Skip to main content Skip to main content

Physics for the 21st Century

Manipulating Light Interview with Featured Scientist Lene Hau

Interviewer: How did you get interested in physics?

LENE: Well, I think my first interest was in some sense in mathematics dating back, I think, to first grade and I had a terrific math teacher. I think having a great teacher that really gets you interested in a subject is super, super important.

Then, when I entered university, I ended up studying some mathematics and physics and initially in my studies I thought I would become a mathematician until I started learning about quantum mechanics. That really switched my interest. I thought quantum mechanics was just absolutely fascinating.

Interviewer: What was it about quantum mechanics you found so compelling?

LENE: There are very non-intuitive phenomena in quantum mechanics that are just like weird. It’s like does nature really behave like this? Then there was also that I had this interest in mathematics and the fact that you can sort of use very mathematical tools to figure out what the allowed energy levels for an electron in a hydrogen atom, for example.

This looks like pure math. Can nature possibly behave in this way? This doesn’t seem to be quite possible but yet there are experiments like looking at the spectrum, for example, the radiation spectrum, the kind of light that hydrogen will send out. That really seemed to indicate that there had to be something to these ideas.

So, even though it seemed like pure math, it seemed like nature really behaved that way. And then there were these crazy, non-intuitive phenomena. We know at each point in time a particle has a particular precision and a particular velocity, but now we’re starting to have this idea that maybe the particle is this wave-like light beam and it might be here or it could actually also be there and it’s actually kind of both places at the same time. That I thought was like, wow; that’s pretty amazing.

Interviewer: Is there a simple way you can explain how you stop light?

LENE: Well, basically, I find light absolutely fascinating and there’s nothing that goes faster than light and light travels almost 200,000 miles per second. That’s pretty mind-boggling and really there’s nothing that goes faster that light. If you start to send light through a window—a piece of glass, glass has a refractive index, and that will tend to slow light down just a little bit.

The whole idea of a refractive index is the idea why lenses work, why eye glasses work, that you have a refractive index for materials. In glass it’s a little bit higher than what it is in free space, in a vacuum or in air. So, basically what happens is when we send light through a piece of glass, there will be a bunch of electrons and molecules in making up the glass and what the light beam does is it starts getting these electrons into vibrations.

Now, the fact that these electrons are getting into vibrations they will themselves radiate a little bit of light because they’ll start acting like little antennas. The light that they sort of send out will kind of add together with the light that you send into the window, so the main effect of all of that will be that light slows down a little bit when it goes through a piece of glass just by like 30 percent or so, not very much.

So, in our lab we started getting these ideas that maybe we can start to get light to go really slow, so slow that you can beat it on a bicycle. This whole idea has always totally fascinated me because, again, there is nothing that goes faster than light and if you can somehow control it, tame it, to the extent that you can get it down on a human scale so you can it beat it on a bicycle, I just that that this is fantastic.

Interviewer: So Are you stopping light by increasing the refractive index?

LENE: If you start cranking this refractive index up more and more to slow light down, in the medium what will happen is you will reflect all the light off the medium before it even enters. What you will do in that way is just create the world’s best mirror. It’s not particularly interesting, so we have to do something totally different and we actually do not create a very large refractive index. That’s confusing to a lot of people because they think that’s what we do.

In glass you partly slow it down a little bit because the refractive index is a little higher than what it is in free space, and that means you also reflect the light off the medium before it even gets in there. On top of that, you also have absorption of light in the medium and what that means is that the amount of light that actually enters the medium, you won’t get all of it out on the other side because some of it will absorbed, gobbled up, by the medium.

The way it gets gobbled up is it goes into heat in the medium; it will heat the glass a little bit and you will never get that heat turned back into light. You will never add it again to your light pulse.

Interviewer: Then how is it possible to stop light, isn’t this a crazy idea?

LENE: Yes. That’s exactly what my colleagues were saying. They said, Aren’t you crazy? Basically, I’m paraphrasing. The thing is that initially I was really dying to get my hands on a condensate because this was a totally new state of matter expected to be a super-fluid state of matter. I just wanted to get my hands on this; I was curious to start to poke at it and see how it would react. What sorts of properties does it have?

The best I could think of was to send light into this condensate and light that is particularly kind of dramatic to send in, namely light that has a wavelength or frequency that’s tuned very precisely such that it matches the particular characteristic frequency of the atoms. Once you get this it’s called a resonance condition. Once you get that resonance condition you get a very strong interaction between atom and laser light.

We cool and create condensates out of sodium atoms and what we do is we send a laser light in with a wavelength such that the light is yellow, so roughly 589 nanometers, so it’s a kind of yellow light like from the sodium street lamps. That wavelength is such that is precisely matches that characteristic frequency of the sodium atoms. So, for example, sodium will absorb a lot of light if we send light in with that particular frequency or wavelength, of that yellow light.

So, sending resonant light in to a condensate, that’s particularly dangerous but also particularly interesting because you will get a very strong interaction. For example, sodium photons will absorb light out of this beam very efficiently but, that’s exactly the dangerous situation to be in because here you have these super cooled atoms and if the atom in that atom cloud just absorbed a single photon that’s enough that, from that single photon absorption, that the atom will get a little kick and then get basically kicked right out of the atom cloud, out of the condensate, and we just lose it. Actually, on its way out that atom will start to bang into the other atoms and actually heat the whole thing and the whole condensate will just basically evaporate.

If you start to send the resonant laser light into a condensate then you should just blow the condensate apart. That’s why my colleagues said, “You’re crazy. This is too dangerous.” The thing is, because you get this very dramatic interaction between atoms and laser light you also have a tremendous sensitivity in terms of probing these condensates, and that’s what I was after and really getting this because it’s really a philosophy.

If you want to probe something, probe it as hard as you possibly can without it totally blowing apart. So, don’t ding it a little bit; ding it a lot and then sort of see what happens.

Interviewer: So if your colleagues didn’t think it would work, why did you think it was going to work?

LENE: I thought it was going to work because if you sort of control the parameters correctly you can use this very dramatic interaction to very, very sensitively probe the properties.

It was sort of in this process of trying to probe condensates with this resonant laser beam that we started realizing, gee, what if you have not one but two precisely tuned laser beams with the exact light properties coming in at the right angles with the right wavelengths and all of that? If you have two of those, those two can kind of together do the right things to the atoms such that you actually might be able to slow light down to bicycle speed.

I mean of course it’s one thing to have the idea that something should work but it was a totally different thing to actually to get it to work in the lab. You set out and you set it up according to how you think it should work but, wow, it didn’t quite work as it was supposed to. What could possibly be wrong here?

When things weren’t working you really had to think about all the time what should I change here? What is going on? It was so intense. I almost walked into the shower with my clothes on because we were just thinking about this all the time.

Interviewer: But you didn’t give up. What happened next?

LENE: Then, eventually, we started to see a little bit of slow-down. Of course, that’s like in the middle of the night, like 4:00 am or something, and you’re sort of looking under your telescope and starting to measure light pulses seemingly having slowed down a little bit. But then, of course, we get nervous. Gee, did one of us bump the knob on the oscilloscope? Maybe it’s an artifact. So, we had to do a control experiment, so that means we had to load the atoms into the system again and cool them down and form a condensate and then send another light pulse in and try to see if it slowed down.

That whole process took roughly two minutes but it was like the longest two minutes in my life, I think. Then, eventually we measured that a second time and sure enough it was slowed down. That was really exciting. It was just a little bit of slowdown but it was like “I think we have something here.” And then it was a question of just keep pushing and then we got it down to airplane speed and then, this was in the summer of ’98, and at that point all of a sudden I had to go to Copenhagen to teach a master class.

I didn’t want to go but I had promised to teach that class so I had to go. I remember I was taking off in the airplane from Boston to Copenhagen and following the speed of the airplane on the big screen there and thinking oh, wow; now we are going faster than my light pulse in the lab. I was calculating if I had sent a light pulse off from Boston at the time I left in the airplane I would arrive in Copenhagen an hour before my light pulse.

So, basically I was in Copenhagen for the week and then rushed back to Cambridge to continue with the experiments. Then, roughly a month or two later we started to slow light down to bicycle speed. I remember that night. That was, again, in the middle of the night and you were just sitting there and you’re just the first person in history being in this regime of nature seeing light go this slow. It was really amazing and an amazing feeling and sort of worth all the hard work that had gone before that.

Interviewer: Can you explain in greater detail the stopping of light in the BEC?

LENE: The stop light came after the slow light, yes. So we are slowing light down by factors of 10100,000,000. We’re not talking like 30 percent in a window; it’s a factor of 10100,000,000. So, from 200,000 miles per second we go down to 15 miles an hour and that’s the kind of slow down.

What happens is that together with the slow down the light pulse also spatially compresses by the same factor as it slows down. So, now we have a teeny, weenie little cloud of cold air cooled to a few billionths of a degree above absolute zero. It’s really, really cold, and the cloud is really only 0.1 millimeter in size so it’s a pretty small cloud that we hold in the vacuum chamber.

Then we send the light pulse in and the light pulse is about a mile long when it starts in free space and then we start to send it into our atom cloud. What happens is the front edge will slow down because now that’s starting to enter the atom cloud but the back edge is still out in free space so the tail is out here, and that will keep going at the normal light speed. So, now the back edge will start to catch up to the front edge so you get this light pulse to compress like a little concerting.

As I said, it compressed by the same factor as we slow it down so from one mile to 0.02 millimeters is less than half a thickness of a hair; that’s how small the light pulse ends up being. At that point it fits totally inside the atom cloud, even though the atom cloud is smaller the light pulse ends up getting even smaller so it fits inside. If we then just let it propagate it will propagate very, very slowly at bicycle speed through the atom cloud and eventually it will start coming out at the other side.

At that point the light pulse will start to exit. The front edge takes off, it speeds back off and the light pulse stretches out and ends up with exactly the same length it had to begin with, about a mile. Then it eventually moves on, accelerates back up and moves on at the normal high light speed, but once you have it slowed down and compressed and contained within the atom cloud what the light pulse actually does is it makes a little imprint, like a little holographic imprint in the atom cloud.

So, it actually changes the internal state of the atoms in the condensate when they form this imprint. That imprint follows along the light pulses as it slowly propagates through the atom cloud and now, when we feel like it, we can actually completely stop that light pulse because it’s completely contained in the atom cloud and we can just actually completely stop it, turn the light pulse off and then just hold onto the holographic imprint. Then, we can freeze that imprint within the atom cloud and then later we can decide to turn that hologram back into light. We revive the light pulse and send it on as if nothing had happened.

Interviewer: What is currently motivating your research?

LENE: Well, basically what really motivates our research now is that we have found this method by which we can completely convert light to matter and back to light with no information loss. So, we have a light pulse; we can create a perfect matter copy of that light pulse with exactly the same properties and the same shape, same information content, and there’s no other way by which you can do that.

What it allows us to do now going forward is we can start to create what is called, for example, quantum networks. It’s kind of the quantum mechanical analog of the information, the optical communication, the fiber optic network that we know today where we love to send data around in optical fibers, high data waves and all of that and download it on our computers. But there is sort of a whole different way of doing that where we can start to send quantum states of light around in optical fibers.

For that purpose, light works great. You can encode information in light and send it out in optical information. It’s great for transport but the problem is if you want to control optical information, if you actually want to change the information or you want to manipulate it or you want to process it and you want to determine where I should route this information, all of that requires control of light and that is quantum state of light.

That’s why you want to take light and turn it into matter form because in the matter form you can manipulate it extremely powerfully which you cannot do while it’s in light form. Then, in matter form you can manipulate it incredibly powerfully and then once you are done you turn it back into light and send it down another optical fiber. So those kinds of possibilities are extremely exciting and it’s all opening up a whole realm of possibilities.

Interviewer: Light is used right now to copy and move data. How will Quantum mechanical light manipulation differ from that?

LENE: So, basically what you do when you normally send light pulses down an optical fiber is they come to a router and you can then turn the optical information into electronic signals. The problem is when you do that you lose a great part of the information. You can’t convert all of the information that is in the light pulse into electronic signals, but with the methods that we have developed you can take the light pulse and turn all the information into matter.

So, there’s absolutely no information loss? The amplitude, phase; statistics: how many photons do I have in the light pulse and quantum mechanically I might actually not have an exact number of photons. I might have 1,000 and then at the same time I actually have 1,001 and 1,002 all at the same time. That statistic, as we call it, is preserved when we turn it into matter. Then we can manipulate it in matter form and then turn it back into light so we preserve all the information. We don’t lose anything and there’s simply no other way of doing that because the methods we are using today lose a great part of the information.

Interviewer: What’s the difference in terms of scale wise between your router, the Bose-Einstein condensate router, and an electronic router?

LENE: You can say at the moment we have a roomful of optics and it’s not exactly a practical system, but that’s the kind of thing that if you’re sort of moving into a new regime that’s how everything starts out, a big prototype system. And, of course, then what would happen, perhaps, is you’d say well, gee I can use this for a particular application.

Could we make a system that is much more practical that can do precisely this? Then you’d say well, yeah, maybe we can. Then you’d sort of say which properties should we focus on, which should we optimize? Then you would try to make a practical system around this set of design criteria. That could be something like implementing the experiment we have with little nanostructures like room temperature chip integrated structures. That’s certainly an exciting possibility.

Interviewer: So you’re trying to compute information without destroying the quantum mechanical weirdness?

LENE: Yes! That’s correct, exactly. You could say we are trying to make a computer. What does a computer consist of? A computer consists of two main ingredients. It has a memory and it has a processor. So, we have to be able to store and hold onto the optical information without destroying any information and, at the same time, we have to be able to process and change it in a controlled fashion.

We want to do both and with these latest results where we turn light into a matter copy and then can start to change the matter copy and then turn it back into light, we can create a process of optical information where we do not destroy any of the quantum information in the light pulses. We preserve them into its matter form and then process that and then turn it back into light.

Interviewer: Where are we at when it comes to Bose-Einstein condensates and manipulating light? Are we at the very beginning of this?

LENE: Yeah. I think we have only seen the tip of the iceberg really because what is really exciting is that for the first time you have light that you can turn into matter form with no loss of information. Light is fantastic for encoding information and transporting that information. But once you have the light information that you turn into matter form, you have extremely powerful processing methods.

That’s when you can start to manipulate, process the information and once you are done processing you turn it back into light. That’s an extremely powerful set-up and absolutely unique. There is no other way where you can turn light into matter and back into light and do the processing in between and do the whole thing on classical states of light, quantum states of light. It’s extremely powerful.

This is a totally new system where we have some absolutely new paradigms that we can pursue. We are just at the beginning of that, I think, very exciting set of possibilities in this whole new area.

Series Directory

Physics for the 21st Century


Produced by the Harvard-Smithsonian Center for Astrophysics Science Media Group in association with the Harvard University Department of Physics. 2010.
  • Closed Captioning
  • ISBN: 1-57680-891-2