The Development of Quantum-Dot Cellular Automata - Field-Coupled Nanocomputing: Paradigms, Progress, and Perspectives - page 10

Another approach that combines single-atom realization with lithographic control
is the STM-base lithography of the Wolkow group [ 71 ]. They have created room-
temperature QCA cells using a remarkable approach involving removing single
electrons from dangling bonds on a silicon surface. As with molecules, the single-
atom sizes easily yield room temperature operation, yet the placement and orientation
of the cells can be controlled lithographically using the STM tip.

3.4 Other Implementations

Nanomagnetic QCA was first introduced by Cowburn's group [ 72 ] and developed
extensively by the Porod group [ 73 ] and the Bokar group [ 74 ]. The mapping from
QCA cells that represent an electric quadrupole to those that represent a magnetic
quadrupole is straightforward. Nanomagnetic implementations are discussed else-
where in this volume.
Some have proposed cell-cell coupling based on an electron exchange interaction
[ 75 ], and indeed the earliest calculations showed a small splitting between the singlet
and triplet spin states [ 76 ]. This approach has two serious drawbacks: the exchange
splitting is quite small, and it is zero if there is not tunneling from cell to cell. If there is
tunneling from cell to cell, the information is no longer localized and spin-wave
solutions predominate.
It is interesting to consider the fundamental question of what sort of systems could
implement QCA action. There are two basic features of QCA that must be satisfied:

1. A bit is to be represented completely by the local state of a cell composed of

atoms.

2. The interaction between cells is through a field, rather than by transport.

The cell's binary information must therefore be represented by the positional or spin
degrees of freedom of the electrons and nuclei in the cell. Nuclear positions could be
used to encode the information—for molecules this would entail a conformational
change, for larger cells we would call it mechanical. Lighter mass electrons have an
advantage in that they can switch positions faster than nuclei. Spin states of either
nuclei or electrons could switch quickly.
The field connecting cells must be electromagnetic because the other candidate
fields are either too short range (the nuclear strong or weak forces) or too weak (the
gravitation force). Direct spin-spin interaction energies are very small, so for magnetic
coupling we need many nuclear or electronic spins acting collectively. Nanomagetic
QCA thus must be sufficiently large that the coupling is adequate at room tempera-
tures. The direct Coulomb interaction is quite strong and allows molecule-to-molecule
coupling between multipole moments of the charge distribution. Since by assumption
there is no transport from cell to cell, the charge of the cell cannot change and higher
moments must be used to encode the information. QCA thus far has used dipole
coupling and quadrupole coupling—the difference being what one chooses to define as
a cell. It is also possible to use contact potentials along the cell surface to coupled
mechanical (nuclear) degrees of freedom. The possibilities then can be seen to reduce
to these categories: