Special Topic : Cold Atoms How cold atoms got hot : an interview with William Phillips

William Phillips of the National Institute of Standards and Technology (NIST) in Gaithersburg, Maryland, shared the 1997 Nobel Prize in physics for his work in developing laser methods for cooling and trapping atoms. Interactions between the light field and the atoms create what is dubbed an ‘optical molasses’ that slows the atoms down, thereby reducing their temperature to within a fraction of a degree of absolute zero. These techniques allow atoms to be studied with great precision, for example measuring their resonant frequencies for light absorption very accurately, so that these frequencies may supply very stable timing standards for atomic clocks. Besides applications in metrology, such cooling methods can also be used to study new fundamental physics. The 1997 Nobel award was widely considered to be a response to the first observation in 1995 of pure Bose–Einstein condensation (BEC), in which a collection of bosonic atoms all occupy a single quantum state. This quantum-mechanical effect only becomes possible at very low temperatures, and the team that achieved it, working at JILA operated jointly by the University of Colorado and NIST, used the techniques devised by Phillips and others. Since then, cold-atom physics has branched in many directions, among them being attempts to make a quantum computer (which would use logic operations based on quantum rules) from ultracold trapped atoms and ions. ‘National Science Review’ spoke with Phillips about the development and future potential of the field.


NSR:
You have said that cold-atom physics has evolved in so many directions in recent years that it is hard to keep track of them all.What, in your view, are some of the most important?Phillips: A particularly fruitful avenue has been the intersection of cold-atom physics with condensed-matter physics.For example, atoms can be put into optical lattices (periodic arrays of microscopic traps made from the interference of intersecting laser beams), so that their motion is analogous to the motion of electrons in a crystal lattice.These trapped atoms offer different opportunities for study compared to electrons.Atoms can be fermions or bosons [having half-integer or integer spin, respectively], while electrons are just fermions.It is easy to measure the momentum distribution and the correlation of the motion of atoms, while it is easy to measure bulk transport properties for electrons.Electrons have two spin states, while atoms could have one, two or many.Atoms can be used to make circuitry analogous to electrical circuits in an emerging field known as atomtronics.
Another satisfying direction has been in the development of atomic clocks.Making better clocks was the original motivation for laser cooling, and has been one of its greatest successes.Originally, the clocks were trapped ions or neutral atoms in an atomic fountain (an arrangement where atoms are launched upward, then fall back, providing a long time for observation).These clocks outperformed anything that had been seen before.Today, the best clocks are cold atoms stored in optical lattices, and have systematic uncertainties of only about two parts in 10 18 , or about one second over the current age of the universe.NSR: What was your initial motivation for getting into this area of physics?What or who inspired you to do so?

INTERVIEW
These advances have made possible amazing experiments, some only dreamed of decades ago and others not even dreamed of.
-William Phillips sodium beam in my lab, and I began a rather long odyssey to cool and trap the atoms.Initially I just thought it would be cool to do such a thing, but when I moved to NIST in the fall of 1978 I had the additional motivation of using cold atoms to make better measurements, and in particular to make better clocks.The initial work was strongly motivated by such practical goals in metrology.NSR: There is clearly a great deal of new physics that can be explored with cold-atom systems.Where do you think some of the most interesting fundamental questions lie?Are there properties of such systems that are still poorly understood?Phillips: Many-body physics is one of the great frontiers of modern physics.When the interactions between the atoms cannot be understood in terms of simple concepts like a mean-field [averaging all the interactions over the whole ensemble], calculating the behavior of the system can easily become intractable.Cold atoms represent a new kind of many-body system, with different opportunities for measurement and control, and we hope that they may help in understanding outstanding mysteries like the origin of high-temperature superconductivity.
Another area of fundamental interest is in exploring the basic symmetries of nature.Cold atoms are a good place to look for the possible existence of permanent electric dipole moments of particles like electrons or nucleons.Their appearance with magnitudes large enough to be seen in experiments that we can imagine doing in the next several years or decades would show that there is physics beyond the scope of the Standard Model.NSR: There are clearly various approaches to building quantum computers, of which the idea of using cold trapped atoms as qubits is just one.What do you feel are the real prospects of coldatom lattices for practical quantum computers, as opposed to other approaches?Phillips: The short answer is that no one knows.In my opinion, we are at such an early stage in the development of quantum computers that it is impossible to predict what a successful quantum computer will look like.Trapped ions are the most advanced system, but are far from making a practical device.Different qubit platforms have different strengths and weaknesses.Atoms and ions often have long coherence times [the time during which they can be kept in a single coherent, interdependent quantum state].Solid state systems are often fast and can be fabricated with techniques familiar from the semiconductor industry.In order to do something like factor large numbers (the killer app of quantum computing), all of these platforms will require massive scaling up from the few-qubit systems demonstrated so far.This will be difficult no matter what kind of qubit is used.Neutral atoms may have a slight advantage because optical lattices can literally hold millions of atoms in the space of a few millimeters.But neutral atoms have significant challenges as well.It may be that a combination of quantum platforms will win the race, exploiting what is best about each different kind of qubit.NSR: What was the impact of the first demonstrations of BEC in cold-atom studies?Phillips: At the time, people in the field spoke of BEC as the 'holy grail' of atomic physics.The demonstration of BEC was seen by its enthusiasts as a difficult quest, but one that would be well worth the effort.Not everyone agreed.Some thought BEC to be an unattainable goal; others wondered what all the fuss was about.
After the achievement of BEC, those negative perspectives soon evaporated.It became clear that the opportunities for measurement and control were quite different in an atomic gas BEC than in other 'super' systems.BECs made possible new kinds of experiments that had not been, and often could not be, done in other systems.Among these were the creation of 'atom lasers' (beams of coherent atoms extracted from a BEC), interference of independent BECs, and generation of new coherent atoms through non-linear four-wave mixing, to name but a few.NSR: What about metrological applications?Have they produced any surprises?Phillips: Cold atoms improve time and frequency metrology both because of the reduction of velocity-related frequency shifts (Doppler shifts and relativistic time dilation) and because slow (or trapped) atoms can be observed for long times, improving clock performance.
One particularly satisfying development has been in optically trapped atoms for frequency metrology.Ordinarily one would think that trapping neutral atoms would be bad for such metrology: atomic clocks work by measuring the transition frequency corresponding to the energy difference between two quantum states (the so-called clock states) in an atom.Trapping fields would typically shift the energies of the two states differently and inhomogeneously, leading to a shift and broadening of the transition.(By contrast, such differential shifts are largely absent in trapped ions because the trapping potential is related to the net electrical charge, which is the same for all states.)It was discovered, however, that one could typically find a 'magic wavelength' for laser trapping of atoms that would lead to no differential shift while providing the advantage of long observation times for trapped atoms.
Other metrological applications of cold atoms include their use in atom interferometers for inertial and gravimetric sensing.This has been a major area of practical application.NSR: Given this big impact of laser cooling on atomic clocks and time-keeping, do we now have the time standards that we need, or are further improvements possible and desirable?Phillips: There are two questions here: 'Do we need better time standards' and 'Can we make better time and frequency standards?'The answer to both questions is 'yes'.The need for time standards has traditionally been driven by scientific needs.Today, the scientific questions that motivate wanting better time There appear to be no fundamental limits to how good atomic frequency standards can get.
-William Phillips standards include whether the fundamental constants of nature are indeed constant.By measuring the ratio of the frequencies of different kinds of atomic frequency standards over long periods of time (several years), we can determine, for example, if the fine-structure constant [which determines the interaction strength between light and matter] is changing in time.The better the frequency standards, the better that measurement will be.
The fact that today the best frequency standards have systematic uncertainties at the level of a few parts in 10 18 is nothing less than astounding.When I first began working on laser cooling, the best standards were at about a part in 10 13 .I remember at that time that David Wineland was speculating about some distant future when 10 −18 might be possible.Now, thanks to laser cooling, as well as optical frequency metrology (the subject of the 2005 Nobel Prize to John Hall and Theodor Hänsch), we have such performance, and it seems to be getting better all the time.There appear to be no fundamental limits to how good atomic (or possibly nuclear) frequency standards can get.

NSR:
To what extent has this been a field driven by advances in practical technology, such as microfabrication and advances in laser optics?Is it fair to say that modern technical developments are enabling us at last to explore questions that were already conceptually recognized many decades ago?Phillips: Technological advances, particularly with lasers and optics, have been crucial to the development of this field.My first experiments with laser cooling used sodium because only for that atom did suitable lasers exist.Today, many of the elements of the periodic table have been laser-cooled because of advances in laser technology.Furthermore, advances in non-linear optics, allowing frequency multiplication and sum-frequency generation, have provided coherent optical sources over a wider range of the electromagnetic spectrum.Advances in optical technol-ogy also spurred the development of frequency-comb techniques, which led to high-precision optical frequency metrology.
Advances in fabrication have certainly affected the development of a number of laser technologies, and have also opened avenues of research like chip-based traps for cold atoms and ions and integrated devices for compact cold atom experiments.
These advances have made possible amazing experiments, some that were only dreamed of decades ago and others that were not even dreamed of.A particularly appealing example is the study and manipulation of single quantum objects, the subject of the 2012 Nobel Prize to David Wineland and Serge Haroche.The founders of quantum mechanics thought such ideas to be only in the realm of thought-experiments, with some even deriding the possibility that one could over do such things in principle.Today, the manipulation of single quantum objects is at the heart of quantum information science, a 21st century marriage of quantum and information technologies.NSR: What problems in this field, either technical or theoretical, would you most like to see solved?When might that happen?Phillips: There is a virtual cornucopia of great problems still to be solved.An important landmark would be to use quantum simulation to solve a quantum many-body problem of significant interest that has not been solved by classic numerical means.One example would be the Fermi-Hubbard model, believed by some to hold the key to high-temperature superconductivity.Among the hurdles to be overcome is getting a Fermi gas of atoms to sufficiently low temperature that one sees the essential groundstate properties.
Related to quantum simulation is quantum computing, for which ultracold ions and atoms are a particularly promising platform.A general-purpose quantum computer of sufficient power to solve interesting problems is a long way off, but one goal within reach is to make an 'immortal qubit'; that is, to create a logical qubit from a number of physical qubits, and perform repeated error correction to maintain the (unknown) state of the logical qubit.That would be an important milestone in making a competent quantum computer.
Nobel laureate William Phillips of the National Institute of Standards and Technology (NIST) in Gaithersburg, MD, USA Phillips: I first learned about the possibility of laser cooling when I was a postdoc at the Massachusetts Institute of Technology in 1978.David Wineland (who received the physics Nobel Prize in 2012) demonstrated laser cooling of trapped ions for the first time; Art Ashkin at Bell Labs wrote a proposal for slowing and trapping sodium atoms.These were my inspirations.I had a Downloaded from https://academic.oup.com/nsr/article-abstract/3/2/201/2460204 by guest on 28 December 2018 Philip Ball writes for NSR from London.Downloaded from https://academic.oup.com/nsr/article-abstract/3/2/201/2460204 by guest on 28 December 2018