Modeling tool speeds optical materials research

Assistant Professor Susan Hagness and graduate student Tae-Woo Lee (30K JPG)

In our everyday experience, the intensity of naturally occurring light is low enough to preclude interactions between photons and the atomic structure of material through which the light wave propogates. However, a high-intensity light wave will interact with certain types of so-called nonlinear optical materials and undergo fundamental changes. For example, through nonlinear frequency conversion, a beam of light that enters crystals of gallium arsenide or lithium niobate can emerge as a different color.

The ability to manipulate light can be further enhanced by artificially engineering additional structural features in a nonlinear optical material. Recent advances in materials technology and nanofabrication techniques have made it feasible to build optical structures with physical dimensions smaller than the optical wavelength.

By understanding the behavior of light in these submicron or nanoscale structures, researchers can make light perform the functions of many microelectronic devices. Ultimately, researchers may use these materials and devices to replace the wires and electrons of integrated circuits with optical fibers and light resulting in processors thousands of times faster than those used today.

Now, thanks to a new simulation tool developed by Assistant Professor Susan Hagness and graduate student Tae-Woo Lee, advances in new devices for optical communication, signal processing and computing will occur more rapidly. The team has developed a computer algorithm for modeling the propagation of light in nonlinear photonic microstructures with unprecedented accuracy and computational efficiency.

"With our new computational algorithm, we can conduct numerical simulations that serve as a virtual lab bench for investigating light propagation in optical materials that exhibit second-order nonlinearities," says Hagness.

"Standard first-principles techniques for solving the governing nonlinear Maxwell's equations consumed vast amounts of computer resources over many hours. Our new technique achieves accurate results on the order of 50 to 70 times faster."

Previously, efficient modeling techniques worked well only for conventional optical devices, which at 500 microns or more are very small themselves. At that size, certain approximations can be made about the behavior of light, which simplify the computations. But when the device itself is no larger than the wavelength of light in use, Hagness says the analysis of wave propagation becomes much more complicated. Techniques for solving simplified versions of Maxwell's equations are no longer accurate. "At a very basic level this new simulation tool helps us better understand the physics. It allows us to conduct numerical experiments and develop new insights about some of the fundamental properties of light in nonlinear photonic microstructures," Hagness says. "And once we have a good understanding of the basic physics, then we can use large-scale computations in the design process to help optimize designs prior to fabrication. When you're fabricating devices on the nanoscale, that is a big motivation."