I did the quantum eraser experiment at home

0:00 In quantum mechanics,

0:01 things behave differently the moment they know you're watching.

0:05 I'm gonna do an experiment where light changes

0:09 from this to this, all because it was being observed.

0:14 But why do quantum objects care if they're being watched?

0:18 This is the mystery that got me obsessed with quantum mechanics years ago.

0:24 But I love this experiment because the three

0:28 variations of it explain three crucial facts about measurements,

0:31 and the third might even be a solution to the measurement problem itself.

0:37 In the double slit experiment, you shine light on two very narrow slits,

0:43 which I've represented here like this, and you can see

0:47 that the light sort of spreads out and it goes through both slits.

0:51 In fact, we know that this is what happens

0:56 even with laser because you can actually see it.

1:00 If light is somehow wavelike,

1:01 then this kind of makes sense because it's like saying the wave will spread out.

1:07 And then it will go through both of these slits and make new little waves.

1:15 Eventually, these two little waves will spread and overlap,

1:22 and as you can imagine, when two waves get in each other's way,

1:25 they interfere with each other,

1:27 and this is what causes the classic double slip pattern.

1:31 This is a double slip.

1:33 Hopefully you can see.

1:34 And now I'm gonna put the double slit

1:37 in front of the laser and turn off the lights.

1:41 Let's see what happens.

1:43 And this is what the pattern looks like.

1:47 And just as a side note, I never get over how beautiful this looks.

1:54 Just crazy.

1:54 And so now you get this double slit interference pattern,

1:58 which looks like a whole bunch of little, little dots.

2:02 Right, like there are these sort of bigger, uh,

2:06 regions and then in each of those regions

2:09 there's like little sub regions as well.

2:12 That's the double slope pattern.

2:14 This all makes sense if you buy into the classical wave theory of light.

2:20 But this isn't classical mechanics, this is quantum mechanics.

2:24 And so many of my comments were like, shouldn't this be impossible?

2:30 How can it be possible to see light going through both slits at the same time?

2:36 And I understand why people were confused because this seems to contradict

2:40 one of the big ideas that I wanna talk about in this video,

2:44 which is that measurements cause collapse of the quantum state.

2:49 So here's what that means.

2:51 Light like this can be in a quantum super

2:53 position of doing multiple things at the same time.

2:56 So for example, this light is in a super position

3:00 of going through the left slit and the right slit,

3:03 and you can see why that's totally fine.

3:06 It'll just lead to this overlap.

3:08 But this rule says that this is fine as long as we don't ever

3:13 try and find out whether the light went through the left or the right,

3:17 because when we do, we cause it to collapse.

3:20 So for example, if I was somehow to measure

3:24 what the quantum state was doing by, let's say,

3:27 looking at the light and seeing it going through the left slip,

3:31 then I would collapse the state.

3:33 And so.

3:34 It would have to only do one of these things,

3:38 and let's say the right one might disappear.

3:41 And so the rule says when you look to see whether the light went left or right,

3:46 then you collapse it to just going through one of the slits.

3:51 So shouldn't that have happened when we were doing

3:53 our experiment and looking at the way light was going?

3:56 Well, don't worry.

3:57 It's not actually a contradiction because of this.

4:00 I suppose this is all the light coming out of each of these two slits.

4:05 Before we measure it,

4:06 and then we use the same technique that I did to see the light,

4:10 which is to use a horrible, carcinogenic, uh,

4:13 smoke machine to make these little droplets of smoke here and there.

4:18 And what this will do is when a bit of light hits one of these smoke particles,

4:24 it will get deflected.

4:25 And now that bit of light will no longer go through.

4:28 But the rest of the light will go on sort of unaffected.

4:31 So some little bits of light will deflect off these smoke particles

4:35 and might even end up in the observer's eye or in the camera lens.

4:40 And so that's why we can kind of overall see the path of this light.

4:45 But it doesn't count as a measurement of most of the light because you can see

4:50 that most of the light goes completely unmeasured

4:52 by the smoke particles because they never encounter any.

4:55 And so that's why the smoke is actually not

4:58 a good measurement of which way the light goes,

5:01 but there has to be a cleverer way of figuring out

5:05 whether the light went left or right and collapsing the state.

5:08 And in fact, there is, here's how we're gonna do it.

5:12 So when we imagine light as a wave,

5:14 we're usually thinking of the, the wave sort of waving up and down, right?

5:18 Like these are all the peaks of the wave and they're sort of up and down.

5:23 But actually light can be a mixture of all kinds of, um,

5:27 directions of these waves.

5:28 Like they could be waving left to right,

5:30 they could be waving like this or like this, and this light

5:34 is just gonna be a huge jumble of all kinds of different directions.

5:38 But what if we used this stuff to filter out the light?

5:42 So let's say that we have some light coming in on some random direction here,

5:48 but we only let through light that is oriented.

5:52 Up and down like this, so only this light would be able to go through.

5:57 Or what if we turn the filter around

5:59 and now it only would allow light going this way,

6:03 so this light would go through, but nothing else?

6:07 Well, we could do that and it wouldn't really do much for us,

6:10 but here's the really clever bit.

6:12 Why don't we do both?

6:13 Why don't we put a filter in this direction over

6:17 just the left slit and a filter in this direction.

6:20 Just over the right slit.

6:22 That means that light that makes it through the left slit has

6:27 to be oriented like this vertically and light that makes it through.

6:31 The right slit has to be oriented like this, which is horizontal.

6:35 What we've just done is we've marked these two slits.

6:39 The light coming out of them are very different,

6:41 and so if we want to know whether a bit of light

6:45 is coming from the left slit or from the right slit.

6:48 All we have to do is put a filter like this, because a filter like

6:53 this will only allow through vertical light

6:56 and it won't allow through this horizontal light.

6:59 And so by measuring this polarization,

7:02 we've also measured which direction the light went through

7:05 and we've collapsed it to only going through one, which is the left sl.

7:10 If our rule that measuring the quantum

7:12 state really causes it to collapse is true,

7:15 then we should see that the light goes from being a double slip pattern.

7:21 To a single slit pattern.

7:22 So if we can put these filters behind the two slits,

7:26 then we will have a way to test where the measurement collapse really happens.

7:30 It was very, very finicky, but I finally got it to work.

7:35 Here are the little filters that go behind the slits.

7:39 So let's see if this really causes measurement collapse.

7:46 This is the new pattern, and you can see that it actually looks quite different.

7:51 Like the big patches are still basically where they were,

7:56 but now there isn't these little lines in them.

7:59 And just to see how different this actually looks, let's turn the camera around.

8:05 You can see that there's basically.

8:09 Only one beam that's really visible, which is the middle beam.

8:12 The tiny little beams that you see in the double slip pattern are mostly gone.

8:18 So observing this light really did change it.

8:22 Just to double check that we really have

8:25 collapsed this light into just going through one slip.

8:28 I compared this pattern to what would happen if

8:31 you did just have one slit to start off with.

8:34 And you can see that they look very, very similar.

8:37 The spacing of these bright bands is pretty much identical,

8:42 which proves our hypothesis that observing the light really collapses it.

8:48 But here's the really trippy bit.

8:51 So this filter here is what allowed us to know

8:55 that the light is going through the left slit.

8:58 But what about if I just remove it?

9:00 Here's what the light looks like with the filter,

9:03 and here's what it looks like without.

9:07 Basically the same.

9:09 It just looks a bit brighter without.

9:13 So what on earth is going on?

9:16 This second version of this experiment leads us

9:19 to our second big idea for this video,

9:23 and it's that measurement doesn't need a conscious observer.

9:27 Measurement happens when information is leaked to the environment.

9:31 To understand why this is true, let's look at our experiment again.

9:36 So remember we said that this bit marks which way?

9:41 What I mean by marked is the information about whether the light

9:46 went left or right has now been transferred or copied onto something else.

9:52 And in this case, it's the polarization of the light.

9:55 So now the polarization of the light contains

9:57 that information about which way the light went, and then.

10:01 That other filter we had, this was technically the measurement.

10:06 When this is here, we know that the light

10:09 that comes out must have gone through the left slit,

10:11 and so we expect to see the single slit interference pattern,

10:15 not the double slit one, so it'll look something like that, whereas

10:20 the double slit one would look like this.

10:24 Now the remarkable thing that we found

10:26 was that we didn't actually need this measurement.

10:28 Even without it, we still only get the single

10:32 slit interference pattern and not the double slit one,

10:35 and it's because marking which way the light went is as bad as a measurement.

10:39 One way to think about this is

10:42 that information got leaked from the particle's position,

10:45 so where it was to its polarization.

10:49 And this is a more general phenomena in quantum mechanics.

10:51 Let's say we have an object that's in a quantum

10:54 super position of doing two different things like our light,

10:58 which was in a super position of left and right.

11:02 Then let's say another particle comes along and marks the two options.

11:07 For example, if the original particle was in state one,

11:11 the new one will lose energy.

11:14 But if the original particle was in state two,

11:17 the new particle will gain some energy.

11:19 So the energy of this new particle

11:22 now carries information about the original particle,

11:26 and so that's what I mean by marking.

11:28 Now let's say that that new particle just flies off with that information,

11:33 what happens to our original object?

11:36 It will look like it collapsed to doing one of those two things.

11:42 The information leaking into the environment will look just like a measurement.

11:47 This is called Decoherence.

11:48 Okay.

11:49 But our story's not over yet because I wanna

11:52 show you the craziest variation on this experiment yet.

11:55 So, so far we said that having these filters on top of the double slit,

12:02 um, marks, which way the, uh, light goes.

12:05 And since that information has leaked into the environment,

12:08 it collapses the light to either going through the left slit or the right slit.

12:13 But if that's true, then how do you explain.

12:16 This here, we got the single slip pattern that we've

12:20 gotten used to, but I can actually recover the double

12:23 slip pattern by putting a filter in front

12:25 of this, but not in the direction that we were doing before.

12:30 If I put it at this 45 degree angle,

12:33 you can see that I actually recover the double slip pattern.

12:37 I just think this is insane.

12:40 How crazy is that?

12:41 You can see the double slip pattern just pop back out.

12:44 It's even crazier when you look at it

12:47 this direction where this is the single slip pattern,

12:50 which we expect because of our observation.

12:53 And here is the filter.

12:56 And look, it brings back the double slip pattern.

13:00 The double slit only exists in front of the filter,

13:05 but behind it, like you can just see the the single slip pattern.

13:12 This is such a bizarre experiment.

13:15 I, yeah, look at that.

13:19 How is it possible that if this is a measurement

13:23 that collapses the light to only going one direction,

13:26 that we're able to make it go through

13:29 both directions later on by putting a filter

13:33 in like this, like we're influencing whether

13:35 the light went through both of the slits here.

13:39 By putting a filter in here.

13:41 That's where the third,

13:42 and I think most interesting lesson from this experiment comes in.

13:47 Measurement collapse is an illusion, at least sometimes, and maybe always.

13:52 So in the case of this experiment that we just did,

13:56 which is called the Quantum Eraser measurement collapse was an illusion.

14:01 So here's what I mean.

14:03 Originally we said that just marking which way this light goes.

14:08 Was enough to cause it to collapse,

14:11 but let's say that it has in fact collapsed to just going left.

14:15 Then it goes through some other filter,

14:18 and then we end up with the double slip pattern.

14:21 How is that possible?

14:23 If the light really had only gone through one of them?

14:27 That would be like saying, suddenly after going through this other filter,

14:31 not only does this light continue,

14:33 but suddenly there's new light coming from the right slit.

14:36 Which is impossible if there wasn't any here.

14:40 So we must have been wrong.

14:42 There must have been some light still coming from the right slip at this point.

14:48 But then how come without this filter,

14:50 we didn't get a double slip pattern, we only got a single slip pattern.

14:55 Well, what this shows is that without this other filter here,

14:59 what happened on this side?

15:01 Is just fake collapse,

15:02 or if you wanna be technical about it, it's called subjective collapse.

15:07 This sounds really bizarre, but it's actually not controversial at all.

15:11 So what happens is when you mark, which way the light goes, it doesn't.

15:18 Actually collapse.

15:19 Both bits of light are still there.

15:22 They'll overlap in space, but even though this is happening,

15:27 they're no longer able to affect each other.

15:29 And remember, it was this interference that caused the double slip pattern.

15:34 And now that that interference can't happen anymore,

15:37 you just get the single slit interference pattern.

15:39 Like I said, I know this sounds weird and it,

15:42 it sounds a little bit mystical when I have to say it

15:45 in English like this, but when you write down the mathematics of it,

15:49 you kind of see why they're no longer allowed to interact

15:52 and it's all because of the fact that they're marked,

15:55 each one is marked in a separate way.

15:57 And those two bits of light,

15:59 now that they have this difference, aren't able to interact anymore.

16:03 If you are interested in actually seeing that maths for yourself,

16:07 then I've actually written it up and the link will be in the description.

16:11 But if you're happy to accept that.

16:13 Then you'll understand why the 45 degree filter fixes everything.

16:17 Remember light that's coming through the left

16:20 slit is vertical and light that's coming through.

16:23 The right slit is horizontal.

16:25 But if I put a 45 degree filter in between,

16:29 what happens is that this light has a 50% chance of going through,

16:34 and this light also has a 50% chance of going through.

16:38 So after this 45 degree filter.

16:41 Firstly, both bits of light will now be oriented 45 degrees,

16:45 so they'll both be oriented kind of like this.

16:48 And any light that's gone through this filter is equally likely

16:53 to have come from the left or from the right now,

16:57 because both of them will have been polarized in the 45 degree direction.

17:02 And so in a way, this filter undoes what this one did.

17:08 It unmarks them.

17:09 In other words, this filter has erased

17:13 the information about which direction the light went.

17:16 You can't just look at the polarization anymore to tell

17:20 did it go left or did it go right?

17:23 And so everything is back to how it was,

17:26 and you end up with the double slip pattern again.

17:30 I love this experiment because it's a really tangible

17:33 example of a more general thing in quantum mechanics.

17:36 So you can have a particle that is in a super position and then

17:42 interacts with some other particle in a way that seems to make it collapse.

17:47 And we might say that the second particle measured the first one,

17:51 but then if you're able to get that second

17:54 particle back and erase the information that was on it.

17:57 You see that the first particle was still in a super position

18:01 and now it acts like it was in a super position the whole time.

18:05 What I really like about that is that this is potentially

18:08 a general story about what measurement really is in quantum mechanics.

18:13 So replace that second particle with a measuring

18:17 device or a human whatever you want.

18:20 And what we see is we have a quantum object and when.

18:25 The person or the machine measures that object.

18:28 It seems to cause that object to collapse, but.

18:32 It actually is recoverable.

18:34 If there was a way to remove all of the information about what

18:40 that particle was doing from your machine or from, you know, your brain,

18:45 the entire environment, every photon that interacted with that, that particle,

18:49 then you are able to see that the particle was never actually collapsed at all.

18:56 This whole thing of sort of fake collapse is called decoherence.

19:00 And like I said, it's not controversial that decoherence happens sometimes.

19:05 Like this example is a really great

19:07 one of decoherence happening and then getting erase.

19:10 So it's not controversial that this can happen.

19:14 What is controversial is, is this, the answer to the measurement problem when

19:19 we talk about measurement is this what's always happening?

19:23 It is so controversial because once you start saying

19:27 that this is the solution to the measurement problem,

19:30 what you mean is that every time you go and measure a quantum particle,

19:36 there are two versions of you.

19:38 One that sees the particle,

19:40 do one thing and the other that sees the particle do the other thing,

19:45 and both still exist.

19:46 You'll feel like.

19:47 You have collapsed to one of those and the particle has collapsed

19:52 to one of those, and every experiment that you try to do

19:56 to get evidence about whether the particle has collapsed to going just

20:00 one way or if it's still going both ways will show you No,

20:04 no, it's definitely collapsed.

20:06 You'll only see the single slit interference pattern.

20:09 But if you were able to somehow undo all of that, then

20:14 you would see that actually there was no collapse at all.

20:18 Really what happened was that you and the particle were in super position,

20:24 which I think is a very uncomfortable thought.

20:26 This solution to the measurement problem is

20:29 called the many world's interpretation of quantum mechanics.

20:32 And I think that at least mathematically

20:35 it's extremely nice because think about it,

20:37 the maths going on here, we already know that it can happen,

20:42 and it does happen in quantum mechanics.

20:44 So you don't need to propose any new mechanism to explain measurement,

20:48 unlike the traditional, um, interpretations of quantum mechanics that do

20:52 make measurement this special and very weird thing.

20:56 In many worlds, it's completely natural.

20:58 This is exactly what you would expect

21:00 to happen if you take quantum mechanics very seriously.

21:03 But I think that it does take a while to let that mats really sink in.

21:09 Um, so if you're interested, I have written up a bit more information about

21:13 this experiment and how it generalizes a little bit.

21:17 So the links in the description.

21:18 I'm also making a follow up video about incorrect interpretations

21:21 of the double slit experiment and this delayed choice quantum array experiment.

21:26 So if you're interested, that should be here somewhere.

21:30 I.