Talk:Mach–Zehnder interferometer

Are there always two detectors in a Mach-Zender? I'm pretty. You only really need one, perhaps depending on the application. -- DrBob 17:03, 7 Apr 2004 (UTC)

According to Optics, Hecht, 4th edition, you need only one. I currently don't have time to rewrite the article though. -- Gerritholl 08:58, 20 Sep 2004 (UTC)

Don't phase shifts only occur during reflection off a material with higher n and never during refraction? If so then there are fewer phase shifts than the article states in the last paragraph. Chad 11/8/2005

There is a phase shift of pi upon reflection on a medium with higher index of refraktion, as you say. There is also never a phase shift caused by the process of refraction itself (off the top of my head, this wouldn't be compatible with the boundary conditions of the fields, but this might be entirely wrong). I think what the author means in this case is a difference in phase compared to what the light would have, had it traversed the same path in vacuum. Basically, the optical path length difference $n_{glass}x-n_{air}x$ -- Osquar F 9 Oct 2006

I think a full quantum description of the action of the Mach-Zehnder interferometer should be added in the main article. Danko Georgiev MD 12:27, 23 January 2007 (UTC)[reply]

This article is sloppy and lacks a bit of rigor. These interferometers are not just used for interrogation of samples but also for distance ranging and some other applications. A Michelson interferometer is a Mach-Zehner interferometer which has been folded upon itself. Also note that Michelson interferometers do have two output ports, the second is where light is reflected back towards the source. I'll delete this comment in the article and add some more stuff in the next few weeks when I have some time. Fincle (talk) 14:58, 4 February 2010 (UTC)[reply]

I agree that this article needs more rigor. Is there a flag that we can set to help it get picked up by one of the clean-up projects? Cavebear42 (talk) 01:41, 13 March 2013 (UTC)[reply]

What is a Mach-Zender modulator ? --68.0.120.35 20:26, 17 April 2007 (UTC)[reply]

With what probability photon reach detector A in right down picture ( http://wiki.riteme.site/wiki/Image:Interferometre_de_Mach-Zender_paradoxe.svg ) ?

This article should be compared to the one on Fabry Perot interferometers. That one has the math worked out and this one is more conceptual. I would imagine the articles would want to be written in a similar form. 136.152.177.188 (talk) 20:40, 30 April 2010 (UTC)[reply]

The drawing, http://wiki.riteme.site/wiki/File:Interferometre_de_Mach-Zender_paradoxe.svg, should be corrected or omitted; the probabilities are not legible. --Ryan Westafer (talk) 22:44, 14 July 2010 (UTC)[reply]

I'd have to agree here. The (very large) picture at the end doesn't seem to fit into the article at all and lacks explanation. It should either be expanded or removed. Cavebear42 (talk) 01:41, 13 March 2013 (UTC)[reply]

I did some cleanup work on the article. I added figures, and modified and provided an explanation for the figure that you found did not fit in. I would appreciate further suggestions for improvement. Thanks! Stigmatella aurantiaca (talk) 12:18, 19 March 2013 (UTC)[reply]

Suggestions for improvement

Here are some points I noticed when reading the article, which might help improving it further:

General
1. Avoid excessive use of "we". In parts this looks a bit like textbook style, did anybody check for copypaste problems?
2. A phase is an angle, not a (wave)length. For example, a phase shift of π corresponds to half a wavelength, but it is π.
3. Bulleted lists are nice where they make sense, but having a bullet in front of each paragraph doesn't help much.
Topics that could be covered/expanded if material is found
1. Other types of beams (electrons, atoms?)
2. Other, not so fundamental applications?
3. History
Lede
Is a bit shaky ("the two beams caused by a sample"?). Could be longer, could use a short introduction for non-experts.
Introduction
1. Start out explaining what it is and how it works, not with specific details (maybe re-arrange with material from the Set-up section). Introduce concepts like appearance of fringes or test cell before using them.
2. "Note also the precise orientation of the beam splitters." No, I don't, how should I? What's precise about it?
3. "when an extended source is used" or when the beam is expanded, like in the figure.
4. Figure 1: The text would fit better in the main article. "we imagine" ??
5. Figure 2: I don't get what that's saying.
6. What is a "well behaved response"?
How it works
1. Find a more encyclopedic section title
2. The beam splitters don't have to be half-silvered, dielectric ones are mentioned later
3. One detector can be enough
4. The second bullet point in set-up is hard to understand
5. In "Caveat": it's not clear which paths are equal or not and what energy conservation has to do with it.

These are of course just suggestions, some points might be not so important and there might be room for improvement elsewhere. I hope it helps anyway. — HHHIPPO 12:11, 14 April 2013 (UTC)[reply]

Suggestion for improvement

The section on Counterfactual measurement should be corrected. The question in the text and the caption of Figure 4 is ill-posed. There is nothing paradoxical about photons moving along the upper path being evenly detected by A and B when the lower path is interrupted. — Preceding unsigned comment added by 187.112.59.69 (talk) 16:51, 1 June 2013 (UTC)[reply]

Counterfactual measurement chapter

Could someone explain what is counterfactual or paradox about this experiment. It has the same simple classical explanation as a double slit where one slit is closed. Mirror M blocks one wave, so there is no (destructive) interference at A anymore, so A can measure photons again. DParlevliet (talk) 20:47, 24 February 2014 (UTC)[reply]

Consider the Elitzur–Vaidman bomb tester. Through straightforward modifications of the illustrated apparatus, it is possible to achieve arbitrarily close to 100% efficiency distinguishing between live and dud bombs, without photons ever being absorbed by live bombs setting them off. Does classical wave mechanics predict anything remotely comparable to this scenario, which involves measuring the properties of the bombs without any interactions with photons? Stigmatella aurantiaca (talk) 02:52, 25 February 2014 (UTC)[reply]

The bomb tester is a another paragraph. I asked about the "Counterfactual measurement". DParlevliet (talk) 19:56, 25 February 2014 (UTC)[reply]

They are very closely related. Stigmatella aurantiaca (talk) 21:12, 25 February 2014 (UTC)[reply]

QM version

In my opinion, it's a necessity to have 2 input-ports in this version. No input is equivalent to the vacuum. No vacuum, no QM.

nonlocalized fringe pattern

The article talks about localized and nonlocalized fringe patterns, but what does that mean? A fringe pattern is shown in eg the flame images in the article. Is it localized? nonlocalized? What would change if it were the other kind? Johnjbarton (talk) 16:51, 18 August 2023 (UTC)[reply]

Possible Mach–Zehnder interferometer Animation

( Moved here from Talk:Double-slit experiment).

I would favor an animation similar to File:Mach-Zehnder photons animation.gif but that did not depict individual photons in transient.

Animation could cycle through two major states: 1. top detector in path and 2. top detector not in path.

State 1 begins with top detector in path. Dark pink ray propagates from laser to first splitter. Two medium pink rays propagate from splitter. Top ray stops at top detector. Bottom ray proceeds to bottom splitter. Two light pink rays propagate to bottom detectors. Top detector flashes one half of the time. Bottom detectors flash one fourth of the time.

State 1 begins with top detector not in path. Dark pink ray propagates from laser to first splitter. Two medium pink rays propagate from splitter. Both rays proceed to bottom splitter/combiner. Two medium pink rays propagate to bottom detectors. Top detector never flashes. Bottom detectors flash one half of the time. Constant314 (talk) 18:27, 5 November 2024 (UTC)[reply]

In the context of this article (Mach–Zehnder interferometer), single-photon detection applies exclusively in the section Mach–Zehnder interferometer § Quantum treatment, and so the proposal is unclearly stated.

I would object on the grounds that it is the process of detection that is being presented in a biased fashion: it entrenches the way of thinking that is implicit in the Copenhagen school of thought (essentially that wavefunction collapse is inherent in any treatment). A single-photon treatment of the Mach–Zehnder interferometer is quite straightforward if the detection in various setups is kept out of the diagram(s). —Quondum 20:08, 5 November 2024 (UTC)[reply]

I'm wondering if instead of little particles, it would help comprehension to represent the photons as little wave packets? Then at the combiner mirror the in-phase packets could be shown with their crests coinciding, combining into one wave packet, and the out-of-phase packets could be shown with their crests opposed, canceling out to zero. This would include the vital wave interference information that is the basis for the operation of the device, making it more understandable. It would also avoid encouraging the erroneous view of particles splitting. --Chetvorno^TALK 21:13, 5 November 2024 (UTC)[reply]

Yes, that is one improvement that I would suggest (colour is used to show phase in similar diagrams). There are other improvements (such as that the wave packets should be stretched out to reflect the level of coherence typically needed in such experiments, and the colour intensity should fade off according to |ψ|, perhaps as an elongated striated Gaussian packet. The wavelength must be made large and slow enough (diagram unfortunately not to scale) to allow individual crests of colour to be tracked by eye). It should be noted that in the single-photon case, the issues with superposition do not seem to arise before detection: such issues are associated with multi-particle cases and entanglement. A classical wave packet seems to be an accurate representation of a single photon, but it fails for more than one photon. I would also suggest replacing the "laser" with a coherent single-photon source (or just a "photon source": a laser inherently produces a large number of entangled photons, and attenuation does not convincingly/obviously emulate a single-photon source.) —Quondum 21:33, 5 November 2024 (UTC)[reply]

I realize that this is only discussion on how to create a reasonable animation, but I want to discourage graphical representations of photons in flight. Ultimately this will lead to the same issues as the previous image. Shaped pulses or wave packets would be appropriate for time-based experiments, but wave packets are intrinsically composed of many energies (wavelengths) while photons are single energy, a single frequency, a single sinusoid from source to absorber. I know this is counter-intuitive, but that is exactly the point. Intuition fails us for quantum transitions propagating at the speed of light. Johnjbarton (talk) 23:16, 5 November 2024 (UTC)[reply]

I agree that there should be no depiction of individual photons in transit. The theory does not tell us where individual photons go. It only tells us the probability of finding them in certain places. That is fine when you are not trying to depict interference. Any attempt of tracing individual photons to interference is ultimately OR. Constant314 (talk) 00:40, 6 November 2024 (UTC)[reply]

I don't believe that this is accurate: the theory does actually ascribe a wavefunction as a function of space and time to a single particle, and this can be depicted. (When we have multiple particles, we cannot depict it readily as we need more dimensions.) But I do not feel it necessary to argue the point if we do not have a diagram; indeed, perhaps we should not even be having this discussion at all, since we should not be trying to figure out interpretations in WP, and there is clearly disagreement. —Quondum 01:04, 6 November 2024 (UTC)[reply]

That is a wise outlook. My initial attempt was to only depict the observables, which are the blinking detectors. Constant314 (talk) 02:33, 6 November 2024 (UTC)[reply]

As to interference, when multiple particles are involved, there is a problem depicting it, but you have self-interference even in free-space propagation; if this is fine, then any setup that impacts propagation is also fine. —Quondum 01:07, 6 November 2024 (UTC)[reply]

The multiple photons was a nightmare. Constant314 (talk) 02:43, 6 November 2024 (UTC)[reply]

I've made this alternative gif following the suggestions discussed above. How do we feel about it? L3erdnik (talk) 21:19, 8 November 2024 (UTC)[reply]

When the upper detector is out, the remaining beam is split equally between the two bottom detectors. Both of the bottom detectors blink half of the time. Otherwise, I find that this animation is acceptable in that nothing is animated except what is observable. Probably needs a note specifying that the "beams" simply depict the optical path and average intensity. Constant314 (talk) 01:46, 9 November 2024 (UTC)[reply]

Would it be possible to add numbers near the detectors synced with the position of the moving detector? 1/2,1/4,1/4 and 0,0,1. Or 0, 0(destructive), 1 (constructive)?

This source

Marshman, Emily; Singh, Chandralekha (2016-03-01). "Interactive tutorial to improve student understanding of single photon experiments involving a Mach–Zehnder interferometer". European Journal of Physics. 37 (2): 024001. doi:10.1088/0143-0807/37/2/024001. ISSN 0143-0807.

uses a different comparison: rather than inserting a detector, they remove the second beam splitter. That removes the destructive interference (the bottom 0) because the path is known by which detector fires. In that condition, the entire path from the laser to the detector could be lit (detector fires path known) or unlit (detector off) and the paths could switch back and forth randomly. Johnjbarton (talk) 02:26, 9 November 2024 (UTC)[reply]

The probabilities - can do. Exchanging removable detector for removing the second beam splitter, however - that leads to some problems. Right now the animation seems to be only (mean to be) used in a MZI section in double split experiment. In that context blocking one arm makes sense, it is equivalent to covering one slit with a detector. However, removing the second beam splitter doesn't have a clear parallel in double slit version. Introducing a phase shifter instead of the upper detector would make sense in that context, but the section would have to be rewritten accordingly. L3erdnik (talk) 03:40, 9 November 2024 (UTC)[reply]

Yes sorry I lost track of that important aspect. In that case a source is

Maries, Alexandru; Sayer, Ryan; Singh, Chandralekha (2020-07-01). "Can students apply the concept of "which-path" information learned in the context of Mach–Zehnder interferometer to the double-slit experiment?". American Journal of Physics. 88 (7): 542–550. doi:10.1119/10.0001357. ISSN 0002-9505.

which is cited a the beginning of the the Double slit section on MZI. Johnjbarton (talk) 17:43, 9 November 2024 (UTC)[reply]

Just wanted to add my opinion of the new animation: I feel it is even less informative than the previous one. I can't imagine anyone who doesn't already know how a Mach–Zehnder works getting much information from it. --Chetvorno^TALK 02:56, 10 November 2024 (UTC)[reply]

What if instead of beams I made is sine waves? John really opposes localization in longitudinal direction... L3erdnik (talk) 13:07, 10 November 2024 (UTC)[reply]

Because of the change in phase upon reflection, sine waves might illustrate the interference that cancels amplitude for the bottom detector. (FWIW, I would also oppose angels pushing planets to illustrate gravity for the same reason: lack of evidence/sources. The folklore of particles is much more difficult for us to shake because it works in many seemingly similar circumstances.) Johnjbarton (talk) 16:57, 10 November 2024 (UTC)[reply]

An animation about modern (quantum) optics and wave-particle duality is just going to be difficult.

Maybe we have two separate issues:

what do we put in the double slit article for MZI?
how to explain MZI?

Maybe we need different answers for these. EG: for Double slit just a pair of static images, MZI and double slit. "See two paths that converge and interfere". For MZI the animated image can be discussed in detail as away of explanation.

I still think that images based on interference (eg candles) are a better place to start. The wave case can have an image but words suffice for the quanta. I wish we had an MZI equivalent to the images in Wave–particle_duality#Electrons_behaving_as_waves_and_particles where the interference image is built up a dot at a time (Or the screen is a blob of dots for the other case). Johnjbarton (talk) 16:28, 10 November 2024 (UTC)[reply]

I want to avoid those funky half photons that sometimes disappear in flight and sometimes become whole photons. I would have no problem with continuous beams showing wave-based interference. I think animating what happens to individual photons is a step to far into OR. Constant314 (talk) 17:54, 10 November 2024 (UTC)[reply]

@L3erdnik: I think using sine waves would be a big improvement, it would provide a visual explanation of how the device works. I would say that including waves is a minimum prerequisite for a useful diagram. I appreciate all your work, I know how hard it is to make diagrams of quantum experiments comprehensible by the general public. --Chetvorno^TALK 20:15, 10 November 2024 (UTC)[reply]

So far the hardest part seems to be to get the people converge on a consensus on what exactly the animation should look like... L3erdnik (talk) 20:52, 10 November 2024 (UTC)[reply]

@L3erdnik: Forget it. You're never going to get a consensus on this, and if you did the resulting animation wouldn't be any use to anyone but physics grad students. Just draw the animation as you think it should be, put it in the article, and make the least changes to it necessary so editors will live with it. --Chetvorno^TALK 22:08, 10 November 2024 (UTC)[reply]

I've made a version [1] with probabilities and sine waves. How about this one? (Also, would the main image update in a while? When I try to use it now it still uses the old version) L3erdnik (talk) 22:29, 20 November 2024 (UTC)[reply]

Sometimes you have to wait a little while and sometimes reload your page. Seems like it is always fixed by the next day. Constant314 (talk) 23:36, 20 November 2024 (UTC)[reply]

In the case with the upper detector out, I thought that both detectors in the bottom corner get equal numbers of photons. Can anyone confirm or rule out? Constant314 (talk) 23:42, 20 November 2024 (UTC)[reply]

The way I understand this, the mirrors introduce a phase shift so with the detector out the bottom detector is in destructive interference and the right one in constructive interference. But of course it is never as simple as all that. The half-silvered mirror plays an important role. Here is ref that discusses exactly the issues:

Zetie, K. P., Adams, S. F., & Tocknell, R. M. (2000). How does a Mach-Zehnder interferometer work?. Physics Education, 35(1), 46.

Johnjbarton (talk) 23:56, 20 November 2024 (UTC)[reply]

Makes sense. Constant314 (talk) 00:11, 21 November 2024 (UTC)[reply]