I have coupled high-power LED light inside 50um and/or 200um multimode fibres and then collimated it using Thorlabs parabolic mirrors. However even using a 20X microscope objective, the incoherent light does not focus sharp. Have you any special procedure or optical system to advice?
Eugenio,
You may want to check the NA of the MM fiber and the microscope objective. For example, while 20X objective may do the work in terms of reducing the spot size to fit the fiber, it may provide a higher NA that the one of the fiber. For example, your MM fiber may have 0.22 NA while it's not uncommon for 20X to have 0.4 NA.
Coherent and incoherent light has to be treated different when when setting up an optical system.
Keep in mind that, in contrast to a laser, you will always just produce images of the incoherent emitting surface of your LED. (Those images will always have lower intensity than the emitting surface. This is due to the limited NA of every optical system.)
If you want to achieve a 10µm spot size with incoherent light the most efficient way might be 1:1 imaging of the LED onto a 10µm pinhole and a second 1:1 imaging onto your desired position. You simply "cut out" a small area from the huge emitting surface to reduce the image that will be produced.
The NA of all the lenses in your system should be maximum, while the smallest NA will define the intensity of your image.
In general use of off-axis-parabolas will increase NA and thus increase brightness, but adjusting them properly is not as easy as with lenses.
When using high quality aberration corrected optics you might end up with a top-head intensity distribution rather than a Gaussian. This can be tailored by an aperture in the flourier plane, but will obviously decrease intensity.
If you want to use multi-mode fibres you might loose some efficiency but gain flexibility. Also here you could set up an imaging system that images the end of the fibre onto your desired object. I have no experience with microscope objectives but I think you should place the fibre-tip at the distance where your objective produces the intermediate image inside the microscope.
1) Check the optical invariant principle. It is a kind of a conservation law. Put simply, it means that you can have 10 micron spot image of 100 micron source, but angles of rays hitting the image plane will be correspondingly 10 times larger. So there is no way of achieveing what you want without using very high numeric aperture objective (even unattainably high!). Or using non-regular optical elements (that do not fall into catadioptric category). See details e.g. at http://www.newport.com/Optics-Fundamentals/604533/1033/content.aspx.
2) However, there is another optical principle which connects the brightness of the source to the intensity in defined spot. You will lose a lot trying to collimate LED to 10 mciron spot and increasing its optical power will not help - important is brightness and for LED it is quite low. The task you are trting to achieve is almost equivavlent to coupling LED to single-mode fiber (well, it is 5-6 micron), you can look for them and estimate how much light can you really get. Consider laser diodes or superluminescent LEDs, if you need a decent light intensity in spot - they are much much brighter even if the optical power could be lower.
To add to Patrick and Shavkats responses: From thermodynamic considerations, any incoherent source that fills both the acceptance angle and the spot size of any optical waveguide, can only couple 1/2kT per degree of freedom, (i.e. per mode). (k = Boltzmans constant and T is the effective temperature.) Playing with the launch optics within this regime has no effect. Only the equivalent temperature of the LED has any influence. For example, at any given wavelelngth a hot radiant surface launches the same power into all single mode fibres, whatever their core size.
With respect to this problem - are there any commercially-available fiber coupled LEDs using fibers with 50-60 micron core? I've only found products that use ~200 microns core fibers.
Assaf
Assaf you got the same problem I have. I have somehow solved this using 2 coupled parabolic mirror lens with different NA. You loose a lot of light but still it remains something.
Thank you to all. Patric I did already the procedure with 10 um pin-hole but you loose so much that is almost unuseful. I shall try again and I shall write you back
If LED active area is much larger than 10-20 fiber diameter simply place fiber close enough to fill its acceptance angle NA.
Assalamualaikum. In case of Gaussian beams, by my experience, expand the gaussian beam before contracting it. An expanded beam contracts to a far smaller spot than a narrow beam. But I wonder if the same principle applies to beams from LEDs.
- Badges
- Science topic
- Similar topics
- Philosophy
- Esthetics
- Interpretation