Single Photon Interference

[Applause]

Previously on Veritasium, we saw how our understanding of light has changed over the centuries. In the late 1600s, Hyans proposed that light was a type of wave, while Newton considered it a stream of particles. This debate appeared to be settled in 1801 by Young's double slit experiment, which showed light passing through two slits produced patterns like water waves. But by 1900, it was clear that light energy was not evenly distributed, as expected for a wave; rather, on the smallest scales, it comes in lumps called quanta or photons.

So the question is, how does this affect the double slit experiment? Here I have a more contained double slit experiment, where there is a laser that fires a beam through a single slit, then through a double slit, and onto a screen where you can see a well-defined interference pattern. There's a series of bright and dark bands, which are much easier to observe than when I use sunlight, because with only one wavelength, there are no other colors involved.

This is what a graph of intensity versus position would look like for the interference pattern. What creates that pattern? Consider the bright spot in the middle; the light from each slit has to travel the same distance to reach that point, and hence both waves arrive in phase. That means crests with crests and troughs with troughs, so they add together and create an interference maximum: a bright spot.

But if you look slightly to the left, there's a dark spot. Now that's because light from one slit has to travel at an angle and has to travel an extra half a wavelength compared to the light from the other slit. This means that when this light is arriving as a crest, the light from the other slit is arriving as a trough, and they cancel each other out.

However, if you go further left, you see another bright spot because now the light from one of the slits has to travel a full extra wavelength compared to the light from the other slit. So again they arrive in phase—crests with crests and troughs with troughs—creating constructive interference, and so we see a bright spot of light.

And that's how the whole pattern is created. But what if I decreased the intensity so much that there wasn't a whole wave of light going through there; there were only single photons? Then how could they interfere with each other, because there's only one in the device at any one time? So would we still see an interference pattern? That is what we're going to find out.

In order to make this work, I had to line up a very faint source, and to see where the light was going, I had to shroud my head in that black cloth. But I finally have the apparatus ready. You can see up here I have a frequency counter, which actually counts the number of photons received per second at the detector. The detector is a photo multiplier tube, which is capable of detecting single photons. It's like a very sensitive eye, like the frog's eye.

Now, right now there's no light passing through to the detector, but there is still a bit of a background reading, and that's because I can't block all the light out of there. Plus, there are cosmic rays passing through this room, which will also set off the detector. I'm plotting a graph of the number of photons counted as a function of position across the detector.

If you have a look, after 1 second, the distribution seems random. There doesn't seem to be any pattern in the arrangement of those photons as they hit the detector. So maybe it's true; a single photon can't interfere with itself because it's just a localized point—it has to go through one slit or the other. But just to be sure, let's add up the results over a period of time and see if any pattern emerges.

[Music]

Look at that! You can clearly see the same interference pattern that we got when we were sending tons of photons through, but we're getting it out of single photons. We're counting up individual photons, and that pattern is emerging as we aggregate the results over time. But how could this be happening? How could a single photon pass through both slits?

Well, if we try to interpret these results in terms of the objects we experience every day, they don't make any sense. A photon is something different from a macroscopic object. It's not a wave, and it's not a particle; it's a quantum mechanical object. Sometimes it seems like it has properties of a wave, and sometimes it seems like it has properties of a particle, but ultimately, it is something totally different from anything we've experienced before, and that's what makes this seem so counterintuitive.

So what is light: wave or particle? The true answer, I think, is neither. Though if you want, you could call it a wave.

[Music]

It is created by the ripples from the two sources interacting with each other, where they meet up—peaks with peaks and troughs with troughs. The amplitude of the wave is increased; that's what we call constructive interference. But if the peak from one wave meets up with the…

At the end of the last video, I asked why it is that the interference pattern is made by blobs rather than thin lines of light, as it was in this experiment. In the comments, I saw two common answers. One was that the blobs were images of the Sun. The other answer was that it was due to Heisenberg's uncertainty principle.

Now those were very thin slits, so it's true Heisenberg's uncertainty principle would be involved, and that light passing through each slit would spread out. But the reason we actually saw blobs was because they were images of the Sun. So if the sunlight comes through one slit, it diffracts out and will spread onto the bottom of the box, and light from the other slit would do exactly the same.

Where those two spreads overlap is where we saw the three main blobs. So there's a diffraction maximum created by one slit and another diffraction maximum created by the other slit, and those two overlap. But due to the light being in different phases, it cancels out in certain positions, creating those separate images of the Sun.