July/August 2004
http://www.coe.berkeley.edu/labnotes/0704
Commuter-friendly version

Dean Richard Newton, flanked by chemsitry Dean Clayton Hancock (left), ITRI President Johnsee Lee and Arun Majumdar, professor of mechanical engineering and director of the ITRI/UC Berkeley Research Center, signs the collaborative agreement creating the ITRI Center. (Peg Skorpinski photo)

"Energy is the single biggest technological issue that will haunt us for the next 50 years," says UC Berkeley mechanical engineering professor Arun Majumdar. "The environmental impact of continuing to use fossil fuels means that we have to look for other ways of producing energy."

To extend the search, UC Berkeley has entered an historic collaboration with Taiwan's largest research organization, the Industrial Technology Research Institute (ITRI). The Berkeley-ITRI Research Center will spur development of powerful energy technologies based on the university's nanoscale innovations--from flexible solar cells fabricated onto plastic to a "bio battery" powered by the glucose in your body.

This panel contains eight plastic solar cells based on nanorods and semiconducting polymers. The shuny ovals are the aluminum back electrodes of the individual solar cells. (UC Berkeley photo)

According to Majumdar, director of the Center, the new collaboration provides ITRI with "immediate access to UC Berkeley's basic research and the Silicon Vallley ecosystem."

"For us, it will help Berkeley continue to bring in the best people from around the world," he says.

The Berkeley-ITRI Research Center is affiliated with the Center for Information Technology Research in the Interest of Society (CITRIS). ITRI will provide $500,000 per year for five years to support ITRIS Fellows, graduate students and post-doctoral researchers working on the Center's projects. Visiting ITRI researchers hosted by the university will work with the ITRI Fellows and affiliated faculty.

"The nanotube in this transmission electron micrograph image has an internal diameter of about 10 nanometers, or 1/10,000 the diameter of a human hair. Nanotubes could be used as the basis for highly efficient energy conversion and storage devices." (courtesy the researchers)

For instance, mechanical engineering professor Carlos Fernandez-Pello is designing microreactors that could power laptop computers for hours longer than today's batteries. The reactors would utilize catalysts developed by Berkeley chemistry professor Gabor Somorjai. Meanwhile, Majumdar and his research group are studying semiconductor nanostructures with thermoelectric properties that convert heat into electricity and grow cold when current flows them. Someday, the materials could lead to power generators that run on waste heat, or even solid-state home refrigerators that are incredibly energy efficient.

Already, he adds, a Berkeley-ITRI collaboration is underway to develop a novel nanotech-enabled battery that converts salt water into electricity. ITRI researcher Ming-Chang Lu contributed expertise in the fabrication of nanofluidic arrays, the tiny "plumbing" system embedded in the device.

"All of the fundamental processes of energy conversion occur at the nanometer scale," Majumdar says. "So if you can manipulate things down at those scales, you might be able to increase the performance of existing devices and even create new methods of converting and storing energy."

Today, about 20 percent of the electricity consumed in the United States is generated by approximately 100 nuclear fission power plants. Nearly half of these commercial reactors first went online in the 1970s. After 30 years of extreme temperatures and constant radiation, how safe are they? That's one of the questions that Brian Wirth, a UC Berkeley professor of nuclear engineering, hopes to answer through materials science.

"Even though I'm in the nuclear engineering department, I really consider myself a materials scientist," Wirth says. "I want to understand the physical processes that are responsible for defect production and evolution in materials so we can predict when nuclear facilities are no longer safe to operate."

In the United States , commercial nuclear reactors have a licensed lifetime of 40years, yet, Wirth says, "most utilities would like to renew their licenses for another 20 years." The key though is determining whether the reactors are up to the task.

One threat is embrittlement of the reactor pressure vessel, the steel cylinder surrounding the nuclear fuel. The constant bombardment of neutrons causes the steel to weaken. A sudden change in temperature could crack the vessel. For example, in the event of an emergency, some reactors are designed to flood the core with cold water. This kind of thermal shock to alreadyembrittled steel is akin to filling a glass that's still hot from the dishwasher with ice water.

"Embrittlement could lead to catastrophic failure of the vessel and potentially a meltdown," Wirth says.

Fortunately, it's hasn't happened yet. That's due in part to stricter rules established by the Nuclear Regulatory Commission, Wirth says. Policy is best though when it's based in science. Using various scientific methods--from positron annihilation spectroscopy, a technique for studying the tiniest defects in solids, to experiments where pieces of steel are irradiated and studied to molecular modeling on supercomputers--Wirth and his colleagues hope to uncover the cause-and-effect between neutron exposure and material failure.

"We're trying to link up models of all the levels, from the interaction of electrons in the crystalline structure up to hundreds, thousands, and billions of atoms at the micron scale," he says.

Wirth already has a good track record. He's continuing a longtime collaboration with G. Robert Odette, his former graduate adviser at the University of Santa Barbara , to study how the copper used to weld the pressure vessel together causes nanoscale defects in the steel.

"It turns out that welding with copper was a terrible idea," Wirth says. "They also used recycled scrap metal from cars that contained copper."

Along with the science of keeping reactors of the past in service, Wirth is also focused on the future of nuclear energy. Nuclear fusion--smashing two hydrogen isotopes together to release energy--promises an effectively limitless fuel supply with orders of magnitude less radioactive waste than fission reactors.

"Fusion is a potentially clean energy source, but right now the commercial reactors are just designs on paper," Wirth says. "In addition to the plasma physics challenges, the materials challenges are so tough that we can't make fusion work yet."

To that end, he's developing multiscale models of how materials change in fusion environments. The greater researchers' fundamental scientific knowledge, he says, the more success they'll likely have in creating new alloys capable of containing the kind of energy that powers the sun.

"Materials are the limiting technological factor in a wide variety of industries," Wirth says. "It's where safety, performance, and economics come together. The more we learn of the fundamental science, the more insight we'll have when we try to create new materials."

Directed by Carlos F. Daganzo, the Volvo Centre will incorpoate research conducted by eight ITS faculty members drawn from the Departments of Civil and Environmental Engineering and City and Regional planning. (David Pescovitz photo)

Imagine that every time you drove into the heart of your city's downtown you had to pay a small fee. How would traffic be affected if that fee shifted predictably by time-of-day depending on the congestion in the city center? This is one of the ideas that UC Berkeley researchers hope to test drive in a new international center dedicated to the future of urban transportation policy and technology. The Volvo Research and Educational Foundations recently awarded $2.4 million over five years to the UC Berkeley Institute of Transportation to establish this Centre of Excellence.

"This is about technology pull rather than push," says Civil and Environmental Engineering professor Carlos Daganzo, principal investigator for the Volvo Centre. "We would like to come up with policies that might make cities better and then try to create the technologies that allow those policies to be put into place."

Data from loop detectors (the six large circles on the road) can be used to measure the speed and number of cars on the freeway. (courtesy PATH)

The Volvo Centre will collaborate with institutes and governments around the world to ensure that "the research is guided by a city's vision of its own future." For example, advanced traffic monitoring systems could lead to dynamic policies for controlling traffic and easing gridlock. These policies, Daganzo explains, would combine conventional traffic engineering tools such as signal timing and parking control with smart pricing schemes. Meanwhile, new wireless data services could help manage traffic flow and inform drivers about what they'll face on the road ahead. Even seemingly simple policy changes--restricting the times that freight delivery trucks can enter urban centers, for example--can have dramatic impact on traffic flow.

"Cities are very complex systems," Daganzo says. "If you do something small in one place, patterns may change all over. Yet despite this complexity, some macroscopic responses can be predicted."

The key to success, he explains, is accurately forecasting whether a new policy will work at the macroscopic level and measuring its performance after it's put into place. That involves building better methods to predict and monitor traffic.

"Right now, people don't have the data to run effective micro-simulations with enough precision to really try out policies," Daganzo says.

To that end, one of Daganzo's pet projects is modeling the ebb and flow of traffic in a bustling city. In his eyes, "city physics" are strikingly similar to the physics of, say, gas in a piston. Physicists, he explains, use mathematics to precisely predict behavior on the macroscopic level without knowing all the microscopic details. Daganzo believes his mathematical models can do the same thing for urban traffic without requiring the location or destination of every car. He's optimistic that his theories are correct, but now he needs to develop them further and, of course, check his work against reality.

Eventually, Daganzo hopes the Centre can work with several cities around the globe to instrument roadways with sensors that collect data on the number and speed of vehicles. Teams of Berkeley researchers are exploring various traffic-monitoring techniques, from strategically-located cameras to "loop detector" sensors buried in the asphalt. Once the cities are instrumented, the researchers' predictions can be compared to data from the real world. Even if those models prove invalid, Daganzo points out, instrumenting cities is still beneficial as it will provide instantaneous data on the effectiveness of new traffic management policies.

"We need the feedback mechanism to tell us whether we're on the right track or not," Daganzo says. 'Then we'll be able to determine whether the policies and technologies we develop at the Centre for one city will work in other places with similar problems."

Soda Hall

In the late 1980s, the Berkeley campus was gearing up to build a much-needed building for the computer science division. Carlo Séquin, the computer science professor heading the building committee, had already caught many flaws in the two-dimensional architectural floor plans. But, he thought, he could do his job much better if he could take a virtual walk through a three-dimensional model of the building.

However, Seth Teller, a Ph.D. student working with Séquin, realized that only a tiny fraction of the polygons in the Soda Hall model are visible from a given vantage point. So at any moment in a walk-through, the rendering program may need only about one percent of the model's polygons. "If you can find a way to avoid sending the other 99 percent through the graphics pipeline, you can make models that run 100 times faster or are 100 times more detailed than before," he says.

Another challenge remained: to get the polygons to the rendering program quickly enough to refresh the scene 30 times every second, which creates a fluid, realistic animation. The Soda Hall model was too large for a computer's memory, so the team was storing it on the RAID disk system created by Randy Katz and David Patterson's architecture group. But data had to be fetched from the disk each time it was needed. To avoid slowdowns, Tom Funkhouser, another of Séquin's Ph.D. students, came up with an algorithm that predicts where a user is likely to move next, and brings the relevant portions of the building into memory before they become visible. Funkhouser sped the Walkthru further by creating models of the furniture at multiple levels of detail, and developing an algorithm to determine the appropriate level of detail for each item in a frame in real time.

The final program could render detailed animations of Soda Hall, complete with rooms, staircases and furniture, in real time. "Seth and Tom created a system that demonstrated to the world how this could be done efficiently and robustly, and this general approach has now been widely adopted," Séquin says.