In the wide weird world of quantum mechanics, the atom laser is out there on the fringe. Instead of photons, it shoots out ultracooled atoms, atoms so cold that they no longer move or interact like particles and instead behave just like waves. Marching coherently, their peaks and troughs all in step, like regular laser light, these wavy atoms form a beam that, unlike laser light, moves so slowly a person can outwalk it.
The beam dribbles out along an arc--like water from a squirt gun. As the beam falls, it accelerates under gravity, and the waves dramatically shorten until they are one-70 000th as long as a light wave.
The first atom laser was built in 1997 by Wolfgang Ketterle and co-workers at the Massachusetts Institute of Technology, in Cambridge. Since then, technophiles have wondered: what can we do with this incredible thing?
How about shrinking lithographic circuit patterns to nearly an atomic scale? Or measuring acceleration so precisely that you could steer a submarine blind from San Francisco to Yokohama? Or detecting gravitational variations so subtle that they show the movement of magma deep inside the earth? These applications won't arrive tomorrow morning, but they're on their way, and they could be big [see sidebar, "What's a Superatom Good For?"].
Thanks to its freakishly short waves, an atom laser can resolve much smaller features than an ordinary laser. To get an idea of just how small, consider today's most exquisitely sensitive measuring stick: the laser interferometer. To make one, you split a laser's beam into two parts that proceed along paths of different length, reflect off mirrors, and finally overlap at a detector.
Because one beam takes a little longer than the other to arrive at the detector, the waves are slightly out of phase. Not all the peaks meet up to reinforce each other in a bright line; some coalesce with troughs and cancel, forming a dark band. The resulting interference pattern turns out to be exquisitely sensitive to phase differences. Keep one mirror steady and accelerate the other, and the pattern will change; measure the change, and you can calculate the acceleration. What you've got is an accelerometer.
An atom laser interferometer would be able to register gravitational fields with great precision. That's because the atoms--unlike the photons in an optical laser interferometer--feel the pull of gravity. Such a detector could pick up even the slight variations in gravitational field strength coming from small hollows in rock formations, like underground oil deposits.
To build an atom laser, you chill atoms of a single kind to near absolute zero and herd them together. This maneuver makes them condense into a quantum mechanical blob in which they all have the same energy and position. The blob is called a Bose-Einstein condensate, because it was predicted in 1924 by Albert Einstein, who extrapolated from the work of Satyendra Nath Bose on the nature of photons.
There's a lot of counterintuitive stuff going on here, but the strangest thing must surely be the way that many atoms occupy a single position. Credit the famous Heisenberg Uncertainty Principle, which states that you can precisely know a particle's position or its energy, but not both. Because we've cooled the atoms to nearly zero, we know their energy very precisely indeed: it's practically zero.
That means we no longer know their position well at all. Instead of sitting in a sphere a few hundredths of a nanometer in diameter, the region in which each atom can exist has ballooned to micrometers in diameter. In that space, millions of them overlap, fall into the same quantum mechanical state, and become one giant superatom.
It took 70 years to create Einstein's blob, because it was necessary to chill the atoms to less than a millionth of a Kelvin--that is, less than a millionth of a degree above absolute zero, the point at which all particle motion stops. The atoms floating in interstellar space are nearly three degrees hotter. In the 1980s, when refrigeration technology finally got good enough to make the goal seem achievable, it sparked a scientific race in the classic style that ended in 1995. That was the year Eric Cornell and Carl Wieman at JILA, a physics laboratory in Boulder, Colo., jointly run by the University of Colorado and the National Institute of Standards and Technology, created the world's--perhaps the universe's--first Bose-Einstein condensate. Four months later, MIT's Ketterle created another. For their efforts, the three men shared the Nobel Prize in Physics in 2001.
Not every atom can become a superatom. Researchers have to pick an element that will resonate at a wavelength that can be produced by an optical laser. One such element is rubidium. According to one method perfected at JILA, scientists place a few grams of rubidium inside a vacuum chamber pierced by six intersecting laser beams [see illustration, ]. Rubidium atoms rise off the sample in a vapor, and the light-pressure of the beams slows the atoms, reducing their effective temperature to about one-10 000th of a Kelvin. The force of the lasers holds the atoms in place at the beams' intersection and away from the room-temperature walls of the chamber.
Next, the scientists turn off the lasers and use a magnetic field to confine the atoms. The field makes the atoms slosh back and forth so that they bang together, transferring momentum at random. Some atoms get more than the average amount of momentum and achieve escape velocity; they leave the trap, carrying energy out of the system. This cools the gas further, much as escaping steam cools your coffee.
As the atoms cool, their velocity approaches zero, their positional uncertainty grows, and they behave less like particles and more like waves. At around a millionth of a Kelvin, the atoms' wave packets are large enough to overlap. Suddenly, the material reorganizes its structure and undergoes a phase transition. The transition is somewhat comparable to that of water when its temperature reaches the point of freezing, boiling, or condensing, except that rather than trying to minimize their energy state, the atoms in the magnetic trap are all falling into the lowest quantum state. And so a Bose-Einstein condensate is formed.
When the magnetic trap is turned off, the superatom falls out of it like a drop of liquid (although it's actually still a gas). Importantly, though, this drop is a wave. It can splash and splatter, but it can also be reflected, refracted, phase-shifted, and focused. The smallest superatom is about 1 micrometer across--10 000 times the diameter of a hydrogen atom--and contains anywhere from 100 to 1000 atoms. The largest ones to date are perhaps half a millimeter in size, with maybe 10 million atoms inside them. They persist for tens of seconds before the heat and chaos of the world tear them apart.
To keep the condensate coming, scientists have found ways to move the stuff out of the production chamber to a reservoir so that they can produce a second batch, and a third. The condensate is conveyed through a waveguide formed by laser light.
The material shares exactly one set of quantum mechanical properties, so it can form a beam whose waves march in step. Ketterle and his colleagues were the first to demonstrate this phenomenon by extracting atoms from a sodium condensate in 1997. This first atom laser was comparable to its optical counterpart but had a lower energy output and a far sparser stream of particles--about a trillionth of the photon output of an optical laser. However, the atoms had a wavelength of about 0.1 nm, versus visible light's wavelength of about 400 to 750 nm.
Apart from supersensitive accelerometers , what other applications for atom lasers are in the works? For starters, if you put an accelerometer into a spin (or just send the atom beam around a circuit), you'll have a gyroscope. Three such gyroscopes will tell you how an aircraft moves along all its axes and, just as important, how it rotates around them. You could figure out which way you are now pointing--something the Global Positioning System cannot tell you.
This system of inertial guidance, a kind of dead-reckoning system, works with the help of a little calculus. You measure the acceleration and rotation rates and then go backward, integrating acceleration twice to get the position and integrating the rotation rate once to get the aircraft's rotation. Do this again and again, and you can draw one vector on the tip of another to plot your progress--all without reference to external phenomena.
The problem with using optical lasers in inertial-guidance systems is the noise in the signal. That noise gets integrated, too, and because integration involves an exponent, the errors compound at a fearsome rate. Even the best military inertial-guidance systems now stay accurate for only a few hours; after that, they will calmly inform you that your airplane is tunneling underground or traveling faster than light.
If you could replace the optical laser with an atom laser, however, the system could in principle stay accurate for years at a time. Thanks to the exponent, wavelengths one-70 000th the length of an optical laser's would yield a signal-to-noise ratio 100 billion times as big. That's why the U.S. Navy is funding work at JILA to build a version that is small and sturdy enough for use in a submarine.
The process of miniaturizing the atom laser--which at present can occupy anywhere from a tabletop to an entire lab filled with vacuum and laser equipment--is well under way, in part because many of the tools have already been developed. For instance, engineers already know how to use electrical fields to control atom beams for the epitaxial deposition of atoms on wafers. Jakob Reichel's Microtrap Group at the Max Planck Institute of Quantum Optics, in Garching, Germany, and at the ecole Normale Superieure, in Paris, was the first to successfully make and move around condensates on a chip, using microscopic magnetic traps to cool and manipulate the atoms.
Other groups are also pursuing these "atom chips," which right now are ungainly, lab-built devices. The chips themselves can fit in the palm of your hand, and the vacuum equipment is not much larger. But when you add in the lasers, the whole setup fills the volume of a few desktop PCs.
Even so, while atom chips cannot yet be commercially manufactured, they're proof that such products are feasible. And when an atom-laser gyroscope can finally fit inside a cellphone, children will ask, "Daddy, how did you ever find your way around when all you had was GPS?"
Indeed, inertial guidance beats GPS not only by telling you which way you are pointing but by actually functioning when you're indoors, underground, underwater, in the canyons of Manhattan, in dense jungle, or in deep space. Those are places where GPS won't work, and they are just where people, and their robots, are the likeliest to get lost.
Another important application for atom lasers is holographic printing. Like optical holography, atom holography gives a true, three-dimensional image of an object, but because of its short wavelength, it can resolve details 70 000 times as fine. And just as atom particles can be made to undergo a phase change to a wavelike state, so can an atom wave revert to particles. That means that the holographic image can deposit itself, like frost forming on a windowpane.
This research is in the fundamental stages right now, but it has exciting possibilities for circuit fabrication. In the future, engineers might be able to deposit a complex integrated circuit pattern, just a few nanometers in size, on a semiconductor surface. Building a circuit like this, an atom at a time, would allow circuit components that are perfect down to an atom.
Conventional holograms are made by splitting a laser beam and sending one half, the reference beam, straight to a piece of film and the other half, the object beam, to the object to be imaged. Because the light bouncing off the object arrives later than the reference beam, it is out of phase with it and hence able to interfere with it. The interference pattern is the hologram, because it can now function as a diffraction grating: shine a laser on it, and it will reconstitute the 3-D image.
So a hologram creates an image by diffracting light the right way. Does that mean that any diffraction grating--a plate or "mask" drilled with thousands of holes--will also act like a hologram? Yes, it will, if it has the right pattern of holes. This is useful in optical holography when you don't want to start with an existing object but instead want to create an image of an object that is in your imagination, like a spaceship. You could use a diffraction grating with the right pattern of holes as a hologram that gives you your spaceship image.
When an atom-laser gyroscope can finally fit inside a cellphone, children will ask, "Daddy, how did you ever find your way around when all you had was GPS?"
This technique is critical to atom holography, because atom waves cannot illuminate an object; the minute they hit something, they revert to particles and stick. (For the same reason, an atom laser beam cannot work in the open air, only in a vacuum.) Instead, you must find a way to perforate a diffraction grating so that it constitutes a hologram.
How do you get the right pattern of holes for your diffraction grating, or holographic mask? You use a mathematical trick known as a Fourier transform, originally developed to describe complex shapes in mathematical terms. Any complex shape may be regarded as the sum of a number of simpler ones, notably sine waves. Such waves are described by a straightforward formula involving frequency and amplitude. A hologram, or any interference pattern created from an object, stores precisely this kind of information about an object, and therefore a hologram may be considered a Fourier transform of the object. Similarly, the 3-D image is the "transform" of the hologram--that is, the solution to its equation.
So if you know the image you want, you can work backward to get the hologram that specifies it. This information can be translated into a pattern for a holographic mask. You can then use electron-beam lithography to punch holes in a grid, making the mask.
Suppose the image you want is the pattern of an integrated circuit. You work backward to get the hologram, then perforate the pattern into a grid to make a mask. Then you shoot an atom laser through the mask. When the Bose-Einstein condensate in the laser beam passes through the mask, it breaks up into wavelets. The wavelets then interfere with each other to create the image of the integrated circuit on the substrate, "frosting" it in the features of the circuit [see illustration, ].
The first images made with atom holography were created by Fujio Shimizu at the University of Electro-Communications, in Tokyo, and his colleagues at the NEC Fundamental Research Laboratories, in Tsukuba, Japan, in 1996. They successfully deposited millimeter-wide letters and even tiny art images.
Since then, Shimizu's researchers have taken holographic masks a step further and developed a neat trick to produce layers in different patterns. Their holographic masks have strips of charged wires around the holes. A potential difference between two strips causes a phase change in the atomic wave, altering the image formed. Switching the charge on different wires thus makes different images without having to drill a new mask each time. The scientists have already built masks with 150-nm precision and are working on designs with lines as fine as 1 nm.
With atom holography, if you were willing to make a circuit out of superatom-suitable elements like rubidium and sodium, you could simply deposit the entire image directly onto the wafer in a single operation, in the solid analog of photolithography. If you needed a material that doesn't easily form a Bose-Einstein condensate--such as gold or aluminum--you could use rubidium atoms to expose a resist. Then you could use the resist to etch gold lines into the wafer with the usual techniques.
In regular patterns like gratings and grids, the lines could be as tiny as one atom wide and positioned to within less than an atom's diameter. This capability would make atom holography especially useful in the production of photonic devices, such as nanoscale Fresnel lenses and diffraction gratings used in optical sensing, telecommunications, and imaging. On a somewhat larger scale--feature sizes of around 5 to 10 nm--these lines can combine to form any desired image, including a complete integrated circuit.
No commercial systems exist yet for making nanoscale circuits with atom holography, but researchers are hopeful that someday they will. And if IBM Corp. was able to show off its atomic-force microscope by spelling out its name in xenon atoms, why shouldn't JILA and MIT go it one better with atom holography? They could imprint a transistor with their logos--or, better yet, a portrait of Einstein and Bose.
About the Author
Wil McCarthy is president of the Programmable Matter Corp., in Lakewood, Colo., which studies materials science applications for quantum dots. He received his B.S. in aerospace engineering from the University of Colorado at Boulder. He has also written many science-fiction novels, including To Crush the Moon, published by Bantam/Random House in March.
"Bose-Einstein Condensation in a Dilute Gas: The First 70 Years and Some Recent Experiments," the Nobel lecture given by Eric A. Cornell and Carl E. Wieman on 8 December 2001, is available online at http://nobelprize.org/physics/laureates/2001/cornell-lecture.html.
Co-Nobelist Wolfgang Ketterle's Nobel lecture, "When Atoms Behave as Waves: Bose-Einstein Condensation and the Atom Laser," is online at http://nobelprize.org/physics/laureates/2001/ketterle-lecture.html. Ketterle's team recently created the first superfluid gas, which is closely related to Bose-Einstein condensation; see "Vortices and superfluidity in a strongly interacting Fermi gas," by M. W. Zwierlein et al., Nature, 23 June 2005, Vol. 435, no. 7045, pp. 1047-1051.
For a discussion of recent advances in creating atom chips, see "Toward Atom Chips," by Jozsef Fortagh and Claus Zimmermann, Science, 11 February 2005, Vol. 307, Issue 5711, pp. 860-61, and "Atom Chips," by Jakob Reichel, Scientific American, February 2005, pp. 48-53.