Physics for the 21st Century logo

Section 6: Implementing These Ideas in the Laboratory

Experimental realization of slow light.

Figure 13: Experimental realization of slow light.

Source: © Reprinted by permission from Macmillan Publishers Ltd: Nature 397, 594-598 (18 February 1999). More info

Figure 13 shows a typical setup that we actually use for slowing light. We hold the atom cloud fixed in the middle of the vacuum chamber with use of the electromagnet and first illuminate the atom cloud from the side with the coupling laser. Then we send the probe light pulse into the cooled atoms. We can make a very direct measurement of the light speed: We simply sit and wait behind the vacuum chamber for the light pulse to come out, and measure the arrival time of the light pulse. To figure out the light speed, we just need to know the length of the cloud. For this purpose, we send a third "imaging" laser beam into the chamber from down below, after the probe pulse has propagated through the atom cloud and the coupling laser is turned off.

Measurement of ultra-slow light in a cold atom cloud.

Figure 14: Measurement of ultra-slow light in a cold atom cloud.

Source: © Reprinted by permission from Macmillan Publishers Ltd: Nature 397, 594-598 (18 February 1999). More info

As the imaging laser is tuned on resonance with the atom's characteristic frequency, and there is only one laser beam present (i.e., there is no quantum interference), the atoms will absorb photons and create an absorption shadow in the imaging beam. By taking a photograph of this shadow with a camera, we can measure the length of the cloud. An example is seen in Figure 13 where the shadow (and the cloud) has a length of 0.008 inches. By sending a light pulse through precisely this cloud, we record the red light pulse in Figure 14. It takes the pulse 7 microseconds (7 millionths of a second) to get through the cloud. We now simply divide the cloud length by the propagation time to obtain the light speed of the light pulse: 71 miles/hour. So, we have already slowed light by a factor of 10 million. We can lower the intensity of the coupling beam further and propagate the pulse through a basically pure Bose-Einstein condensate to get to even lower light speeds of 15 miles/hour. At this point, you can easily beat light on your bicycle.

Figure 15 illustrates how the light pulse slows in the atom cloud: associated with the slowdown is a huge spatial compression of the light pulse. Before we send the pulse into the atom cloud, it has a length of roughly a mile [its duration is typically a few microseconds, and by multiplying the duration by the light speed in free space (186,000 miles per second), we get the length]. As the light pulse enters the cloud, the front edge slows down; but the back edge is still in free space, so that end will be running at the normal light speed. Hence, it will catch up to the front edge and the pulse will start to compress. The pulse is compressed by the same factor as it is slowed, so in the end, it is less than 0.001" long, or less than half the thickness of a hair. Even though our atom clouds are small, the pulse ends up being even smaller, small enough to fit inside an atom cloud. The light pulse also makes an imprint in the atom cloud, really a little holographic imprint. Within the localized pulse region, the atoms are in these dark states we discussed earlier. The spatial modulation of the dark state mimics the shape of the light pulse: in the middle of the pulse where the electric field of the probe laser is high, a large fraction of an atom's state is transferred from 1 to 2. At the fringe of the pulse, only a small fraction is in 2; and outside the light pulse, an atom is entirely in the initial state 1. The imprint follows the light pulse as it slowly propagates through the cloud. Eventually, the light pulse reaches the end of the cloud and exits into free space where the front edge takes off as the pulse accelerates back up to its normal light speed. In the process, the light pulse stretches out and regains the length it had before it entered the medium.

It is interesting to note that when the light pulse has slowed down, only a small amount of the original energy remains in the light pulse. Some of the missing energy is stored in the atoms to form the holographic imprint, and most is sent out into the coupling laser field. (The coupling photons already there stimulate new photons to join and generate light with the same wavelength, direction, and phase). When the pulse leaves the medium, the energy is sucked back out of the atoms and the coupling beam, and it is put back into the probe pulse.

Figure 15: Light slows down in a Bose-Einstein condensate.

Source: © Chien Liu and Lene V. Hau. More info

Try to run the animation again and click on the screen when the pulse is slowed down, compressed, and contained in the atom cloud: The light pulse stops. You might ask: Could we do the same in the lab? The answer is: Yes, indeed, and it is just as easy—just block the coupling laser. As mentioned earlier, when the coupling intensity decreases, the light speed also decreases, and the light pulse comes to a halt. The atoms try to maintain the dark state. If the coupling laser turns off, the atoms will try to absorb probe light and emit some coupling light, but the light pulse runs empty before anything really changes. So, the light pulse turns off but the information that was in the pulse is not lost: it is preserved in the holographic imprint that was formed and that stays in the atom cloud.

Before we move on, it is interesting to pause for a moment. The slowing of light to bicycle speed in a Bose-Einstein condensate in 1999 stimulated a great many groups to push for achieving slow light. In some cases, it also stopped light in all kinds of media, such as hot gases, cooled solids, room temperature solids, optical fibers, integrated resonators, photonic crystals, and quantum wells, and with classical or quantum states of light. The groups include those of Scully at Texas A&M, Budker at Berkeley, Lukin and Walsworth at Harvard and CFA, Hemmer at Hanscom AFB, Boyd at Rochester, Chuang at UIUC, Chang-Hasnain at Berkeley, Kimble at Caltech (2004), Kuzmich at Georgia Tech, Longdell et al. in Canberra, Lipson at Cornell, Gauthier at Duke, Gaeta at Cornell, Vuckovic at Stanford, Howell at Rochester, and Mørk in Copenhagen, to name just a few.