The Development of Quantum-Dot Cellular Automata - Field-Coupled Nanocomputing: Paradigms, Progress, and Perspectives - page 13

(molecular cells are *1 nm square), and to exploit the clocking paradigm wherein all
communication is via shift registers. Kogge and Niemier have been early pioneers of
this [ 83 - 88 ], suggesting a universal clocking floorplan which supports general data-
flow. Tougaw et al. [ 89 ] and Niemier et al. [ 90 ] have explored programmable QCA
logic arrays, a promising area. Exploring QCA architectures and circuit ideas is a very
active area, as witnessed by other contributions to this volume.

4.6 Lithograph and Self-assembly

QCA circuits need to have a designed layout that reflects the circuit function; entirely
regular arrays do not have interesting behavior. As a consequence the information
contained in the circuit layout must be imposed and this is usually done, as with all
extant semiconductor circuits, through lithography. It is possible, however, to have
self-assembly take care of constructing the dots and forming the dots into cells, and
perhaps even forming the cells into lines or other functional groups. For molecular
QCA, this is particularly promising because bottom-up self-assembly of molecules
into supra-molecular structures is a common, albeit demanding, strategy. Building in
some level of self-assembly from below, where the cell size is about 1 nm, and
imposing circuit structure from above using lithographic techniques, which can reach
below 10 nm, is an appealing match of technologies.
Another approach that has received some attention is to use DNA or PNA [ 91 ]
self-assembled structures as ''molecular circuit-boards.'' DNA structures with sur-
prising amount of inhomogeneous patterning have been synthesized using Seeman
tiles [ 92 ], or the more recent DNA origami techniques [ 93 ]. The long-range concept
would be to engineering attachment sites in the DNA scaffold which would covalently
bond appropriate QCA molecules or supramolecular assemblies [ 94 - 97 ]. The geo-
metric information that defines the circuit layout would in this way be expressed
through the sequencing of base-pairs that self-assemble into the scaffold. Many issues
remain, of course, including the requirements of geometric matching to the DNA
repeat distance and the polyanionic nature of DNA, which could interfere with QCA
operation. PNA scaffolds are neutral and could potentially solve this problem.

4.7 Energy Dissipation

QCA has two fundamental motivators: ultra-small devices in large functional-density
arrays, and low power dissipation. Power dissipation has been a major driver in every
stage of the evolution of microelectronics. Adiabatic switching between instantaneous
ground states allows the absolute minimum dissipation of energy to the environment.
As Landauer [ 22 ] and Bennett [ 98 ] showed, there is no fundamental lower limit to the
amount of energy that needs to be dissipated as heat in order to compute a bit of
information. If information is erased, however, a minimum about of energy equal to
k B T log(2) must be dissipated. The combination of these two ideas is known as
Landauer's Principle (LP) and is connected to the Maxwell Demon [ 99 , 100 ]. Though
there is a substantial consensus on the correctness of LP, it has come under criticism