Taming Choas? German Scientists Force Checkerboard Patterns From Turbulence, Suggesting Better Control Strategies

The research, published in the 18 May issue of the journal, Science, may suggest new strategies for controlling a broad range of complex systems--such as the catalytic reactions that take place in cars, chemical factories, and the atmosphere. Because chemical waves also exist inside living cells, the study may ultimately prove useful for biotechnology, too.

How much control can scientists exert over chaos? Images of the Science experiment show blue and red stripes forming tidy, alternating rows, replacing a jumbled array of random turbulence.

These colorful images demonstrate a fundamental experiment in control: They reveal how the movement of (red) carbon monoxide molecules can be choreographed-during an oxidation reaction that releases carbon dioxide-to form regular patterns with (blue) oxygen atoms.

But, the principles at work in this simple system could someday be applied to far more complex industrial and environmental problems.

In addition to stripes, the Max-Planck team actually forced oxygen and carbon monoxide to form fractal-like designs across a platinum surface, similar to a checkerboard pattern. These "cluster" patterns repeated twice as fast as the random oscillations they replaced-a phenomenon called "period doubling," which demonstrates a high degree of control.

"Repeating cluster patterns have never been seen in this type of system before," said Harm Hinrich Rotermund, a co-author on the Science study, with Gerhard Ertl, Alexander S. Mikhailov and others. "Chemical turbulence can be very effectively controlled by our approach."

In theory, said Ertl, director of the department of physical chemistry at the Fritz Haber Institute: "It may be possible to increase the selectivity of various reactions, to improve the yield of a desired product. There may be other applications, too, from biology to climate studies."

Such findings represent the latest advance in "control theory." A field initially developed for industrial and military applications, control theory has been used for steering physical and chemical systems, for example, and for stabilizing missile trajectories.

"Control theory plays a role in simple problems, like regulating your heater with a thermostat, as well as more complicated issues, such as using a computer to optimize the fuel-air mixture in your car," explained Phillip Szuromi, a Science editor. "The central problem, in control theory, is the amount of time you have to control the system. A child shooting at toy ducks in a barrel faces a problem involving control theory. He's got to anticipate where the duck will be. To make that prediction, he must identify and measure properties that influence what the system can do."

Max-Planck scientists describe their control concept as "global delayed feedback." Using this principle, it's possible to tweak various aspects of a system, depending on information returned at regular time intervals by various components located "globally" throughout the entire system. Feedback can be used to create a control signal, which regulates system performance.

To put their theories to the test, Rotermund, Ertl and Mikhailov--with lead Science author Minseok Kim, as well as Matthias Bertram, Michael Pollmann and Alexander von Oertzen--examined the effect of global delayed feedback on carbon monoxide oxidation.

Inside a vacuum chamber, carbon monoxide (CO) molecules and oxygen (O) atoms adsorbed from the gas phase onto a catalytic, or chemically reactive platinum surface. Once adsorbed, the carbon monoxide molecules began diffusing and reacting with oxygen to produce carbon dioxide (CO2), which desorbed. Similar reactions help eliminate toxic CO from car exhaust.

Throughout this process, various two-dimensional patterns emerged as a result of surface changes induced by the reaction. Scientists manipulated these events by changing carbon monoxide pressures, which changed the reaction across the substrate.

Meanwhile, they captured all this activity using a camera-like device called a PEEM (photoemission electron microscope). High-energy light from the microscope causes photoelectrons to be ejected from the surface species, thus making it possible to identify them. An image of these emissions then revealed each of the system's chemical components, which appeared as brightly colored patterns.

This type of imaging technology can't be used in real-world settings, where high pressures may interfere with electron emissions. But, Rotermund said, another technology, previously developed at Max-Planck, could be coupled with the global delayed feedback strategy to control various systems outside the experimental vacuum chamber. Known as EMSI, for ellipsomicroscopy for surface imaging, the system is impervious to pressure changes. (See Science, 27 October 1995, pages 608-610.)

"The edge between chaos and high organization is very thin," Mikhailov said. "This research allows us to better understand the mechanisms of organization, and the functioning of biological, social and economic systems. Our results might also be used to engineer complex systems."