Research uses clusters to explore how biomimetic composite materials meet and greet

June 12, 2012

To the computer savvy, "interface" might bring to mind the way a computer operating system like Windows presents the world inside the machine to its users. Business types might hear interface and think of meeting to explore mutual interests.

One doesn't get face-to-face with the interfaces that interest Vikas Tomar, however. These meetings take place at the level of atoms. The Purdue aeronautics and astronautics professor's Interfacial Multiphysics Lab is helping develop new and improved composite materials with a variety of potential uses, in part by using the way nature builds composites at the nanoscale - bone, for example - as a guide.

Purdue's Hansen and other community cluster supercomputers are important tools in Tomar's research, which focuses on biomimetic ceramic and polymer composites that could find application in fields ranging from medicine (improved replacement hips) to energy (engines capable of burning fuel hotter and, hence, more efficiently, as well as next-generation nuclear power generators).

He and his students use a mix of computationally demanding simulations and proprietary physical experiments developed in his lab to study the places where the "faces" of materials meet at an atomic level — the interfaces of composites. Their aim is to uncover how those interfaces might be manipulated to influence functional and mechanical properties together or separately, for instance, to make a material that performs better at higher temperatures or is more flaw tolerant.

Tomar works with regular, inorganic engineering materials but also "biomimetics," in which researchers draw on the ways natural materials work to design and engineer new materials. One of the model materials he uses, tropocollagen-hydroxyapatite nanocomposite, is made up of a ceramic, hydroxyapatite, and tropocollagen, a polymer akin to a protein common in skin, muscles and other parts of the body

Among other things, the research has helped advance understanding of osteogenesis imperfecta, known as brittle bone disease, a genetic bone disorder related to a collagen protein deficiency.

Moreover, from a design and development standpoint these materials are interesting for the variety of properties they can take on with just slight chemical changes at the nanoscale. Skin is thin and stretchy. Bone is thick and mass bearing. Muscle is somewhere in between.

From a nanoscale perspective, a mere centimeter of a new material can contain thousands of trillions of atoms whose vibrations and interactions Tomar's lab tries to understand. While even today's most powerful supercomputers are unable to simulate that fully, Tomar has developed a mix of techniques that yield accurate simulation results supplemented by experimental results in his lab to fill in gaps left by the computational limitations. He uses a suite of molecular dynamics and finite element codes, some of which he developed originally and others he extended from NWChem, developed by Pacific Northwest National Laboratory, and DL_POLY, developed by British researchers. His methodology allows him to examine materials from scales of a few hundred atoms all the way up to meters.

At Notre Dame, before he came to Purdue, Tomar's lab had its own high-performance computing cluster, but he's sold on Purdue's shared Community Cluster Program.

"I've run my own cluster and what I've seen is that once you start running the codes there are significant issues," Tomar says. "With the excellent support staff we have and the amount of money we invest, I think it's the best deal anybody can get anywhere. I just have to log in and run my codes. It takes a lot of these troubles away from me so I can focus on research proposals and research papers and also invest a lot of time in teaching rather than repairing a cluster."

More Information

Originally posted: July 1, 2014  4:09pm