Talk:Hanbury Brown and Twiss effect

temporal coherence
Note that many of the first radio interferometers in the 1940s were intensity interferometers. Also note that all astronomical interferometers use light which is not temporally coherent, so the lack of temporal coherence for the light in the first optical intensity interferometers was not unusual. Rnt20 13:26, 27 September 2005 (UTC)

light suffers from distortion in the atmosphere, so unless you have a corrected wavefront you lose the temporal accuracy. Robin48gx (talk) 22:48, 16 February 2008 (UTC)

figure caption
One might add a little explanation to the figure on the right that this is (similar to) the setup H B and T used in their experiment published in january '56 in Nature with the title "Correlation between photons in two coherent beams of light". The setup for the Sirius thing was using the searchlight mirrors and no beam-splitter as stated correctly in the article. Since I'm not so much into interferometry this was a little confusing to me.

Philipp Struck

Photons and HBT
The page at it currently stands gives me the impression that there is anti-correlation between the signals at the two detectors. The whole point of the Hanbury Brown - Twiss experiment, with light from a high-pressure mercury arc lamp is that the signals at the detectors are correlated, not anti-correlated. This is only practically measurable, as I understand it, with narrowband incoherent light. It is hard or impossible to detect any correlation with a laser. Robert Hanbury Brown, in his 1972 book "The Intensity Interferometer" rejects the particle notion of light. I think he uses "photon" to mean the generation and deposition of (implicitly unquantized) electromagnetic radiation (emr) by matter, not to indicate that light can be regarded as a stream of "photon" particles. On page 7 he writes: "These difficulties about photons troubled physicists who had been brought up on particles and had not fully appreciated that the concept of a photon is not a complete picture of light. Thus many people are reluctant to accept the notion that a particular photon cannot be regarded as having identity from emission to absorption."

On page 6 he refers to the work of Forrester, Gudmundsen and Johnston (FGJ), also in 1955, (http://prola.aps.org/abstract/PR/v99/i6/p1691_1) who used the interference between two closely spaced wavelengths of a Zeeman split emission line to generate an electrical beat frequency on a photo-cathode. This seems to be a very clear instance of two emr signals summing to produce a modulated total signal. The FGJ experiment creates a modulated electromagnetic wave at the photo-cathode, leading to an electrical signal, at about the 10GHz resonance of their 3cm cavity. This is not explicable by two separate beams of "photons", with different frequencies, since such photons in the pure particle sense are not, or were not, regarded as being capable of interfering with each other - the two beams should both produce a continuous current independently.

It seems to me that the Hanbury Brown - Twiss effect, which is correlation between the signal of photo-detectors receiving a split version of a narrowband incoherent light source, can't be explained with conventional "photons", and can only be explained with the emr being a classical wave. Both the FGJ experiment and the Hanbury Brown - Twiss experiment constitute arguments against the conventional view of emr being quantitized as "photons". I think that the current state of this page does not reflect this, and by explaining the effect solely in terms of particles (photons, which are bosons) hides the fact that there is a debate and presents one view as if there was no debate.

Robin Whittle


 * You are right to complain that the article did not make explicit that Hanbury-Brown and Twiss themselves observed a correlation. Is the current revision better?  A positive correlation was also observed in the Mercury experiment you described (referenced in the page as Morgan and Mandel 1966).  However, the experiment pictured does in fact result in an anticorrelation, as described in the 1986 paper.  This result can be intuitively understood as simply demonstrating the fact that a single photon can only be observed at a single detector.  Anticorrelation, unlike the correlation found by Hanbury-Brown and Twiss, cannot be understood classically.


 * There is perhaps an interesting discussion of the history of Dirac's famous statement that a photon interferes only with itself. Although the development of quantum field theory has made it clear that photons can interfere with each other, it was at one time widely believed that this was not the case.  Hence the trouble introduced by the simple interference effects (FGJ) you describe.  I'm not particularly well versed in this history - maybe there should be another page for it.  If you accept that photons are able to interfere with one another, HBT and FGJ are suddenly much more accessible from the quantum point of view.  You might imagine that two nearby photons, being bosons, tend to interfere constructively, causing their wave packets to be amplified in the place where they overlap.  For this reason, two photons which are close to each other tend to be detected right next to each other, since that's where their amplitudes are greatest.  Similarly, you might imagine that fermions tend to destructively interfere, so that the amplitudes of their wave packets are smallest where they overlap.  This is just the Pauli exclusion principle.  That's all very informal, but I hope it helps conceptually.


 * I think the section titled "Wave Mechanics" does a good job explaining the effect purely as a classical one. I don't think there's a need to emphasise it any more.  Finally, I think I ought to emphasise that there is no debate concerning the existence of photons.  They are not some construct whose existence is controversial in any way, they are established physics.  In fact, a careful consideration of the century-old photoelectric effect should essentially convince you that the energy of light must be delivered in a corpuscular manner, and that each corpuscle must have energy hf.  Particularly note the fact that increasing the intensity of the incident light does not increase the energy of the emitted electrons, but increasing the frequency does.  Since the classical energy of light is given only by its intensity and not its frequency, it's hard to believe that this can be explained classically without heavy modification.  Combined with Planck's solution for the blackbody radiation curve and the anticorrelation effect described in the 1986 paper, I hope you are convinced.


 * --Hyandat 08:44, 4 March 2006 (UTC)


 * The 'quantum' section states that Ugo Fano showed correlation between amplitudes, but doesn't give any actual formulas, which makes it hard to understand. The formulas are needed.  The beat-frequency for the zeeman effect is also a rather interesting phonomenon, deserving exposition here or elsewhere. User:Linas (talk) 20:10, 1 December 2013 (UTC)

I agree there is no debate in the field about the existence of photons. However, there was much debate about whether the photoelectric effect could be described by the interaction of classical waves with quantized atoms (i.e. with energy levels). For instance, see the writings of Marlan Scully and Lamb. The conclusion is that the photoelectric effect does not require the existence of photons, although usually in undergrad it is still given as proof of their existence. The Hanbury, Brown, Twiss interferometer performed with single photons at the input (or possibly the Lamb shift) is the true conclusive proof of photons. BTW: Loved the article - especially the history. --J S Lundeen 20:19, 21 March 2006 (UTC)


 * Note: the HBT experiment with a sharp number state, n=1, does not imply anything about localized photon particles. The lack of coincidence counts is required by conservation of energy.  After the first detection, n=0.  —Preceding unsigned comment added by 67.169.157.55 (talk) 08:11, 6 October 2008 (UTC)

It is misleading to think of point-particle photons when describing “that-which-evolves”: that which is governed solely by a wave equation. The particulate nature associated with the electromagnetic field pertains solely to the processes of emission and absorption. Einstein’s explanation of the photoelectric effect was a drastic departure (“It was as if the ground had been pulled out from under one, with no firm foundation to be seen anywhere, upon which one could have built.” [Schilpp: Einstein]). And it is understandable, in spite of his comment in the second paragraph of his paper indicating that this phenomena relates primarily to the “emission and transformation of light” involved in its interaction with matter, that he would come to the conclusion (in the next paragraph): “… the energy of a light ray spreading out from a point source is not continuously distributed over an increasing space but consists of a finite number of energy quanta which are localized at points in space, which move without dividing, and which can only be produced and absorbed as complete units.” There were others that disagreed. Planck's assistant Max von Laue (among others) strongly advocated a perspective that the quantum of the electromagnetic field related only to emission and absorption of light, i.e., "only when it is exchanging energy with matter".

The subsequent failure to make the distinction between the theoretical domain of that-which-evolves (wave-like) and the domain in which it becomes know to us: absorption of discrete quanta in a detection process, has lead to ill-formed questions in quantum theory. A prime example of this is the Pfleeger and Mandel 2-laser experiment in which the concept of individual photons being emitted from each laser led to the belief that any demonstration of interference between the two lasers would contradict Dirac’s statement that photons cannot interfere with each other, but can interfere only with their self. Interference was demonstrated. Then Dirac must have been wrong. Well, not really. After this result was obtained, the basic idea leading to the experiment was reformulated into the superposition of the possible paths by which each single detected photon could have reached the photodetector – equivalent to the use of Huygen’s principle to determine the quantum or classical field amplitude at the photodetector. And, interference is to be expected after all. The original premise of the experiment was found to be in error; it arose from misconception rooted in the picture of individual photons as that-which-evolves -- a picture that contributes nothing to the comprehension of the process of evolution. (John A. Wheeler has supported the perspective that phenomena involving the electromagnetic field can be treated as a classical field, and the annihilation and creation operators inserted later, if needed.)

Further, in this experiment, the authors attenuated the laser radiation so that the rate of photodections was of the order of or less than one per transit time from laser to detector, and claimed that only one photon was in the apparatus at a time. However, the frequency stability of the lasers when data was taken was such that the coherence length of the single photon wave packets would have to be of the order of 6000 meters, which means that there had to be thousands of overlapping single-photon wave packets present at the time of data collection in order to produce the indicated counting rate. It is difficult to create a comprehensible picture of individual evolving photons in this context that is useful in any sense.

As others have stated, the HBT phenomena are essentially classical. The use of the name “photon-bunching” for classical intensity variations serves no useful purpose here – though one might expect such to arise in experimental domains focused on quantum aspects of systems with small photon number.

My preference would be to not call other intensity correlation devices "HBT devices", but rather "intensity correlation" devices, and if relevant mention any analogies to the HBT work. And, it is important, I think, to acknowledge the innovative aspects of HBT (“Is it possible to get an interference pattern from incoherent sources?” “Yes!  But, it requires lack of ‘second-order coherence’: requires intensity fluctuations”). And, equally important is the further work that was prompted directly by HBT: the development of Roy Glauber’s fully comprehensive quantum theoretical description of the electromagnetic field, from which it is easily seen that the field of optics prior to HBT was devoted solely to first-order coherence, while the HBT phenomena required lack of second order coherence (intensity fluctuations) in the field.

(It is worthwhile to note that Glauber’s work can also be expressed classically by simply replacing the quantum correlation functions involved [or equivalent density operators] with their classical counterpart. The radiation from an ideal laser and the carrier waves from radio or TV stations correspond to complete coherence: factorization of all orders of correlation functions.  One can use either the quantum formulation or the classical formulation in the description of the field for any of these.  [In the laser, however, the generation of the field requires a quantum description of the atomic transition that produces a narrow band field and the population inversion that amplifies it, along with the resonant cavity that supports primarily only one mode.]  This is not to support a semi-classical theory, but to show that there is no requirement for a picture of a particle, and good cause to avoid such, in the evolution of the quantum electromagnetic field.)

I hope that this will be helpful to those doing the actual editing, and that it will result in a more comprehensive historical perspective.

-- A. Einstein, Ann. Physik 17, 132 (1905). Translation A.B. Arons and M.B. Peppard, Am. J. Physics 33, 5 (1965).

A. Einstein, in Albert Einstein Philosopher-Scientist, P. A. Schilpp, Ed. (Harper, New York, 1951) p. 45.

Max von Laue to Albert Einstein, June 2, 1906, Collected Papers of Albert Einstein, vol. 5.

R.L. Pfleegor and L. Mandel, Phys. Rev. 159, 1084 (1967).

P. A. M. Dirac, Quantum Mechanics (Oxford University Press, London, 1958), 4th ed., p. 9.

Roy Glauber, “The Quantum Theory of Optical Coherence”, Phys. Rev. 130, 2529 (1963).

Roy Glauber, “Coherent and Incoherent States of the Radiation Field” Phys. Rev. 131, 6 (1963).

--James R. Johnston 21:27, 27 September 2008


 * Hi James, ... interesting read, but "those doing the actual editing" need to be expert to accomplish your hopes and goals for the article, which really means that only you are capable of getting this to happen. You should do the edits. User:Linas (talk) 20:29, 1 December 2013 (UTC)

Please amend the heading
The page: Hanbury-Brown and Twiss effect (and associated pages) The problem: Hanbury Brown is not hyphenated, but I don't know how to fix the heading of the article. Starrylady 23:54, 27 January 2007 (UTC)
 * I moved the page. Now we have to change all the links from other pages... Akriasas (talk) 10:32, 8 February 2008 (UTC)

Photon bunching merger proposal
I suggest that photon bunching be merged into this article - as far as I can tell, photon bunching is just a simple case of the HBT effect? (In any case, the photon bunching article is orphaned at the moment as well). Akriasas (talk) 10:35, 8 February 2008 (UTC)


 * its not photon bunching, its photons in synchronisation. read quantum electrodynamics by Feynman if you want to understand this. Robin48gx (talk) 22:46, 16 February 2008 (UTC)


 * Not sure of the difference between bunching and synchronisation. In any case the quantum optics jargon is "bunching". I agree that photon bunching should be merged —Preceding unsigned comment added by Christoph.westbrook (talk • contribs) 16:52, 12 April 2008 (UTC)


 * New to this site, I have proofread some of the text and made tiny typographical corrections, eg spellings. ericraymondfranklin@googlemail.com


 * HBT and photon bunching are different things, and the two pages should therefore not be merged. QuantumCyclops (talk) 14:41, 2 January 2009 (UTC)


 * The photon bunching article is currently (Dec 2013) an awful stub. It's badly written, uses terible grammar, makes a multitude of dubious, inchoherent statements, and says nothing at all that would suggest that the effect is in any way different than what is described here. Perhaps photon bunching is something different, but you won't get it from that article. Thus, I will redirect to here, until such time that someone can write a decent article on photon bunching. OK? User:Linas (talk) 19:36, 1 December 2013 (UTC)

Intensity formula is incorrect
With reference to the section on wave mechanics, the formula given for intensity is i1 = E^2 sin^2 wt. However, detectors (e.g. PMT) are generally not fast enough to detect frequency of optical waves. We are detecting instead the time average and as such both i1 and i2 should be just proportional to the square of the E-field instead of sin^2 wt.

However, since HBT effect is essentially a second-order correlation effect, the final form of the correlation formaula is correct. Perhaps it may be clearer if the intensity correlation be written as  instead of -- Tzu Hao Goldbeaver (talk) 03:29, 8 September 2008 (UTC)


 * The problem goes deeper than this, and the final form of the correlation formula (in August 2011 at least) is not correct. The article states that the HBT effect is equivalent to a classical effect which supposedly allows the phase (of g^(1)) to be measured. That is, the final form of the correlation function, even with the averaging of sinusoidal oscillations, contains the optical phase phi. This is completely untrue, and it is untrue because it assumes that intensity measurements give a square-of-sinusoid signal. This is certainly untrue at optical frequencies using photon counters, which is what the HBT effect is all about. The HBT effect is a measurement of g^(2), which equals |g^(1)|^2 for thermal fields. The classical analgoue only makes sense for fields with fluctuating intensities, and for real *intensity* measurements, not E^2 measurements pretending to be intensity measurements. I hope to amend the article soon. Vortimer (talk) 02:19, 29 August 2011 (UTC)


 * So, can someone fix this? User:Linas (talk) 19:50, 1 December 2013 (UTC)

Photon statistics and the HBT effect
The article contains a false statement that the HBT effect is due to boson nature of photons: "...the results of a given experiment depend on whether the beam is composed of fermions or bosons". Actually, the HBT effect is due to interference between waves of close frequencies, which results in intensity fluctuations. However, no wave effect, neither interference nor diffraction, may be derived from photon statistics, be it quantum (Bose-Einstein), classical, or intermediate statistics. It means that photon statistics has nothing to do with the HBT effect. For details see, for example, M.O.Scully and M.S.Zubairy, "Quantum Optics". —Preceding unsigned comment added by 81.5.96.132 (talk) 10:26, 20 February 2010 (UTC)


 * I disagree, in that it would make a difference if photons were fermions (but then there would be no classical limit of light as waves). Vortimer (talk) 02:13, 29 August 2011 (UTC)

Please reconsider your disagreement to the previous paragraph. The HBT effect does not rely on the particles being bosons. It may actually be possible to set up an experiment with electrons from a narrow, fluctuating source, where close detectors show correlation of the intensities and more widely separated detectors show less correlation. You will see that the correlation of the close detectors will be positive, not negative, even though the particles this time are fermions.Robertfromcal (talk) 19:41, 20 June 2016 (UTC) Never mind. I thought it some more and read the articles. The positive correlation does occur because the identical particles are bosons.Robertfromcal (talk) 18:53, 14 July 2016 (UTC)

Dead link
the external link * http://physicsweb.org/articles/world/15/10/6/1 is dead and from that citation it is difficult to figure out what it pointed at

Netrapt (talk) 01:42, 16 April 2010 (UTC)

Photon anti-bunching from a single atom
Another explanation for why emission from a single atom, to a single detector, is anti-bunched:

Two transitions (the transition to an excited state, then the transition to a lower state) are required to emit one photon. Each transition occurs according to a Poisson process, but the convolution of two Poisson processes is no longer Poisson, it is anti-bunched.

This somehow is not captured by the text in the article which reads:

"The above treatment also explains photon antibunching:[11] if the source consists of a single atom, which can only emit one photon at a time, simultaneous detection in two closely spaced detectors is clearly impossible. Antibunching, whether of bosons or of fermions, has no classical wave analog."

Spope3 (talk) 03:26, 20 February 2024 (UTC)


 * The Kimble et al interpretation has been disputed or at least revised (AFAICT):
 * Origin of Antibunching in Resonance Fluorescence
 * Lukas Hanschke, Lucas Schweickert, Juan Camilo López Carreño, Eva Schöll, Katharina D. Zeuner, Thomas Lettner, Eduardo Zubizarreta Casalengua, Marcus Reindl, Saimon Filipe Covre da Silva, Rinaldo Trotta, Jonathan J. Finley, Armando Rastelli, Elena del Valle, Fabrice P. Laussy, Val Zwiller, Kai Müller, and Klaus D. Jöns
 * Phys. Rev. Lett. 125, 170402 – Published 23 October 2020
 * Light of Two Atoms in Free Space: Bunching or Antibunching?
 * Sebastian Wolf, Stefan Richter, Joachim von Zanthier, and Ferdinand Schmidt-Kaler
 * Phys. Rev. Lett. 124, 063603 – Published 13 February 2020
 * Bunching and antibunching properties of various coherent states of the radiation field
 * M. H. Mahran and M. Venkata Satyanarayana
 * Phys. Rev. A 34, 640 – Published 1 July 1986
 * The Henry et al paper ref[10] discusses fermion antibunching and boson bunching, which conflicts with the claim about photon (bosons) antibunching.
 * I think we should remove the comments about antibunching here, as this topic is much more complicated that the text implies. Johnjbarton (talk) 16:30, 20 February 2024 (UTC)

Resolution of HBT compared to that from diffraction
Hey, how about comparing measuring stellar diameters to other methods? Like diffraction limits in a telescope? Dawes' limit and such? Betaneptune (talk) 18:18, 21 June 2024 (UTC)