I am setting up a test to measure power and spectrum of laser dies. To avoid precise alignment, and also due to lasers' high output power, I have decided to use an integrating sphere to do the power measurement.
I have also been trying to use the same sphere to sample light and couple it into a 50um NA=0.22 multimode fibre for spectral measurement. As you can imagine, the coupling efficiency of bare fibre is quite low. So I tried other options including:
1- attaching a collimator to the fibre and put the collimator at the sphere's port,
2- trying to use a lens to collimate the light coming out of the port and then focus it into the fibre using another lens.
However, none of these methods gave me a significant improvement over bare fibre directly connected to the sphere's port.
The sphere is 2-inch diameter and port diameter is 0.5 inch. The wavelength is 1310nm.
Is there any other way that I can get better results? Thanks.
Dear Mojtaba Mansour Abadi ,
I havn't seen any collimating optics for picking up light from an integrating sphere by an optical fiber. All of them were coupled via a butt joint type of connection.
I think, a collimating optics will not improve the pick up efficiency because of the huge divergence of the light beam in the sphere (NA=1). Please note that the fiber NA is only 0,22.
Please see for example Fig. 4 of the following link:
Article Absolute determination of photoluminescence quantum efficien...
The colleagues there make use of a diffuser in front of the fiber in order to scramble the light beam directions further more. No hint onto a collimating optics there.
See also:
Article Optical investigation of bovine grey and white matters in vi...
They make use of a 400µm fiber... no collimating means.
Even in the catalogues of optical manufacturers for integrating spheres there are no fiber optics collimators listed as accessories for the fibers.
Best regards
G.M.
You are asking how to undo what the integrating sphere is doing. The obvious answer is don’t use the integrating sphere. Of course, then you lose the homogenization. Basically, you can’t have your cake and eat it too.
Collimating the laser, using a beam splitter to reduce the power, and using another collimator to couple into the detector fiber is a common approach. This requires alignment, as you note, but is it really all that hard?
Personally, I think the integrating sphere is doing exactly what you want: A) uniformly sampling the Entire beam without concern for alignment, and B) reducing the power on the detector. Your only problem is that it did B too well.
I should mention: you want to avoid limiting the NA you sample out of the integrating sphere. Generally you can’t collect the whole hemisphere at the output port, but that’s the direction you want to go to be assured of best homogenization and uniform beam sampling. You definitely don’t want to collect with a collimator. The 0.22 NA of the fiber is probably enough angular sampling, but even more would be better.
Anyhow, given that you don’t want to restrict the NA your only choice is to increase the ratio between collection area and the total interior area of the sphere. You either need a larger area detector, or a smaller integrating sphere.
Do you really have to use a fiber coupled detector? Could you get a bare detector that you could push right up against the output port? Usually the detectors are bigger than the 50 um fiber core. What’s more they collect efficiently from a larger NA. Both the larger area and larger NA will increase your collected light with the added benefit of further assuring the uniform sampling.
if the bare detector collects too much light, going the other direction is much easier. You can always just reduce the area by putting an aperture in front of the detector.
One final note: The integrating sphere also smears the laser in time. If you want to look at a short pulse, an integrating sphere isn’t a good option.
Let's answer your question from an illumination optics perspective :
The etendue of the sphere port is about (pi * r)^2~1600 mm2sr.
The etendue of your fiber is (pi*25µm*0.22)^2=0.0003 mm^2sr.
So maximum efficiency is 0.2 ppm, independent of any lens system you apply. The reason for this limit is the second law of thermodynamics.
Thanks Gerhard Martens for your answer and recommendations.
I already have tried the diffuser and it did not help that much. I am going to check with a larger fibre though. It seems like the recommanded option by everyone.
Thanks Mike Albert for your help. Unfortunately, I have to use a fibre since I am interested in SMSR measurement and for that I am using a spectrum analyser to look at the centre peak and other side modes of a laser. I will try with a larger fibre core size to see if I can get a better results.
Thanks Henning Rehn for your solid analysis. As others suggest and your analysis shows, I'll try a larger core fibre may to improve the coupling.
- Badges
- Science topic
- Similar topics
- Filing