University of Southern California physicist Stephan Haas and electrical engineer Anthony Levi, have teamed up to build a computer program they expect will lead to a more systematic approach for the design of ultra-small devices — and spur the creation of new applications.
The duo’s interdisciplinary collaboration promises to advance quantum theory, computational modelling and the way engineers design new systems.
At the same time, the researchers are defining a new field they call adaptive quantum design that bridges the abstract realm of quantum physics and the more practical concerns of nano-scale engineering.
Quantum physics reveals a universe very different from what people expect based on everyday life, perception and experience.
“When things get very small, to the atomic or subatomic level, the classical laws of Newton don’t apply anymore. Quantum mechanics takes over,” said Haas, an associate professor of physics and astronomy in the USC College of Letters, Arts and Sciences.
Haas focuses on quantum theory. He studies the fundamental properties of matter, investigating the behavior of molecules, atoms and even smaller particles, as well as the consequences of their myriad interactions such as superconductivity and magnetism.
A basic tenet of quantum theory is that at very small scales, things begin to show a dual nature: particles (atoms, electrons, photons, etc.) behave like waves and vice-versa.
“Because of that, quantum theory predicts quite exotic effects,” Haas said. One example is entanglement, what Einstein called “spooky action at a distance.”
“Entanglement predicts that particles like electrons can become paired in such a way that, even at relatively far distances from each other, one’s behaviour affects the other’s,” Haas said.
Haas is best known for building sophisticated and highly complex computer models that predict how millions of atoms — a number well below the amount in a pinch of salt — interact and move in the quantum world.
“As a scientist, I study fundamental questions about how the quantum world works,” he said. Those understandings are proving important to the development of applications for nanotechnology — “such as how to design and build tiny switches, light filters, computer chips and other devices out of molecules or atoms.”
Levi, a professor in the USC Viterbi School of Engineering, said, “If nanotech is to succeed, it must have design tools similar to the sophisticated software used to design circuits in electronics.
“We want to create the design tool for nanotechnology,” he said.
Although they are still improving it, their quantum design software already has produced viable new designs for atomic, laser and millimeter wave (used in wireless communication) components.
Many of the computer-generated designs look nothing like anything a person would ever think up. They look random. And that’s exactly the point, said Levi, a leading researcher in fiber optics, electronics and the design of new technologies and systems.
“A computer is unbiased by past experiences or standard ideas of design, so instead of designing something ad-hoc, it is systematic. It searches through a near infinite number of geometric shapes or configurations to find the one that best fits,” said Levi, who holds a joint appointment in the USC College physics and astronomy department.
“What’s novel is that we’re doing everything in reverse,” he said. “Like the TV show ‘Jeopardy,’ we start with the solution.”
Once the computer comes up with the optimal design, Levi can manufacture a physical version to test in experiments. Data from the experiments are used to help improve the software.
The search for the optimal configuration — like looking through many haystacks for one needle, Levi said — is an extensive one and would not have been possible without the computational power of USC’s High Performance Computing Center.
Calculations that could take a regular desktop more than one year can be done in a single day using the HPCC resource. And that’s key.
“What we’re doing would not have been possible five years ago because we just didn’t have the computer power,” said Haas, who leads development of the software.
For Levi, the intellectual appeal of the project lies in creating software that can come up with better, novel, solutions.
“I’m interested in how the machines can teach us,” he said. “We’ve had to think hard about why the machine chooses certain designs.”