Measurement of the mean value of the number of photons?

Click For Summary

Discussion Overview

The discussion revolves around calculating the mean number of photons (MNP) using an avalanche photodiode (APD) in Geiger mode, particularly in the context of quantum optics experiments. Participants explore methods to estimate MNP when conventional power measurement tools are unavailable, focusing on the implications of using an APD and the challenges associated with dead time and photon counting.

Discussion Character

  • Exploratory
  • Technical explanation
  • Debate/contested

Main Points Raised

  • One participant suggests estimating MNP by counting detections over a time period ##T##, emphasizing that ##T## should be much greater than the APD's reset time.
  • Another participant questions how to design the time period ##T## for different light sources and requests an example.
  • It is proposed that the number of photons per second could be estimated by dividing the laser power by ##h\nu##, with a caveat about the accuracy of this method.
  • Concerns are raised about the APD's effectiveness for measuring laser power, noting that it performs best at low light levels and that dead time can lead to underestimation of photon counts.
  • One participant suggests using a simple photodiode as a temporary substitute for a power meter, contingent on calibration with a working meter.
  • Another participant challenges the idea of neglecting dead time, explaining that multiple photons arriving during this period will not be counted, which complicates accurate measurement.
  • Participants inquire about the specifics of the light source, including its type, power, and wavelength, to better tailor measurement strategies.
  • It is noted that if the dead time is known precisely, it can be accounted for, but the probability of multiple photons arriving during this time should be minimized to reduce uncertainty.

Areas of Agreement / Disagreement

Participants express a range of views on the effectiveness of using an APD for measuring MNP, particularly regarding the impact of dead time and the design of measurement periods. There is no consensus on the best approach, and multiple competing strategies are discussed.

Contextual Notes

Limitations include the dependence on the characteristics of the light source, the need for precise knowledge of the APD's dead time, and the challenges of accurately calibrating measurements without a functioning power meter.

Pattarasak
Messages
10
Reaction score
0
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.
 
Physics news on Phys.org
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.
 
  • Like
Likes   Reactions: Pattarasak
Mentz114 said:
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?
 
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.
 
  • Like
Likes   Reactions: Pattarasak
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)
 
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.
 
  • Like
Likes   Reactions: Pattarasak
Cthugha said:
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?
 
Pattarasak said:
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.
 
  • Like
Likes   Reactions: Pattarasak
Cthugha said:
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.
 
  • #10
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?
 
  • Like
Likes   Reactions: Pattarasak

Similar threads

  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 0 ·
Replies
0
Views
442
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 2 ·
Replies
2
Views
1K
  • · Replies 10 ·
Replies
10
Views
9K
  • · Replies 17 ·
Replies
17
Views
2K
  • · Replies 1 ·
Replies
1
Views
1K
  • · Replies 4 ·
Replies
4
Views
3K
  • · Replies 38 ·
2
Replies
38
Views
7K
  • · Replies 58 ·
2
Replies
58
Views
5K