NEWS AND ANALYSIS The revolution that has not stopped A decade after the creation of the first Bose–Einstein condensate, research into the properties of ultracold atoms is about to enter a new era. Peter Rodgers reports

At 10.54 a.m. on Monday 5 June 1995 a JILA tween atoms play a central role in the cooling group of physicists at the JILA laboratory of atoms, but the exclusion principle pre- in Boulder, Colorado, created something vents identical fermions getting too close that had never been seen before – a Bose– to one another. One way to overcome this Einstein condensate made of atoms. Eric problem is to mix bosons and fermions in the Cornell, Carl Wieman and colleagues had same trap and rely on collisions between the made a new state of matter by cooling a gas two species – which are not forbidden by the of rubidium-87 atoms to a temperature of exclusion principle – to do the cooling. This just 170 nK. This was so cold that the de approach is known as sympathetic cooling. Broglie wavelength of the atoms was com- However, the absence of collisions can be parable with the distance between them, an advantage in some experiments, accord- causing the atoms to condense into the same In the beginning – the first Bose-Einstein ing to Massimo Inguscio of the University quantum ground state. The breakthrough condensate was created at the JILA lab in 1995. of Florence in Italy. When using beams of in Boulder kick-started an intense period ultracold atoms to measure small forces in of research into the properties of ultracold further by producing the first Fermi conden- interferometer experiments, for instance, atoms that continues to this day, and has sate, again in potassium-40 (see Physics World fermionic atoms give better results because implications for topics as diverse as super- March 2004 pp23–24). A Fermi condensate collisions between bosons reduce the pre- fluidity and the early universe. might seem like a contradiction in terms but cision of the measurements. Within months of the first observation, it is not: two fermionic atoms can join forces Bose–Einstein condensation had also been to form a molecule or the atomic equivalent Building bridges seen in lithium-7 by Randy Hulet’s group of a Cooper pair, just like electrons form Most researchers working on ultracold at Rice University in Texas and in sodium- pairs in the Bardeen–Cooper–Schrieffer atoms believe that the next big goals in the 23 by Wolfgang Ketterle and co-workers at (BCS) theory of superconductivity.The mo- field are to make connections with other the Massachusetts Institute of Technology lecules and the pairs both behave as bosons areas of physics. The observation of super- (MIT). And by the time Cornell, Ketterle because they both have integer spin. fluidity in a Fermi gas is widely seen as the and Wieman shared the 2001 Nobel Prize Most of these experiments rely on a phe- next milestone, and it is common knowledge for Physics, condensates had also been seen nomenon known as a Feshbach resonance in the community that Ketterle and co- in hydrogen, helium and potassium-41. to control the interactions between the workers have just seen conclusive evidence Moreover, a whole new frontier of research atoms. First demonstrated in condensates for this in the form of quantized vortices in a was opening up – the study of ultracold by Ketterle and co-workers in 1998, the rotating gas of lithium-6 atoms. This should Fermi gases. Feshbach resonance has become an essen- confirm less direct evidence for superfluidity “What has happened in the past year with tial technique in many experiments on seen at a number of other labs. Fermi gases has only been topped by the dis- ultracold atomic gases, including the con- Another challenge is to explore the cross- covery of the first condensates in 1995,” says densation of molecules. “In the early days over between the BEC region in which mo- Ketterle. “But this is just the beginning – of BEC the strength of the interactions lecules are condensed and the BCS regime research into Fermi gases is now connecting between the atoms was fixed,” says Rudolf in which pairs of atoms are condensed (see to a long list of intellectual challenges in con- Grimm of the University of Innsbruck in Physics World March pp43–47). “A number densed-matter physics, such as superfluidity, superconductivity,magnetism and so on.” “Fermi gases have been a Fermions and bosons All atoms, indeed all particles, fall into one of two fundamental classes depending on rich source of new results.” the value of their intrinsic angular momen- tum or “spin” in quantum units – atoms Austria. “The discovery of Feshbach reson- of experiments are under way but we need 1 3 with spins of /2, /2 and so on are fermions, ances changed this completely and led to a some better theory of the phenomena to whereas those with integer spins are bosons. new world of experiments.” guide them,” says Keith Burnett, a theorist The difference becomes clear at very low Fermi gases have proved to be a very rich at Oxford University in the UK. temperatures: large numbers of bosons can source of new results. Indeed, since quarks, The race is also on to explore the prop- collapse into the same quantum state to electrons and all the fundamental matter erties of a Fermi gas trapped in an optical form a condensate, whereas the Pauli exclu- particles in nature are fermions, degenerate lattice – a 3D landscape in which the energy sion principle prevents fermions from doing Fermi gases have the potential to be even varies to form a perfect lattice of peaks and the same (see pages 13 and 14). Rather, as a more interesting than their bosonic counter- troughs. If the lasers producing the lattice Fermi gas is cooled, more and more of the parts. However, this comes at a cost. “In gen- are intense enough, atoms can be trapped at lowest energy levels fill up until all the levels eral, fermions are more challenging to work the peaks (or troughs). Such an experiment below the Fermi energy are occupied. The with than bosons,” says Randy Hulet at Rice, would basically reproduce what happens first such “degenerate” Fermi gas was pro- “due to the difficulty in cooling them and inside a crystal, with the atoms playing the duced in potassium-40 by Deborah Jin and measuring their temperature. These chal- role of the electrons and the lattice repre- Brian DeMarco in Boulder in 1999. lenges make any experiment more difficult.” senting the ions. The advantage of the ultra- Last year Jin and colleagues went a step One of the problems is that collisions be- cold approach, however, is that it is possible

8 physicsweb.org P HYSICS W ORLD J UNE 2005 NEWS AND ANALYSIS Where are they now?

When Eric Cornell, Wolfgang Ketterle and improvements was quite challenging OULDER

Carl Wieman shared the Nobel prize in 2001, B technically. However, I found learning about they became three of the best-known business and the ‘how to win friends and physicists in the world. Cornell and Ketterle influence people’ side of the job especially OLORADO AT remain very active in condensate research, C challenging – business skills are something although Cornell was diagnosed with that a physics education does not prepare necrotizing fasciitis last October and NIVERSITY OF you for!” After considering a career as a subsequently had to have his left arm and /U cabinet maker, Anderson recently set up his BBOTT shoulder amputated. He returned to his lab A own company, Vescent Photonics, to make EN part-time in April, with plans to search for the K sensors based on a new type of diode laser. dipole moment of the electron. “The doctors Anderson’s co-workers also followed tell me I am lucky to be alive,” he said, “and I careers in industry. Jason Ensher is a am much inclined to agree with them.” Famous five – Carl Wieman (left), Mike Matthews, systems engineer for Ball Aerospace & Wieman, meanwhile, claims to have Mike Anderson, Jason Ensher and Eric Cornell. Technologies, while Mike Matthews currently retired from thinking about big goals in works for 3M in Texas. “It was a great physics. “Nowadays I am satisfied to work on what works and so on are combined with experience to attend the Nobel-prize physics problems that seem interesting to effective uses of technology, it is possible to ceremony in 2001,” says Matthews, “but I me but are not aimed at big goals,” he says. vastly improve science education without it don’t regret leaving the field of BEC “However, I am thinking about big goals in requiring a great deal of additional cost or specifically. However, I do miss the almost science education and how to reach them. faculty time,” he says. boundless curiosity and the energy that you My main conclusion is that current science And what happened to the graduate find at a place like JILA.” education is failing badly at achieving the students and postdocs who worked on the Two of the co-authors on the MIT paper – needs of the 21st century – it needs to be original BEC experiments at Boulder and Ken Davies and Marc-Oliver Mewes – are effective and relevant to a large fraction of MIT? Michael Anderson, the first author on management consultants, while a third, the population, and not just the small fraction the Boulder paper, joined a company called Michael Andrews, has worked for Lucent and going into science.” Meadowlark Optics, where he worked his start-ups. The other three have remained in Wieman, who was named US professor of way up to be vice-president in charge of R&D academic research: Dallin Durfee and Dan the year last year for “his unwavering and manufacturing. Stamper-Kurn have faculty positions at dedication to undergraduate teaching”, is “I worked there until the autumn of 2001, Brigham Young University and the University convinced that science education requires a when the bottom fell out of the telecoms of California at Berkeley, respectively; while scientific approach. “When practices based market,” he recalls. “Developing reliable Klaasjan van Druten is project leader at the on good data, disseminating and duplicating manufacturing processes and product University of Amsterdam. to control the interactions between the tain tasks much faster than is possible on a had been seen in experiments built to pro- atoms, which is impossible with the elec- classical computer. However, there is a lot of duce a quark–gluon plasma – a state con- trons in an ordinary solid (see Physics World competition from other approaches. taining free quarks and gluons that last April 2004 pp25–29). “Condensates have lots of potential quan- existed just a fraction of a second after the In late 2001 Immanuel Bloch, then based tum bits,” says Chris Monroe of the Uni- Big Bang (see pages 23–24). in Munich, and co-workers managed to trap versity of Michigan, “but it remains very “The basic connection between a quark– ultracold rubidium atoms in an optical lat- difficult to address and control individual gluon plasma and our strongly interacting tice and observe phase transitions between neutral atoms at the same level as ions.” Fermi gas is that they both obey nearly per- the normal superfluid state of a Bose con- Other groups are hoping to show that the fect hydrodynamics, where perfect means densate and a state known as a Mott insu- electron has a small electric dipole moment, zero damping and zero viscosity,” says lator. However, to date only two groups – as predicted by certain extensions of the Thomas, who has recently started working Inguscio and co-workers in Florence and Standard Model of particle physics. These with Ed Shuryak, a nuclear theorist at Stony Tilman Esslinger’s group at ETH Zurich – experiments currently involve making ex- Brook University,on these topics. More gen- have managed to repeat this feat with a tremely precise measurements on atoms erally, Thomas sees strongly interacting Fermi gas. such as caesium and thallium, or on dipolar Fermi gases as a way of testing the theoret- “The next steps will be to understand the molecules such as ytterbium fluoride. ical methods that are used in other areas of influence of atom–atom interactions in the Another line of research involves experi- physics where strong interactions dominate. lattice,” says Esslinger. “Various phases that ments in which the Feshbach resonance is Indeed, a recent paper on string theory ref- have been predicted by theorists may then used to make the scattering length – which erenced the Duke group’s 2002 experiment. become available in the experiment. From describes the interactions between the And this is not the only connection be- the experimental point of view, the chal- atoms – much longer than the average dis- tween ultracold gases and strings. Only last lenge is to reach low enough temperatures tance between them. Such gases are said month a group of theorists at Utrecht Uni- and to extract the information from the to be strongly interacting. In 2002 John versity in the Netherlands proposed that experimental observables.” Thomas and co-workers at Duke University superstrings could be made in the laboratory in North Carolina produced the first Fermi by trapping an ultracold cloud of fermionic Thinking outside the box gas that was both strongly interacting and atoms inside a vortex in a Bose condensate While most researchers are looking to the degenerate. However, when they released (see page 3). Given the progress that has been parallels with condensed-matter systems, a the gas from the trap – which was shaped made over the past 10 years, it should sur- small number are moving in a completely like a cigar – they noticed something un- prise nobody if someone actually manages different direction. For instance, some groups usual: it expanded more rapidly in the nar- to do the experiment. are exploring the use of condensates to make row direction than in the long direction. Further references and links are available quantum computers that can perform cer- It then emerged that similar behaviour at physicsweb.org/articles/world/18/6/8

P HYSICS W ORLD J UNE 2005 physicsweb.org 9