Silicon detector response at 1064nm

What's going on is that the absorption depth for light in silicon is strongly wavelength-dependent, whereas the thickness of the diode is fixed. Out at 1064, the photon energy is close to the bandgap, so the effect is magnified--at low temperature much more of the light passes through the depletion zone of the junction without being absorbed.

Integrating spheres will help a little. You can do somewhat better by using specially designed red-enhanced photodiodes with highly reflective back-side metal and a multipass geometry, where light reflected from one diode bounces off others before escaping. However, this requires a very thin contact region--light absorbed in the highly-doped region near the back contact is mostly lost.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com

Thank you for the quick reply.

I am working on using a standard dirt cheap LED as a photo detector

I guess it will have similar inherent properties that the absorption is affected by temperature?

Cheers

Klaus

t that the responsibility begins to drop (about 950 nm for a Hamamatsu S133

6 diode). At 1064nm there is about a .6% / degree C increase in responsivit y as the temperature increases.

ing less? What I am trying to find out: If I use an integrating sphere, wil l that cancel out the temperature effect because I will eventually absorb a ll the light available? Or is there another effect, that I am overlooking?

OK. Where does the strong T dependence come from? I was going to guess that the x-tal expands with T, larger atomic separation makes the band gap (energy) decrease.. so more absorption at low energy. Is that mostly correct or is there more going on?

George H.

I think that's mostly right Klaus, but I do know that some LED's 'go the wrong way' when you cool them. (You can google LED's in LN2) Most LED's go to shorter wave lengths when cooled..

George H.

I think it's also an initial/final state effect--T smears out the band gap. To get photocurrent you need a filled valence state and an empty conduction band state, separated by the energy of the photon. For h nu < bandgap - a few * kT, that becomes exponentially rare.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com

int that the responsibility begins to drop (about 950 nm for a Hamamatsu S1

336 diode). At 1064nm there is about a .6% / degree C increase in responsiv ity as the temperature increases.

cting less? What I am trying to find out: If I use an integrating sphere, w ill that cancel out the temperature effect because I will eventually absorb all the light available? Or is there another effect, that I am overlooking ?

Right, that too. GH