Recently in Computational physics Category

On 30 August 2000, as the sun beat down on Texas’s largest city, ozone concentrations soared to unhealthy levels. Usually in summer, as the city heats up, sea breezes blowing in from nearby Galveston Bay and the Gulf of Mexico refresh the air. But the prevailing winds over Houston, although mild, tend to counteract the sea breeze. Thus, if the breeze collides with the prevailing winds, stagnation sets in over the city and pollutants can build up. Now a numerical study led by Fei Chen of the National Center for Atmospheric Research suggests that the materials of the urban environment are partly to blame for ozone pollution. Chen and colleagues validated their computer model by comparing their simulation of the August 2000 pollution event against extensive data collected in the Texas Air Quality Study 2000. Then, to understand how various environmental features affect the development of the sea breeze, they simulated conditions that were wetter or dryer than normal, and in one simulation they replaced the urban landscape with cropland. The substitution of crops for concrete had the greatest impact on boosting the sea breeze and reducing periods of stagnation. (It also increased the efficacy of the nighttime land breeze that blows pollutants out to sea.) Compared with green space, the researchers found, the urban environment is hotter. That effect actually tends to enhance the sea breeze, but the enhancement is more than offset by the frictional damping from Houston’s buildings. (Fei Chen et al., J. Geophys. Res. [Atmospheres], in press, doi:10.1029/2010JD015533; image from http://www.utexas.edu/research/ceer/texaqs/visitors/photos.html.)—Steven K. Blau

When a transition-metal compound is subject to high pressure, its electronic spin state can change, which in turn can change the compound's material properties. That spin-state crossover is of geophysical relevance because of the iron-bearing minerals in Earth's lower mantle. But the most abundant mantle mineral—Fe-bearing magnesium silicate perovskite (Pv)—is a challenge to study, since it contains three nonequivalent types of Fe atom: Not only can Fe replace either Mg or Si in the crystal lattice, but Fe replacing Mg can be either ferrous (Fe²⁺) or ferric (Fe³⁺). Experiments on spin states under pressure probe the electron configuration indirectly, via its effect on nuclear energy levels, so computational studies are necessary to connect experimental measurements with the correct interpretations. Last year, an experimental study of ferric Fe in Pv yielded results that were at odds with the computational studies to date. Now, Renata Wentzcovitch and colleagues at the University of Minnesota have verified the experimental results computationally and predicted their geophysical consequences. The researchers found that ferric Fe that replaces Si undergoes a spin-state crossover at a pressure somewhere between 40 and 70 GPa, equivalent to a depth between 1100 and 2000 km and consistent with the 50–60 GPa crossover pressure measured experimentally. Since that transition causes the unit cell to shrink in volume by about 1%, it has a significant effect on the mineral’s bulk modulus and thus on the speeds of seismic waves and on mantle convection. (H. Hsu et al., Phys. Rev. Lett. 106, 118501, 2011.)—Johanna Miller

Since Andre Geim and Konstantin Novoselov first touched off the graphene “gold rush” in 2004—their pioneering work earned them this year’s Nobel Prize in Physics—researchers have been pursuing ways to scale up its production. Among graphene’s remarkable properties is its roughly 100-GPa tensile strength, which is 40 times greater than the value for steel. That, however, is for defect-free graphene sheets; when formed by chemical vapor deposition, a proven industrial technique, graphene sheets contain crystallites separated by grain boundaries (see the news story in Physics Today, August 2010, page 15). Now, a computational study by Rassin Grantab and Vivek Shenoy at Brown University and Rodney Ruoff at the University of Texas at Austin reveals that graphene sheets with highly misaligned boundaries are actually stronger than slightly misaligned ones. As the image shows, misaligned grain boundaries consist of repeating pairs of 5- and 7-member rings separated by hexagonal rings. In simulations of the stress–strain curves as a function of the misalignment, the researchers found that, surprisingly, tensile strength increases with increasing misalignment angle. According to their model, stress failure begins at critical bonds within the 7-member rings; and critical bond length, which decreases with increasing misalignment angle, is proportional to initial material strain. In one simulation, a graphene sheet with a boundary angle of 28.7° and strained by 15% resisted stress up to 95 GPa; conceivably, it might be more efficient for researchers to engineer controlled defects into a graphene sheet rather than trying to make a perfect one. (R. Grantab, V. B. Shenoy, R. S. Ruoff, Science 330, 946, 2010.)—Jermey N. A. Matthews

As famously predicted by Hendrik Casimir in 1948, parallel conductors in a vacuum will attract each other because the conductors impose boundary conditions that affect the vacuum energy of the electromagnetic field (see the article by Steve Lamoreaux in Physics Today, February 2007, page 40). In general, the Casimir force depends on the shape of the conductors and its value is notoriously difficult to calculate, but research groups worldwide have been developing increasingly applicable computational techniques. Now a team at MIT has shown how tabletop measurements might provide the key information needed for the general calculation. The Casimir force may be expressed as an integral over frequency (ω) of correlation functions that involve electric and magnetic fields. In principle, those frequency-dependent correlations can be obtained in a suitably scaled tabletop experiment from measurements of how an antenna at one point responds to a current generated at a distant point. In practice, such measurements won’t work because the integrand oscillates wildly with ω. The integrand becomes well behaved—it decays and doesn’t oscillate—if the integration is performed in the complex plane, but real antennas respond to real frequencies. The key observation made by the MIT team is that their mathematical expressions always involve ω in the combination εω², where ε is the permittivity. Thus, the researchers predict, a force integral with real vacuum permittivity and complex contour can be calculated from a tractable number of antenna measurements made at real ω in a medium of complex permittivity—for example, salt water. (A. W. Rodriguez et al., Proc. Natl. Acad. Sci. USA, in press, doi:10.1073/pnas.1003894107.)—Steven K. Blau

The partitioning of the continuous visible spectrum into a small number of basic colors is done differently in different languages. But the variation is less than would be expected by chance, as statistical analysis of the World Color Survey's data set has shown. Several computational approaches have been taken toward understanding how languages’ color categories develop. Among them is the work of Andrea Baronchelli (Polytechnic University of Catalonia, Barcelona, Spain) and his collaborators. They performed computer simulations in which individuals in a population, beginning with no words to describe colors at all, were tasked with describing different colors to one another. The individuals independently invented words and categories and, based on the success or failure of their communications, adjusted their categories and vocabularies to match those around them. Eventually, each population came to a near-consensus, as shown in two examples in the top panel of the figure. Now, the researchers have revised their model to include a real property of human vision, the “just noticeable difference” (JND; shown in the bottom panel), or wavelength resolution. In the new simulations, individuals were not required to distinguish between colors that a human couldn't tell apart. The categories produced by the JND-based simulations clustered together in color space to the same degree as the World Color Survey results did. The researchers hope that the quantitative agreement between their simple model and empirical data will pave the way for greater use of synthetic modeling in studying language development. (A. Baronchelli et al., Proc. Natl. Acad. Sci. USA, in press, doi/10.1073/pnas.0908533107.) —Johanna Miller

Boron's next-door neighbor in the periodic table, beryllium, forms a simple metal lattice at 0 K. Boron's other next-door neighbor, carbon, forms another simple structure at 0 K, graphite. As for boron itself, its most stable form at 0 K is unknown. Compounding the mystery, the lowest-energy phase that experimenters have found, the β-rhombohedral phase, is stunningly complex and defect riddled: Each hexagonal unit cell has 423 atomic sites; on average only 320 of them are occupied. Now, Tadashi Ogitsu of Lawrence Livermore National Laboratory and his collaborators have explained why the stable β-rhombohedral phase has so many empty sites. If boron were simple, the defects—vacancies and interstitial atoms—would disappear as boron attained its perfect crystalline structure. But according to Ogitsu's calculations, which he carried out on a Livermore supercomputer, the defects actually stabilize the β-rhombohedral phase. It turns out the defect sites in the crystal are arranged in a particular geometric configuration, a double-layer expanded kagome lattice (see figure). Ogitsu and his collaborators realized that the problem of how boron atoms fill empty sites is essentially the same as another problem: how antiferromagnetically coupled spins align themselves on an expanded kagome lattice, whose ground state is degenerate and disordered. Like spin ices, and ordinary water ice, boron's β-rhombohedral phase is geometrically frustrated. Ogitsu notes that the hopping of defects between nearly degenerate configurations can also account for some of boron's peculiar and long-puzzling transport properties. (T. Ogitsu et al., Phys. Rev. B, in press.)—Charles Day

Designers of transportation networks have to weigh the cost of serving customers against the need for an efficient, robust system. Natural organisms, too, confront tasks in which they need to balance competing desiderata. As it forages for food, for example, a slime mold must balance cost (that is, the amount of protoplasm it extrudes), efficiency, and the ability to withstand injury. Remarkably, as recently reported by Atsushi Tero and colleagues from Japan and the UK, the molds do as well as transportation engineers in balancing their analogous competing needs. Panel a of the figure re-creates a 17-cm-wide map of the principal cities served by the Tokyo railway system with a slime mold (yellow) at the location of Tokyo and food flakes (white) representing other cities. In about a day’s time, the slime mold finds where the nourishment is and generates a protoplasm network with the food flakes as nodes. Standard metrics for analyzing transportation networks reveal that the mold’s foraging network and the Tokyo railway system perform similarly. Perhaps more significantly, Tero and company imitated slime-mold networks in numerical simulations that don’t incorporate detailed biochemistry. Instead, they include a feedback step in which tubular links carrying a large protoplasm flux grow wider and flux-poor links contract. By tweaking their simulation parameters, the researchers could nudge the network toward, for example, greater cost efficiency. With optimal parameters, they could even improve upon the work of slime molds and human engineers. (A. Tero et al., Science 237, 439, 2010.) —Steven K. Blau

The tunable elasticity and porosity of colloidal gels lead to some interesting applications, among them tissue scaffolding and drug delivery. Conventionally, colloidal particles interact and assemble under entropic and electrostatic forces to form predictable structures. But greater control can be achieved from an approach developed by Paul Clegg, Michael Cates, and their collaborators at the University of Edinburgh in the UK. The researchers disperse silica particles in the single-phase region of two partially miscible solvents—water and the organic base 2,6-lutidine. When the solution is heated above a critical temperature, the solvents separate and the particles become trapped at the liquid–liquid interfaces. The bulky particle domains then jam together and arrest the solvent separation, forming a two-phase network the researchers call a bijel. But cool the solution and remix the solvents too soon and the distinct structure disappears, as shown in movie 1 and the two left images in which the colloids appear green, the water black, and the lutidine red. Now the researchers have discovered an approach to stabilize the bijel structure. When the phase-separated solution is allowed to sit for at least 24 hours before it is cooled, the bijel surprisingly keeps its shape, as shown in the two right images and movie 2. From Monte Carlo simulations, the researchers deduce how the resulting network of colloidal monolayers, or monogel, stays intact: the particles become compressed by capillary forces, remain attracted by van der Waals forces, and are kept from collapsing into each other by repulsive electrostatic forces. (E. Sanz et al., Phys. Rev. Lett., in press.) —Jermey N. A. Matthews

Movie 1

Movie 2

In a single cell, thousands of simultaneously occurring biochemical reactions carry out such functions as converting and storing energy and regulating nutrient levels; together, those processes make up the cell’s metabolic network. Computational biology involves, among other things, the linking of metabolic pathways to form a metabolic network model, a promising tool for preclinical drug studies and other medical research. However, such computational models do not traditionally include the function-determining structural details of a network’s macromolecules; for example, an enzyme’s ability to catalyze reactions and regulate the cell’s response to external stimuli is determined by its three-dimensional configuration. Now, an international team led by Adam Godzik at the Burnham Institute for Medical Research in California has taken a rare step and introduced atomic-level protein structural data to the metabolic network model of an ancient thermophilic bacterium, Thermotoga maritima, shown in this optical microscope image. The image also shows schematics of proteins in their 3D configurations, which, when they were expressed in the reconstructed metabolic network, helped the research team solve the puzzle of how proteins evolve when their cell networks grow larger.

It turns out that only 37% of T. maritima’s proteins are essential to the formation of its metabolic network; those “core-essential” proteins adopt the bulk—61%—of the bacterium’s relatively few unique 3D configurations. The finding suggests that the core-essential proteins evolved their structure to perform additional functions in distinct pathways. (Y. Zhang et al., Science 325, 1544, 2009.)—Jermey N.A. Matthews

When a person’s head strikes, or is struck by, another object, it accelerates. As it begins to decelerate, the brain slams into the skull, then bounces off and oscillates until the impact energy dissipates. The resulting shear and compressive strains can lead to brain damage. But in battlefield explosions, just the acoustic waves alone can cause soldiers traumatic brain injuries. To better understand that process, Lawrence Livermore National Laboratory's William Moss and Michael King and the University of Rochester’s Eric Blackman compared numerical simulations of a head colliding with a wall to one being struck by an explosion’s blast waves. Despite accelerating the head at less than half the rate of the wall collision, the simulated blast produced on the brain surprisingly comparable pressure spikes—ranging up to 3 bars—and even larger pressure gradients. Apparently, those mechanical loads are delivered by the skull, which ripples under the pressure of blast waves—the rippling motion is indicated in the image by velocity vectors. The researchers confirmed the role of the skull’s elasticity by making it 1000 times stiffer in their simulations and observing a fivefold drop in the pressure spikes. The simulations also revealed that helmets in contact with the head can impart an additional mechanical load to the skull and helmets that allow for an air cushion geometrically focus and increase the magnitude of blast waves. (W. C. Moss et al., Phys. Rev. Lett., in press.)—Jermey N. A. Matthews

Quantum computing is a goal that both excites and challenges researchers, who are working on a wide variety of physical realizations of the basic building block: the quantum bit, or qubit.

One type is the superconducting qubit made from one or more Josephson junctions. The biggest advantage of superconducting qubits is their strong coupling to microwave signals, which can control the qubits and mediate their interactions. The greatest limitation is their short coherence lifetime.

Despite that limitation, recent experiments have demonstrated the kind of precise control that will be needed to make progress toward a viable quantum computer.

In one experiment, Max Hofheinz, John Martinis, Andrew Cleland and colleagues from the University of California, Santa Barbara, showed that they could impose on a microwave resonator any desired superposition of photon-number states. (M. Hofheinz et al., Nature 459, 546, 2009.)—Barbara Goss Levi

In an effort to reduce the pervasive smog in Beijing (see photo), Chinese authorities imposed measures to restrict traffic and close factories around the city during the 2008 Olympics. Were those efforts successful in reducing total atmospheric aerosol? Climate scientists Jan Cermak and Reto Knutti at ETH Zürich in Switzerland attempted to find out. They began by comparing absolute values of aerosol optical thickness—transmittance measurements from the Moderate-Resolution Imaging Spectroradiometer aboard NASA's Terra satellite—for the years 2002–08. They found that within a 150-km radius of Beijing, the average 2008 AOT value was more than 14% lower than the previous years. But what would it have been without the mandated emissions reductions? To answer that question, the researchers used a neural network approach: With data from the preceding six summers, they trained a model to predict AOT as a function of relative humidity, wind velocity, and precipitation. The model then predicted that within a 500-km radius of the city, AOTs in 2008 would have been 10%–14% higher than the actual observed values; the model was less accurate when larger regions were analyzed. Although the magnitude of the reductions was lower than expected, the emissions restrictions did have a statistically significant local impact. (J. Cermak, R. Knutti, Geophys. Res. Lett. 36, L10806, 2009, doi:10.1029/2009GL038572. Photo by Michael Silverman, 6 August 2006.)—Jermey N. A. Matthews

Before they form snowflakes and other hexagonal crystals, water molecules nucleate in smaller configurations. Determining the structure of those precursors—even in the outwardly simple case of water on a clean metal surface—is an area of ongoing interest and controversy. For example, at submonolayer coverage on a copper (110) surface, water molecules form chains that can grow to many tens of nanometers in length but are just 1 nm wide. The chains’ structure has been a mystery, since no arrangement of water molecules into hexagonal units entirely explains the experimental data. Now, Andrew Hodgson and colleagues of the University of Liverpool in the UK have collaborated with Angelos Michaelides’ group at University College London to find the structure. Michaelides and postdoc Javier Carrasco ran calculations on some 50 possible chain structures; they found that the most energetically stable one also gave the best fit to the Liverpool group’s high-resolution scanning tunneling microscopy images (as shown in the top panel) and vibrational spectra. That structure (bottom panel) is an arrangement of pentagons, not hexagons. The water molecules shown in red and yellow are perpendicular to the Cu surface—the hydrogen atoms pointing up are responsible for the bright spots in the STM images, and the ones pointing down (not visible in the figure) interact with the Cu atoms. The researchers suggest that nonhexagon arrangements might be involved at other water–metal interfaces where the structure of water is unknown. (J. Carrasco et al., Nat. Mater., doi:10.1038/nmat2403.) — Johanna Miller

Soft biological tissue is often subjected to forces that affect the tissue’s geometry, and finite elasticity provides a robust theoretical framework for analyzing the mechanical behavior of those tissues. Although the theory can accommodate anisotropic, nonlinear, and inhomogeneous processes subjected to large stresses and strains, its complexity makes many problems intractable. For growing tissue, though, the slow addition of cells generates shape- or size-changing stresses that are small enough to model successfully (see PHYSICS TODAY, April 2007, page 20). So, too, are simple geometries for tissues in equilibrium, even after those tissues are subjected to large stresses. Two recent papers have looked at applying the theory to those cases in thin elastic disks. In one recent study, Julien Dervaux and Martine Ben Amar (both of École Normale Supérieure, Paris) looked at anisotropic growth rates: If growth was faster in the radial than in the circumferential direction, the disk became conelike, while a reversal of rates generated saddle shapes. A separate study by Jemal Guven (National Autonomous University of Mexico) along with Martin Müller (ENS) and Ben Amar looked at excessively large circumferences for a given radius. Using the fully nonlinear theory, the researchers found an infinity of quantized equilibrium states for an ever-increasing perimeter at fixed radius. The ripples around the edge grew in size and number—not unlike the flower petals shown here—eventually crowding together enough to touch, like the ruffled collar in a portrait by Rembrandt. For more on the elasticity of thin sheets, see the article in PHYSICS TODAY, February 2007, page 33. (J. Dervaux, M. Ben Amar, Phys. Rev. Lett. 101, 068101, 2008; M. M. Mueller, M. Ben Amar, J. Guven, Phys. Rev. Lett., in press.) — Stephen G. Benka

In plasma physics as in fluid dynamics, turbulence remains one of the most challenging fundamental problems to understand. The nonlinear processes that lead to characteristic turbulence power spectra observed in the solar wind—the plasma that flows from the Sun out through the solar system—are poorly understood. So too are the dissipation mechanisms by which plasma turbulence transfers its energy to plasma electrons and ions. Most research in plasma turbulence has assumed that dissipation is weak and the plasma may be approximated as a fluid. But interest has increased in the so-called short-wavelength regime, in which dissipation plays out. Recently, an international team performed the first kinetic, particle-in-cell simulations of decaying short-wavelength whistler-mode turbulence in a collisionless plasma, using parameters similar to those of the solar wind near Earth. (Whistlers acquired their name from World War I radio operators who frequently heard what they thought were outgoing artillery shells, brief whistling sounds that decreased in frequency.) Limited not by any approximations but only by computing resources, the researchers found steep power-law magnetic fluctuation spectra consistent with those observed in space. In addition, they found and analyzed anisotropies in the turbulence whereby stronger initial fluctuations generated more magnetic energy perpendicular to the background magnetic field than along it. Because of the anisotropies, the whistlers were found to be more compressible than expected. The physicists also demonstrated the first simulation results of whistler turbulence dissipation by showing signatures of two well-known types of wave–particle interactions. (S. Saito et al., Phys. Plasmas, in press.) —Stephen G. Benka