Photon bunching in quantum optics

[yosofun]

The Hanbury Brown Twiss effect, aka the photon bunching effect, states that photons (from a thermal or coherent source) arrive in pairs instead of individually.

The experimental setup consists of having a source go through a 50-50 beam splitter, where one path goes through a variable time delay.

It is found that for no time delay, the intensity is twice as that for significant time delay (i.e., greater than coherence time). This implies that photons arrive in pairs rather than individually.

Now, my question: what happens if the photons are split in more than 2 ways. Would the intensity be n times for n splits? Would this imply that photons (from thermal/coherent sources) are actually molecules?

 

[Physics Monkey]

There seems to be some confusion here. Photon bunching occurs for chaotic light but not for coherent light. For example, consider chaotic light with a Lorentzian spectrum: classical and quantum theory predict the same degree of second order coherence. In particular, the classical inequality g^{(2)}(\tau) \leq g^{(2)}(0) is satsified. The interpretation is of course different, but photon bunching is equivalent to the intensity fluctuations of a classical field.

It is non-classical effects like photon anti-bunching g^{(2)}(\tau) > g^{(2)}(0) or photon number squeezing g^{(2)}(0) < 1, both of which violate classical inequalities, that truly distinguish the quantum from the classical.

Having said all that, I still haven't answered your question. Could you please describe your hypothetical experiment in a little more detail?

Hope this helps.

 

[yosofun]

Hi, I believe photon bunching also occurs for coherent light, since although the coherence function is 1 for coherent states, the coherent states are also prone to arrive randomly, as per the Poisson Distribution. Thus, even for low intensity, the coherent sources are likely to produce bunched pairs.

my experiment is basically a variation of the Hanbury Brown Twiss experiment, a.k.a, the famous photon bunching experiment. In the HBT experiment, a single beam-splitter is used to divide up the beam into two. One beam is subject to a variable time delay. The result was that when the variable time delay is 0, the intensity was 2 times as great (when compared to the case where one beam is time delayed). What happens if multiple splitters were used to divide the beam into n. Would the intensity be n times?

There seems to be some confusion here. Photon bunching occurs for chaotic light but not for coherent light. For example, consider chaotic light with a Lorentzian spectrum: classical and quantum theory predict the same degree of second order coherence. In particular, the classical inequality g^{(2)}(\tau) \leq g^{(2)}(0) is satsified. The interpretation is of course different, but photon bunching is equivalent to the intensity fluctuations of a classical field.

 

[yosofun]

further comments?

 

[Physics Monkey]

Hi again yosofun, photon bunching is actually associated with super-Poissonian statistics, not Poissonian statistics. Lasers high above threshold tend to have Poissonian photon statistics while chaotic light sources are super-Poissonian i.e. they exhibit bunching. It is a hallmark of Poissonian processes that probabilities of events at different intervals are independent of each other.