The Bose-Einstein Condensate Three years ago in a Colorado laboratory, scientists realized a long-standing dream, bringing the quantum world closer to the one of everyday experience

by Eric A. Cornell and Carl E. Wieman

n June 1995 our research group at long quest to prove Einstein’s theory. But the Joint Institute for Laboratory most of the fascination now stems from Astrophysics (now called JILA) in the fact that the condensate offers a ATOMIC TRAP cools by means of two I different mechanisms. First, six laser Boulder, Colo., succeeded in creating a macroscopic window into the strange minuscule but marvelous droplet. By world of quantum mechanics, the theory beams (red) cool atoms, initially at room cooling 2,000 rubidium atoms to a of matter based on the observation that temperature, while corralling them toward temperature less than 100 billionths of elementary particles, such as electrons, the center of an evacuated glass box. Next, the laser beams are turned off, and the a degree above absolute zero (100 bil- have wave properties. Quantum me- magnetic coils (copper) are energized. Cur- lionths of a degree kelvin), we caused chanics, which encompasses the famous rent flowing through the coils generates a the atoms to lose for a full 10 seconds Heisenberg uncertainty principle, uses magnetic field that further confines most their individual identities and behave as these wavelike properties to describe of the atoms while allowing the energetic though they were a single “superatom.” the structure and interactions of matter. ones to escape. Thus, the average energy The atoms’ physical properties, such as We can rarely observe the effects of of the remaining atoms decreases, making their motions, became identical to one quantum mechanics in the behavior of the sample colder and even more closely another. This Bose-Einstein condensate a macroscopic amount of material. In confined to the center of the trap. Ulti- (BEC), the first observed in a gas, can ordinary, so-called bulk matter, the in- mately, many of the atoms attain the low- be thought of as the matter counterpart coherent contributions of the uncount- est possible energy state allowed by quan- of the laser—except that in the conden- ably large number of constituent parti- tum mechanics and become a single entity sate it is atoms, rather than photons, cles obscure the wave nature of quan- known as a Bose-Einstein condensate. that dance in perfect unison. tum mechanics, and we can only infer its Our short-lived, gelid sample was the effects. But in Bose condensation, the experimental realization of a theoretical wave nature of each atom is precisely in construct that has intrigued scientists phase with that of every other. Quan- ever since it was predicted some 73 years tum-mechanical waves extend across ago by the work of physicists Albert Ein- the sample of condensate and can be stein and Satyendra Nath Bose. At ordi- observed with the naked eye. The sub- nary temperatures, the atoms of a gas microscopic thus becomes macroscopic. are scattered throughout the container holding them. Some have high energies New Light on Old Paradoxes (high speeds); others have low ones. Ex- panding on Bose’s work, Einstein he creation of Bose-Einstein con- showed that if a sample of atoms were Tdensates has cast new light on long- cooled sufficiently, a large fraction of standing paradoxes of quantum me- them would settle into the single lowest chanics. For example, if two or more possible energy state in the container. In atoms are in a single quantum-mechan- mathematical terms, their individual ical state, as they are in a condensate, it wave equations—which describe such is fundamentally impossible to distin- physical characteristics of an atom as guish them by any measurement. The its position and velocity—would in ef- two atoms occupy the same volume of fect merge, and each atom would be- space, move at the identical speed, scat- come indistinguishable from any other. ter light of the same color and so on. Progress in creating Bose-Einstein con- Nothing in our experience, based as densates has sparked great interest in the it is on familiarity with matter at nor- physics community and has even gener- mal temperatures, helps us comprehend ated coverage in the mainstream press. this paradox. That is because at normal At first, some of the attention derived temperatures and at the size scales we from the drama inherent in the decades- are all familiar with, it is possible to de-

40 Scientific American March 1998 Copyright 1998 Scientific American, Inc. The Bose-Einstein Condensate scribe the position and motion of each we can expect to find the atom. As a er distinguish one atom from another. and every object in a collection of ob- collection of atoms becomes colder, the The current excitement over these jects. The numbered Ping-Pong balls size of each wave packet grows. As long condensates contrasts sharply with the bouncing in a rotating drum used to se- as each wave packet is spatially sepa- reaction to Einstein’s discovery in 1925 lect lottery numbers exemplify the mo- rated from the others, it is possible, at that they could exist. Perhaps because tions describable by classical mechanics. least in principle, to tell atoms apart. of the impossibility then of reaching the At extremely low temperatures or at When the temperature becomes suffi- required temperatures—less than a mil- small size scales, on the other hand, the ciently low, however, each atom’s wave lionth of a degree kelvin—the hypothe- usefulness of classical mechanics begins packet begins to overlap with those of sized gaseous condensate was consid- to wane. The crisp analogy of atoms as neighboring atoms. When this happens, ered a curiosity of questionable validity Ping-Pong balls begins to blur. We can- the atoms “Bose-condense” into the and little physical significance. For per- not know the exact position of each lowest possible energy state, and the spective, even the coldest depths of in- atom, which is better thought of as a wave packets coalesce into a single, mac- tergalactic space are millions of times blurry spot. This spot—known as a wave roscopic packet. The atoms undergo a too hot for Bose condensation. packet—is the region of space in which quantum identity crisis: we can no long- In the intervening decades, however, MICHAEL GOODMAN

The Bose-Einstein Condensate Copyright 1998 Scientific American, Inc. Scientific American March 1998 41 Bose condensation came back into fash- putting the “super” in superfluidity. than the rate for hydrogen atoms. These ion. Physicists realized that the concept Not until the late 1970s did refrigera- much larger atoms bounce off one an- could explain superfluidity in liquid he- tion technology advance to the point other at much higher rates, sharing en- lium, which occurs at much higher tem- that physicists could entertain the no- ergy among themselves more quickly, peratures than gaseous Bose condensa- tion of creating something like Einstein’s which allows the condensate to form tion. Below 2.2 kelvins, the viscosity of original concept of a BEC in a gas. Lab- before clumping can occur. liquid helium completely disappears— oratory workers at M.I.T., the University Also, it looked as if it might be rela- of Amsterdam, the University of British tively easy and inexpensive to get these Columbia and Cornell University had atoms very cold by combining ingenious to confront a fundamental difficulty. To techniques developed for laser cooling achieve such a BEC, they had to cool and trapping of alkali atoms with the the gas to far below the temperature at techniques for magnetic trapping and which the atoms would normally freeze evaporative cooling developed by the re- into a solid. In other words, they had to searchers working with hydrogen. These create a supersaturated gas. Their ex- ideas were developed in a series of dis- pectation was that hydrogen would su- cussions with our friend and former persaturate, because the gas was known teacher, Daniel Kleppner, the co-leader to resist the atom-by-atom clumping of a group at M.I.T. that is attempting that precedes bulk freezing. to create a condensate with hydrogen. Although these investigators have not Our hypothesis about alkali atoms yet succeeded in creating a Bose-Ein- was ultimately fruitful. Just a few stein condensate with hydrogen, they months after we succeeded with rubidi- did develop a much better understand- um, Wolfgang Ketterle’s group at M.I.T. ing of the difficulties and found clever produced a Bose condensate with sodi- approaches for attacking them, which um atoms; since that time, Ketterle’s benefited us. In 1989, inspired by the team has succeeded in creating a con- hydrogen work and encouraged by our densate with 10 million atoms. At the own research on the use of lasers to trap time of this writing, there are at least and cool alkali atoms, we began to sus- seven teams producing condensates. pect that these atoms, which include ce- Besides our own group, others working sium, rubidium and sodium, would with rubidium are Daniel J. Heinzen of make much better candidates than hy- the University of Texas at Austin, Ger- drogen for producing a Bose conden- hard Rempe of the University of Kon- sate. Although the clumping properties stanz in Germany and Mark Kasevich of cesium, rubidium and sodium are not of Yale University. In sodium, besides superior to those of hydrogen, the rate Ketterle’s at M.I.T., there is a group led at which those atoms transform them- by Lene Vestergaard Hau of the Row- selves into condensate is much faster land Institute for Science in Cambridge,

EVAPORATIVE COOLING occurs in a magnetic trap, which can be thought of as a deep bowl (blue). The most energetic atoms, depicted with the longest green trajectory arrows, escape from the bowl (above, left). Those that remain collide with one another frequently, apportioning out the remaining energy (left). Eventually, the atoms move so slowly and are so closely packed at the bottom of the bowl that their quantum nature becomes more pronounced. So-called wave packets, representing the region where each atom is likely to be found, become less distinct and begin to overlap (below, left). Ulti- mately, two atoms collide, and one is left as close to stationary as is allowed by Heisen- berg’s uncertainty principle. This event triggers an avalanche of atoms piling up in the lowest energy state of the trap, merging into the single ground-state blob that is a Bose- Einstein condensate (below, center and right).

Copyright 1998 Scientific American, Inc. MICHAEL GOODMAN LASER COOLING of an atom makes use of the pressure, or force, exerted by re- peated photon impacts. An atom moving FORCE against a laser beam encounters a higher frequency than an atom moving with the VELOCITY same beam. In cooling, the frequency of the beam is adjusted so that an atom moving into the beam scatters many more photons than an atom moving away from the beam. The net effect is to reduce the speed and thus cool the atom.

FORCE net and thus is subjected to a force when placed in a magnetic field [see illustra- VELOCITY tion on opposite page]. By carefully con- trolling the shape of the magnetic field and making it relatively strong, we can

MICHAEL GOODMAN use the field to hold the atoms, which move around inside the field much like Mass. At Rice University Randall G. the laser light relative to the frequency balls rolling about inside a deep bowl. Hulet has succeeded in creating a con- of the light absorbed by the atoms [see In evaporative cooling, the most ener- densate with lithium. illustration above]. getic atoms escape from this magnetic All these teams are using the same In this setup, we use laser light not bowl. When they do, they carry away basic apparatus. As with any kind of only to cool the atoms but also to “trap” more than their share of the energy, refrigeration, the chilling of atoms re- them, keeping them away from the leaving the remaining atoms colder. quires a method of removing heat and room-temperature walls of the cell. In The analogy here is to cooling coffee. also of insulating the chilled sample fact, the two laser applications are sim- The most energetic water molecules leap from its surroundings. Both functions ilar. With trapping, we use the radiation out of the cup into the room (as steam), are accomplished in each of two steps. pressure to oppose the tendency of the thereby reducing the average energy of In the first, the force of laser light on the atoms to drift away from the center of the liquid that is left in the cup. Mean- atoms both cools and insulates them. In the cell. A weak magnetic field tunes the while countless collisions among the re- the second, we use magnetic fields to in- resonance of the atom to absorb prefer- maining molecules in the cup apportion sulate, and we cool by evaporation. entially from the laser beam that is out the remaining energy among all pointing toward the center of the cell those molecules. Our cloud of magneti- Laser Cooling and Trapping (recall that six laser beams intersect at cally trapped atoms is at a much lower the center of the cell). The net effect is density than water molecules in a cup. he heart of our apparatus is a small that all the atoms are pushed toward So the primary experimental challenge Tglass box with some coils of wire one spot and are held there just by the we faced for five years was how to get around it [see illustration on pages 40 force of the laser light. the atoms to collide with one another and 41]. We completely evacuate the These techniques fill our laser trap in enough times to share the energy before cell, producing in effect a superefficient one minute with 10 million atoms cap- they were knocked out of the trap by a thermos bottle. Next, we let in a tiny tured from the room-temperature ru- collision with one of the untrapped, amount of rubidium gas. Six beams of bidium vapor in the cell. These trapped room-temperature atoms remaining in laser light intersect in the middle of the atoms are at a temperature of about 40 our glass cell. box, converging on the gas. The laser millionths of a degree above absolute Many small improvements, rather light need not be intense, so we obtain zero—an extraordinarily low tempera- than a single breakthrough, solved this it from inexpensive diode lasers, similar ture by most standards but still 100 problem. For instance, before assem- to those found in compact-disc players. times too hot to form a BEC. In the pres- bling the cell and its connected vacuum We adjust the frequency of the laser ence of the laser light, the unavoidable pump, we took extreme care in cleaning radiation so that the atoms absorb it random jostling the atoms receive from each part, because any remaining resi- and then reradiate photons. An atom the impact of individual light photons dues from our hands on an inside sur- can absorb and reradiate many millions keeps the atoms from getting any cold- face would emit vapors that would de- of photons each second, and with each er or denser. grade the vacuum. Also, we made sure one, the atom receives a minuscule kick To get around the limitations imposed that the tiny amount of rubidium vapor in the direction the absorbed photon is by those random photon impacts, we remaining in the cell was as small as it moving. These kicks are called radiation turn off the lasers at this point and acti- could be while providing a sufficient pressure. The trick to laser cooling is to vate the second stage of the cooling pro- number of atoms to fill the optical trap. get the atom to absorb mainly photons cess. This stage is based on the magnet- Incremental steps such as these helped that are traveling in the direction oppo- ic-trapping and evaporative-cooling but still left us well shy of the density site that of the atom’s motion, thereby technology developed in the quest to needed to get the evaporative cooling slowing the atom down (cooling it, in achieve a condensate with hydrogen under way. The basic problem was the other words). We accomplish this feat atoms. A magnetic trap exploits the fact effectiveness of the magnetic trap. Al- by carefully adjusting the frequency of that each atom acts like a tiny bar mag- though the magnetic fields that make

The Bose-Einstein Condensate Copyright 1998 Scientific American, Inc. Scientific American March 1998 43 up the confining magnetic “bowl” can atoms scatter this light out of the beam, exact center of the trap and had exactly be quite strong, the little “bar magnet” casting a shadow that we observe with zero energy. According to the uncer- inside each individual atom is weak. a video camera. From this shadow, we tainty principle, we cannot know both This characteristic makes it difficult to can determine the distribution of veloc- these things simultaneously. push the atom around with a magnetic ities of the atoms in the original trapped Einstein’s theory requires that the field, even if the atom is moving quite cloud. The velocity measurement also atoms in a condensate have energy that slowly (as are our laser-cooled atoms). gives us the temperature of the sample. is as low as possible, whereas Heisen- In 1994 we finally confronted the need In the plot of the velocity distribution berg’s uncertainty principle forbids them to build a magnetic trap with a narrow- [see illustration on opposite page], the from being at the very bottom of the er, deeper bowl. Our quickly built, nar- condensate appears as a dorsal-fin- trap. Quantum mechanics resolves this row-and-deep magnetic trap proved to shaped peak. The condensate atoms conflict by postulating that the energy be the final piece needed to cool evap- have the smallest possible velocity and of an atom in any container, including oratively the rubidium atoms into a thus remain in a dense cluster in the our trap, can only be one of a set of dis- condensate. As it turns out, our partic- center of the cloud after it has expand- crete, allowed values—and the lowest ular trap design was hardly a unique ed. This photograph of a condensate is of these values is not quite zero. This solution. Currently there are almost as further proof that there is something lowest allowed energy is called the zero- many different magnetic trap configu- wrong with classical mechanics. The point energy, because even atoms whose rations as there are groups studying condensate forms with the lowest pos- temperature is exactly zero have this these condensates. sible energy. In classical mechanics, minimum energy. Atoms with this ener- “lowest energy” means that the atoms gy move around slowly near—but not Shadow Snapshot of a “Superatom” should be at the center of the trap and quite at—the center of the trap. The un- motionless, which would appear as an certainty principle and the other laws of ow do we know that we have in infinitely narrow and tall peak in our im- quantum mechanics are normally seen Hfact produced a Bose-Einstein con- age. The peak differs from this classical only in the behavior of submicroscopic densate? To observe the cloud of cooled conception because of quantum effects objects such as a single atom or smaller. atoms, we take a so-called shadow snap- that can be summed up in three words: The Bose-Einstein condensate therefore shot with a flash of laser light. Because Heisenberg’s uncertainty principle. is a rare example of the uncertainty prin- the atoms sink to the bottom of the mag- The uncertainty principle puts limits ciple in action in the macroscopic world. netic bowl as they cool, the cold cloud is on what is knowable about anything, Bose-Einstein condensation of atoms too small to see easily. To make it larg- including atoms. The more precisely is too new, and too different, for us to er, we turn off the confining magnetic you know an atom’s location, the less say if its usefulness will eventually ex- fields, allowing the atoms to fly out free- well you can know its velocity, and vice tend beyond lecture demonstrations for ly in all directions. After about 0.1 sec- versa. That is why the condensate peak quantum mechanics. Any discussion of ond, we illuminate the now expanded is not infinitely narrow. If it were, we practical applications for condensates cloud with a flash of laser light. The would know that the atoms were in the must necessarily be speculative. Never-

50 NANOKELVIN: CONTINENT-WIDE THERMOMETER is 4,100 kilometers (2,548 miles) APPROXIMATE long and shows how low the temperature must be before a condensate can 0.683 TEMPERATURE OF 0.7 0.6 form. With zero degrees kelvin in Times Square in New York City and 300 de- NEARLY PURE grees assigned to City Hall in Los Angeles, room temperature corresponds to CONDENSATE San Bernardino, Calif., and the temperature of air, frozen solid, to Terre Haute, 400 NANOKELVIN: Ind. The temperature of a nearly pure condensate is a mere 0.683 millimeter BOSE-EINSTEIN from the thermometer’s zero point. CONDENSATE 0.5 FIRST APPEARS mm IN RUBIDIUM 0.02 KELVIN: 300 KELVIN LOWEST-TEMPERATURE 5.4 293 KELVIN: HELIUM REFRIGERATOR ROOM 6 5 4 3 2 1 0 TEMPERATURE 3 KELVIN: INTERGALACTIC MILLIMETERS SPACE 77 KELVIN: 273 KELVIN: AIR FREEZES 4 KELVIN: WATER FREEZES 216 KELVIN: HELIUM SOUTH POLE LIQUEFIES WINTERTIME TIMES SQUARE BUILDING 42ND STREET & BROADWAY NEW YORK CITY ARIZONA- RARITAN, N.J. LOS ANGELES CALIFORNIA TAOS, N.M. TERRE HAUTE, IND. BORDER SAN BERNARDINO PLAINFIELD, N.J.

41ST STREET & 8TH AVENUE NEW YORK CITY MARK GERBER

44 Scientific American March 1998 Copyright 1998 Scientific American, Inc. The Bose-Einstein Condensate theless, our musings can be guided by a SHADOW IMAGE of a forming Bose-Einstein condensate was processed by a com- striking physical analogy: the atoms puter to show more clearly the distribution of velocities of atoms in the cold cloud. Top that make up a Bose condensate are in and bottom images show the same data but from different angles. In the upper set, many ways the analogue to the photons where the surface appears highest corresponds to where the atoms are the most closely that make up a laser beam. packed and are barely moving. Before the condensate appears (left), the cloud, at about 200 billionths of a degree kelvin, is a single, relatively smooth velocity distribution. Af- The Ultimate in Precise Control? ter further cooling to 100 billionths of a degree, the conden- sate appears as a region of almost sta- very photon in a laser beam travels tionary atoms in the center of the distri- bution (center). After still more cooling, Ein exactly the same direction and only condensate remains (right). has the same frequency and phase of oscillation. This property makes laser light very easy to control precisely and leads to its utility in compact-disc players, laser printers and other ap- pliances. Similarly, Bose condensation represents the ultimate in precise con- trol—but for atoms rather than photons. The matter waves of a Bose condensate can be reflected, focused, diffracted and modulated in frequency and amplitude. This kind of control will very likely lead to improved timekeeping; the world’s best clocks are already based on the os- cillations of laser-cooled atoms. Appli- A cations may also turn up in other areas. JIL In a flight of fancy, it is possible to imagine a beam of atoms focused to a spot only a millionth of a meter across, “airbrushing” a transistor directly onto an integrated circuit. But for now, many of the properties both a fluid and a stripes of high density and low density. of the Bose-Einstein condensate remain coherent wave is go- Our group has looked at how the in- unknown. Of particular interest is the ing to exhibit behav- teractions between the atoms distort condensate’s viscosity. The speculation ior that is rich, which the shape of the atom cloud and the now is that the viscosity will be vanish- is a physicist’s way of saying that it is manner in which it quivers after we ingly small, making the condensate a going to take a long time to figure out. have “poked” it gently with magnetic kind of “supergas,” in which ripples and Meanwhile many groups have begun fields. A number of other teams are swirls, once excited, will never damp a variety of measurements on the con- now devising their own experiments to down. Another area of curiosity centers densates. In a lovely experiment, Ketter- join in this work. on a basic difference between laser light le’s group has already shown that when As the results begin to accrue from and a condensate. Laser beams are non- two separate clouds of Bose condensate these and other experiments over the interacting—they can cross without af- overlap, the result is a fringe pattern of next several years, we will improve our fecting one another at all. A condensate, alternating constructive and destructive understanding of this singular state of on the other hand, has some resistance interference, just as occurs with inter- matter. As we do, the strange, fascinat- to compression and some springiness— secting laser radiation. In the atom cloud, ing quantum-mechanical world will it is, in short, a fluid. A material that is these regions appear respectively as come a little bit closer to our own. SA

The Authors Further Reading

ERIC A. CORNELL and CARL E. WIEMAN are both fel- New Mechanisms for Laser Cooling. William D. Phillips and Claude lows of JILA, the former Joint Institute for Laboratory Astro- Cohen-Tannoudji in Physics Today, Vol. 43, pages 33–40; October 1990. physics, which is staffed by the National Institute of Stan- Observation of Bose-Einstein Condensation in a Dilute Atomic dards and Technology (NIST) and the University of Colorado. Vapor. M. H. Anderson, J. R. Ensher, M. R. Matthews, C. E. Wieman Cornell, a physicist at NIST and a professor adjoint at the uni- and E. A. Cornell in Science, Vol. 269, pages 198–201; July 14, 1995. versity, was co-leader, with Wieman, of the team at JILA that Bose-Einstein Condensation. Edited by A. Griffin, D. W. Snoke and produced the first Bose-Einstein condensate in a gas. Wieman, S. Stringari. Cambridge University Press, 1995. a professor of physics at the university, is also known for his Observation of Interference between Two Bose Condensates. studies of the breakdown of symmetry in the interactions of M. R. Andrews, C. G. Townsend, H.-J. Miesner, D. S. Durfee, D. M. Kurn elementary particles. The authors would like to thank their and W. Ketterle in Science, Vol. 275, pages 637–641; January 31, 1997. colleagues Michael Anderson, Michael Matthews and Jason Bose-Einstein Condensation. World Wide Web site available at http:// Ensher for their work on the condensate project. www.colorado.edu/physics/2000/bec

The Bose-Einstein Condensate Copyright 1998 Scientific American, Inc. Scientific American March 1998 45