News

by Eric J. Lerner

pdf version of this article

Hot writing

In the search for new ways to create circuits and other patterns at nanometer scales, one actively researched idea has been to use arrays of atomic force microscope (AFM) tips to write the patterns directly onto substrate surfaces. In this approach, a tiny AFM cantilever is maneuvered over a surface to deposit the coating molecules (the ink) wherever it touches the material. Researchers have developed arrays of such cantilevers, which potentially could create complex patterns at scales smaller than 100 nm, without the complicated process used in conventional photolithography of creating masks and projecting them onto substrates.

However, this dip-pen nanolithography, or DPN (see The Industrial Physicist, December 2000, p. 28), has had two major drawbacks. For one, the pen writes as long as it is in contact with the surface. Lifting the pen off the surface loses registration; that is, it is not easy to find the exact place to resume writing. Second, the resulting patterns cannot be measured with the same tip that writes them. An uncoated tip with no ink on it must be used, which also necessitates time-consuming re-registering of the tip with the tiny pattern.

Researchers at the Naval Research Laboratory (NRL) in Washington, DC, and Georgia Institute of Technology (Atlanta, GA) have developed a simple solution to these problems that could move DPN closer to practicality. They use materials with high melting temperatures as the inks, and heat the tiny cantilever to melt the ink and allow the writing. When the heating stops, so does the writing, even if the tip remains in contact with the surface (Appl. Phys. Lett. 2004, 85, 1589).

To test the idea, the team used octadecylphosphonic acid (OPA), a material that melts at 99 °C, and wrote on a surface of mica, which is easy to prepare. When they raised the temperature of the cantilever to 122 °C, full deposition of the OPA started instantly. However, when they turned the heat off, writing continued for about 2 min. Because the cooling time for the cantilever tip was only about 10 ns, the researchers concluded that the heat content of the mica, an excellent insulator, kept the OPA molten. On silicon or metal substrates, more likely practical targets with higher heat conductivity, cooling time could be reduced to around 10 µs, making writing practical with an array of several thousand cantilevers.

“Our next step is to write a complete circuit with the device,” says team member Paul Sheehan of NRL, who worked with William King of Georgia Tech and Lloyd Whitman of NRL. “We also intend to use practical writing materials such as a semiconducting polymer.” Although the technology might be used eventually for writing entire microchips, early applications would involve repairing masks used for conventional photolithography.

Blackout clears the air

When 100 power plants in northeastern North America shut down in the blackout of August 2003, the outage had one beneficial effect. Suddenly, the air over the region contained fewer pollutants from the effluents of the power plants. Although this decrease may offer cold comfort to the tens of millions of people inconvenienced by the blackout, it did present an opportunity for scientists to see how much pollution fossil- fuel plants produce. A recently published study by a University of Maryland research team provided a surprising answer: even more than previously believed (Geophys. Res. Lett. 2004, 31, L13106).

Fossil-fuel power plants produce a number of pollutants, particularly nitrogen oxides and sulfur dioxide. “Nitrogen oxides combine with volatile organic compounds in the presence of sunlight to produce ozone, the main component of smog,” explains team leader Lackson Marufu of the university’s department of meteorology. Sulfur dioxide creates sulfuric acid when dissolved in water, and it can be oxidized to produce sulfates, which precipitate into particles that, in turn, generate haze. These pollutants, generated both in the Northeast and Midwest, are carried by winds to the Eastern Seaboard, where they reduce visibility, adversely affect health, generate acid rain, and contribute to climate change.

There are, of course, other large sources of air pollution, especially the exhaust from automobiles. To quantify the contribution made by the power plants, the Maryland team sent two light aircraft aloft on August 15, 2003, a day after the start of the blackout, when most power plants remained shut down. One flight made its measurements outside the blackout area—over Cumberland, Maryland, and nearby parts of Virginia—and the other over Selinsgrove, Pennsylvania, in the heart of the blackout region. The rural sites were chosen to minimize the contribution from urban, nonpower- plant sources, such as automobiles. The planes flew in vertical spirals to sample various altitudes and measured ozone, carbon monoxide, sulfur dioxide, light scattered by suspended particles, and particle counts. The team compared the results with survey findings obtained a year earlier over Selinsgrove on August 4, 2002, a day when the power plants were running normally and weather conditions were similar to those on August 15, 2003.

(University of Maryland Department of Meteorology)

“What surprised us was the extent and speed with which the pollution levels fell,” says Marufu. Over Selinsgrove, sulfur dioxide dropped by 92%, ozone by more than 50%, and the light scattered by particles by more than 70%. Visual range increased from 25 km to a crystal-clear 65 km. In contrast, the pollution levels over Cumberland were essentially the same as the 2002 readings from Selinsgrove. The team confirmed that the reduction resulted entirely from the power-plant shutoff and not from any fall in contributions from cars. For one thing, lightabsorbing particles, emitted only by autos— in contrast to the light-scattering particles emitted by power plants—actually increased during the blackout, and on-the-ground traffic counts confirmed that traffic levels were almost unaffected by the blackout. The reductions are greater than could be anticipated from previous estimates of power-plant pollution. “We do not have figures for the Northeast alone, but nationally, the Environmental Protection Agency [EPA] estimates that power plants produce 69% of sulfur dioxide pollution and 22% of nitrogen oxide,” says Marufu. “Clearly, the blackout measurements show that, at least in the Northeast, this is an underestimate.” Although some plants upwind of Selinsgrove (about a quarter of normal) were still operating on August 15, the measurements indicated that power plants contribute at least 90% of sulfur dioxide and at least 50% of nitrogen oxide (and, thus, ozone) pollution in the region. This may mean that EPA is seriously underestimating power-plant pollution. “The good news is that pollution cleared up in 24 hours,” points out Marufu. “This means that if we could ever get clean power, pollution will go away immediately.”

Fighting big blackouts

No one would argue that regional blackouts are a good way to reduce pollution, or should happen at all. But procedures designed to prevent the spread of the August 2003 blackout actually helped it to spread. An electrical grid, overstressed by long-distance transfers of power, collapsed when local events, such as a power line sagging into a tree, shifted power onto lines already near capacity, which caused an expanding cascade of failures. Such a network collapse can occur on the Internet as well, if a concerted attack overloads key nodes (see The Industrial Physicist, December 2000, p. 14).

Analysts pointed out at the time of the 2003 blackout that changes in regulations could greatly reduce the long-distance energy flows that overstressed the system, and that greater communication among utilities could help to prevent the cascade of a regional blackout (see The Industrial Physicist, October/November 2003, pp. 8–13).

Other researchers asked whether general strategies could be used to stop local failures in a network, even large ones, from producing huge cascades of failure. Adilson E. Motter of the Max Planck Institute for the Physics of Complex Systems might have found such a strategy, which works with simulations of simplified networks (Phys. Rev. Lett. 2004, 93, 098701-1). These networks consist of nodes connected to each other through a system of edges, which could be either power lines or information- carrying lines. Once a failure occurs, an algorithm would prevent the spread of the failure by intentionally removing certain nodes and edges.

In the present- day electrical grid, controllers do shed load with deliberate blackouts when the system nears capacity. But no systematic way exists to determine which loads to shed, and in practice, automatic systems that shut down overloaded lines (edges) to protect power plants (nodes) have often worked to spread the blackouts, not to contain them. The algorithm devised by Motter is based on two simple rules—nodes near the initial failure carr ying small loads are removed, and edges carrying overly excessive loads are also removed. Excess loads are defined as the increase in load that has occurred as a result of the initial failure. The removal of the nodes carrying small loads tends to isolate the failure and prevent it from spreading. Shutting down the edges whose loads have increased most severely tends to spread the excess load more evenly among the remaining edges and, thus, prevents additional nodes from becoming overloaded.

Motter emphasizes that more work needs to be done before his strategy could be implemented on real networks. “My model is the simplest model that allows for comprehensive analysis, and it is not intended to be realistic for electrical networks,” he says. “Modeling such realistic networks is the next step.”

Bacteria stir things up

Bacteria tend to concentrate into assemblies mixed with polymers that they exude. These biofilms, which help to protect their inhabitants from outside attack by antibiotics, are a major cause of tooth decay and infections, as well as a contributor to corrosion. But how do these dense concentrations form? University of Arizona physicists and mathematicians have proposed an intriguing answer, which could lead to ways to combat the formation of such films. The bacteria concentrate themselves and generate adequate oxygen flow by swimming in correlated patterns that create water jets and vortices larger than the bacteria themselves (Phys. Rev. Lett. 2004, 93, 098103-1).

Aerobic bacteria require oxygen, in contrast to anaerobic bacteria such as the one whose toxin causes botulism, or the E. coli that live in digestive tracts. Because aerobic bacteria inside water droplets swim toward the droplet surface where oxygen is most abundant, they can begin to concentrate themselves in densities that approach a billion bacteria per cubic millimeter. Being 10% heavier than water, the bacteria slide down the edge of the drop and concentrate themselves further. If that were the whole story, further concentration would be limited by the rate at which the oxygen diffused passively into the water and the rate the bacteria consumed it.

Bacteria, however, overcome this limitation by creating fluid flows in the water. An individual bacterium swimming 10 to 30 µm a second or so does not create much of a ripple in the flow of the surrounding water. With billions of bacteria, the situation is different. As a random local concentration of bacteria starts swimming toward the surface, its collective action sets up flows in the water, which draw in more bacteria and increase the flows. When the bacteria reach the surface, their weight carries them downward, creating a return flow—a process somewhat similar to thermal convection. The net result is to cause upward- and downward-moving plumes. In the most concentrated regions, jets of cells and vortices arise. Some have velocities several times that of the bacteria’s swimming speeds, and widths of more than 100 µm, 100 times the diameter of the rodshaped bacteria.

“This phenomenon is only possible because tens of thousands of bacteria are swimming in parallel,” points out John Kessler of the University of Arizona’s department of physics. “Those swimming in parallel with the jet can move faster than those swimming at an angle to the flow, because they are swept along by the current. So the bacteria naturally tend to align along the lines of flow.”

The accelerated flow of the water inside the droplet hastens the transport of oxygen away from the surface and allows large bacterial populations to sustain themselves. Equally important, the vortices concentrate the bacteria, allowing them to chemically sense each other’s proximity. “A sufficient number of cells signaling to one another— causing changes in behavior as well as changes in physiology and gene expression— is called quorum sensing,” says Kessler. “It appears to be an important ingredient in the development of significant collective behaviors, including the formation of biofilms.”