COMPUTING WITH QUANTUMDOT CELLULAR AUTOMATA - Bio-Inspired and Nanoscale Integrated Computing - page 126

1

0 0

1

Figure 4.15. Wires created with rotated and nonrotated cells do not interact. We
can exploit this to create coplanar crossovers, if the effect of one cell onto
another of the same orientation is able to traverse the gap.

et al. were able to show that the progress of these domains through the junction
could be interlaced, as shown in Figure 4.16. Although this building block resolves
the problems associated with the previous coplanar crossover and represents the
most promising approach to wire crossing, it comes at a significant cost associated
with the additional area overhead; the requirement for four additional clocking
phases; and the extra complexity associated with the clock timing.

4.9.3. Multilayer Crossover
Previous work has examined the possibility of building multilayer QCA [64]. The
use of multilayer QCA is quite attractive as a means of wire crossing but it also
introduces the possibility of a 3D computer since it is quite reasonable to assume
that the additional layers could perform logical functions. The application of
multilayer networks as a means of signal crossing was first explored in [62]. Many
of the circuits presented later in this chapter use this approach to achieve signal
crossing. Using these multilayer QCA cells, we can effectively cross signals over
on another layer; such a technique has been simulated to show significantly less
sensitivity to cell displacements [63]. Using this method, multilayer QCA circuits
can potentially consume much less area than planar circuits (Figure 4.17).
This building block has area and displacement tolerance advantages. How-
ever, the physical realization of this building block is technically very challenging
and it is difficult to predict whether such a multilayer QCA technology will be
feasible in the future.