Predicting Material Performance

By Geneva Serrer Ringel

Transportation vehicles, from minivans to fighter jets, continuously require new materials with greater structural integrity and improved mechanical properties for improved performance, increased personal safety and reduced costs. The development of these high performance materials can be expensive and time consuming - two factors that are not compatible with the accelerated product development cycle and the economic climate of the transportation industry.

The Center for the Accelerated Maturation of Materials promises a new method for the development of mature high performance materials that can meet the demands of industry; its initial focus has been the aerospace and automotive sectors, both extremely important to Ohio. By combining computational resources with experimental characterization and validation, advanced and improved materials can be more rapidly and cheaply developed while simultaneously providing the quality assurance necessary for commercial use.

Gathering the Data

In 1998, Hamish Fraser, Ohio Regents Eminent Scholar and professor of materials science and engineering, formed the Center for the Accelerated Maturation of Materials in order to develop a new approach to the maturation of advanced materials. With funding from the National Science Foundation and the Department of Defense, Fraser assembled a team of international researchers to explore the acceleration of the maturation process by developing and applying computational techniques.

Hamish Fraser, Ohio Regents Eminent Scholar and professor of materials science and engineering, formed CAMM to develop a new approach to the maturation of advanced materials.     Photo Geneva Ringel

The maturation of a material is the process of generating enough information for engineers to confidently predict its performance and mitigate the risk associated with its use in components. Traditionally, engineers have relied on experimental performance data to provide this information, a long and expensive process. For turbine engine disk materials, the maturation process entails an extensive test matrix that can take as long as 10 years and $35 million.

For every new alloy developed, no matter how small the change in chemical composition, the entire process has to be completed. Even though the cost of the maturation process is high, the cost of material failure is even higher.

“Financial concerns associated with liability require high performance materials developed for commercial applications to have a high degree of maturity,” explained Fraser.

In addition to increasing development costs, the traditional maturation process does not coincide with product development cycles. New aircraft engines are being designed to meet the higher temperature and higher load needs of commercial and military customers in as little as 25 months. Waiting 10 years to obtain performance data on a new material for that engine isn’t economical. Instead, engineers select from existing materials that meet the accelerated timeline and compromise on the performance criteria demanded by the product.

New Approaches to Materials Maturation: From Inception to Implementation

Advances in high performance materials have been enabled by developments in materials characterization, especially that involving electron microscopy. Engineers are able to directly observe and fully characterize the microstructural features of a material from the atomic scale. As a result, materials scientists have discovered that the mechanical properties of an alloy are determined by its microstructure, the nature of the constituent phases (a phase is a homogeneous portion of a system that has uniform physical and chemical characteristics) and the manner in which they are spatially arranged. Modifications to the microstructure, either through the different elements present in the alloy or the application of heat and mechanical stress, can result in enhanced performance or alter the way the material fails.

This knowledge has enabled researchers to develop new materials with specific properties by understanding and manipulating an alloy’s microstructure. CAMM wanted to apply the same approach to predicting the performance of materials and accelerating the maturation process.

“By understanding the microstructures of the material and relating them to properties, one can predict material performance,” said Mary Juhas, associate director of CAMM and senior assistant dean for outreach and special programs.

CAMM’s goal was to use computers to generate these relationships between the microstructure of an alloy and its properties in order to predict materials performance. Other materials, such as electronic materials, had previously incorporated computational techniques to aid in process design, but these techniques had contributed little to the development of structural materials. The wide variety of phases present within an alloyed material and a lack of theory to guide computational development had limited the ability to apply computational techniques.

Achieving Results

The properties of all materials ultimately depend on their component atoms. The long-term goal of CAMM is to base the computational tools on accurate first-principles calculations, using quantum mechanisms to describe electrons, spins and nuclear motions. But at present this is simply too difficult to be accomplished. Piecemeal progress, however, is possible under various simplifications.

“Interestingly, when we first started the Center, no industry partners would believe us. We said that if we don’t start now, we will never have computational tools. Now, CAMM’s idea has been a major government funding thrust,” explained Yunzhi Wang, associate professor of materials science.

The computational tools that are being developed at CAMM are based on a combination of first-principles calculations, coarse-grained mesoscale microstructure and dislocation level models, continuum-level phenomenological frameworks and/or rules-based approaches such as neural networks. The specific types of tools that are being developed are dictated by the focus on specific materials problems.

“The materials are so complex that there is no established theory about how the atoms interact,” explained Yunzhi Wang. “Material properties are averaged over a material area, but you have to know how all the atoms interact because defects that can cause material failure are localized. There are so many combinations possible for these defects, it is hard for theory to make a prediction of where they will occur. Mother nature tries to tell us, but can’t speak.

"The answer is embedded in the data. If you can identify the cause and the result, neural networks and fuzzy logic can identify functional relationships between the microstructure and performance.”

Conducting a process that substitutes for theory is no small task. CAMM researchers use highly specialized facilities for advanced characterization to examine material down to the atomic level in order to identify its microstructural features - the cause. At the same time, critical experiments are performed to provide quantified observations of the materials performance - the results.

Once the cause and results are identified and analyzed, sophisticated databases relate the material’s microstructural features to its performance. In order to make the data accessible and easily available to engineers making materials selection, CAMM also had to develop neural networks capable of analyzing the information in the database to identify functional dependencies and ultimately predicting the performance of materials – an efficient method of data mining.

This new paradigm for materials maturation provides the necessary information for engineers to confidently select new advanced materials for use in commercial applications. As a result, better materials will enable better products.

CAMM’s method also delivers considerable savings in terms of time and resources. Fraser explained that the CAMM method has the potential to reduce the cost of developing new materials, such as aluminum, titanium and nickel-based superalloys, by a factor of five and shorten the development cycle to as little as 30 months.

“It is the intimate coupling of the computational activity with the experimental characterization and validation that is the key to CAMM’s success,” said Fraser.

Computational techniques may also provide further advantages to the development of advanced materials in the near future. Building on the success of its neural network, CAMM is developing computational tools that will allow engineers to simulate the performance of a material on a computer. This could impact not only the development of new transportation products, but the way in which their performance is tested. These advances may also one day have impact our everyday life. Just imagine test driving a new car, or even a fighter jet, from your home computer.

Secondary Benefits of CAMM Pete Collins, materials science graduate student, uses LENS™ to engineer graded materials based on prealloyed Ti powders with a variety of composition profiles. Structures created with LENS™ are engineered layer-by-layer while controlling their composition and microstructure. This means that material properties can be controlled as well as localized. In addition to accelerating the maturation process, knowledge obtained through CAMM contributes to many related research activities, for example, the laser engineered net-shaping (LENS™) research conducted by Fraser’s research group. This technology use lasers to deposit powdered metals to create structures during a direct laser deposition process. LENS™ has significant potential benefits to the aerospace industry because it can result in savings in cost and weight while improving performance. The elemental powder blends in LENS processing, which result in significant savings in terms of raw material costs, exploit the thermodynamic enthalpy of mixing to aid the laser deposition process.

For more information about CAMM please contact Hamish Fraser at fraser.3@osu.edu

CAMM wishes to acknowledge the financial support it has received from the National Science Foundation, the Air Force Office of Scientific Research, the Air Force Research Laboratory, the Defense Advanced research Projects Agency, and several partnering companies.