Measurement of the mean value of the number of photons?

[Pattarasak]

Hi everyone, I do experiment in field of quantum optics and I want to calculate the mean of the number of photons (MNP). In the most of papers, MNP can be calculated by using average power which is measured by power meter. But my power meter is defective. For now, I have only the avalanche photodiode (APD) working in the Geiger mode. How should I do to use the APD to calculate MNP? Did you have paper to suggest me? thank you.

 

[Mentz114]

You are trying to estimate the mean of a Poisson distribution. Counting the number of detections in a time period $T$ ( where $T$ is much greater than the APD reset-time) will give you suitable data. The period will depend on the source, naturally.

 

[Pattarasak]

You are trying to estimate the mean of a Poisson distribution. Counting the number of detections in a time period $T$ ( where $T$ is much greater than the APD reset-time) will give you suitable data. The period will depend on the source, naturally.

thank you for your answer, but I just confuse how to design T for any source. Can you show me for an example?

 

[mfb]

Can you directly use it to detect photons with a reasonable dead time? Can you reduce the intensity before letting it hit the diode?

To reduce the statistic uncertainty you want to detect many photons, where suitable numbers of "many" depend on your setup.

 

[Mentz114]

I think the number of photons per second can be estimated by dividing the power of the laser by $h\nu$ where $\nu$ is the light frequency.
You must check everything I propose - I cannot guarantee total accuracy !

Poisson distribution (Wiki)

 

[Cthugha]

APDs are not exactly good for measurements of laser power as you need to work at really low light levels to get into the regime, where they give you count rates which are linearly proportional to the number of photons present. If any photons arrive during the dead time of the detector, they will not be registered and the photon number will be underestimated. Attenuating the light beam strongly is one possibility, but you would need to know the attenuation factors of filters very precisely to get good accuracy.

Are there any other labs nearby, which still have a working power meter? In that case you might be able to use a simple photo diode as a substitute for the time being and calibrate the photo current with respect to the power meter from the other lab.

 

[Pattarasak]

Are there any other labs nearby, which still have a working power meter? In that case you might be able to use a simple photo diode as a substitute for the time being and calibrate the photo current with respect to the power meter from the other lab.

No, there are not other labs that have a good power meter. Can we neglect a dead time of the detector by writing a module to detect only the correct pulse?

 

[Cthugha]

Can we neglect a dead time of the detector by writing a module to detect only the correct pulse?

No, not really. I also do not understand what you mean by "only the correct pulse". The problem with APD dead times is simple. If one photon arrives, it triggers the avalanche, which has some finite duration. Afterwards the diode "resets" to the original state. If there is only one photon arriving during this time, you will only get one avalanche. If there are two photons arriving, you will only get one avalanche. If there are 2789564 photons arriving, you will only get one avalanche. If you are sure that the possibility of two photons arriving is small, this works well, but that requires knowledge of the mean power of the light field, which was the starting point of the problem.

However, it might help if you could tell us a bit about the light source you intend to use. Is it a laser, white light or some spectral lamp? Is it cw or pulsed? Do you have a rough estimate of the power it should be emitting? What is the wavelength? There are different strategies for measuring optical power that work well for different light sources.

 

[Pattarasak]

However, it might help if you could tell us a bit about the light source you intend to use. Is it a laser, white light or some spectral lamp? Is it cw or pulsed? Do you have a rough estimate of the power it should be emitting? What is the wavelength? There are different strategies for measuring optical power that work well for different light sources.

thank you for the information about dead time. For all I know, the light source is pulse laser source,with wave length 808 nm, which the repetition rate can be varied from 5 MHZ- 80 MHZ.

 

[mfb]

If you know the dead time very precisely you can take this into account, but in practice you want the probability of more photons arriving during the dead time to be small to avoid a large uncertainty from this.

A well-known attenuation makes that possible - if you put multiple not-too-strong filters in series you can even measure their absorption rate with the APD in dedicated measurements.

What is the rough expected power of your laser where you want to measure it? Just as order of magnitude?
What is the dead time of the APD?