19:50 Thinking of photons as discrete transactions is so insightful - the most understandable explanation of photons and interference patterns I have encountered in my 67 years!
You really demystified the wave particle duality in this one. The fact that they are probable interactions that represent a wave was the a ha moment for me.
Fantastic video! I started watching this and as soon as I got to the astronomy part I rewound and re-watched it with my wife, who is preparing to perform this exact measurement at an observatory.
I am a new PhD student in biomedical imaging, and I’m currently taking a statistical optics course. Really enjoyed the video! We literally just derived the van Cittert-Zernike theorem in my class from the propagation of coherence.
You're the first person I've heard clearly articulate that the photon interactions are quantized, rather than the field everywhere. Would've been really helpful when I was first wrapping my head around this topic!
The real world (particles in the way of light, or our eyes) see it as quantised but the wave itself must be not since repeating the same single photon experiment brings different peaks everywhere that show that a wave must be present to have these interactions measured and not single photons. Although they say interaction with light collapses the wave (function) and then forth it looks like single particles. So the quantised interaction removes some energy out of the non localised wave? Until it's energy gone and dissappears from all potential points of interaction at once? It's hard for me still to grasp this faster than light wave collapse of light. Can Huygens maybe explain that simpler?
I could not understand the basic concepts of quantum electrodynamics before, but seeing little people slap each other finally helped me. Thanks Huygens Optics!
Great video! I as I saw some one point out sodium lamps and spectral filtering, there's another way that has been used to create what I'll define as pseudothermal light at a specific wavelength: namely a rotating diffuser. This technique was shown I think in the 60s, but you can read about it more thoroughly in goodmans statistical optics. The main idea is to consider a single frequency laser and look in the far field to see the speckle: it should be stationary. If we add several frequencies, each with their own speckle, we see that the speckles will sort of add together and when you detect youll average over the various speckle configurations leading to what looks like a speckle free beam. Now, if you take a laser and put a diffuser in front, youll end up with a random speckle pattern in front due to the random phases imparted to different spatial points of the beam on the diffuser. As the diffuser turns, the speckle pattern will change as the pattern of hills and valleys change on the diffuser. By rotating, and averaging over all realizations of the diffuser, youll end up averaging over all of the speckle patterns and thus something similar to the LED. I call this pseudothermal because the light is not coming from a blackbody source. This is easily implemented with a cheap RC car motor and some plastic (scotch tape works pretty well usually, but it depends on the wavelength of light and how big the small deviations in the surface). Ive used this (well using a random piece of plastic) in the lab to get g^2(0)=2 within experimental error. Outside if this, I want to point out a couple of other things. I think you glossed over the Hanbury-Brown-Twiss conteoversy too fast. There's a ton of interesting history and actually this discovery lead to what Ill call the second revolution in quantum optics. It took a painful derivation by Roy Glauber to show that indeed the g^2(x,t) is related to the first order coherence (the g^1(x,t)). The formulation to show this lead directly to the second revitalization of quantum optics. Some other stuff that might be fun to try. First, you should try doing homodyne tomography of different quantum states of light. Effectively, this is taking your weak quantum state of light (thermal, coherent state, single photon) and interfering it at a beamsplitter with a strong laser. By subtracting the currents in either arm (called balancing), you can get access to reconstructing the quadrature state of the light. Most of these will be gaussian unless you can get a nonclassical light source. On this note, you should also do the famous Hong-Ou-Mandel interference experiment that actually (and directly) shows the "particle" nature of light. In this case, take your beamsplitter and interfere two photons at the input ports (one at each), if the photons are truely identical (same polarization, spatial mode, and frequency distribution) then it turns out that the photons will always leave at the same port and never at different ports. Then as you change the states to become more distinguishable by, e.g., delaying one photon by some time, as the mode overlap between the states decreases, the coincidences will begin to increase. Specifically, if you have detectors at the outputs, then youll never see coincidences. Further, this cannot be due to a simple wavelike interference effect because it isnt phase dependent. You do not have an envelope with oscillations under it in this effect. If you did the same thing with coherent states (i.e. laser light), there is an oscillation effect that is phase dependent. Thermal light, also is phase independent, but the dip in coincidences will be smaller.
i love the explanation that photons are the interactions with the em field. this is so much easier to think about compared to the way i am taught quantum physics in school thank you!
I showed several coworkers your video on the photo-multiplier. We work as construction inspectors. We sometimes use a Nuclear Density Gauge. These devices use a photo-multiplier. It was really cool to see the way a device works that is very difficult to understand.
AT 1:47.... finally able to comprehend what wave function, oscilloscopes and wave diagrams, the wave itself; and, what they represent after watching hours and hours of videos discussing electrical engineering or electronics, photons and quantum effects, the science of sound, etc. etc. So, thank you for finally bringing a clear picture and understanding into my knowledge base!
The Hanbury Brown and Twiss effect seemed so wild to me when I first heard about it. But you're right that you can't argue with experimental results. I also appreciate that you were able to do all this stuff in software. So much more pleasant than doing this with a crate of analog electronics and a mess of coax cables.
Amazing video as always! I've been doing g2 measurements in our lab for a few years now, albeit with single photon sources and ~$300k worth of detectors and timing electronics. It is unbelievable how much youve accomplished with just a scope and some PMTs! I have a recommendations that might help you resolve photon bunching. You could replace the broadband LED with an atomic spectral line. If you have a sodium or neon discharge, you could use a diffraction grating to send a sinlge emission line to the input of your HBT setup. Most of those spectral emission lines are very narrow (sub-GHz), which corresponds to coherence times in nanoseconds. Ive used this technique to measure argon spectral lines in the near-IR to calibrate a spectrometer and it works very well. The next logical step of course is to construct a spontaneous parameteic down-conversion based single photon source and characterize that 😋! You could even do Hong-Ou-Mandel interference, the wackiest measurement ever! Cannot wait for your next upload!
As a hobby, I'm working on re-framing many basic elements of physics modelling using a finite-axiom approach (including the photon) and so far it's been interesting, enlightening, great fun, and often points to quite a different set of models as commonly used in 'standard' physics. It's interesting to see how you rationalise the photon in the standard system of models - importantly, for me, it is the measurements that counts and I greatly enjoy the way that you focus on the measurements with such great detail. Ultimately any model has to be able to reflect and predict experimental measurements. It seems to me you are searching for understanding. I would simply add caution to both the acceptance and understanding of current models and build a confidence in your own ideas. You are one of the few people I would like to share my initial ideas with (and real results) with. Mostly my ideas are likely to go nowhere or be simply wrong and will be likely laughed at, but as I progress It's been such fun. However, if the results go well I hope to publish the work - even if it's just to show how not to do it - LOL. Anyway thank you for showing the results of the experiments - especially as it means I can get a strong feel and sense of the measurements and the process etc. This has really helped me. Thank you for sharing your work.
A major component of the pulse height variations from photomultipliers is due to the variation in the number of secondary electrons generated at the first dynode. This variability is described by the Poisson distribution, which says that if n_ave is the average number of secondary electron ejections from a dynode, then the probability of n secondary electrons being ejected for a given incident primary electron is (1/n!) * (n_ave)^n * exp(-n_ave). This distribution is broad when n_ave is small, but becomes narrower the larger n_ave becomes. This source of variability in the pulse heights can be reduced by increasing the voltage drop between the photo-cathode and the first dynode, which increases the secondary electron yield.
But increasing delta V between cathode and 1st Dynode also increases the noise. In the 1980's I worked on the detection of Cerenkov radiation from ultra low radiation sources with a single photon emission rate of 10 to 20 counts per minute and photomultipliers specifically selected for ultra low dark counts of < 1 count per minute (when cooled to -50 degC and using discrimination electronics - see later). The multipliers NEVER saw any day light and were kept in the dark even when powered off. We had several older pmt's which had increased dark counts (because of a loss of ultra high vacuum in the tube and with these we were able to demonstrate that a single brief exposure to room light (neon tubes) caused some residual fluorescence / phosphoresence response in the cathode but room light from tungsten filament lamps showed a lower effect on the dark count. I spent endless hours in the darkest lab with walls painted black, every surface covered with black cloth. The experimental setup was contained in a thick black canvas curtain covered box (for access to the samples). Once the canvas was washed and the experiment went to hell, because of the washing power used to clean the canvas contained a whitener i.e. a flourophore. In terms of pulse counting, we used threshold discriminators to eliminate low energy signals stemming from dark counts which typically had only a signal intensity of 1/10th of a single photon event. Since Cerenkov radiation (from beta decay of 32P) is a truly random event we were able to use statistical analysis to get to the real single photon count rates and eliminate any spurious noise. The other 'nice' aspect of working with a radioactive substance is that it has a natural decay rate (half life) and thus we were able to use measurement over several half life times to extrapolate back to zero and eliminate the last remnants of dark counts. I am still proud of these measurements conducted as a diploma student for my dissertation.
@@robertrjm8115 Was there a published paper that is online? Sounds fascinating. I saw the brighteners becoming an issue as soon as you said washed. Perhaps dry-cleaning would have left less residue. Scary how normal things like opening a box can ruin a sensitive detector.
@@KallePihlajasaari Hi Kalle, I am sorry, but NO there has been no publication on the issue of brightners throwing a spanner into our work. Also, this was in the early 1980's and very few publications of that era have been digitised. Also, our work was medically related (diagnostics of malignant melanoma of the eye by a 32P uptake test) and although our work did succeed in detecting melanomas the issue was that it required the injection of a fairly high dose of a radioactive substance and the potential side effects were considered to be too severe to be acceptable as a routine diagnostic tool.
This is so mathematically logical and practically logical and useeful at the same time it deserves an investigation of voltage drop to average secondary electron yield depending on the dynode material by itself.
I'm not a physicist, and I've never "truly" understood the wave/particle duality, but I feel absolutely cheated that it took until now hearing about looking at the photons as interactions. That's way more intuitive, and my understanding about several concepts have changed significantly from that single insight
Thank you so much for this video! I just started working with a single photon source research team as an undergraduate and I was having trouble wrapping my head around the g(2) measurement - it makes a lot more sense now
Thank you very much. This has answered several long standing questions I have had about how interferometry is used in astronomy. I don't think I really understand it yet, but I am much closer than before. I have asked questions on several relevant videos and never got an answer that usefully answered any of the issues I had.
This is fantastic- I had a student working on approximately the same setup to measure g2 with similar PMTs. I think what could lead to better data is using coincidence measurement for longer accumulation times (specifically, triggering coincidence events every time two pulses overlap with a specified time window). We originally tried this with NIM modules. Remember, HBT claimed better timing resolution for their tabletop experiment than their detractors because of the care they took in the experiments and the coincidence techniques they used. For the delay, if the data were collected as pure coincidences (and not as single counts from both detectors), the experiment could be performed in the same manner as the original HBT demonstration by translating one of the detector pairs relative to the other. The chances of measuring coincidences will be greatest when the two detectors are equidistant from the beamsplitter, and this probability should drop as the delay is adjusted. For 5 fs of temporal resolution, you would need 1.5 microns of spatial translation, which is possible with a flexure stage or PZT actuator.
There might be a misunderstanding here; the adjustment in the Hanbury Brown and Twiss demonstration was orthogonal to the beams (to sample the same or different spatial modes), rather than along the beams. The longitudinal adjustment, i.e. distance to the beam splitter, does not matter (if there is no dispersion in the medium). All it does is to shift the arrival times, but this can also be done digitally in the data analysis after the fact. A shift vanishingly small compared to the temporal resolution of the single-photon detectors (as 5 fs would be) does not matter.
@@DavidNadlinger It seems that you have some experience in the topic. Since one week I wanted to show bunching effect using halogen lamp and two SPCM Thorlabs single photon detectors. The halogen lamp is fiber coupled and output is collimated . I am using PBS in the experiment. I am using timetager and my histogram is always flat without any peak in g(0). Do you have an idea what I am doing wrong. Do I need use some pinholes to get point source. Is temporal coherence matters here?
@DavidNadlinger Fair point and I definitely missed it! Reread the first page of the paper, the distance could vary by a cm either way. Thank you for your correction! I still think that the small collection window compared to the coincidence rate could be the reason it was so difficult to see the correlation here. HBT collected coincidence data over the course of an hour for their source/detector combo, any comparable length data set using 10 ms intervals would take 360,000 samples assuming a similar true coincidence rate.
i'll come back to this properly after escaping the ADHD well i'm currently stuck in (but comments, engagement, etc.) 😅 good pointer to sigview, this looks very useful your experiments continue to be among the most enlightening and approachable explorations of wave mechanics i've come across. i particularly appreciate the practical tips and "dead ends" you sprinkle around, showing your thought process, and contrasting positive/negative examples worth pondering
I'm familiar with the bunching effect derived from the wavefunction symmetry constraint. But I can't believe I hadn't heard of the Hanbury Brown & Quintin effect! Thanks for sharing!
Very interesting experiment! I agree that temporal resolution is key, because within 10 ns, plenty of oscillations and fluctuations in the density function occur. So whether two waves are synchronous or not doesn't make much of a difference... Btw, people like you who say "the math is complicated but the essence is simple" are both good scientists and good educators.
I enjoy this channel because you’ve landed on a similar intuitive interpretation of quantised photons that I have. While I can see the flaws in it (no probabilistic element) I love the slap analogy!
You have an enviable approach to explaining things and presenting your work, which is beautifully done and very nicely illustrated. I am delighted to hear someone talking sense when it comes to quantum electrodynamics. The term ‘photon’ has become so abused that I meet few scientists who use it properly. Many physicists also utterly fail to make their ideas understandable by using fancy terminology and complex maths. Nature is quite forgiving in this way though, which is why the misconception persists. I was relieved to see you use the highest time resolution to give you the chance of observing the weak temporal correlations of fluctuations in the incoherent light (I was wondering what your sample rate was from the outset). Not sure about the huge amplitude fluctuations of the pulses from your photomultiplier. I guess I need to watch your other video to understand that…… Signing off, I hope you continue to enjoy making these fabulous videos. I thank you by showing a short clip of me at work in my ultrafast IR spectroscopy lab and analysing data. My strange little video (filmed during the pandemic) is intended as a snapshot showing a couple of minutes of working days that can last more than 14 hours doing research trying to figure basic things out about light interactions with molecules. th-cam.com/video/Vy0mSO7euFA/w-d-xo.htmlsi=0TyWBYky8ycg5l3-
19:30 waaaaaaaait.... exactly this is what i think about all the "photon" theory. The EM field itself is not quantized, just the energy transfer! This needs much more attention and should be published as research.
So about this.. the part where quantization will be necessary is in the third part (anti bunching) which was not discussed in this video. I think the goal of this video was to show that many phenomena, which usually get attributed to effects in quantum physics, are not as non-classical as one thinks, and in fact also can have a nice classical or at least semi classical view - which I enjoy very much. That at some point classical theories fail is consensus in the physics community and there are enough experiments to indicate that. But all those quantum optics effects are very hard to understand since they involve quite in depth mathematical understanding.
only problem you cant send more energy to electron by raising the amplitude of light wave, bases on that peoples agree that there is something in light, not only in electrons. like, you need different wave length to transfer more energy and move electron in higher position, but you cant do it by increasing the amplitude of some infra red wave length. but... maybe it is explained by core of the atom is waving with the electron on low frequencies. idk.
@slimeball3209 I guess the main problem is that most people view the electron and the EM field as 2 separate things, whereas they are not: The electron's charge is an integral part of the total EM field and therefore any energy transactions to the electron involve the field as a whole. In a sense, the electrons basically just add some local impedence to the field. The other question you might ask is: where does the total photon energy come from if it is not contained locally in the field? I tend not to worry too much about, since as mentioned above, it is a transaction with the field as a whole. Apparently quantum physicist are not bothered by that question either, because QED assumes spontaneous creation of virtual particles and photons. On top of that, an energy content of 10^120 Joules per cubic meter of vacuum is needed to make the theory work.
@@HuygensOptics no, i would not ask this question, i imagine photons as delayed movement of electron, so it includes polarisation, but still dont understand what is happening when we impact the atom with different frequencies of field, and why is need higher frequency to make electron jump higher, but we cant use amplitude of lower frequency to make this electron jump at the same "height" like, if 1 electron goes up, so in some time other electron will see and also go up, some electrons are locked to surface of material, so they moving in projection, and at some angles we will see it as polarisation, especially at high angles, because in that case we have two projections that flattens the initial movement down to 1 direction. it is all about waves still, but i dont understand what is going on in electro chemical side. i questioning the mechanic of breaking the electron energy level barrier, and how the high frequencies can do it, but lower cant, even with 10x amplitude and the same ascent of higher frequency. and i too far from exact math of this phenomenons, so i cant even make valuable question about total energy of interaction ect.
@@slimeball3209 Because an electron is also a wave, a standing wave, trapped around the nucleus of an atom. To jump higher, a standing wave must increase its frequency, and only light of a higher frequency can excite it to do this. Light of a lower frequency will simply pass through the atom as if it were transparent. Here's an example, not very perfect. A ball is bouncing on a surface. Assuming there is no friction, the ball will continue to bounce to the same height. If the surface moves up and down slowly, regardless of the amplitude, the ball will continue to bounce to the same height. But if the surface starts bouncing up and down at a higher frequency, it can transfer its energy to the ball, causing it to bounce higher. Again, this example is not perfect, perhaps you can come up with a better one.
I love this video, thank you for sharing the results of your hardwork, and for sharing your frustration with the language qm physics uses to describe the concepts being worked with - paricles instead of energy exchanged.
This video (and many of your videos) amazes me with the depth of complexity in concepts can be presented while still being understandable. There are few sources of information that make me feel both enlightened and overwhelmed at the same time the way your videos do. I know this would be a departure from the things you are working on, but I would really love to see an explanation of polarization including terminology, filtering, detection, and how the concepts relate to the wave propagation, refraction, reflection, and diffraction. I have never seen any presentations that talk about polarization concepts in a field of waves.... it's usually only diagrammed in terms of a single wave 'ray' I think. Also, your mixed in 'bye bye' at the ending got me laughing. Thank you for that.
Looking into polarization and single photons is definitely on my list. Unfortunately, I'm not sure if I can make a video about the subject any time soon due to a lack of time.
Single photon emission and anti bunching is the logical next step! You just mentioned them in passing. The explanation of the interactions being quantized, not the field, made some things clear, but now I'm no longer sure about my understanding of actual single photon events (where there is only one photon emitted) like from nano dots or nanoparticles.
Oh rad! Another video of yours. I always end up going though some old quantum & optics notes, because there's always something that suddenly starts making sense... "the probability of discrete energy transactions is governed by a field intensity distribution" klicked for me. Anyway. Thanks for this amazing content!
Great video as always. Thanks for sharing. . . . . . . . . I continue to assert that light possesses mass. If light is interpreted merely as interactions (energy transfers), it implies that it must collide with something to provoke a chain reaction of energy transfer until it reaches a point of absorption. This behavior suggests that, like ordinary matter, light could also exhibit intrinsic mass properties. Furthermore, I believe there is matter in the vacuum of space, even if we do not perceive it directly. The crucial difference is that our vibrational frequency does not coincide with what is necessary to physically detect this matter, just as we sense water or earth. Analogous to the limitation of sound propagation in space due to the absence of a suitable medium, light, in principle, should also have limitations imposed by barriers, fields, or specific zones of influence, even if space appears to be empty. When we consider electromagnetic waves, we know that, for energy propagation, there is traditionally an assumption of the need for a medium (which historically led to the hypothesis of 'luminiferous ether,' later refuted). However, we may not yet possess the necessary technology to adequately detect the media or fields that could mediate these energetic interactions.
pretty astute to use the plastic box as freezer compatible by keeping the setup small enough... a lot more efficient and manageable the Peltier modules, and more confortable than working in Siberia. I suppose the "classical" approximation is convenient as long as one do not look into the atomic level (requiring enough precision to know where the interaction occurs), so that one is reduced to look into the phenomenon statistically (over a large number of occurences, even if temporally of discrete events) rather than studying each single interaction along the path from emission to detection (at the atomic level, including rarer events like absorption by beamsplitter etc. which are discarded by the classical observer). Thank you for your continuing enthusiastic work, brings back memories, keep it up!
one of the distinct qualities of photon interactions that result in them being described as being also particle like, in comparison to how you describe it, is the implications that quantized energy transfer has for conservation of energy. If we think of three hydrogen atoms floating in space, and consider one of them to have an electron in an energy state above its lowest potential: when it drops down from its higher orbital state it releases a photon. If the photon were entirely described by only wavelike mechanics from the moment the electron is done transitioning to a lower energy state, then the waveform defining it would be, in the simplest case, some isotropic, spherical wavefront that propagates out equally in all directions. However if that was the case then it becomes highly unintuitive how we could be confident that only *one* of the other hydrogen atoms could have an interaction with the wave, resulting in that hydrogen's atom transitioning to a higher energy state. This needs to happen otherwise the single emission of the first atom could potentially result in two equally energetic adsorptions, resulting in the breaking of conservation of energy. In this thought experiment it becomes far easier to explain the behavior of the interactions in the form of a particle, as a particle of light would inherently have the needed properties that maintain conservation of energy. I would love to hear your thoughts on this issue, I am definitely not an authority on this topic and perhaps theres something I am missing, but this really seems like a very hairy problem that gets at the core of the unintuitive nature of wave/particle duality.
I don't think that conservation of energy is really a thing on short time scales. It's not an issue in QM either given the uncertainty relationship between time and energy and the fact that it only works by assuming virtual particles and photons popping out of nowhere. The way I look at this is that basically all energy visible to us (matter EM radiation) is borrowed from the vacuum. And so an energy quantum absorbed is in the background a transaction with the vacuum energy. The vacuum will settle the total energy bill later. In a sense, it's the EM wave (which by the way can be strongly localized both in time and space) that sets the transaction in motion, and not so much what supplies the actual energy locally to an electron.
*I would love to see a Video about how to choose the right optical elements.* So when to use plano-convex, plano-concave, convex-concave, concave-concave, convex-convex -lenses and also other types like Spherical, Aspheric, Achromatic or Cylindrical -lenses and Fresnel-lenses.
If you can get your hands on a TAC and MCA (time-to-amplitude converter; multi-channel analyzer), you could improve time resolution by 2-3 orders of magnitude. It would require minimal modification to your existing setup and give you proper TCSPC ability (time-correlated single photon counting).
A cost-effective way to improve the temporal resolution here might be to use an inexpensive direct-conversion SDR instead of an oscilloscope, to perform the binning much closer to hardware. The core principle of Direct Conversion Software Defined Radio is the idea that, given a signal that only includes frequency content within a given bandwidth of some center frequency, you can fully reconstruct that signal from pairs of 'quadrature' samples, i.e. taken a quarter-wave apart, sampled at the bandwidth frequency -- you don't actually need a _continuous_ nyquist-approved sampling rate unless the lower frequency limit goes all the way to DC. Even an inexpensive SDR like the HackRF One can set a center frequency anywhere from 1 MHz to 6 GHz, meaning a quadrature sampling time that can be swept anywhere from 42 ps to 250ns. The setup I'd be tempted to build here: Get a pair of HackRF One SDR's, set up synchronized with a common refclk. Build a trigger pulse generator circuit to condition the PMT's output, and use the SDRs to capture that output. From computer control, set the SDRs to produce synchronized pairs of I/Q sample pairs, setting the 'operating frequency' to the frequency with a quarter-wave equivalent to the tau value you want to measure g(2)(tau). Then count the rate of sample pair pairs where one tube has a detection in it's samples I value and the other only in its sample's Q value. I suspect you'd need to do some additional analog signal processing on the PMT tubes first -- mostly, to clip the lengths of those pulse tails, though possibly you might want to also modify the SDRs to remove frequency filtering components from the signal path. If you're interested in this approach though I'd be happy to share more details -- but this is already super long for a TH-cam comment as-is lol.
Thanks for this suggestion, I will definitely look into this method. I can also measure the signal directly on the anode over 50 ohm, which drastically reduces the pulse width.
You said that g(2) was the second order correlation function. In what sense is it second order? What would be first or zeroth order? I greatly enjoyed the video and learned a lot. Thank you!
6:10 you mention that you use _reflective_ attenuators. 1) Might these interfere with your experiment? 2) Also, if you stack a number of reflective filters back-to-back (or even combined with a laser-cavity) will they influence the total attenuation and/or operation of the laser? (or are they mounted at an angle, to dump the reflected light into a black surface off to one side? It doesn't look like this, to my (inexperienced) eye. Thanks (again) for a wonderful explanation of difficult topic!
It's very nice to see experiments like these popularized in this way. Regarding whether em fields are made out of photons. I would say that, ultimately physics is about models. If the photon model doesn't help in this case, you don't have to use it. As a researcher in quantum optics, the situations in which the photon picture is really necessary are rare.
An excellent presentation of a complex subject. However I am wondering if some of the bunching effects could be a result of the white light LED. I know the bunching is real but the LED source used here could have an effect on the level of the observed bunching. I assume this was a blue LED covered in a flourophore with a broad 'white' emission spectrum. There are several processes in the flourophore runnning in parallel to create the white light emission spectrum, some of these could create down stimulation from an excited state and some could be due to multi photon interaction in the flouresence material. To eliminate these effects it might be better to use a pure thermal emitter such as a tungsten filament lamp which as a near perfect random emission rate which can then be cut down to the appropriate single photon level. Also, maybe cutting down on the spectral band with of the white light (maybe 50 nm wide ?) would make the to results between the coherent green laser and the non coherent source more comparable. I have sent extracts from your videos to my former colleagues and they have used them in some of their lectures - and of course fully acknowledged you as the source and told the student to visit you channel.
Very interesting videos. Very helpfull to understand the PMT signal processing. From your point of view how do you understand the mysterious correlations between the detections from Alice and Bob in the Bell EPR experiments? Do you plan to make the Bell experiments with entangled polarized photon sources? It would be very fascinating. I would be curious about the correlation function outputs...
Is it possible for a light beam to be collimated, but not coherent? (Ex: a straight beam of light, constant cross-sectional area, but having thermal frequency distribution)
Temporally incoherent is of course possible in a collimated beam, spatially incoherent is infeasible because having different frequencies in the same plane would make the waves change direction and spread out immediately.
Just a brilliant video. As previously said by other commenters, it it would have been possible to press like twice or even more, I would really want to do it. Instead, I will just comment it. Treat it as a triple like. Thanks!
Awesome video. I'm glad there is someone else on youtube demystifying the quantum woo. Probability and statistics are CENTRAL to quantum theory, and modern physicists are arguably using this fact to confuse even themselves, knowingly or (most commonly) unknowingly. You might have heard about a central result of QM, Bell's theorem. I've got a couple videos on demystifing that as well if you're interested. You might get some inspiration to reproduce Aspect et al. results using optical instruments.
I never understood the bell setup due to all this mystifying crap to a point where i could say it makes sense. Could you share the video to general population?
@lucelxebinog it's on my channel. The experimental setup is very easy to understand. I use a coin toss analogy for those unfamiliar with the quantum mechanical formalism, but I also go in depth as to why ultimately Bell's theorem does not provide a complete description of the setup and why it fails to prove what everyone says it does. There is also a video that goes in depth on the history of hidden variable theories that also explains what the formalism means. Hope that's understandable to you, if you're left with any doubts feel free to comment under the videos
Masterfully done. I would be keen to see the anti-bunching measurements using HBT. I reckon you could get your hands on a suitable nitrogen vacancy diamond sample.
It sounded like you were collecting data on a star using your own telescope. If so, I'd like to suggest that if you have a full spectrum modified camera, or any camera with no uv-ir blocking filter to try again capturing data in as far into the infrared as possible. Maybe longer wavelengths could help? Fascinating video. Thank you!
13:00 well, photons are localized, but to varying degrees. What exactly a photon looks like *depends on your system*, which is something I think that even most practicing physicists aren't really aware of.
Single photon detections out of a PMT are approximately gaussian. If you look at Cherenkov light, where you can get something like 6 photons on average, then you get a Poisson dist. of number-of-photons convolved with the gaussian PMT response (iirc, then the N photons is a gaussian w/ variance ~N)...that results in a Polya distribution..which is so unwieldy idk if it helps.
Its curious how reflection and refraction is so difficult to explain with quantum mechanics, while interaction between light and matter is still there (surely there's still at least momentum exchange when light hits the atoms?)
Great video, I learned a lot! A few Questions: How many wavelengths is single photon worth of energy? How long does it take to be absorbed? What happens to the electron over the timeframe that it is being absorbed?
So tl;dr is, that likelyhood of getting multiple detections at the same time is higher with temporarily incoherent light, because the field has more energy on average in specific time frames? Also using this to measure star diameter is pure genius.
You have some of the easiest to digest videos on their subjects I've seen and as someone only very casually into optics, I really appreciate this and all of the content you've made so far. You rock. Also, what do you use to generate your animations of light sources and interference patterns?
"All wave animations in this video were produced using a Python script supplied by @DiffractionLimited . Thank you very much Manuel for supplying me with this tool."
Ik ga binnenkort een AAS slopen. Heb je belangstelling voor de optische bank. Grote Czery-Turner monochromator met een PMT? Eventueel kan ik ook wel enige holle cathode lampen missen.
I was over my head when you were just* grinding lenses. But I still love your videos. So, not just an elaborate way to justify getting a much, much faster scope? (5fs)
What if I say that: The electric field of light is best described as a statistical quantity extracted from many photon detections. It does not make sense to question "what is the electric field of a single photon?"
Something that confuses me: suppose a photon is emitted in one place and absorbed in another. After it's emitted, it creates an EM wave in 3d and therefore there's a probability distribution of where it could be detected. But after the photon is absorbed, does the entire 3d wave instantly disappear everywhere in 3-space? It has to, otherwise there's a nonzero chance the same photon can be absorbed twice, which sounds like it would violate conservation of energy (but making a wave instantly disappear feels physically impossible too...)
This is one of the reasons why QM and GR don't go well together: You either violate energy conservation or the maximum speed of information. I choose to "violate" the first law, because at very short time scales, energy conservation is not really a factor.
Your point is exactly what Einstein was expressing in his "spooky action at a distance" statement. How does the photon that is "everywhere" get absorbed right here.
What I like about this experiment, and is probably already obvious to you, is that it demonstrates how photons are not the "bullets" we imagine, just the interactions between field and electrons, as you say. If photons were "bullets", there would be 100% bunching all the time: source emits a single photon, which gets "split" (a troublesome concept for a single photon), and then each detector should detect that photon at the same time, or offset by a constant time difference based on the imprecision of placing the detectors. We don't see that, we just see statistically random photons, indicating that the photons are just independent interactions at each detector. It implies that there is no "quantum weirdness" behind the double-slit experiment. The photon isn't a "bullet" that somehow goes through both slits or has to choose a slit. The electromagnetic field goes through the slits, and the "photon" appears when that field interacts with matter. This is huge. It contradicts/disproves the Copenhagen interpretation. Even Feynman said that no one understood how to interpret the double-slit experiment, back in the day. Is this the current understanding of the physics community, now, or is it more of a personal model/interpretation that hasn't been widely published yet? I suspect it hasn't been widely published, otherwise the "quantum eraser" experiments wouldn't be quite so popular, as the "it's just an electromagnetic field with quantized interactions" interpretation foils any "reverse causality" interpretation of the quantum eraser. One final note on your results at the end: I suspect that the bunching you finally observe at the end of the video might be a statistical artifact of the limited time resolution, that if you had femtosecond resolution, the 10 ns bunching would disappear from the data.
Yes this is a personal interpretation. It does have it's weak points, such as energy conservation considerations on short time scales. But I think the main thing to take away is that there is no field quantization, just energy quantization in interaction. The latter follows directly from the electron being localized wave energy that can only occupy certain vibrational modes when confined. So to me all visible energy (matter, radiation) is basically borrowed from the vacuum. That is by the way also the only way QED can work only slightly different: here the solution is in proposing virtual photons, particles and a virtually endless reservoir of vacuum energy. As for the results at the end of the video, I'm not 100% convinced that this truly is the bump, but this the result of the analysis I got: if you just focus on exact overlap within the dataset, you find a correlation.
@@HuygensOptics Thanks for your reply! I very much favor your personal interpretation. I suspect that the main problem with convincing others is that there is a lot of extra math to work out so that it (at least mostly) agrees with QED, and maybe if it gets that far, experiments that resolve any remaining disagreements (that might disprove some aspects of QED).
I'm unsure how you think a single photon on a beamsplitter would display bunching. It should rather display antibunching as glossed over in the video. And, indeed, it would be troubling if a photon on a beamsplitter were to cause both detectors to be set off. That would mean there was energy at both detectors leading to energy conservation violation. Rather, what you see instead is a chance (depending on the reflection and transmission coefficients of the beamsplitter and the detectors efficiency) to measure the photon at one detector or the other. This means that there is NO chance to measure a coincidence between the detectors leading to a g^2(0) to be less than 1 (specifically 0 for a true single photon). Now, more specifically, this means that when @hyugensoptics is measuring these statistically random events it cannot be from a single photons. We know in this case that the source if photons is coming from a laser that has been attenuated to the single photon level (meaning that there is on average one photons worth of energy in the field). This means that the quantum state of the field generated by a laser is not a single photon even if it is attenuated. Rather the state is called a coherent state. And, if you go through the actual math of calculating the g^2 of the state, you would find that the g^2 would be exactly 1 as demonstrated by the video. So, at the moment, this experiment has not shown anything that is contrary to known quantum theory. In fact, if anything, it has proved many parts of it. If it's any consolation, I also really dislike when people make the assumption that a single photon and a weak coherent state are the same. Especially when trying to do science communication. But, there are good reasons to do this. Mainly, single photon sources are really hard to make and setup for demonstrations, then one also has to explain the distinction between a weak state and a true single photon, which can take a lot of time (eveb I've compressed this discussion a lot). But, I disagree with this practice just on general communication points.
Thanks for this amazing video. Yes, you're right I think, it's the low resolution ~5 ns that makes it hard to distinguish. In research labs, we use Time-Correlated Single Photon Counter, a specialised hardware that has a temporal resolution of few fs for these purposes. I used this device for fluorescence imaging of proteins. They are based on FPGAs. Anyone aware of any simple DIY-able design of TCSPC please let me know.
Are you sure the resolution was "few fs", and not e.g. some tens of picoseconds? The first would be equivalent to a bandwidth of ~100 THz, i.e. basically only achievable using optical means. Some tens of GHz of bandwidth are doable using modern time-domain electronics without too much work; 10 ps-level jitter should be within reach given state-of-the-art low-jitter photodetectors.
@@DavidNadlinger Yes, you're very right. I got the time units wrong. The device I used is a Swabian Instrument's time tagger ultra. The specs says that it can tag "photons" with a RMS jitter of 8 ps at best with simultaneous sampling of 10+ channels. A performance mode offers 3 ps jitter with 3/4 channels. I was just playing with this device, and I remember its python API offered histogram and coincidence measurements.
@@ParswaNathTheo Yes, I am also using Swabian Time Tagger with resolution of 8 ps. But still could not show bunching effect of an halogen lamp. Any ideas? Do I need match optical path lengths after beamsplitter up to microns ?
@krzysztofswitkowski5364 have you tried filtering the spatial, frequency, and polarization so you only have a single mode? For true thermal sources, the more modes you look at the less coherence you'll get. Also, you may need to do a cross correlation to find your delay between the two detectors. It doesn't have to be path delay, you can do this in post.
This explanation of the double slit experiment is so much more logical than quantum superposition and wave/particle duality. Have you ever experimented with nonlinear down-conversion crystals & polarizers with regard to Bell tests & entanglement? Would love to hear a reasonable explanation of what's actually going on there.
It seems to me the expectation of reasonable-ness of quantum processes is quite unreasonable, given that these things just don't behave in familiar ways. But they do behave predictably, based on established theory.
@@Matthew.Morycinski I'm ok with unfamiliar. But focusing exclusively on patterns and using those patterns to make predictions is to focus on "what". I'm also interested in the "why" and "how" which I feel I haven't heard good explanations for yet.
@@BryanTheYeti That's a sensible pursuit but watch out for occasions when answers acceptable to you are nowhere to be found. "The Universe is under no obligation to make sense to you."
Thanks Jeroen. An unrelated but naive question - did the reflection of the ND attenuator affect the laser operation at all - instability, noise, mode hopping etc? (guessing it was intentionally misaligned so not to not be normal to the optical axis?)
Yes indeed, I did place them in under a slight angle especially the first one which was OD=4 and highly reflective. The operation of this particular laser diode was remarkebly stable. I did try others, like cheap red diode lasers. But on short time scales these were quite unstable and displayed all kinds of unpredictable instabilities.
When a photon interaction occurs and this street energy transfer from the wave to the electron happens what happens to the wave itself? How is the knowledge of that energy transfer propagated in space? From the wave point of view is this interaction almost like an anti-light source, which has a wave that back propagates, or something else entirely?
@@youngbloodbear9662 Do you mean "charge field"? The charge of the electron propagates as a wave in the space medium that travels out spherically from the electron.
@ there is an electric field proportional to the probability of an electron reaction, when that interaction happens, how does the field that caused it react?
You might get back a bit of resolution by using kernel density estimation instead of histograms. You still have the nuisance parameter of kernel width, but there are good ways to pick one. It should be fairly easy to implement in your code too. You just have to generate gaussians centered at each of your time differences instead of adding them to buckets.
The way you understand fotons and their detections is the same way I make the sense of all of quantum mechanics. For me there are no particles, only waves and the quantum measurement is just energy/momentum exchange event between the waves that reshapes the waves.
thus way the essence of a wave itself remains above all others, leaving the photon (= electromagnetic) field and electrostatic field too in the area of "something inexpressible"... semiclassic attempt of description "wave is wave, photon is energy of interaction" was rejected tens years ago.
@@tapdapy There's nothing (semi-)classical about this. The math is exactly the same as the most up to date math of quantum mechanics. Photon is the complex probability wave evolving according to Schrodinger's equation, it's not ever anything else. Measurement is energy/momentum exchange event that reshapes the waves that take part. Electrostatic field (and magnetic field that is a consequence of special relativity applied to electrostatic interaction) is just the field that a photon is a wave of. Traditionally calculations in quantum mechanics done regarding measurement are done as if they were narrowing down measured object to a single point (or at least something smaller than the threshold of measurement apparatus) but that's just artifact of our capability that those are the values we are most interested in. You can collide very fuzzy electrons with very fuzzy electrons and measure the result afterwards and the quantum mechanics math can describe results as well. The only difference is in interpretation. Most mainstream one now is that at some point those fuzzy electrons must have experienced wave function collapse and then they collided. While mine just says that they experienced some energy/momentum exchange and got reshaped not necessarily to something point like. The whole idea of duality is silly and ancient. Elementary particles are just objects unlike anything we see in macro scale. They are not really particles ever but if they are tight together in very large numbers they tend to behave like one with specific momentum and location. Their wave-like math that describes their individual evolution in time is obscured and the Schrodinger's equation simplifies to Newtonian kinematics. However the way the energy/momentum exchange works between a pair of them works same in macro and microscale. That's why historically we believed that they might be a small version of macroscopic objects. But I think keeping this idea is harmful and it's actually this mainstream interpretation that postulates that sometimes elementary particles are actual pointlike particles is semi-classical and will go out of fashion at some point.
Is your setup thermally stable? Small path differences at 100nm level in time between the the photo multipliers will have an effect on your measurements. As you noted histograms are sensitive to the bin size. It might be useful to use the Cumulative distribution functions (CDF) of the coherent and non-coherent photon sources. I think of two options to compare these. One is to fit an analytical function to both and then look at the derivatives of these to look for peaks. The second one is plot both CDF's on the same graph and look for deviations between the two.
It is very annoying that QM is taught as wave-particle duality when it is very obvious that the underlying fundamental object is the field which evolves via the wavefunction. Particles are an invention, and past high school physics they do more harm than good at building an intuition. It took me years to understand this and it's very frustrating.
@2:11 (Still frame) Perfect image what i see without glasse with one only: astigmatism + some remaments (unknown source) on lens - without those color strpis obvuously but those mangling shapes inside... striking resemblance
One question about a beam splitter: In the process of sending/splitting a photon in two directions, does it duplicate the photon or is it the same photon that is later detected at two detection points at the same time? Also, does the beam splitter detect the photon as well (as being some kind of observer in the path of being observed by the detectors?). I just got here from watching the newest hologram video by 3Blue1Brown, and I only trust asking this question at Huygens Optics. Simpler question: Is the photon the same at detection point 1 and detection point 2, or two different ones (duplicated by the beam splitter)?
There were at least a dozen points in this presentation where I wanted to hit the like button. Unfortunately I already had.
Likes are quantized interactions of the appreciation-field ?
I think you left some lip marks on his ass lol
We need more like buttons for this channel 😆
@@BartSliggers perfection
I actually felt the same way before reading your comment. Every time I got mind-blown, I tried to hit the like button but failed.
19:50 Thinking of photons as discrete transactions is so insightful - the most understandable explanation of photons and interference patterns I have encountered in my 67 years!
Crazy that this caliber of content is available to us for free. Learned a lot - thanks.
You know what else is free? Parkour Civilization movies 1 and 2
You really demystified the wave particle duality in this one. The fact that they are probable interactions that represent a wave was the a ha moment for me.
This was a very coherent video.
14:54 and we saw a relatively *short* interaction
Fantastic video! I started watching this and as soon as I got to the astronomy part I rewound and re-watched it with my wife, who is preparing to perform this exact measurement at an observatory.
I am a new PhD student in biomedical imaging, and I’m currently taking a statistical optics course. Really enjoyed the video! We literally just derived the van Cittert-Zernike theorem in my class from the propagation of coherence.
You're the first person I've heard clearly articulate that the photon interactions are quantized, rather than the field everywhere. Would've been really helpful when I was first wrapping my head around this topic!
The Professor is so wonderful. To be invited to the show. Awesome
The real world (particles in the way of light, or our eyes) see it as quantised but the wave itself must be not since repeating the same single photon experiment brings different peaks everywhere that show that a wave must be present to have these interactions measured and not single photons. Although they say interaction with light collapses the wave (function) and then forth it looks like single particles. So the quantised interaction removes some energy out of the non localised wave? Until it's energy gone and dissappears from all potential points of interaction at once? It's hard for me still to grasp this faster than light wave collapse of light. Can Huygens maybe explain that simpler?
I could not understand the basic concepts of quantum electrodynamics before, but seeing little people slap each other finally helped me.
Thanks Huygens Optics!
that was a hard take - even though I personally agree, this is some shaaade thrown at particle physicists xD
A lightwave is a potential probabilistic slap into any electrons face.
Great video! I as I saw some one point out sodium lamps and spectral filtering, there's another way that has been used to create what I'll define as pseudothermal light at a specific wavelength: namely a rotating diffuser. This technique was shown I think in the 60s, but you can read about it more thoroughly in goodmans statistical optics. The main idea is to consider a single frequency laser and look in the far field to see the speckle: it should be stationary. If we add several frequencies, each with their own speckle, we see that the speckles will sort of add together and when you detect youll average over the various speckle configurations leading to what looks like a speckle free beam. Now, if you take a laser and put a diffuser in front, youll end up with a random speckle pattern in front due to the random phases imparted to different spatial points of the beam on the diffuser. As the diffuser turns, the speckle pattern will change as the pattern of hills and valleys change on the diffuser. By rotating, and averaging over all realizations of the diffuser, youll end up averaging over all of the speckle patterns and thus something similar to the LED. I call this pseudothermal because the light is not coming from a blackbody source. This is easily implemented with a cheap RC car motor and some plastic (scotch tape works pretty well usually, but it depends on the wavelength of light and how big the small deviations in the surface). Ive used this (well using a random piece of plastic) in the lab to get g^2(0)=2 within experimental error.
Outside if this, I want to point out a couple of other things.
I think you glossed over the Hanbury-Brown-Twiss conteoversy too fast. There's a ton of interesting history and actually this discovery lead to what Ill call the second revolution in quantum optics. It took a painful derivation by Roy Glauber to show that indeed the g^2(x,t) is related to the first order coherence (the g^1(x,t)). The formulation to show this lead directly to the second revitalization of quantum optics.
Some other stuff that might be fun to try. First, you should try doing homodyne tomography of different quantum states of light. Effectively, this is taking your weak quantum state of light (thermal, coherent state, single photon) and interfering it at a beamsplitter with a strong laser. By subtracting the currents in either arm (called balancing), you can get access to reconstructing the quadrature state of the light. Most of these will be gaussian unless you can get a nonclassical light source.
On this note, you should also do the famous Hong-Ou-Mandel interference experiment that actually (and directly) shows the "particle" nature of light. In this case, take your beamsplitter and interfere two photons at the input ports (one at each), if the photons are truely identical (same polarization, spatial mode, and frequency distribution) then it turns out that the photons will always leave at the same port and never at different ports. Then as you change the states to become more distinguishable by, e.g., delaying one photon by some time, as the mode overlap between the states decreases, the coincidences will begin to increase. Specifically, if you have detectors at the outputs, then youll never see coincidences. Further, this cannot be due to a simple wavelike interference effect because it isnt phase dependent. You do not have an envelope with oscillations under it in this effect. If you did the same thing with coherent states (i.e. laser light), there is an oscillation effect that is phase dependent. Thermal light, also is phase independent, but the dip in coincidences will be smaller.
Why nobody replied to this comment?
Wow.
@@ignastaraskevicius6697 It is too deep.
A like just to later read more about this.
I've studied the subject for years and watched countless lectures but never encountered it explained in this manner. It's intriguing
i love the explanation that photons are the interactions with the em field.
this is so much easier to think about compared to the way i am taught quantum physics in school
thank you!
I showed several coworkers your video on the photo-multiplier. We work as construction inspectors. We sometimes use a Nuclear Density Gauge. These devices use a photo-multiplier. It was really cool to see the way a device works that is very difficult to understand.
AT 1:47.... finally able to comprehend what wave function, oscilloscopes and wave diagrams, the wave itself; and, what they represent after watching hours and hours of videos discussing electrical engineering or electronics, photons and quantum effects, the science of sound, etc. etc. So, thank you for finally bringing a clear picture and understanding into my knowledge base!
The Hanbury Brown and Twiss effect seemed so wild to me when I first heard about it. But you're right that you can't argue with experimental results. I also appreciate that you were able to do all this stuff in software. So much more pleasant than doing this with a crate of analog electronics and a mess of coax cables.
Amazing video as always! I've been doing g2 measurements in our lab for a few years now, albeit with single photon sources and ~$300k worth of detectors and timing electronics. It is unbelievable how much youve accomplished with just a scope and some PMTs!
I have a recommendations that might help you resolve photon bunching. You could replace the broadband LED with an atomic spectral line. If you have a sodium or neon discharge, you could use a diffraction grating to send a sinlge emission line to the input of your HBT setup. Most of those spectral emission lines are very narrow (sub-GHz), which corresponds to coherence times in nanoseconds. Ive used this technique to measure argon spectral lines in the near-IR to calibrate a spectrometer and it works very well.
The next logical step of course is to construct a spontaneous parameteic down-conversion based single photon source and characterize that 😋! You could even do Hong-Ou-Mandel interference, the wackiest measurement ever!
Cannot wait for your next upload!
Yes! I'm 30 seconds in and already intrigued.
As a hobby, I'm working on re-framing many basic elements of physics modelling using a finite-axiom approach (including the photon) and so far it's been interesting, enlightening, great fun, and often points to quite a different set of models as commonly used in 'standard' physics. It's interesting to see how you rationalise the photon in the standard system of models - importantly, for me, it is the measurements that counts and I greatly enjoy the way that you focus on the measurements with such great detail. Ultimately any model has to be able to reflect and predict experimental measurements. It seems to me you are searching for understanding. I would simply add caution to both the acceptance and understanding of current models and build a confidence in your own ideas. You are one of the few people I would like to share my initial ideas with (and real results) with. Mostly my ideas are likely to go nowhere or be simply wrong and will be likely laughed at, but as I progress It's been such fun. However, if the results go well I hope to publish the work - even if it's just to show how not to do it - LOL. Anyway thank you for showing the results of the experiments - especially as it means I can get a strong feel and sense of the measurements and the process etc. This has really helped me. Thank you for sharing your work.
A major component of the pulse height variations from photomultipliers is due to the variation in the number of secondary electrons generated at the first dynode. This variability is described by the Poisson distribution, which says that if n_ave is the average number of secondary electron ejections from a dynode, then the probability of n secondary electrons being ejected for a given incident primary electron is (1/n!) * (n_ave)^n * exp(-n_ave). This distribution is broad when n_ave is small, but becomes narrower the larger n_ave becomes.
This source of variability in the pulse heights can be reduced by increasing the voltage drop between the photo-cathode and the first dynode, which increases the secondary electron yield.
But increasing delta V between cathode and 1st Dynode also increases the noise.
In the 1980's I worked on the detection of Cerenkov radiation from ultra low radiation sources with a single photon emission rate of 10 to 20 counts per minute and photomultipliers specifically selected for ultra low dark counts of < 1 count per minute (when cooled to -50 degC and using discrimination electronics - see later).
The multipliers NEVER saw any day light and were kept in the dark even when powered off. We had several older pmt's which had increased dark counts (because of a loss of ultra high vacuum in the tube and with these we were able to demonstrate that a single brief exposure to room light (neon tubes) caused some residual fluorescence / phosphoresence response in the cathode but room light from tungsten filament lamps showed a lower effect on the dark count.
I spent endless hours in the darkest lab with walls painted black, every surface covered with black cloth. The experimental setup was contained in a thick black canvas curtain covered box (for access to the samples). Once the canvas was washed and the experiment went to hell, because of the washing power used to clean the canvas contained a whitener i.e. a flourophore.
In terms of pulse counting, we used threshold discriminators to eliminate low energy signals stemming from dark counts which typically had only a signal intensity of 1/10th of a single photon event. Since Cerenkov radiation (from beta decay of 32P) is a truly random event we were able to use statistical analysis to get to the real single photon count rates and eliminate any spurious noise. The other 'nice' aspect of working with a radioactive substance is that it has a natural decay rate (half life) and thus we were able to use measurement over several half life times to extrapolate back to zero and eliminate the last remnants of dark counts. I am still proud of these measurements conducted as a diploma student for my dissertation.
@@robertrjm8115 Was there a published paper that is online? Sounds fascinating. I saw the brighteners becoming an issue as soon as you said washed. Perhaps dry-cleaning would have left less residue.
Scary how normal things like opening a box can ruin a sensitive detector.
@@KallePihlajasaari Hi Kalle, I am sorry, but NO there has been no publication on the issue of brightners throwing a spanner into our work. Also, this was in the early 1980's and very few publications of that era have been digitised. Also, our work was medically related (diagnostics of malignant melanoma of the eye by a 32P uptake test) and although our work did succeed in detecting melanomas the issue was that it required the injection of a fairly high dose of a radioactive substance and the potential side effects were considered to be too severe to be acceptable as a routine diagnostic tool.
This is so mathematically logical and practically logical and useeful at the same time it deserves an investigation of voltage drop to average secondary electron yield depending on the dynode material by itself.
I'm not a physicist, and I've never "truly" understood the wave/particle duality, but I feel absolutely cheated that it took until now hearing about looking at the photons as interactions. That's way more intuitive, and my understanding about several concepts have changed significantly from that single insight
Thank you so much for this video! I just started working with a single photon source research team as an undergraduate and I was having trouble wrapping my head around the g(2) measurement - it makes a lot more sense now
Thank you very much. This has answered several long standing questions I have had about how interferometry is used in astronomy. I don't think I really understand it yet, but I am much closer than before. I have asked questions on several relevant videos and never got an answer that usefully answered any of the issues I had.
This is fantastic- I had a student working on approximately the same setup to measure g2 with similar PMTs. I think what could lead to better data is using coincidence measurement for longer accumulation times (specifically, triggering coincidence events every time two pulses overlap with a specified time window). We originally tried this with NIM modules. Remember, HBT claimed better timing resolution for their tabletop experiment than their detractors because of the care they took in the experiments and the coincidence techniques they used.
For the delay, if the data were collected as pure coincidences (and not as single counts from both detectors), the experiment could be performed in the same manner as the original HBT demonstration by translating one of the detector pairs relative to the other. The chances of measuring coincidences will be greatest when the two detectors are equidistant from the beamsplitter, and this probability should drop as the delay is adjusted. For 5 fs of temporal resolution, you would need 1.5 microns of spatial translation, which is possible with a flexure stage or PZT actuator.
There might be a misunderstanding here; the adjustment in the Hanbury Brown and Twiss demonstration was orthogonal to the beams (to sample the same or different spatial modes), rather than along the beams. The longitudinal adjustment, i.e. distance to the beam splitter, does not matter (if there is no dispersion in the medium). All it does is to shift the arrival times, but this can also be done digitally in the data analysis after the fact. A shift vanishingly small compared to the temporal resolution of the single-photon detectors (as 5 fs would be) does not matter.
@@DavidNadlinger It seems that you have some experience in the topic. Since one week I wanted to show bunching effect using halogen lamp and two SPCM Thorlabs single photon detectors. The halogen lamp is fiber coupled and output is collimated . I am using PBS in the experiment. I am using timetager and my histogram is always flat without any peak in g(0). Do you have an idea what I am doing wrong. Do I need use some pinholes to get point source. Is temporal coherence matters here?
@DavidNadlinger Fair point and I definitely missed it! Reread the first page of the paper, the distance could vary by a cm either way. Thank you for your correction!
I still think that the small collection window compared to the coincidence rate could be the reason it was so difficult to see the correlation here. HBT collected coincidence data over the course of an hour for their source/detector combo, any comparable length data set using 10 ms intervals would take 360,000 samples assuming a similar true coincidence rate.
i'll come back to this properly after escaping the ADHD well i'm currently stuck in (but comments, engagement, etc.) 😅 good pointer to sigview, this looks very useful
your experiments continue to be among the most enlightening and approachable explorations of wave mechanics i've come across. i particularly appreciate the practical tips and "dead ends" you sprinkle around, showing your thought process, and contrasting positive/negative examples worth pondering
This video definitly shed some light on this topic in a very coherent way! enjoyed it very much as random chemist
I'm familiar with the bunching effect derived from the wavefunction symmetry constraint. But I can't believe I hadn't heard of the Hanbury Brown & Quintin effect! Thanks for sharing!
Very interesting experiment!
I agree that temporal resolution is key, because within 10 ns, plenty of oscillations and fluctuations in the density function occur. So whether two waves are synchronous or not doesn't make much of a difference...
Btw, people like you who say "the math is complicated but the essence is simple" are both good scientists and good educators.
I enjoy this channel because you’ve landed on a similar intuitive interpretation of quantised photons that I have. While I can see the flaws in it (no probabilistic element) I love the slap analogy!
You have an enviable approach to explaining things and presenting your work, which is beautifully done and very nicely illustrated. I am delighted to hear someone talking sense when it comes to quantum electrodynamics. The term ‘photon’ has become so abused that I meet few scientists who use it properly. Many physicists also utterly fail to make their ideas understandable by using fancy terminology and complex maths. Nature is quite forgiving in this way though, which is why the misconception persists. I was relieved to see you use the highest time resolution to give you the chance of observing the weak temporal correlations of fluctuations in the incoherent light (I was wondering what your sample rate was from the outset). Not sure about the huge amplitude fluctuations of the pulses from your photomultiplier. I guess I need to watch your other video to understand that…… Signing off, I hope you continue to enjoy making these fabulous videos. I thank you by showing a short clip of me at work in my ultrafast IR spectroscopy lab and analysing data. My strange little video (filmed during the pandemic) is intended as a snapshot showing a couple of minutes of working days that can last more than 14 hours doing research trying to figure basic things out about light interactions with molecules.
th-cam.com/video/Vy0mSO7euFA/w-d-xo.htmlsi=0TyWBYky8ycg5l3-
19:30 waaaaaaaait.... exactly this is what i think about all the "photon" theory. The EM field itself is not quantized, just the energy transfer! This needs much more attention and should be published as research.
So about this.. the part where quantization will be necessary is in the third part (anti bunching) which was not discussed in this video. I think the goal of this video was to show that many phenomena, which usually get attributed to effects in quantum physics, are not as non-classical as one thinks, and in fact also can have a nice classical or at least semi classical view - which I enjoy very much.
That at some point classical theories fail is consensus in the physics community and there are enough experiments to indicate that. But all those quantum optics effects are very hard to understand since they involve quite in depth mathematical understanding.
only problem you cant send more energy to electron by raising the amplitude of light wave, bases on that peoples agree that there is something in light, not only in electrons.
like, you need different wave length to transfer more energy and move electron in higher position, but you cant do it by increasing the amplitude of some infra red wave length.
but... maybe it is explained by core of the atom is waving with the electron on low frequencies. idk.
@slimeball3209 I guess the main problem is that most people view the electron and the EM field as 2 separate things, whereas they are not: The electron's charge is an integral part of the total EM field and therefore any energy transactions to the electron involve the field as a whole. In a sense, the electrons basically just add some local impedence to the field.
The other question you might ask is: where does the total photon energy come from if it is not contained locally in the field? I tend not to worry too much about, since as mentioned above, it is a transaction with the field as a whole. Apparently quantum physicist are not bothered by that question either, because QED assumes spontaneous creation of virtual particles and photons. On top of that, an energy content of 10^120 Joules per cubic meter of vacuum is needed to make the theory work.
@@HuygensOptics no, i would not ask this question, i imagine photons as delayed movement of electron, so it includes polarisation, but still dont understand what is happening when we impact the atom with different frequencies of field, and why is need higher frequency to make electron jump higher, but we cant use amplitude of lower frequency to make this electron jump at the same "height"
like, if 1 electron goes up, so in some time other electron will see and also go up, some electrons are locked to surface of material, so they moving in projection, and at some angles we will see it as polarisation, especially at high angles, because in that case we have two projections that flattens the initial movement down to 1 direction.
it is all about waves still, but i dont understand what is going on in electro chemical side. i questioning the mechanic of breaking the electron energy level barrier, and how the high frequencies can do it, but lower cant, even with 10x amplitude and the same ascent of higher frequency.
and i too far from exact math of this phenomenons, so i cant even make valuable question about total energy of interaction ect.
@@slimeball3209 Because an electron is also a wave, a standing wave, trapped around the nucleus of an atom. To jump higher, a standing wave must increase its frequency, and only light of a higher frequency can excite it to do this. Light of a lower frequency will simply pass through the atom as if it were transparent.
Here's an example, not very perfect. A ball is bouncing on a surface. Assuming there is no friction, the ball will continue to bounce to the same height. If the surface moves up and down slowly, regardless of the amplitude, the ball will continue to bounce to the same height. But if the surface starts bouncing up and down at a higher frequency, it can transfer its energy to the ball, causing it to bounce higher. Again, this example is not perfect, perhaps you can come up with a better one.
I love this video, thank you for sharing the results of your hardwork, and for sharing your frustration with the language qm physics uses to describe the concepts being worked with - paricles instead of energy exchanged.
I hope I have the knowledge and clarity of mind you have they day I reach your age. I'm overwhelming impressed.
This video (and many of your videos) amazes me with the depth of complexity in concepts can be presented while still being understandable. There are few sources of information that make me feel both enlightened and overwhelmed at the same time the way your videos do.
I know this would be a departure from the things you are working on, but I would really love to see an explanation of polarization including terminology, filtering, detection, and how the concepts relate to the wave propagation, refraction, reflection, and diffraction. I have never seen any presentations that talk about polarization concepts in a field of waves.... it's usually only diagrammed in terms of a single wave 'ray' I think.
Also, your mixed in 'bye bye' at the ending got me laughing. Thank you for that.
Looking into polarization and single photons is definitely on my list. Unfortunately, I'm not sure if I can make a video about the subject any time soon due to a lack of time.
Single photon emission and anti bunching is the logical next step! You just mentioned them in passing. The explanation of the interactions being quantized, not the field, made some things clear, but now I'm no longer sure about my understanding of actual single photon events (where there is only one photon emitted) like from nano dots or nanoparticles.
I can always learn a lot from your video.
great design from the experiment to theory and to the analysis.
The presentation is also wonderful👍
Oh rad! Another video of yours. I always end up going though some old quantum & optics notes, because there's always something that suddenly starts making sense... "the probability of discrete energy transactions is governed by a field intensity distribution" klicked for me.
Anyway. Thanks for this amazing content!
Great video as always. Thanks for sharing.
.
.
.
.
.
.
.
.
I continue to assert that light possesses mass. If light is interpreted merely as interactions (energy transfers), it implies that it must collide with something to provoke a chain reaction of energy transfer until it reaches a point of absorption. This behavior suggests that, like ordinary matter, light could also exhibit intrinsic mass properties. Furthermore, I believe there is matter in the vacuum of space, even if we do not perceive it directly. The crucial difference is that our vibrational frequency does not coincide with what is necessary to physically detect this matter, just as we sense water or earth.
Analogous to the limitation of sound propagation in space due to the absence of a suitable medium, light, in principle, should also have limitations imposed by barriers, fields, or specific zones of influence, even if space appears to be empty. When we consider electromagnetic waves, we know that, for energy propagation, there is traditionally an assumption of the need for a medium (which historically led to the hypothesis of 'luminiferous ether,' later refuted). However, we may not yet possess the necessary technology to adequately detect the media or fields that could mediate these energetic interactions.
pretty astute to use the plastic box as freezer compatible by keeping the setup small enough... a lot more efficient and manageable the Peltier modules, and more confortable than working in Siberia.
I suppose the "classical" approximation is convenient as long as one do not look into the atomic level (requiring enough precision to know where the interaction occurs), so that one is reduced to look into the phenomenon statistically (over a large number of occurences, even if temporally of discrete events) rather than studying each single interaction along the path from emission to detection (at the atomic level, including rarer events like absorption by beamsplitter etc. which are discarded by the classical observer).
Thank you for your continuing enthusiastic work, brings back memories, keep it up!
Thanks for another impressive video and the extensive effort in its setups! Bravo!
the animations get better with every video. love it.
I really enjoyed the real time experiment showcase. It was indeed very "elucidating". Lol.
one of the distinct qualities of photon interactions that result in them being described as being also particle like, in comparison to how you describe it, is the implications that quantized energy transfer has for conservation of energy.
If we think of three hydrogen atoms floating in space, and consider one of them to have an electron in an energy state above its lowest potential:
when it drops down from its higher orbital state it releases a photon. If the photon were entirely described by only wavelike mechanics from the moment the electron is done transitioning to a lower energy state, then the waveform defining it would be, in the simplest case, some isotropic, spherical wavefront that propagates out equally in all directions.
However if that was the case then it becomes highly unintuitive how we could be confident that only *one* of the other hydrogen atoms could have an interaction with the wave, resulting in that hydrogen's atom transitioning to a higher energy state. This needs to happen otherwise the single emission of the first atom could potentially result in two equally energetic adsorptions, resulting in the breaking of conservation of energy.
In this thought experiment it becomes far easier to explain the behavior of the interactions in the form of a particle, as a particle of light would inherently have the needed properties that maintain conservation of energy.
I would love to hear your thoughts on this issue, I am definitely not an authority on this topic and perhaps theres something I am missing, but this really seems like a very hairy problem that gets at the core of the unintuitive nature of wave/particle duality.
I don't think that conservation of energy is really a thing on short time scales. It's not an issue in QM either given the uncertainty relationship between time and energy and the fact that it only works by assuming virtual particles and photons popping out of nowhere. The way I look at this is that basically all energy visible to us (matter EM radiation) is borrowed from the vacuum. And so an energy quantum absorbed is in the background a transaction with the vacuum energy. The vacuum will settle the total energy bill later. In a sense, it's the EM wave (which by the way can be strongly localized both in time and space) that sets the transaction in motion, and not so much what supplies the actual energy locally to an electron.
*I would love to see a Video about how to choose the right optical elements.*
So when to use plano-convex, plano-concave, convex-concave, concave-concave, convex-convex -lenses and also other types like Spherical, Aspheric, Achromatic or Cylindrical -lenses and Fresnel-lenses.
If you can get your hands on a TAC and MCA (time-to-amplitude converter; multi-channel analyzer), you could improve time resolution by 2-3 orders of magnitude. It would require minimal modification to your existing setup and give you proper TCSPC ability (time-correlated single photon counting).
So... the Hanbury Brown & Twiss effect requires spatial coherence (in the area where the two telescopes are), but no temporal coherence?
If your source is temporally coherent, how would you detect temporal decoherence as a function of telescope separation?
imagining photons as particles is like imagining of shipwrecks roaming around ocean waiting for a ship to hit
Very very fine video and experiment. Way beyond me but I loved every second. But please update windows you had me clicking that icon
A cost-effective way to improve the temporal resolution here might be to use an inexpensive direct-conversion SDR instead of an oscilloscope, to perform the binning much closer to hardware.
The core principle of Direct Conversion Software Defined Radio is the idea that, given a signal that only includes frequency content within a given bandwidth of some center frequency, you can fully reconstruct that signal from pairs of 'quadrature' samples, i.e. taken a quarter-wave apart, sampled at the bandwidth frequency -- you don't actually need a _continuous_ nyquist-approved sampling rate unless the lower frequency limit goes all the way to DC.
Even an inexpensive SDR like the HackRF One can set a center frequency anywhere from 1 MHz to 6 GHz, meaning a quadrature sampling time that can be swept anywhere from 42 ps to 250ns.
The setup I'd be tempted to build here: Get a pair of HackRF One SDR's, set up synchronized with a common refclk. Build a trigger pulse generator circuit to condition the PMT's output, and use the SDRs to capture that output. From computer control, set the SDRs to produce synchronized pairs of I/Q sample pairs, setting the 'operating frequency' to the frequency with a quarter-wave equivalent to the tau value you want to measure g(2)(tau). Then count the rate of sample pair pairs where one tube has a detection in it's samples I value and the other only in its sample's Q value.
I suspect you'd need to do some additional analog signal processing on the PMT tubes first -- mostly, to clip the lengths of those pulse tails, though possibly you might want to also modify the SDRs to remove frequency filtering components from the signal path. If you're interested in this approach though I'd be happy to share more details -- but this is already super long for a TH-cam comment as-is lol.
Thanks for this suggestion, I will definitely look into this method. I can also measure the signal directly on the anode over 50 ohm, which drastically reduces the pulse width.
You said that g(2) was the second order correlation function. In what sense is it second order? What would be first or zeroth order?
I greatly enjoyed the video and learned a lot. Thank you!
Incredible, learned so much from this. Thank you.
Just wow. Thank you for making such great videos!
6:10 you mention that you use _reflective_ attenuators.
1) Might these interfere with your experiment?
2) Also, if you stack a number of reflective filters back-to-back (or even combined with a laser-cavity) will they influence the total attenuation and/or operation of the laser? (or are they mounted at an angle, to dump the reflected light into a black surface off to one side?
It doesn't look like this, to my (inexperienced) eye.
Thanks (again) for a wonderful explanation of difficult topic!
It's very nice to see experiments like these popularized in this way. Regarding whether em fields are made out of photons. I would say that, ultimately physics is about models. If the photon model doesn't help in this case, you don't have to use it. As a researcher in quantum optics, the situations in which the photon picture is really necessary are rare.
need to be watched twice
An excellent presentation of a complex subject.
However I am wondering if some of the bunching effects could be a result of the white light LED. I know the bunching is real but the LED source used here could have an effect on the level of the observed bunching.
I assume this was a blue LED covered in a flourophore with a broad 'white' emission spectrum. There are several processes in the flourophore runnning in parallel to create the white light emission spectrum, some of these could create down stimulation from an excited state and some could be due to multi photon interaction in the flouresence material.
To eliminate these effects it might be better to use a pure thermal emitter such as a tungsten filament lamp which as a near perfect random emission rate which can then be cut down to the appropriate single photon level.
Also, maybe cutting down on the spectral band with of the white light (maybe 50 nm wide ?) would make the to results between the coherent green laser and the non coherent source more comparable.
I have sent extracts from your videos to my former colleagues and they have used them in some of their lectures - and of course fully acknowledged you as the source and told the student to visit you channel.
Very interesting videos. Very helpfull to understand the PMT signal processing. From your point of view how do you understand the mysterious correlations between the detections from Alice and Bob in the Bell EPR experiments?
Do you plan to make the Bell experiments with entangled polarized photon sources?
It would be very fascinating. I would be curious about the correlation function outputs...
Is it possible for a light beam to be collimated, but not coherent? (Ex: a straight beam of light, constant cross-sectional area, but having thermal frequency distribution)
Temporally incoherent is of course possible in a collimated beam, spatially incoherent is infeasible because having different frequencies in the same plane would make the waves change direction and spread out immediately.
Just a brilliant video. As previously said by other commenters, it it would have been possible to press like twice or even more, I would really want to do it. Instead, I will just comment it. Treat it as a triple like. Thanks!
Your outro song made my evening
Awesome video. I'm glad there is someone else on youtube demystifying the quantum woo.
Probability and statistics are CENTRAL to quantum theory, and modern physicists are arguably using this fact to confuse even themselves, knowingly or (most commonly) unknowingly.
You might have heard about a central result of QM, Bell's theorem. I've got a couple videos on demystifing that as well if you're interested. You might get some inspiration to reproduce Aspect et al. results using optical instruments.
I never understood the bell setup due to all this mystifying crap to a point where i could say it makes sense. Could you share the video to general population?
@lucelxebinog it's on my channel. The experimental setup is very easy to understand. I use a coin toss analogy for those unfamiliar with the quantum mechanical formalism, but I also go in depth as to why ultimately Bell's theorem does not provide a complete description of the setup and why it fails to prove what everyone says it does. There is also a video that goes in depth on the history of hidden variable theories that also explains what the formalism means. Hope that's understandable to you, if you're left with any doubts feel free to comment under the videos
Masterfully done. I would be keen to see the anti-bunching measurements using HBT. I reckon you could get your hands on a suitable nitrogen vacancy diamond sample.
It sounded like you were collecting data on a star using your own telescope. If so, I'd like to suggest that if you have a full spectrum modified camera, or any camera with no uv-ir blocking filter to try again capturing data in as far into the infrared as possible. Maybe longer wavelengths could help? Fascinating video. Thank you!
13:00 well, photons are localized, but to varying degrees. What exactly a photon looks like *depends on your system*, which is something I think that even most practicing physicists aren't really aware of.
I was just looking at that same wikipedia page last week and shaking my head lol
Then again, I suppose it does depend on your definition of "particle".
Single photon detections out of a PMT are approximately gaussian. If you look at Cherenkov light, where you can get something like 6 photons on average, then you get a Poisson dist. of number-of-photons convolved with the gaussian PMT response (iirc, then the N photons is a gaussian w/ variance ~N)...that results in a Polya distribution..which is so unwieldy idk if it helps.
Dang, what a great video.
Fascinating !
Its curious how reflection and refraction is so difficult to explain with quantum mechanics, while interaction between light and matter is still there (surely there's still at least momentum exchange when light hits the atoms?)
Great video, I learned a lot!
A few Questions:
How many wavelengths is single photon worth of energy? How long does it take to be absorbed? What happens to the electron over the timeframe that it is being absorbed?
When there is a two peak overlap (simultaneous detection at both detectors) doesn't that mean more than one photons are at play?
If your definition of a photon is the exchange of an amount of energy, then yes.
This is so good I subscribed.
the intensity distribution at min 19 is surprisingly similar to an ofdm spectrograph
Modern day newton, maxwell, tesla! I get incredible amounts of enjoyment from these videos its hard to put into words!
At 30::42 you mention or display 5 femtosecons.. That is the key. Nano versus Femto. You need more money. I love it.
So tl;dr is, that likelyhood of getting multiple detections at the same time is higher with temporarily incoherent light, because the field has more energy on average in specific time frames?
Also using this to measure star diameter is pure genius.
You have some of the easiest to digest videos on their subjects I've seen and as someone only very casually into optics, I really appreciate this and all of the content you've made so far. You rock.
Also, what do you use to generate your animations of light sources and interference patterns?
"All wave animations in this video were produced using a Python script supplied by @DiffractionLimited . Thank you very much Manuel for supplying me with this tool."
Ik ga binnenkort een AAS slopen.
Heb je belangstelling voor de optische bank.
Grote Czery-Turner monochromator met een PMT?
Eventueel kan ik ook wel enige holle cathode lampen missen.
Never heard of photon bunching before
I was over my head when you were just* grinding lenses. But I still love your videos. So, not just an elaborate way to justify getting a much, much faster scope? (5fs)
wonderful
What if I say that: The electric field of light is best described as a statistical quantity extracted from many photon detections. It does not make sense to question "what is the electric field of a single photon?"
Something that confuses me: suppose a photon is emitted in one place and absorbed in another. After it's emitted, it creates an EM wave in 3d and therefore there's a probability distribution of where it could be detected. But after the photon is absorbed, does the entire 3d wave instantly disappear everywhere in 3-space? It has to, otherwise there's a nonzero chance the same photon can be absorbed twice, which sounds like it would violate conservation of energy (but making a wave instantly disappear feels physically impossible too...)
This is one of the reasons why QM and GR don't go well together: You either violate energy conservation or the maximum speed of information. I choose to "violate" the first law, because at very short time scales, energy conservation is not really a factor.
@@HuygensOptics Thanks for answering! I'm not sure I really understand, but this gives me some more context to help look stuff up
Your point is exactly what Einstein was expressing in his "spooky action at a distance" statement. How does the photon that is "everywhere" get absorbed right here.
Hanbury Brown & Twiss represented me in my mesothelioma case
What I like about this experiment, and is probably already obvious to you, is that it demonstrates how photons are not the "bullets" we imagine, just the interactions between field and electrons, as you say. If photons were "bullets", there would be 100% bunching all the time: source emits a single photon, which gets "split" (a troublesome concept for a single photon), and then each detector should detect that photon at the same time, or offset by a constant time difference based on the imprecision of placing the detectors. We don't see that, we just see statistically random photons, indicating that the photons are just independent interactions at each detector.
It implies that there is no "quantum weirdness" behind the double-slit experiment. The photon isn't a "bullet" that somehow goes through both slits or has to choose a slit. The electromagnetic field goes through the slits, and the "photon" appears when that field interacts with matter. This is huge. It contradicts/disproves the Copenhagen interpretation. Even Feynman said that no one understood how to interpret the double-slit experiment, back in the day. Is this the current understanding of the physics community, now, or is it more of a personal model/interpretation that hasn't been widely published yet? I suspect it hasn't been widely published, otherwise the "quantum eraser" experiments wouldn't be quite so popular, as the "it's just an electromagnetic field with quantized interactions" interpretation foils any "reverse causality" interpretation of the quantum eraser.
One final note on your results at the end: I suspect that the bunching you finally observe at the end of the video might be a statistical artifact of the limited time resolution, that if you had femtosecond resolution, the 10 ns bunching would disappear from the data.
Yes this is a personal interpretation. It does have it's weak points, such as energy conservation considerations on short time scales. But I think the main thing to take away is that there is no field quantization, just energy quantization in interaction. The latter follows directly from the electron being localized wave energy that can only occupy certain vibrational modes when confined. So to me all visible energy (matter, radiation) is basically borrowed from the vacuum. That is by the way also the only way QED can work only slightly different: here the solution is in proposing virtual photons, particles and a virtually endless reservoir of vacuum energy.
As for the results at the end of the video, I'm not 100% convinced that this truly is the bump, but this the result of the analysis I got: if you just focus on exact overlap within the dataset, you find a correlation.
@@HuygensOptics Thanks for your reply! I very much favor your personal interpretation. I suspect that the main problem with convincing others is that there is a lot of extra math to work out so that it (at least mostly) agrees with QED, and maybe if it gets that far, experiments that resolve any remaining disagreements (that might disprove some aspects of QED).
I'm unsure how you think a single photon on a beamsplitter would display bunching. It should rather display antibunching as glossed over in the video. And, indeed, it would be troubling if a photon on a beamsplitter were to cause both detectors to be set off. That would mean there was energy at both detectors leading to energy conservation violation. Rather, what you see instead is a chance (depending on the reflection and transmission coefficients of the beamsplitter and the detectors efficiency) to measure the photon at one detector or the other. This means that there is NO chance to measure a coincidence between the detectors leading to a g^2(0) to be less than 1 (specifically 0 for a true single photon). Now, more specifically, this means that when @hyugensoptics is measuring these statistically random events it cannot be from a single photons. We know in this case that the source if photons is coming from a laser that has been attenuated to the single photon level (meaning that there is on average one photons worth of energy in the field). This means that the quantum state of the field generated by a laser is not a single photon even if it is attenuated. Rather the state is called a coherent state. And, if you go through the actual math of calculating the g^2 of the state, you would find that the g^2 would be exactly 1 as demonstrated by the video. So, at the moment, this experiment has not shown anything that is contrary to known quantum theory. In fact, if anything, it has proved many parts of it.
If it's any consolation, I also really dislike when people make the assumption that a single photon and a weak coherent state are the same. Especially when trying to do science communication. But, there are good reasons to do this. Mainly, single photon sources are really hard to make and setup for demonstrations, then one also has to explain the distinction between a weak state and a true single photon, which can take a lot of time (eveb I've compressed this discussion a lot). But, I disagree with this practice just on general communication points.
Moral of the story: select the right data and present it in the right way.
Thanks for this amazing video. Yes, you're right I think, it's the low resolution ~5 ns that makes it hard to distinguish. In research labs, we use Time-Correlated Single Photon Counter, a specialised hardware that has a temporal resolution of few fs for these purposes. I used this device for fluorescence imaging of proteins. They are based on FPGAs. Anyone aware of any simple DIY-able design of TCSPC please let me know.
Are you sure the resolution was "few fs", and not e.g. some tens of picoseconds? The first would be equivalent to a bandwidth of ~100 THz, i.e. basically only achievable using optical means. Some tens of GHz of bandwidth are doable using modern time-domain electronics without too much work; 10 ps-level jitter should be within reach given state-of-the-art low-jitter photodetectors.
@@DavidNadlinger Yes, you're very right. I got the time units wrong. The device I used is a Swabian Instrument's time tagger ultra. The specs says that it can tag "photons" with a RMS jitter of 8 ps at best with simultaneous sampling of 10+ channels. A performance mode offers 3 ps jitter with 3/4 channels. I was just playing with this device, and I remember its python API offered histogram and coincidence measurements.
@@ParswaNathTheo Yes, I am also using Swabian Time Tagger with resolution of 8 ps. But still could not show bunching effect of an halogen lamp. Any ideas? Do I need match optical path lengths after beamsplitter up to microns ?
@krzysztofswitkowski5364 have you tried filtering the spatial, frequency, and polarization so you only have a single mode? For true thermal sources, the more modes you look at the less coherence you'll get.
Also, you may need to do a cross correlation to find your delay between the two detectors. It doesn't have to be path delay, you can do this in post.
This explanation of the double slit experiment is so much more logical than quantum superposition and wave/particle duality. Have you ever experimented with nonlinear down-conversion crystals & polarizers with regard to Bell tests & entanglement? Would love to hear a reasonable explanation of what's actually going on there.
It seems to me the expectation of reasonable-ness of quantum processes is quite unreasonable, given that these things just don't behave in familiar ways. But they do behave predictably, based on established theory.
@@Matthew.Morycinski I'm ok with unfamiliar. But focusing exclusively on patterns and using those patterns to make predictions is to focus on "what". I'm also interested in the "why" and "how" which I feel I haven't heard good explanations for yet.
@@BryanTheYeti That's a sensible pursuit but watch out for occasions when answers acceptable to you are nowhere to be found. "The Universe is under no obligation to make sense to you."
Thanks Jeroen. An unrelated but naive question - did the reflection of the ND attenuator affect the laser operation at all - instability, noise, mode hopping etc? (guessing it was intentionally misaligned so not to not be normal to the optical axis?)
Yes indeed, I did place them in under a slight angle especially the first one which was OD=4 and highly reflective. The operation of this particular laser diode was remarkebly stable. I did try others, like cheap red diode lasers. But on short time scales these were quite unstable and displayed all kinds of unpredictable instabilities.
@@HuygensOptics great, thanks for the reply. Bye bye... bye bye... :)
@@HuygensOptics Les' Lab has a neat video about lasers reflecting back in themselves: th-cam.com/video/ad4hxXz4bwA/w-d-xo.html
When a photon interaction occurs and this street energy transfer from the wave to the electron happens what happens to the wave itself? How is the knowledge of that energy transfer propagated in space? From the wave point of view is this interaction almost like an anti-light source, which has a wave that back propagates, or something else entirely?
The photon 'wave' is actually absorbed by the electron giving the electron more energy and causing it to 'jump' to a higher orbital.
@ but now does the change field propagate?
@@youngbloodbear9662 Do you mean "charge field"? The charge of the electron propagates as a wave in the space medium that travels out spherically from the electron.
@ there is an electric field proportional to the probability of an electron reaction, when that interaction happens, how does the field that caused it react?
24:00 omg this was so helpful.
Thank you
Please keep making videos.
You might get back a bit of resolution by using kernel density estimation instead of histograms. You still have the nuisance parameter of kernel width, but there are good ways to pick one. It should be fairly easy to implement in your code too. You just have to generate gaussians centered at each of your time differences instead of adding them to buckets.
The way you understand fotons and their detections is the same way I make the sense of all of quantum mechanics. For me there are no particles, only waves and the quantum measurement is just energy/momentum exchange event between the waves that reshapes the waves.
thus way the essence of a wave itself remains above all others, leaving the photon (= electromagnetic) field and electrostatic field too in the area of "something inexpressible"... semiclassic attempt of description "wave is wave, photon is energy of interaction" was rejected tens years ago.
@@tapdapy There's nothing (semi-)classical about this. The math is exactly the same as the most up to date math of quantum mechanics. Photon is the complex probability wave evolving according to Schrodinger's equation, it's not ever anything else. Measurement is energy/momentum exchange event that reshapes the waves that take part. Electrostatic field (and magnetic field that is a consequence of special relativity applied to electrostatic interaction) is just the field that a photon is a wave of. Traditionally calculations in quantum mechanics done regarding measurement are done as if they were narrowing down measured object to a single point (or at least something smaller than the threshold of measurement apparatus) but that's just artifact of our capability that those are the values we are most interested in. You can collide very fuzzy electrons with very fuzzy electrons and measure the result afterwards and the quantum mechanics math can describe results as well. The only difference is in interpretation. Most mainstream one now is that at some point those fuzzy electrons must have experienced wave function collapse and then they collided. While mine just says that they experienced some energy/momentum exchange and got reshaped not necessarily to something point like.
The whole idea of duality is silly and ancient. Elementary particles are just objects unlike anything we see in macro scale. They are not really particles ever but if they are tight together in very large numbers they tend to behave like one with specific momentum and location. Their wave-like math that describes their individual evolution in time is obscured and the Schrodinger's equation simplifies to Newtonian kinematics. However the way the energy/momentum exchange works between a pair of them works same in macro and microscale. That's why historically we believed that they might be a small version of macroscopic objects. But I think keeping this idea is harmful and it's actually this mainstream interpretation that postulates that sometimes elementary particles are actual pointlike particles is semi-classical and will go out of fashion at some point.
Is your setup thermally stable? Small path differences at 100nm level in time between the the photo multipliers will have an effect on your measurements.
As you noted histograms are sensitive to the bin size. It might be useful to use the Cumulative distribution functions (CDF) of the coherent and non-coherent photon sources. I think of two options to compare these. One is to fit an analytical function to both and then look at the derivatives of these to look for peaks. The second one is plot both CDF's on the same graph and look for deviations between the two.
It is very annoying that QM is taught as wave-particle duality when it is very obvious that the underlying fundamental object is the field which evolves via the wavefunction. Particles are an invention, and past high school physics they do more harm than good at building an intuition. It took me years to understand this and it's very frustrating.
@2:11 (Still frame) Perfect image what i see without glasse with one only: astigmatism + some remaments (unknown source) on lens - without those color strpis obvuously but those mangling shapes inside... striking resemblance
Hi Jeroen, please tell how you synchronize the detector signals in time
One question about a beam splitter: In the process of sending/splitting a photon in two directions, does it duplicate the photon or is it the same photon that is later detected at two detection points at the same time? Also, does the beam splitter detect the photon as well (as being some kind of observer in the path of being observed by the detectors?). I just got here from watching the newest hologram video by 3Blue1Brown, and I only trust asking this question at Huygens Optics. Simpler question: Is the photon the same at detection point 1 and detection point 2, or two different ones (duplicated by the beam splitter)?
H-BT effect is used by the heavy ion smashers to measure the size of the nuclear mess, but here they use thermal radiation of pions, not photons.