Professor at the École Polytechnique and at the Institut d’Optique, Palaiseau; Distinguished Senior Researcher at Centre national de la recherche scientifique (CNRS), France
Breaking the Wall of Quantum Weirdness. How Experiments Reveal Photon Schizophrenia.
On November 9th, I was amazed that everything happened in a pacific way.
I am glad to be here, but I must say that it is a very difficult task to come after such a speech.
So, I am not sure I should thank the organizers (Laughter in audience). I am going to talk about quantum physics, the weirdness of quantum physics. My only excuse is that we already have had examples of quantum physics changing the society. You all know that the conceptual revolution of quantum physics, which happened at the beginning of 20th century, ultimately led to the invention of transistor and integrated circuits, and lasers, ie the basic components of the information society. So, thinking about quantum weirdness, thinking about these quantum mysteries, can be useful for the society; it will be my excuse for presenting this curiosity-driven research.
I am going to present wave-particle duality about light. Probably you know that there were many different models of light across the ages. I have no time to discuss ancient Egypt model of light, but I will recall the big conflict between Huyghens, who thought that light was made of waves, and Newton, who was defending a model of particles.
The issue between the models came in the 19th century with Young, Fresnel, and Maxwell: light is a wave. At the end of the 19th century, all optical phenomena such as interference, diffraction, propagation, could be interpreted by modelling light as a wave.
Then a young man came, you have heard of him (a photo of Einstein young is shown), and he said, “Light is made of quanta”. (Licht quanten), which are elementary grains of energy. With this model, he was able to make quantitative predictions about the photoelectric effect- and nobody liked his idea. When he was elected in the Prussian Academy of Science, in 1911, the report wrote that he had done so many good things that they were going to forgive him for his mistaken representation of light and his mistaken law of the photoelectric effect. The great experimentalist Millikan then endeavoured to demonstrate that Einstein was wrong. But at the end he found him right, and Einstein received the Nobel Prize for this model of light being particles and his law of the photoelectric effect.
So far so good! But, it is not because Einstein received the Nobel Prize for light being a particle, that we have to forget about light showing behaviour like interference, diffraction that can be understood only by assuming that it is a wave. So the question remains: Is it a particle or a wave?
Einstein himself was obsessed by the problem, and addressed it as early as 1909 in a famous conference in Salzburg. He came with the conclusion that one doesn't have to choose; light is both a wave and a particle. One decade later, going one step further, Louis de Broglie came with the reciprocal idea that something you think is a particle – for instance an electron, is also a wave. Of course, all of this is easy to say in words, but very difficult to understand with images.
How do we represent wave particle duality in textbooks? We suppose we can emit individual particles, individual photons, passing two holes. If we have a detector scanning far behind the holes, we see bright and dark fringes. We cannot then escape the conclusion that light must be described as a wave; passing simultaneously both holes, and yielding the fringes in the overlap of the beams associated with the holes. Only a wave description of light allows us to understand interference.
But what happens if we have single photons falling on the double holes? Do we observe interference with single photons? Will each photon pass simultaneously the two holes? Surprisingly, although the question was raised at the beginning of the 20th century, the answer came only in 1985 when my PhD student Philip Grangier and I built a first source of single photons, and we could demonstrate the wave-like behaviour even in the single photon regime. This implies, in a sense, that a single photon can travel through the two holes simultaneously.
But how could we prove that we had a true source of single photons? We used two detectors and put one behind each hole. And what we observed is that only one detector would fire at a time, we never observed the two detectors firing simultaneously. Everything happened as expected for a single particle, which will pass either one hole or the other one. At this point, we had proven than our source was emitting light in individual photons. And then, removing the two detectors, and replacing it by one detector in the overlap of the beams emerging from the two holes, we could observe interference. It was single photon interference.
This is the source which allowed us to observe this in 1985. You see it was a very complicated source, and here you see the interference pattern we obtained. Now I will rather present the way such an experiment is done with modern equipment. The big room full of lasers has been replaced by a single molecule excited by one laser, at the focus of a microscope. A single photon is emitted, and sent into an interferometer, which is called a "Fresnel double-prism interferometer".
The idea is the following: if you have a wave, a part of the wave will be deflected down here, and a part of the wave will be deflected up here, the two parts will overlap and then separate. What do I expect for a single particle? It will pass either up or down, and I will never observe any coincidence between two detectors placed in the two beams that have separated. We have done the experiment, and it works as I just said, indeed showing that we have produced single photons, well separated from each other. But now let us put a CCD camera here at the overlap between the two beams. If the single photons show interference, it will manifest itself by being detected in fringes observed on the CCD here. I cannot resist showing it.
Here, we are going to see each photon arriving on the CCD as a red spot. So, lets play it, and you see photons arriving, red spots appearing at a slow rate. (I must accelerate, otherwise a gentleman will come and chase me out)... Now, after accumulation of enough photons, you see what happens. There are bright fringes, lines where almost all photons are detected, and dark fringes, lines where photons are almost never detected. This is interference with single photons.
Why do I say it is true interference with single photons? Look- we did a first experiment, and the first experiment told us that the photon is detected either here or there, there is no double detection, so there is indeed only one photon at a time, and it chooses between going up or down. But in second experiment, the observation of interference shows without any doubt that a wave passed simultaneously up and down. But look again, it's the same source, the same double prism, that are used in both experiments.
How can it be that, in the same piece of equipment, light goes either up or down in the first experiment, and goes on both paths simultaneously in the second experiment? Well, these contradictory images are emblematic of the weirdness of wave-particle duality.
When you have a problem understanding quantum mechanics, you are advised to go and read Niels Bohr, and ask the question: “What would Niels Bohr have said about that?” The answer is that case makes little doubt, Niels Bohr would have said, “Look, let us be serious; you have to choose: either you do the first experiment or you do the second one. The end parts of the two experiments are different, you have to choose between them. And the photon will adjust its behaviour according to the kind of equipment you choose”. In other words, if you ask the question, “Are you a particle”, light answers, “Yes, I am a particle.” If you ask the question, “Are you a wave?”, it answers, “Yes, it is a wave.” "Aha! you say, I understand: when the photon arrives at the end of the first (common) part of the apparatus, it checks the second part and says to itself: “I see, this is an interferometer; I have to behave as a wave.” (Laughing in audience). And in the other case of course, it thinks: "That system will try to determine whether I will go up or down. In order to be able to answer, I must behave as a particle which goes on one side only."
Then comes John Archibald Wheeler, a very smart physicist. He then asks you: "If you believe that such a thing does exist, that the photon adjusts its behaviour to the type of measurement you do, what do you think would happen if the choice was not yet done before the photon exits from the first part of the experiment, which is the same in both cases?” And you realize that he is right, it is possible, at least in principle, to wait for the photon quits the first part before inserting one or the other piece of equipment, according to the question you want to ask. Actually you can delay the choice until the moment when the photon arrives here. You can decide then either to introduce a CCD camera to check interference, or to remove it and to let the photon proceed to the two "which path detectors".
This so-called "Wheeler’s delayed choice experiment" has been carried out recently by a team at ENS Cachan, in collaboration with Institut d'Optique. It uses a long interferometer, 50 metres long, so that there is enough time to perform the delayed choice. The photon enters, travels, travels, and only at the very last moment you decide either to recombine the two beams in order to look for interference, or not to recombine them and let the photon go to the two "which path detectors". The result we found was the same as usual: at the end, quantum mechanics wins; we get the result corresponding to the measurement which is chosen, although the photon could not know in advance what measurement would be chosen.
So, what can we conclude? I let you appreciate this conclusion of John Wheeler, in the case where Quantum Mechanics would win: “Thus one decides the photon shall have come by one route or by both routes after it has already done its travel.”
As Feynman writes in his famous lectures on physics: wave particle duality is one of the “great mysteries” of quantum mechanics. We cannot escape it, experimental facts force us to accept it. And the delayed choice experiment teaches us that Bohr’s complementarity is not as naïve as presented in some elementary books of quantum mechanics. There is certainly some truth in Bohr’s complementarity, but it is much more subtle than we thought.
To finish, I would emphasize that trying to understand better quantum mechanics lead physicists farther. Questioning the foundations of quantum mechanics, in the last decades lead us to what is called 'quantum information'. Understanding better the two great mysteries of quantum mechanics, wave particle duality for a single particle, and entanglement of two particles (another story), leads to the fantastic idea of quantum computer, which does not yet work, and to quantum cryptography, which is already operational, and that you can buy from small start up companies.
This is an example of a scheme of quantum cryptography, based on single photon behaviour. The scheme is due to Gilles Brassard and Charles Bennet. It is a story about two lovers, Alice and Bob, who want to exchange secret messages. (It is more fun than bankers, isn't it?) The lovers want to exchange messages being sure that there is no eavesdropper on the line able to read the travelling message. The idea of quantum cryptography is that the secrecy of the communication is based on fundamental properties of quantum mechanics: the fact that when you have a single photon, if someone tries to observe it, there will be a slight change that you can detect. The spy leaves a footprint.
I thus want to conclude that breaking the wall of weird concepts may eventually lead to applications. And before leaving the stage, I want to show the people who have done the experiment. Actually it is a family affair: Philippe Grangier was my first PhD student, Jean-François Roch was the first PhD student of Philippe, and the scheme continue to the next generation with the two PhD students Vincent Jacques and E Wu. We were all delighted to be four (scientific) generations involved in carrying out one of the most spectacular demonstration of quantum weirdness, Wheeler's delayed choice experiment.