Two-dimensional electron gases provide 'mini laboratories' for fundamental physics as well as the basis for the semiconductor industry. Devices are ultimately limited by the flow of electrons, and a direct visualization of their flow could lead to better designs and applications. To this end, Michael Jura and co-workers use scanning gate microscopy to observe the electrons flowing through samples with different levels of disorder. They find that although the electrons flow along narrow branches, as expected, electrons in the cleanest sample – which still include a few scattering sites – show no hard scattering from impurities and defects. Moreover, the trajectories are stable to changes in the initial injection sites and injection angle. This stability is not expected for classical chaotic motion and quantum mechanical simulations are necessary to explain the results. [Letter p841]

[Article p790] ; [News & Views p756]

On 4 October 1957, the world changed. A series of beeps across the sky confirmed that the Soviet Union had won the first leg of the space race, by putting an artificial satellite into orbit. In his Perspective article, Joe Burns shares his memories of the Sputnik launch, and discusses its impact not only on US politics in a cold-war climate, but also on science: the huge boost in influence and funding for science and engineering; and the subsequent strides in understanding the physics of the cosmos, made possible through space-borne instrumentation. In an accompanying Commentary, as plans are drawn up for a return to the Moon, Mike Lockwood considers the science case for building a base there, and whether a human presence is justified.

[Perspective p664] ; [Commentary p669]

The final stages of planet formation take place within a disc of dust and gas surrounding a young star. Part of the disc spirals inwards onto the central star, dragging along the fledgling planets and setting their final orbits, but what drives this gas flow is still unclear. Eugene Chiang and Ruth Murray–Clay study the recently discovered 'transitional' discs, in which the innermost regions around a star are swept clean of dust. They propose that the dust-free condition enables a coupling between disc magnetic fields and rotating gas, owing to the ionization of the gas in the inner disc by X-rays emitted from the star. The ionized gas then activates a magnetic instability at the rim, which drives gas towards the central star while radiation pressure from the star pushes out any dust accompanying the infalling gas.

[Letter p604]

Superconductors, superfluids and supersolids can be defined in terms of how they respond to rotation. According to the London law, a rotating superconductor will generate a magnetic field that depends on fundamental constants alone. And for superfluids composed of neutral particles, such as helium-4, rotation velocities above a certain threshold will result in the formation of vortices; the quantization of the superfluid velocity within a vortex is known as Onsager-Feynman quantization. Both of these laws would be broken by a two-component superconductor, propose Egor Babaev and Neil Ashcroft. They show that for liquid metallic hydrogen, in which Cooper pairs can be formed through electron pairing and proton pairing, the superfluid velocity quantization becomes fractional and the generated magnetic field no longer depends only on fundamental constants but on density as well.

[Letter p530]

In the laboratory, under the highest currently attainable pressure, hydrogen solidifies but remains insulating even though theoretical calculations suggest it should be metallic (and perhaps superconducting). Higher-pressure studies will settle this question. In the meantime. the actual atomic structures of the highpressure phases of hydrogen are the subject of debate. Part of the problem is the paucity of experimental data to constrain theoretical calculations, for which huge tracts of phase space must be searched for possible structures. To reduce the effort, Chris Pickard and Richard Needs have come up with a computational approach that optimizes the enthalpy as a function of the atomic configuration. Their candidate structure for phase III hydrogen is not only stable and insulating but agrees with available experimental evidence, thus revising the phase diagram of the simplest element.

[Letter p473] ; [News and Views p452]

The dynamic response of a periodic structure to any wave-like excitation — be it optical, acoustic or electronic — is governed by the coupling between the eigenstates of the structure. A complete description of these eigenstates involves not only real space, with which physicists and non-physicists alike are familiar, but the inverse of this space — so called reciprocal- or k-space. Although techniques exist to characterize the electronic eigenstates of atomic crystals in k-space, the same cannot be said for the optical eigenstates of photonic crystals. Rob Engelen and colleagues have now developed a near-field optical microscopy technique that enables them to track the temporal evolution of a pulse of light in k-space from one eigenstate to another as it passes through a complex photonic-crystal structure.

[Letter p401]

All physicists know that light carries both linear and angular momentum. What is perhaps less well known, however, is that its angular momentum can be broken down into spin and orbital components. Spin angular momentum is associated with polarization, whereas orbital angular momentum arises from a more complex combination of the phase and amplitude profiles of an optical field. Although the spin momentum is the predominant property used in optical-based quantum information applications, orbital momentum is potentially more powerful for encoding and processing such information in high-dimensional quantum spaces. In this issue, Gabriel Molina-Terriza and colleagues review progress in the generation, understanding and use of the orbital angular momentum of light.

[Progress Article p305]

Percolation theory has taught us how to transmit information efficiently through a network of connected nodes. But is the same concept helpful for quantum information, where the task is to distribute entanglement through a network? Yes and no, say Antonio Acín and colleagues. They show that the design of quantum networks can be related to problems studied in 'classical' percolation theory, but they also present examples for which classical strategies fail to yield optimal results, calling for an approach that exploits explicitly the quantum nature of the networks. Acín et al. also argue that the distribution of entanglement through quantum networks defi nes a novel type of critical phenomenon, and that understanding — and using — the associated phase transitions could enable quantum communication over longer distances than had been thought possible. (Artwork by Beata Wehr)

[Letter p256]

At the 1987 March Meeting of the American Physical Society, thousands attended a special session — the 'Woodstock of physics' — to hear about the new copper-oxide-based superconductors. But the vital question has remained unanswered: what causes the electrons to form pairs? Lattice vibrations or magnetic excitations? In this issue, Baptiste Vignolle et al. use neutrons to map spin excitations in unprecedented detail, arguing for magnetically mediated superconductivity; Dennis Newns and Chang Tsuei, however, present a theory based on a two-phonon mechanism. Eschewing both phonons and magnons, Krzysztof Byczuk and co-workers offer an explanation for the origin of the 'kink' in photoemission data, taken as evidence by both sides — they say the kink may have nothing to do with superconductivity.

Letter p163 | Letter p168 | Article p184 | News & Views p148

A comprehensive toolbox has been developed over the past few years for electrostatically manipulating the motion of molecules and for storing them. One advance was the construction of a storage ring for neutral polar molecules. Confining particles in rings rather than traps (as is common in atomic physics) is useful because, for example, circulating particles can be made to interact repeatedly with electric fields and with other particles at well-defined times and locations. However, keeping the molecules in bunches as they circulate — rather than allowing them to occupy the entire ring — remained a challenge, hindering the full exploitation of the approach. Cynthia Heiner et al. now report the construction and operation of a molecular synchrotron in which bunches are maintained — the first synchrotron for neutral particles.

Letter p115 | News and Views p77

When a medium is irradiated with a laser pulse so intense that it forces the medium's electrons to move in synchrony with the laser's electric field, the exact phase of this field with respect to the pulse envelope — known as the carrier-envelope hase (CEP) — plays an important role in determining how the medium responds. Measuring the value of this phase is challenging, and usually requires averaging over many pulses. In this issue, Charles Haworth and colleagues show that by analysing the high harmonics generated by the interaction of an intense femtosecond laser pulse with a gaseous medium, they can determine the CEP of a single pulse. Moreover, they suggest that such an approach could soon enable individual attosecond pulses emitted during a particular optical half-cycle of the driving field to be isolated.

Article p52 | News and views p19