Spring 2008
In This Issue:
Materials Modeling and Simulation
Janna Maranas studies polymers and biological systems with a combination of computer simulation and experimentation that is unusual among scientists, who tend to practice either theory or experiment, but rarely both together. By incorporating tools that range from test tubes to computers to a nuclear reactor, her students learn to attack problems that provide both basic knowledge and practical applications.
Trained in a computational approach while earning her Ph.D., Maranas continued computational research on the dynamics of proteins and polymers after joining the faculty at Penn State. But it was a field without a great deal of experimental research to validate her simulations. “I wanted to know if I was right, and I never knew,” she says. “It occurred to me that no one was really doing this kind of research experimentally, so maybe I should.”
To study polymers in motion, she taught herself how to do neutron scattering experiments using the nuclear reactor at the National Institute for Standards and Technology (NIST), located in Gaithersburg, MD. One of her experiments incorporating both neutron scattering and molecular simulation involves an unusual desert plant known as the Resurrection Plant.
The Resurrection Plant can survive without water for years in the dry Southwest, shriveling up during drought and growing when the rains arrive. The plant curls up into a tight ball, preserving a small amount of water in its center.
Scientist knew that the cells of the plant secrete a substance called a disaccharide that helps preserve the cells, but it was not clear how it worked. Using neutron scattering, Maranas and collaborators were able to observe that a sugar molecule immobilized the molecules called lipids that make up the outside of the cell. The immobility acted as a barrier to preserve water inside the cell. Then using molecular simulation, she was able to watch the process in more detail, observing a transition occurring within the lipid membrane that she could not have observed experimentally. “I can know how fast the lipids are moving in an average sense in the experiment, but I can’t know, for instance, that they are moving more at the bottom than at the top. That’s what you get from the computation,” she explains. Understanding how living cells can survive desiccation, as the Resurrection Plant does, has practical uses in the pharmaceutical industry.
Modeling Polymer Membranes for Lithium Ion Batteries
Current lithium ion batteries use a liquid membrane to move ions between the anode and cathode. The liquid requires a heavy casing and often contains toxic materials to improve the conductivity of the ions. A light, flexible and non toxic solid polymer membrane would be a highly desirable improvement for lightweight laptops and flexible cell phones, especially when it comes to their disposal.
Maranas and her graduate students Susan Fullerton, Kokonad Sinha, Kan-Ju Lin, and Wenjin Shi are using both experimental and computational instruments to improve the ability of a polymer called PEO to conduct ions at room temperature. Adding nanoparticles to a polymer can increase its conductivity, but the reason why is not understood. Some people believe that the polymer itself is in motion and that it grabs onto the ion and carries it. Her group has set out to see if that is the case.
So far they have completed the experimental observations using neutron scattering. They show that the polymer is indeed moving faster as a result of confinement by the added nanoparticles. They are now carrying out the simulations to see if the lithium ions are grabbed by the polymer and carried across the membrane to the other side. This requires a great deal of preliminary work to set up the appropriate force fields – the effects that the electrical charges have on one another.
Maranas says, “These days it’s not so hard to go into the literature and find what you need to set up computation. Almost any polymer, any protein, any lipid molecule I want to simulate I can find several force fields where the parameters are available that I can go in and try and see which one does a better job with my experimental data.”
But as more inorganic nanoparticles are added to the model, the simulations grow increasingly complicated. The right balance among lithium, nanoparticles, and the polymer has to be carefully selected. “I tend to think of it this way,” she says. “When you do an experiment you get the data out very easily. Then you spend a lot of time thinking about what it means. All the work is in the interpretation. When you do computation, a lot of the work is up front. But once you get the data, it’s pretty clear.”
Different Shapes Give Different Characteristics
Sometimes, really remarkable things happen when you change the shape of nanoparticles, according to Maranas and other researchers looking into the phenomenon. While doing simulations on a glass-forming material, boron oxide, her group happened upon a puzzling result. In a simulation in which the boron oxide particles were in the shape of a sphere, hexagonal rings of atoms similar to benzene rings formed on the interior of the sphere. However, when they changed the shape of the particle to a cube, all of the rings formed at the surfaces of the cube. In the simulation, the cube had a structure almost similar to a crystal.
It is already well known that materials at the nanoscale can have different optical and mechanical properties than their bulk counterpart. As electronics and biomaterials shrink to nanoscale, the shape of nanoparticles will also need to be taken into account, if indeed shape can affect surface properties, and considering that materials are mostly surface at the nanoscale.
In her spare time, graduate student Susan Fullerton, who is conducting the lithium battery experiments, is doing simulations on different shaped nanoparticles to see how their shape impacts the mobility, or other properties, of the polymer membrane she is studying.
Of the polymer mobility research, Maranas says, “The thing about nanoparticles in a polymer is they have an influence larger than you would expect. You put in a small amount and you get a large change in the property of the polymer. I don’t think people understand that well.”
Thinking Inside the Box
Simulations are performed inside a virtual box, a cube of space on a computer screen. The box represents a volume of space with, in the case of molecular dynamics, molecules moving around and interacting. Computer time increases rapidly with the number of particles interacting inside the box.
The typical size of her simulation box is 79 angstroms cubed, Maranas says. If the typical separation between nanoparticles is 50 angstroms, there is only room in the box for the atoms and molecules that make up one nanoparticle. Many polymers, including protein molecules, are so big that it often takes a supercomputer to model just one of them. “If you could do simulations in atomistic details of 20, 30, or 50 proteins, new areas of science would be stimulated,” Maranas says.
Until that happens, researchers rely on a technique called coarse grain modeling. Instead of using all the atoms in a simulation, they use only a few. For instance, in the lithium battery she replaces 14 atoms with one, a “blob” of atoms. “With the coarse grain model you can get the structure right, but you can’t get the dynamics – how they move – right. Specifically, they always move too fast. So we’ve been working on figuring out the specific way they are too fast, and we’ve been able to get some realistic results.”
Applied Science
Although she sees herself as doing basic science, there are practical applications for her group’s research that could benefit industry. “When I think about the things that people in industry might be interested in, I think about what kind of properties you get if you mix 2 or 3 or 4 or 10 polymers together,” Maranas says. Often there are simple rules that govern what kind of properties result from mixing polymers, usually based on the amount of each polymer. But one area where this doesn’t work well is with the dynamic properties, how the polymers move, which can be very important to industry. For instance, to get polymers into the shape of a car bumper, they have to be heated enough so the polymers’ molecules move around enough so they can be molded. The more they have to be heated the more energy it takes. Finding a polymer mix that can be molded at lower temperatures could have great cost benefit.
Maranas says, “I do both simulation and experimentation with the mixing of polymers. That’s one case where computation can start asking detailed questions. There are a lot of physical explanations as to why polymers don’t follow those simple mixing rules, but testing those ideas was not possible before simulation. When you mix two polymers together and measure the dynamics, even though the molecules are completely mixed together in some really intimate way, they still have their own individual dynamics. You can tell the difference – a moves differently from b. It’s not the same as when they were pure, but they’re still distinct. And the extent to which they’re distinct depends on the mixture.”
As well as studying the dynamics of polymers using computers and nuclear reactors, Maranas and her students also go into the lab and make things. “I taught myself essentially how to do experiments as a faculty member,” she says. “Now if you go into my lab we can make polymers and we can do all kinds of stuff. I can even go into my lab and blow glass.” For Maranas and her students, it’s a nice mix of theory and practice.
Contact:
Dr. Janna Maranas is associate professor of chemical engineering and materials science and engineering. She is on the faculty of the Penn State Center for the Study of Polymeric Systems, http://www.csps.psu.edu/.
