Lene Hau and condensed matter physics, transcript
Lene Hau, Mallinckrodt Professor of Physics and of Applied Physics, Harvard University: Well, thank you very much. So, well, as you heard a little bit of what I’m going to talk about is definitely changing gears here. Some of you might know a very exciting area in physics in the recent years has been the studies and uses of ultra-cold atoms. These days we can cool atoms down to micro Kelvin and nanoKelvin, that’s a few millions to a few billions of a degree above absolute zero, so that’s pretty freezing cold. And we typically use lasers to do that job that’s sort of new techniques that are being developed. And when matter gets that cold, if you can sort of collect the bunch of particles or atoms, cool them down, they tend to form extremely strange states of matter.
Watch the full video here
Hau: And a really good example of that is supposed Einstein Condensate; it’s a state of matter that Bose and Einstein predicted some eighty years ago. And it’s only in recent years it’s actually been possible to create this state of metal clearly in the lab. First seen in the summer of ‘95 by Eric Cornell and Carl Wieman. And I should say also, both Einstein condensate, before we get back to more precisely what they are, but Bose-Einstein Condensate are also the reason explaining why we have this strange phenomena such as super fluidity and superconductivity. This is idea of superconductivity that you can pass crayons and certain materials of low temperatures, through electrical currents with no resistive losses. And that’s of course an issue of great importance to help solve our energy problem. If we can start to pass electrical power over long distances, with minimal losses, that will be tremendous. So Bose-Einstein condensate that whole idea is really behind this phenomenon. Now in my lab, what we have really been focusing on is using this Bose-Einstein condensate to manipulate light to the extreme. And I will tell you a bit about how we go about slowing light to the speed of a bicycle, just as you heard in the introduction. And I’ll also tell you some of our own more recent stuff, where we an stop and extinguish with the light pausing one part of space and then we can regenerate the light pulse in a completely different location and send it back on its way.
So to do these kinds of experiments, of course, we need some cold atoms. And with the kind of temperatures we’re talking about, you are not going to do that with your sort of standard household kitchen refrigerator. We need a very special atom cooling set up. And just to show you a few pieces here, the whole set up is typically two by three meters, to give you the scale. And of course, first of all we need some atoms. We use sodium atoms at the moment. So we have an atom source where we take a chunk of sodium metal and then heat it to a high temperature, typically hundreds of degrees centigrade, to get a good high vapor pressure. So we can get a good emission rate of atoms out of this work. And that’s great that works. The only problem is that because of this high temperature, the atoms will come out with very high velocities. Really, so high of a velocity that we effectively use the atom source as soon as the atoms come out of the source we hit them head on with yellow laser beam that we send up through the system. And then we use radiation pressure to slow these atoms down very dramatically and very rapidly. And then we can load them effectively into what we call an optical molasses, created in the middle of vacuum chamber by three pairs and counter propagating laser beams to the exact right frequency relative to the characteristic absorption frequency of these sodium atoms. And when we hit that right frequency these laser beams will really act as molasses, a viscous medium for these atoms. And we’ll tempt their motion down and we can cool the atoms quickly down to microKelvin.
This is what it looks like in real life and in the lab when we are laser cooling. As you can see we have a room full of optics, lasers, yellow laser beams running all over the place, criss-crossing the lab. And this is a laser table in daylight and quite a few men and women help with building setup as you can imagine. Over on the side here we have our high vacuum chamber, where the main action takes place. And I have a closer view of that. This is where we create our Bose-Einstein condensates and also where we slow light. And we have vacuum flange windows sitting all over this vacuum chamber. You can see one here. And with our laser tool it’s actually really easy because we can just look through one of these windows and see these cold atoms by eye. It looks like a little bright sun, freely suspended in the middle of the vacuum chamber about one centimeter in diameter. The reason we can see this is because with the laser cooling process, the atoms keep absorbing laser photons. And then they reemit spontaneously so the atoms will flow as yellow.
Well so, with laser cooling we get down to microKelvin but we are not totally happy. We want to get down to nanoKelvin. So, at this point we turn the lasers off and then we turn on the electro-magnets. We use the fact that the atoms actually act as small magnets. They have a small magnetic dipole moment so we can track them in the magnetic field minimum. And then once we have that happily trapped, we start an episode of cooling process. Where we can selectively kick out the hardest and most energetic atoms in the sample, and then the rest of the trapped atoms we trapped with the magnet will calibrate down to a lower temperature. Then we kick the next set of hot atoms out and then we keep going. And eventually we start to cool down to just a few nanoKelvin. And at that point all the atoms will start to pile in to a very particular quantum mechanical state, the lowest possible energy state allowed by quantum mechanics for atoms to act in this magnetic potential here. And at this point, where we start to have a whole bunch of particles in the very same quantum mechanical state, that’s when we form these Bose-Einstein condensates. They start to act in lockstep. They face lock. They’re all described by the exact same wave function. And the cloud actually turns super fluid. So now this is kind of the most exciting part of the experiment because now we have this condensate trapped by the magnet. We have brought the five million atoms into it and their size is roughly a hundred microns. And now we can start to shoot laser beams towards this condensate and start to create some slow light.
Now how do we create slow lights? So I have to show you this one because I’ve so far been focusing mostly on the whole motion of the atom and how we try to dim that motion out during a cooling process to get the atoms to a stand still. Now I have to turn my attention to the internal structure of an atom. Now sodium is an alkali atom so that means it has some nucleus and some tightly bound electrons and then a very, very loosely bound electron, the valence electron. And that valence electron, again according to quantum mechanics, could only have very particular energy levels. So those are exactly the energy levels I’m showing here. We see there’s a lowest energy levels here, we call it the one state. And that’s actually where we find all the atoms after the cooling process. But we notice there’s another internal energy level a two state over here and a little bit higher in energy. And what we do with all the atoms in the one state is we shine a laser beam on them, we call it the coupling laser. It’s really a yellow laser beam. It’s tuned to a frequency such that the frequency matches the energy difference between an upper energy state and this state two. And with this coupling laser, we now reengineer the optical properties of the atom cloud in such a way that when we now send a laser pulse in, we call it the probe pulse, its tuned to a frequency matching this transition indicated by blue here. We can now slow this light pulse down dramatically.
What’s happening is that by shinning that coupling laser beam on the Bose-Einstein condensate we are dramatically changing the refractive index of the atom cloud. Now glass for example has a refractive index that’s a little bit larger than one, a little bit larger than what it is in free space. So that means if I send light to a window it laterally slows down a little bit, by about 30%. And you would say “Well gee, if I want a slow light down a lot I should create a medium with a really high refractive index.” no, no, that doesn’t work at all because if I did, I would just create the world’s best mirror and reflect all the light back, I will never get any light into the cloud. So what I should rather do, and that’s what we are doing, is we should maintain refractive index that’s very close to one, close to what it is in free space. But then we generate a very rapid variation of that refractive index as a function of frequency of this whole laser field. And the steeper we can make that slope, the slower we can get a light pulse to go when we send it in to the atom cloud. Because the signal, well what we call the whole velocity of a light pulse, is inversely proportion to that slope. And so if we now turn the intensity of the coping laser way down, we can create a really steep slope and with it incredibly slow light. So with the intensity of the coping laser we can directly control the light speed.
And what’s going on here is we are really playing with quantum mechanics because with these two laser beams we are bringing the atoms into a quantum mechanical super position state. So instead of the atom just being in state one, that’s where we started out, we bring it in to the superposition state of one and two, where the atom is a little bit in one and at the same time a little bit in state two. And we drive it into actually a matching superposition state, that we call the dark state, where we have the exact right amount of the atom in both states. And that right amount is such that the amount of the atom we find in state two related to one is determined by the ratio of the strength of the probe laser relative to the coupling laser. And we talked about that magic superposition state as a dark state. When the atoms find that state, they will not absorb any laser light from any of the laser beam, so we have no losses but the probe pulse will go incredibly slowly. So you can start to see that by forming this superposition states the light pulse will actually start to write it’s information into the atoms by changing their internal states. And that idea is extremely important for all our experiments.
So let me show you the set up that we use, and the typical geometry for slow light. We can run different geometries but this is a typical one. So we have the condensate trapped in the middle of the vacuum chamber. We illuminate it from the side with the coupling laser and then we launch the probe pulse and send that in along the long dimension of the atom clouds. We want to measure the light speed, so we simply sit and wait behind the vacuum chamber for this light pulse to come out. And after it has come out of the system we send a third laser beam up from down below. And now the atoms will create an absorption shadow. And we can image that absorption shadow on to our CCD camera and take a snapshot of the clouds. And I’ll show you an example up here; a cloud that is being cooled to four hundred fifty nanoKelvin is about two hundred microns long or 0.02 millimeter. And if I now send the light pulse into exactly this atom cloud, this is what we will record in our experiment; we have the photo multiplier signal as a function of time in microseconds. We have a blue pulse here, that’s a reference pulse, just recorded with no atoms of the system just to set the zero point of the time axis. And then we let atoms into a system, cool them down, and launch another light pulse. And now get the red pulse out. And notice this guy is delayed by seven microseconds. And that was in a cloud that was only two hundred microns long. Now you just divide those two numbers out. And you immediately get, in this case, a light speed of 32 meters a second. We just slowed light by a factor of ten million. And now by just turning the intensity of the coupling laser further down, we can get the light speed even slower. And we can actually get it down to fifteen miles an hour. And that’s certainly at a point where I can claim that you beat light on your bicycle.
So at this point let me give you a real sense of, before we go on, what it feels like to be a light pulse going through one of our Bose-Einstein condensate. So in this little animation here, we have the condensate illuminated with the coupling laser. Then we send the light pulse into the atom cloud. And at the same time I have a reference pulse running here, just for comparison. So we can see what happens when the light pulse starts to enter the atom cloud. The front of this will enter first, of course. And that will start to slow down. But the back end, that’s still out in free space. So the back end will be running at normal light speed so I’ll start to catch up with the front. And the whole thing will start to compress. And back to the mouse, so now we’ll continue, and the more it slows down the more it compresses. And it shocks along. And now it starts to accelerate, and just stretches back outward when it exits the cloud.
Of course in our experiments, this is an animation done for light speed reaction effect. And then we’re talking in the experiment’s reduction of effect is of ten to a hundred million. So that means the light pulse starts out being typically a kilometer long and compresses to an only twenty microns or 0.02 millimeters, so less than half the thickness of a hair. So even though our atom clouds are small, the light pulse ends up getting even smaller so it fits inside the atom cloud. And I should also point out that the light pulse creates an imprint of itself in that condensate, really like a little holographic imprint. And the reason to this that, within the localized light pulse region, the atoms are transformed in to this superposition that state that I talked about. So in the middle of the light pulse, we have a high intensity. So a lot of the atoms are transferred from that initial one state to a two state. But outside the light pulse the atoms are in their original one state and that imprint follows a long as the light pulse slowly propagate through it. And it exits out.
So now, with all these in mind, let me just give you an example of the kind of fun you can have with this system. These are much more recent experiments, so instead of just creating one condensate of state on atom, let’s create two separate condensates. So they’re really separately created in two different atom traps. They have never seen each other before. And what I will now do is I will send a coupling laser in from the right and then send the light pulse into the left, the first condensate there, and slow that down. And what I mean with that, I have again a little animation here, so coupling laser comes, in light pulse comes in and it slows down, it compresses. And I turned off the coupling laser. The light pulse will then come to a grinding halt and turn itself off. But the information that was in the light pulse will not be lost, because that was already imprinted in the atoms through that little holographic dark imprint. And that imprint will actually stay in the atom cloud but it will start moving now. It will start moving at a speed of about 200 feet an hour. So we can see it exits the first condensate, and moves out in to free space. And we have photographed this whole process in the lab and you see it over here in the right. We start with the initial two condensates of state one atoms. We then send the light pulse into the first condensate, slow it down, compresses it, stop it. And we turned it off. And then 1.3 milliseconds later, sure enough we can see a little matter imprint, a little cloud of atoms out there in free space. So what we have actually created out here in free space, that’s a matter imprint. What we have created out here is really a perfect matter copy, in the form of a little atom pulse, of the light pulse that we have already extinguished in that first condensate. And you might ask “Gee, why is it boomerang shaped?” well the reason is that when the light pulse slows down at this first condensate. This light pulse will slow down most along the sensor line of the condensate where the density is the highest. So the light pulse itself will develop a boomerang shape as it’s slowing down. And that’s perfectly mimicked by that matter copy. And now as time moves on, the matter copy moves on. And then it’ll eventually reach the second condensate. And if we don’t do anything well, the matter copy will go through that second condensate and come out on the other side. Now on the other hand once that matter copy is imbedded in the second condensate. If we then turn on the coupling laser again, low and behold out comes the light pulse. So what we have now done is we have stopped and extinguished the light pulse in one condensate, the first one here. And then we generated it from a completely different condensate in a different location. So let’s look at the whole process here, so coupling laser comes in, light pulse comes into the first condensate, which headed into a matter copy. A matter copy travels across to the second condensate. And then we turn it back into light, and it just moves off.
Now, with the more you think about it, the weirder this is really. Because, here you have two sets of atoms, you have the matter copy and then the second condensate. These two sets of atoms have never seen each other before, so how on earth can they, together, figure out to regenerate this light pulse? The secret is we are dealing with Bose-Einstein condensate. And you can kind of regard the atoms and the matter copy, when we illuminate it with a coupling laser, that the atoms will act as little radiating antennas. But under normal circumstances, these little antennas would be completely randomly faced out, just to each their thing. And just scan some random light, and so there will be no information content to it. Now in the presence of that second condensate that second condensate will face a lot of these antennas, so they all radiate in unison. They cooperate and can regenerate this light pulse so we can get the information back out into the light field. So with this system, we actually have a unique system whereby we can take light, turn it to matter and back into light with absolutely no loss of information. And of course once you have light in matter form it’s easy, really easy, to manipulate. And you could actually grab on to that matter copy, put it on the shelf and then later let it go and turn it back into light. Or while you’re holding on to it, why not start manipulating, massaging it, change its shape, its information content. Whatever changes you make to the matter copy, will be in the light pulse when you turn it back into light. And now you might say “Gee, how long could we hold on to this guy?” well clearly a couple of milliseconds, but lately we can actually hold on to it for several seconds, and we are actually approaching minutes. And to set the scale here, in a matter of just a few seconds light can go back and forth to the moon.
With this system we really have a new paradigm for optical information processing. Of course, we love to encode information in light and send it down optical fibers because we have high bandwidths with high data transfer rates. And with the system we can now take that information, particular notes, turn it in to matter, massage it and really do very powerful processing. And then, once we are done, turn it back in to light and send it down another optical fiber. And I would say that so far we have really only started to touch the tip of the iceberg of this system. And to maybe tie back to the talk just before, we actually have experiments going on in the lab where we are combining nanotubes with whole atoms to try to start to squeeze this whole structure and these kind of properties down to the nanoscale, but I’m out of time here. Thank you.