

[Zachariah]

OK but is there any common series resistance in the circuit? Power supplies and lack of decoupling are very common faults. Sometimes they are very subtle.
I don't see the cause as being due to interference because you would have to expect fringe patterns that would move about as the equipment is physically distorted. You would need a fantastic level of alignment to see partial cancellation all across the field of view. This is in addition to the other reasons from other people, above.
A diagram (complete) would be an advantage in chasing something as odd as this.

Well actually i don't say this is definitely an interference, i encountered this phenomena and i tried to know why it occurs and i did not find any other explanation but the possibility of having an optical interaction. Same electrical circuit is used in all other combination i experimented and gave good results. The existence of the interference does not imply that we must have a pattern of it, in case of partial coherence we will have a blurred screen instead of fringes.

 

[Zachariah]

Here is one that's along the same lines as, but not exactly identical to, the optical crosstalk conjecture. When the spectrums of the two LEDs overlap, that could be similar to a 3-mirror optical cavity when used with a LASER source.

Quote from: http://opticalengineering.spiedigitallibrary.org/article.aspx?articleid=2195791

"The power spectrum of the laser changes due to interaction between the lasing field and the small backscattered field, which re-enters the laser cavity with results described by Eq. (1)."

Thank you very much, i hope this will help.

 

[Charles Link]

Experimentally, there are so many factors that could give an erroneous experimental result. One of them is in coupling to the fiber (this is unlikely, but possible), the heating of the entrance of the fiber is causing its acceptance area to increase. In any case, it is extremely unlikely that any kind of optical interference is occurring between 2 LED's. The optics and interference principles all say "no" to any interference phenomena of this kind. Along the lines of what Tom G. said, perhaps light from one LED is reflecting off the fiber and/or around it and into the other LED, perhaps heating it or otherwise disrupting its performance.

 

[sophiecentaur]

in case of partial coherence we will have a blurred screen instead of fringes

Agreed. But the mean value that the blurred screen shows would be the sum of the two powers - conservation of energy demands that.
We have a problem here, in that you know the details of your experiment but we don't. Can you give more information about what you have actually done? A diagram is easy to produce with a drawing package and it can be attached with no effort. At present, we are thrashing about in the dark, asking you questions about stuff that isn't relevant because we don't know what you know. It is not an efficient method of fault finding.

 

[Zachariah]

Experimentally, there are so many factors that could give an erroneous experimental result. One of them is in coupling to the fiber (this is unlikely, but possible), the heating of the entrance of the fiber is causing its acceptance area to increase. In any case, it is extremely unlikely that any kind of optical interference is occurring between 2 LED's. The optics and interference principles all say "no" to any interference phenomena of this kind. Along the lines of what Tom G. said, perhaps light from one LED is reflecting off the fiber and/or around it and into the other LED, perhaps heating it or otherwise disrupting its performance.

Can the optical components (Lenses..) be responsible for such aberration?

 

[Zachariah]

A diagram is easy to produce with a drawing package and it can be attached with no effort

The set up is so simple: Two lights are mixed in an integrant sphere which is coupled with an optical fiber, at the end of the optical fiber i placed a camera for readout.
Precautions already had been taken into consideration:
- Sufficient supply power.
- No external noise.
- All parameters (camera parameters, distances, input power) are not changed during the experiment.
The experiment is about multispectral discrimination, I used LEDs with separated then overlapped spectra, in case of distinct spectra i had no problem but when i use LEDs with overlapped spectra this aberration occurs.

 

[Charles Link]

I would not rule out something such as the solid state photosensor getting slightly saturated at a particular wavelength and thereby behaving slightly non-linearly. In the case of a photodiode, the response depends upon having sufficient minority carriers to absorb the photon. In any case, that is more likely than optical interference.

 

[Drakkith]

at the end of the optical fiber i placed a camera for readout.

What kind of camera?

 

[Zachariah]

What kind of camera?

basler scout sca640-70gm lens mounted.

 

[Tom.G]

I would not rule out something such as the solid state photosensor getting slightly saturated at a particular wavelength and thereby behaving slightly non-linearly.

A way to rule out camera saturation is to run the experiment with different exposure settings, or even an external f-stop.

 

[sophiecentaur]

The set up is so simple:

There must be something about the setup that is not as simple as you think. If it were simple then you would not have this mysterious result. Unless the World is about to end and all our Physics is wrong, there must be a practical reason for this. In the absence of more details, I can only suggest changing one parameter / component at a time and seeing which change makes the Physics come right again.
As things stand, PF is just blundering about in the dark, making random suggestions. This is not efficient fault finding.