Five COE faculty win NSF CAREER Awards
The National Science Foundation has awarded five UW-Madison College of Engineering faculty with 2006 Faculty Early Career Development Awards (CAREER). The NSF CAREER awards, among the most prestigious given to faculty members who are just beginning their academic careers, are granted to creative projects that integrate research and education effectively. Faculty recipients include Assistant Professor of Electrical and Computer Engineering Irena Knezevic, Assistant Professor of Chemical and Biological Engineering Christos Maravelias, Assistant Professor of Biomedical Engineering Kristyn Masters, Assistant Professor of Soil Science and Civil and Environmental Engineering Joel Pedersen and Assistant Professor of Industrial and Systems Engineering Shiyu Zhou.
Just how fast will field-effect transistors for computing and other digital electronic applications eventually become? The very newest and tiniest of these voltage-controlled switches can toggle between their 0 and 1 states every picosecond, or one trillionth of a second — entering what is called the terahertz range. “The question is, 'Can you go beyond this?'” says Electrical and Computer Engineering Assistant Professor Irena Knezevic. “Because when transistors switch this quickly, they become very hard to characterize.”
With her five-year, $400,000 award, Knezevic will explore this question as she develops new quantum-mechanical theory and simulators capable of describing nanoscale devices operating on ultra-short time scales. The basic problem she will address lies in the behavior of electrons, whose flow within transistors controls the “on” and “off” states. In transistors that switch between these states much more slowly than once every picosecond, like those in today's computers, electrons act as individual entities. Thus, says Knezevic: “You can mathematically describe transistors one electron at a time.”
But when transistors switch at time scales of a picosecond or less, electrons begin displaying strange collective properties. “Now, you must treat all the electrons as one system,” she explains, “and that requires some heavy mathematical artillery and changes in the simulation software you use to predict electron behavior.” A major outcome of the proposed research will be a web-based virtual nanoelectronics laboratory (VNL), where graduate students will be able to access custom-made software and supporting materials to learn about solid state nanoelectronics and semiconductor electronic transport.
“The VNL will literally allow students to play on a computer with nanodevices — to make them, vary their features and see how they behave,” says Knezevic. With help from the UW-Madison Materials Research Science and Engineering Center, she also plans to create a version of the VNL suitable for high school students.
Chemical and Biological Engineering Assistant Professor Christos Maravelias will apply his $408,285 award toward modeling and optimizing the pharmaceutical R&D and supply chain. His goal is to enable pharmaceutical companies to manage more effectively their R&D pipelines to reduce the cost of developing new drugs.
To address these challenges, Maravelias will study four problems. First, he will consider the problem of project selection and portfolio management for the R&D pipeline. He will use simulation-based optimization to account for the uncertainty in clinical trials. Second, he will address the problem of resource planning for new product development using a two-stage approach: In the first stage, he will generate resource profiles for different realizations of uncertainty; in the second stage, he will solve a stochastic programming model to determine the optimal levels of in-house resources and outsourcing. Third, he will develop a stochastic programming approach for capacity planning under uncertainty in the outcome of clinical trials, the launch date of a new drug, and the market conditions.
Fourth, Maravelias will develop methods for effectively integrating production planning and scheduling in the pharmaceutical industry. The percentage of graduates in chemical engineering that join bio-related and pharmaceutical companies has increased significantly the last 10 years. At the same time, optimization methods are becoming very popular in the chemical industry. Maravelias is integrating into the curriculum both the subject and the methods used in his research. At the undergraduate level, he is developing case studies to be used in the senior process design course. He also is developing a new graduate-level course on optimization methods for chemical engineers. In addition, Maravelias will be involved as a research mentor in diversity-related programs and other initiatives to attract students from underrepresented groups.
Capitalizing on her overall research theme of quantifying cell-material interactions to advance the design of smarter materials, Biomedical Engineering Assistant Professor Kristyn Masters is using her $400,000 award to study the mechanical mechanisms involved in heart valve disease. Ultimately, she hopes to create a tissue-engineered heart valve. The ways in which material properties influence cell function are not well understood; as a result, researchers often are unable to control or direct cell function in the manner needed to create functional engineered tissues. “What we need to learn is how the cells interact with these materials in order to make the materials bioactive,” she says.
To do that, Masters essentially is taking several steps backward. She is studying how cell function changes in response to controlled systematic changes in biomaterial scaffold properties, and will apply the findings to building bioactive materials. In particular, Masters is studying the heart valve cells known as valvular interstitial cells, or VICs. Already she has learned that VICs are very responsive to both the mechanics and protein composition of their environment. As a result, she says, the types of extracellular matrices and protein environment present for VICs can tell VICs whether or not to become diseased. “We can actually form calcification in our in vitro cultures,” she says. “And calcification is one of the main causes of native valve failure — it becomes so stiff that it won't open right or close right.”
Masters is learning that engineered tissue is no different. “We've found that how stiff the material we culture these cells on can dictate how calcified they become, and the type of proteins present on the material's surface can dictate how diseased these cells become,” she says. On material surfaces, cells appear flat and spread out. They use intracellular fibers to grip and even pull those materials. The stiffer the material, the harder the cell pulls. “We are going to be measuring how the stiffness of the material affects the forces generated by the cell,” says Masters.
In her lab, there is growing evidence that how much the cell pulls on the material correlates to how many disease-related factors it produces. “And this pulling can actually initiate this disease program within the cell,” she says.
For Masters, the challenge is to determine what first triggers the cells to become diseased. “Does it start with a mechanical insult, or does it start with something else, where the cells decide to remodel their extracellular matrix, which then leads to mechanical stiffness, which then leads to more disease,” she says. “It's pretty much a chicken and egg problem.”
The answer might help Masters and others design better materials on which to engineer tissue. But this laboratory-based understanding of how material properties and the extracellular matrix regulate disease also could bring to light how heart disease occurs naturally within the body. “I didn't mean to develop a tool to be able to block the progression of heart valve disease, but that might be a consequence of our investigation,” she says.
Joel Pedersen, an assistant professor of soil science and civil and environmental engineering, will use his five-year, $400,036 grant to develop a highly sensitive quantitative method for measuring prion proteins — the rogue proteins thought to cause chronic wasting disease and other related diseases.
Prions are difficult to quantify, says Pedersen, because researchers often need to measure them at very low levels. “Methods that have been used in the past are mostly semi-quantitative,” he says. Pedersen will use his measurement method to study how prions interact with specific inorganic and organic components of soils. The work is key to understanding scrapie transmission in sheep and chronic wasting disease transmission in deer and elk, he says. “In both of these prion diseases, an environmental reservoir of infectivity appears to contribute to the transmission,” he says. “Understanding prion interaction with soil is also important for evaluating strategies to deal with prion-contaminated waste, such as infected carcasses — especially since landfilling is being considered as an option.” One of Pedersen's goals is to develop a quantitative understanding of prion interaction with soil that also will provide insight into the protein's mobility and bioavailability. “Understanding the role of soil in disease transmission may help in developing strategies to manage the disease,” he says.
The information age has transformed nearly every aspect of society, and manufacturing is no exception. Due to greater computing power and the extensive use of sensors, more data on manufacturing processes are available than ever before. Manufacturing has also become increasingly complex: Not only are individual operations — like emerging micro-fabrication techniques — more intricate, but manufacturers often must link tens to hundreds of them together to make a single product.
With his five-year, $545,600 award, Industrial and Systems Engineering Assistant Professor Shiyu Zhou now aims to bring the core approaches of industrial engineering in line with these revolutionary changes. Through a variety of educational and outreach activities, he also seeks to train future industrial engineers in analysis methods that can address the rapid, data-rich manufacturing environment of today and tomorrow. “The problem today isn't that we don't have the data to diagnose problems,” Zhou says. “The problem is we have too much data and it's hard to make sense of right now. That's why we need new tools.”
In developing these tools, Zhou will focus on one critical aspect of quality control, which is finding and fixing the root causes of variation in manufacturing processes. It's no easy task: In complex, multi-step processes, variation at any one stage could have arisen there (called local variation), or it could have emerged during an earlier step and been transmitted down the line (called propagated variation). “To quickly find the root causes, we first need to distinguish between these two levels of complexity,” says Zhou. To do so, he will develop graphical modeling techniques that can handle large and complex data sets. Once these methods narrow the cause of variation to a particular manufacturing step, he will then employ “self-improving” techniques that can learn patterns, or signatures, in product quality data indicating specific problems or errors.
It's all part, says Zhou, of a major research direction right now. “The question is, 'How do we fully exploit advances in technology and all the available data to revolutionize manufacturing,'” he says.