June 2010 Archives

Matthias Kaschube of Princeton University and his collaborators have applied concepts from pattern formation and continuum dynamics to address a key question in neuroscience: Do neurons retain their roles in a growing brain? Although an adult human’s brain is four times as large as a baby human’s, it has roughly the same number of neurons. The extra volume accommodates the developed brain’s greater number of blood vessels, nonneuronal cells, and—crucially for memory and intelligence—interneuron connections. To understand the development of those connections, Kaschube and others study ocular dominance (the brain’s preference for input from one eye over the other) in the primary visual cortex (the brain’s principal image processor). Several factors make OD a convenient model system. Not only are OD signals readily induced and tracked; the neurons responsible for OD are grouped in recognizable rows of columns. Data gathered from kittens by Kaschube’s collaborators—Karl-Friedrich Schmidt and Siegrid Löwel of Friedrich Schiller University in Jena, Germany—conflict with the simplest growth model in which OD columns and their separations both expand by the same factor as the visual cortex. Rather, the columns increase in number and keep their separations while the rows lose their original straight configuration and become wavier. Kaschube and his graduate student Wolfgang Keil account for the pattern change by invoking the zigzag instability, which arises when stripes in isotropic systems are stretched. In the simple growth model, individual OD neurons would continue to carry signals from the same part of the visual field of view to the same part of the visual cortex. In Kaschube and Keil’s model, the reorganization of OD neurons implies that individual neurons continuously and flexibly change their signal-carrying roles. (W. Keil, K.-F. Schmidt, S. Löwel, M. Kaschube, Proc. Natl. Acad. Sci. USA, in press.)—Charles Day

It’s not just the spilled oil: Since the Industrial Revolution, Earth’s oceans have been soaking up unprecedented levels of atmospheric carbon dioxide as well as runoff from manufacturing processes. In a new review, Woods Hole Oceanographic Institution scientist Scott Doney describes the impact of human activities on the open ocean and coastal waters. Observations and models show that up to 30% of the carbon released by fossil-fuel combustion is absorbed by the ocean; with that rise in subsurface CO₂, a weak acid in seawater, the ocean’s pH plummets. Measurements of CO₂ and pH collected over the past two decades off Hawaiian shores show that the ocean is becoming more acidic—30 to 100 times faster than what geological records report. The shift in the ocean’s acid-base equilibrium due to excess CO₂ also hinders the precipitation of calcium carbonate, which makes up the shells and skeletons of many marine species. Yet another concern is hypoxia—dangerously low levels of subsurface oxygen caused by climate-warming-induced degradation of organic matter—which leads to anaerobic respiration conditions and produces even more CO₂. And yes, oil spills, even naturally occurring seeps from oil wells, release pernicious hydrocarbons into the marine environment. Most such industrial chemicals are “woefully undersampled,” writes Doney, whose summary calls on the oceanographic community to better coordinate its monitoring of the ocean’s biogeochemical cycle. (S. C. Doney, Science 328, 1512, 2010.)—Jermey N. A. Matthews

When CERN’s Large Hadron Collider is fully operational, it will accelerate countercirculating protons to energies of 7 TeV, the highest particle energy ever achieved by human ingenuity. By macroscopic standards, 7 TeV (10⁻⁶ J) is tiny. But each beam produced at the LHC will ultimately include some 3 × 10¹⁴ protons; should it go awry, it could seriously damage the LHC and the delicate particle detectors that the accelerator hosts. For that reason, as described in a new paper by Robert Appleby and colleagues at CERN, the LHC includes an elaborate safety system that regulates the beams, monitors them, and dumps them out of harm’s way if they go off course. To give an idea of the system’s complexity, the authors note that 17 distinct subsystems must continually give a virtual OK to a central processer, or else the beam will be dumped. In addition to describing the LHC’s safety features, Appleby and colleagues calculate how the beams would respond to various possible accidents—improperly set magnetic fields, for example—and ask if the protection system would respond quickly and accurately enough to avert disaster. Although no system can protect against all failures, the researchers conclude that the accelerator and its associated detectors are as safe as could reasonably be expected. In addition to the simulated studies, scientists are testing the protection system with real 3.5 TeV beams whose intensity is gradually being ramped up. (R. B. Appleby et al., Phys. Rev. ST Accel. Beams. 13, 061002, 2010.)—Steven K. Blau

Newtonian fluids, such as water, are described by the Navier-Stokes equations. But many everyday fluids lack a similar complete description, and researchers still seek better observations and models of their flow. Yield stress fluids (YSFs), a subset of non-Newtonian fluids that includes toothpaste and mayonnaise, hold their shape under low stress but flow under high stress. Some YSFs are also thixotropic, meaning their viscosities decrease with time during continued flow. Thixotropy in a YSF can result in heterogeneous flow—confinement of the fluidlike behavior to part of the material, which flows more and more easily, while the rest remains solid—an important phenomenon to understand and control when handling YSFs industrially. Now, Sébastien Manneville, of the École Normale Supérieure de Lyon, and colleagues have unexpectedly observed similar localized flow in a nonthixotropic YSF subjected to a shear stress. The observed behavior was transient, but it lasted a surprisingly long time: more than a day in one of their trials, several hours in others. Many of the researchers’ observations, such as the power-law dependence of the transient duration on the shear rate, remain unexplained. Even so, the data indicate that nonthixotropic YSFs are more complicated than was previously assumed, and they exemplify the importance of distinguishing between transient and steady-state behavior in YSF experiments. (T. Divoux et al., Phys. Rev. Lett. 104, 208301, 2010.)—Johanna Miller

In photonic devices and networks, optical waveguides routinely need to be coupled, split, and switched; several methods are in use to fabricate those junctions. A promising technique for making three-dimensional junctions in transparent materials, such as silica glass, is to write them with femtosecond lasers. Under optimal laser-writing conditions, a focused spot photoexcites the glass through nonlinear absorption and locally changes the refractive index. Slowly moving the material stretches the spot into a line of altered index that acts as a waveguide. For creating junctions, however, it has proven difficult to precisely position the material for each branch. Now, researchers at Japan's Kyoto University have sidestepped that issue by introducing parallel laser writing of multiple branches in three dimensions. The key is computer-generated holograms. With a CGH and a single laser, multiple beam spots can be focused at precise locations in the glass. As the material is moved, CGHs are sequentially swapped in and out to change the spots' locations. Using 256 CGHs, the researchers fabricated a 20-mm-long continuous waveguide that split and spread into four branches about 85 µm apart. A schematic is shown on the left, and the output from a single incident 635-nm laser beam is shown on the right. Optimization of the method is under way. (M. Sakakura et al., Opt. Express 18, 12136, 2010.)—Stephen G. Benka

In 1954 Princeton University’s Robert Dicke predicted a remarkable phenomenon: A dense cloud of excited atoms in a light field, he argued, could decay by spontaneously emitting coherent and highly polarized photons—an effect he termed superradiance. By subtly altering the Hamiltonian, researchers in the early 1970s realized that the phenomenon need not be restricted to transient pulses, and they made their own prediction: When light and matter interact strongly enough, even at zero temperature, they can exhibit a steady-state superradiant phase. By confining a Bose–Einstein condensate of some 10⁵ rubidium atoms driven by a standing-wave laser beam in an optical cavity, Tilman Esslinger and his colleagues at ETH Zürich have now observed the predicted quantum phase transition. As shown in the sketch of their experiment, if the laser light exciting the BEC is intense enough to spawn superradiant photons along the cavity axis, the photons’ repeated reflections establish a field that interferes with the laser field to form a square-patterned potential. Because the condensate atoms produce the superradiant light, they are active participants, since atoms and photons dynamically influence each other’s motion through the coherent exchange of momentum. The upshot is that above some critical laser power, the atoms still exhibit superfluid behavior but become self-ordered into a crystalline lattice. Interestingly, although the Zürich group’s system is open, laser driven, and dissipative—far from the closed equilibrium system that Dicke considered—his Hamiltonian still captures the essential physics. (K. Baumann et al., Nature 464, 1301, 2010.)—R. Mark Wilson

Just as accelerating charge produces electromagnetic radiation, accelerating mass is predicted to produce gravitational radiation. The effect of a gravitational wave’s alternating distortions of space could be detected by a Michelson interferometer, but gravity’s weakness means that extraordinary sensitivity is needed to observe even a relatively intense wave. The Laser Interferometer Space Antenna (LISA) is a proposed mission to achieve that sensitivity with an interferometer 5 million km on a side, its vertices located on three spacecraft orbiting the Sun. But the inevitable fluctuations in laser frequencies introduce noise a billion times more intense than the signal from the gravitational waves researchers hope to see. The solution, a technique called time-delay interferometry (TDI), is to reduce noise not through better stabilization of the physical components but by signal processing. In essence, the phase gained by light traversing each arm is subtracted from the phase from the same arm offset by the round-trip time for the other arm; subtracting those two differences then yields a quantity unmarred by laser frequency fluctuations. In two more steps, noise due to clock error can likewise be eliminated. Now, researchers at NASA’s Jet Propulsion Laboratory have demonstrated TDI in a laboratory experiment designed to mimic LISA’s noise environment. They’ve shown that the technique can indeed reduce laser frequency noise and clock noise by the necessary nine orders of magnitude. (G. de Vine et al., Phys. Rev. Lett. 104, 211103, 2010.)—Johanna Miller

Superfluid liquids—ultracold, zero-viscosity liquids that creep over vessel walls—manifest a kind of Bose–Einstein condensation. They are subject to quantum constraints that might be thought to thwart both the onset and dissipation of turbulence. Before superfluid turbulence was first seen in 1958, Richard Feynman posited the generation and reconnection of quantized vortex lines, tornadolike topological defects that would allow turbulent eddies to form and dissipate in superfluids. The quantized line integral of the fluid velocity around any loop enclosing a single vortex line would be h/m, where m is the mass of the relevant boson. For the next half century, the existence of such vortex lines was assumed but never directly seen. But techniques recently developed in Daniel Lathrop’s lab at the University of Maryland render the vortex lines and their reconnection events visible in superfluid helium-4. Injected hydrogen gas forms micron-sized hydrogen-ice particles that attach themselves to the vortex lines and scatter illuminating light, allowing the Maryland team to film reconnections. The close-ups show a reconnection event at 25-millisecond intervals in turbulent superfluid ⁴He. Lathrop and coworkers, having digitally filmed 40 000 such events at about 80 frames per second, report that the dynamics of reconnection is dominated by h/m, “the quantum of circulation.“ The tracer particles also make it possible to view the overall distribution of velocities in the turbulent superfluid at much finer spatial resolution than was previously possible. On that scale, a strong high-velocity tail in the velocity distribution makes turbulence in superfluids look quite different from classical turbulence. (M. S. Paoletti, M. E. Fisher, D. P. Lathrop, Physica D, in press, doi: 10.1016/j.physd.2009.03.006.)—Bertram Schwarzschild