Major Areas of Opportunity - Implications of Emerging Micro-and Nanotechnologies - page 38

investigation of alternative schemes such as cellular automata. Quantum comput-
ing provides a tantalizing hint of orders-of-magnitude increases in computing
power for some important operations such as factoring large numbers, but today
it remains beyond reach and is the subject of intense worldwide investigation.
The second theme is the evident superiority of the human brain over even the
most advanced computers for some computational tasks such as image recogni-
tion. Biomimetic investigations have long sought to capture the processing of the
brain, which involves analog functions and massive storage. In general it may be
anticipated that alternative approaches to computing involving analog or other
yet-to-be-invented methods may someday have an impact on computing.
Cellular Automata
The term “cellular automata” (CA) generally refers to a group or population
of interacting cells in which each cell (automaton) behaves according to a well-
defined rule set. Each cell has a limited number of well-defined states. The cells
are dynamic, and the properties of each cell evolve with time in a manner depen-
dent on the states of neighboring cells. Complex phenomena may be simulated
with an appropriate set of criteria, much as fluid behavior “solves” the Navier-
Stokes equation for interacting fluid “cells” or for population behavior by means
of interactions among individual members of a group.
Quantum-dot cellular automata (QDCA) represent one such implementation
of this basic idea. They consist of arrays of quantum dots in which the state of an
array exists in various electronic configurations. The array evolves in time de-
pending on the states of the interacting cells. A group at the University of Notre
Dame has theoretically and experimentally investigated one such model exten-
sively. 48,49 A recent review provides an excellent overview of this and related
work. 50 The fundamental cell is a square array of four quantum dots (each dot at
a corner of a square) and two electrons. The preferred locations of the two
electrons are at opposite corners of the square. Since there are two ways to
arrange these electrons, each having the same energy, this basic unit may repre-
sent a bit (one or zero) depending on the arrangement taken by the two electrons.
By judicious placement of these cells, the behavior of wires and conventional
logic gates may be mimicked. QDCA evolve with time in a continuous manner
rather than in discrete time steps. Logical operations may be demonstrated; more
complex behavior depends on the array geometry and size.
Input to a QDCA consists of initiating the state of several cells with neigh-
boring charged wires. The circuit then relaxes with time to the lowest ground
state. The output is read by measuring the polarities at output cells; this represents
the solution to the problem (as defined by the geometry of the array). The behav-
ior of these arrays has been simulated and demonstrated at the micrometer level
with lithographically fabricated metallic islands, and logical operations have been
demonstrated. Because of the large size of the dots, these arrays must be cooled to