9

I have this LED and I want to characterize this photodiode.

The LED should produce light of 370 nm, but I measured it and it is around 358 nm, and at that wavelength the quantum efficiency (\$\eta\$) of the photodiode is 0.5141.

Now, I feed the photodiode signal to this ADC, I got data and I turned it into a histogram.

If you are familiar with photon statistics then you know that for poissonian light the mean of those histograms should be equal to their sigma squared, for example in one case I have:

\$\mu\$ = 19,462.2

\$\sigma\$ = 135.785

\$\sigma^2\$ = 18,437.56

close enough I guess.

If I understood correctly this book(eq. 5.56 and 5.57) then the number of photons should be given by this equation:

\$n_{photons}\$ = \$\frac{\mu}{\eta}\$

which in this case would give me some 37,856.83 photons, which seems like a very low number...

Then I started decreasing the voltage and getting things like this:

\$\mu\$ = 7123.51

\$\sigma\$ = 135.7

\$\sigma^2\$ = 18961.29

This is called super-poissonian light. Here to get the photon count I need the general formula (eq. 5.57):

\$\sigma_{counts}^2\$ = \$\eta^2 (\sigma_{photons})^2 + \eta(1+\eta)\mu_{photons}\$

in this case I know 2 of the values but I also have 2 unknowns and as far as I can see no other equation or formula that I could use.

So basically I'm wondering:

1.- If I'm using the formula for the amount of photons correctly

2.- What should I do when the light becomes super-poisonian?

  • 7
    How did you convince yourself that shot noise is the only noise source in your system? If you want to convince us, you should share the circuit that connects the photodiode to the ADC. And probably also share the complete set of settings given to the ADC when you collected your data. – The Photon Aug 10 '19 at 19:55
  • 2
    There are a few explanations but more information about your setup would help determine which is most applicable. I would ask yourself these questions:
    1. How much of the light from the led is incident on the photodiode?
    2. How are you measuring the photocurrent? i.e. post a circuit diagram
    3. Are you calculating this assuming a single wavelength? Leds can have a fairly wide output spectrum
    Usually photon counting is done with APDs or photomultiplier tubes, not VIS PDs. you are operating way out of your PDs spectral range, so I would also look into a different PD
    – Luke Gary Jun 04 '21 at 00:55

1 Answers1

1

Using an LED to characterize a photodiode is very complex, the non-uniform irradiation of the source, relatively wide spectral irradiation, angular response of the receiver, refraction, all causes effects that are difficult to quantify.

A better approach would be to use a high grade laser, as a coherent light source is much easier to predict.

It is important to focus the laser using optics to have the beam within the photodiode area.

You can then measure the laser power output, after the optics, using a calibrated thermopile, from which you can calculate the photon output of the laser using the formula:

enter image description here

Damien
  • 7,827
  • 1
  • 12
  • 29