The grant was announced today (Wed., Sept. 25) by NSF's Information Technology Research initiative, which supports research at the interface between engineering and science, as well as research on the implications of information technology for society.

The problem with the technologies that proved quantum computing possible is that they will not scale up to the size of quantum processors, which involve hundreds or thousands of quantum bits, or qubits, said project leader K. Birgitta Whaley, professor of chemistry. Her group will look beyond these technologies to explore the feasibility of solid state quantum computers, which are scalable.

"Our goal is to find a material that would make a practical quantum nanoprocessor," she said.

Quantum processors are being hailed as the ultimate computer, potentially able to do millions or billions of calculations simultaneously, unlike today's computers, which calculate sequentially. Quantum nanoprocessors would be a thousand times smaller than today's silicon-based microprocessors, breaking into a new realm beyond Moore's Law. Moore's Law is an observation by Intel co-founder Gordon Moore that the number of transistors on a computer chip - an indication of the chip's computing power - doubles roughly every 18 months.

The astounding promise of quantum computing is based on the fact that each quantum bit is entangled with every other bit in the computer, so that a single manipulation affects them all. Though these capabilities aren't necessary for word processing or simple math calculations, they would speed immensely data encryption and decryption or quantum mechanical calculations.

"There are lots of theoretical papers on quantum computing and quantum information processing, but the real task is how to experimentally implement these ideas," said Michael F. Crommie, associate professor of physics. "We'd like to exploit condensed matter, since engineers have had so much success making solid state devices that are everywhere today - the chips in cell phones, computers and even washing machines. We want to create comparable devices taking advantage of quantum mechanical effects."

Over the grant's five years, the group also will evaluate new techniques needed to control and measure qubits on the scale of atoms. Among these are ultra-low temperature scanning tunneling microscopes, which can pick up and move single atoms on the surface of a material, and superconducting quantum interference devices, or SQUIDS, sensitive enough to measure magnetic emanations from the brain.

"We are working at the interface between theory and experiment," said Whaley, a theoretical chemical physicist. "We plan to explore new architectures, synthesize nanostructures optimal for quantum operation, and learn how to control and manipulate entanglement to do quantum information processing."

The fundamental building block of any digital computer is the bit - something that can exist in either of two states that represent the binary units 0 or 1. In today's infant quantum computers, the quantum bit, or qubit, is usually the up or down spin of an atomic nucleus or electron.

What makes quantum computers unique is that, at very low temperatures, quantum weirdness sets in. Qubits, normally restricted to the states 0 or 1, suddenly can be in limbo between the two. Only when probed or measured do they settle into a 0 or 1. In addition, qubits interact with one another in a coherent way that hides their individual states until some probe forces them to disentangle and settle into a state of 0 or 1. Though complicated to manipulate, qubits theoretically can do wonders far beyond the capability of today's digital computers.

Last year, scientists at IBM and Stanford University created a seven-qubit computer employing designer molecules in a test tube. The spins of these molecules - the qubits - were detected with nuclear magnetic resonance. Employing a special algorithm, the computer performed a simple factorization of 15 - in other words, it found two numbers that, when multiplied, equal 15. The best encryption schemes are based on 100+ digit numbers that are hard to factor even with today's most powerful computers.

Other simple quantum computations have involved qubits represented by the spin of an atom captured by lasers in an atom-trap or the polarization of light waves, or photons. Theoretically, an array of qubits could be manipulated to perform all the logical operations of today's silicon-based computer processors.

The UC Berkeley team will investigate alternative, solid state qubits: the nuclear spin states of phosphorus atoms immobilized in silicon; electron spin states of atoms near the surface of thin magnetic layers, what the researchers refer to as atomic-scale spin valves; the spin states of molecules encaged in nanotubes like peas in a pod; and the circulating current in microscopic superconducting rings.

The UC Berkeley experimental physicists investigating these possible qubit systems are J. C. Seamus Davis, Crommie, Alex Zettl and John Clarke - all from the Department of Physics.

Whaley and control theory expert Shankar Sastry, professor and chair of the Department of Electrical Engineering and Computer Sciences, will explore the theory and control of quantum computers, develop new efficient procedures to manipulate and extract information from qubits, and find ways to minimize the tendency of the world to creep in and destroy quantum states. Called decoherence, this becomes a major problem when qubits are packed into a solid material.

"Condensed phase materials have an exciting potential because of their scalability - we could take advantage of the nanofabrication techniques used in Silicon Valley - but the condensed phase causes greater decoherence problems than for isolated atoms in the gas phase, because qubits can now interact with so many other entities surrounding them," Whaley said. "One of the big questions of quantum computing is whether decoherence will kill you."

She believes that the problem of decoherence can be mitigated by using error-correcting methods and more bits to build the computer.

"We will look at this problem from a different perspective, using control theory techniques to minimize decoherence," she said.

One goal is to design logic circuits around the physical properties of each type of qubit, rather than try to use the same approach for all systems.