NANOSCALE DEVICES: APPLICATIONS AND MODELING - Bio-Inspired and Nanoscale Integrated Computing - page 46

wires can have very high mean free paths (the average distance they travel before
being scattered), and even ballistic transport (collision free transport) can happen
over lengths of up to several micrometers. This could lead to very high operation
speed. Transistor action in these devices can happen as a result of both direct
modulation of the channel due to the gate or the modulation of the tunnel barriers
at the source and drain contacts. These nano transistors are currently the subject
of active research, both experimentally and theoretically. For instance, silicon
nanowire transistors have been reported to have higher on-off ratios and smaller
subthreshold slope (approaching the theoretical limit) than planar silicon-on-
insulator FETs [21]. FETs based on other types of nanowires, including germa-
nium, zinc oxide, and gallium nitride, have also been investigated [3, 22, 23].
Similarly, high performance FETs based on carbon nanotubes, including ballistic
transport devices, have been demonstrated [24-26]. Applications for logic circuits
and high frequency operation (GHz range) have also been investigated [27, 28].
For a theoretical study and modeling of nanowire and nanotube transistors, the
reader is invited to consult [29-32].
2.4.2.2. Exploiting the Hidden Offerings of Nanomaterials. If the length
of the channel in a nanowire or nanotube transistor is made very short, the device
will naturally become a quantum dot or single electron transistor. Obviously this
can be achieved if high resolution patterning is possible, and it is one way of
making nanoelectronic devices with new materials [33, 34]. Note that in this
approach we are using only the fact that these materials are one-dimensional or
quantum wires, i.e., confinement in two directions is already in place. However,
the size of the quantum dot achieved in this way is still determined by our ability
to pattern electrodes on the nanotube with very high spatial resolution. Therefore,
this kind of quantum dot typically has a length on the order of 100 nm and, as
discussed before, room temperature operation is not easy. Here we will discuss
briefly how some of the less obvious properties of these nanomaterials can be
exploited to make new types of devices that are less dependent on traditional
patterning techniques.
Consider a single-walled carbon nanotube. It is a quantum wire and can be
made into a quantum dot by creating only one additional degree of confinement in
it. Using first-principles calculations, the authors in [35] predict that if the cross
section of a nanotube is deformed from circular to elliptical, its electronic structure
(both valence and conduction bands) can be altered significantly, changing the
nanotube from semiconducting to metallic and vice versa. This property was used in
[36] and two kinks were created in a nanotube using AFM manipulation (Fig. 2.11)
to induce potential barriers and isolate a dot as small as 20 nm, which exhibited
room temperature operation. The device was also studied theoretically in [37].
Another device suggested in [38] is based on the geometrical idea that a dot
can be defined by crossing two lines. Here the induced mechanical deformation
at the intersection of two nanotubes crossing each other (Fig. 2.12) creates a
local potential well with a width of about 1.5 nm, potentially enabling room
temperature operation.