Page 42 - Principles and Applications of NanoMEMS Physics

28 Chapter 1

The system consists of a steel chamber which is equipped with pumps, to
− 11
create a very low pressure environment, typically about 10 Torr, and a
growth chamber containing several vacuum furnaces, called effusion cells or
K-cells, from where a variety of atomic or molecular materials evaporate.
The target wafer, on which growth is to occur, is placed inside the chamber
where it is held at a high, controlled temperature and under high vacuum,
and rotated to achieve uniformity over the wafer.
Growth occurs when heating of the K-cells causes the various materials
in them to evaporate, thus forming atomic beams that land on the wafer
surface. The properties of the growing layers are controlled by a number of
parameters, particularly, K-cell temperature, which controls beam intensity
or atomic/molecular flux, and substrate temperature, which controls the
dynamics of the atoms once these reach the wafer surface, see Figure 1-
26(b). In particular, the arriving atoms evolve according to the following
competing mechanisms: 1) Immediate absorption to the surface, i.e., they
“stick” wherever they land; 2) Migration across the surface, i.e., move
around before coming to a resting place which may not preserve the
crystalline structure; 3) Incorporation into the crystal lattice; and 4) Thermal
desorption, i.e., they reevaporate from the surface. To achieve good
crystal quality, such a set of flux and substrate temperature parameters must
be discovered that the arriving atoms have sufficient energy to move to the
appropriate position on the surface, without re-evaporating, and be
incorporated on the crystal.
The MBE technique is very versatile in that it allows the composition of
the layers to be fine tuned. This is accomplished by equipping the K-cells
r
with shutters which, th ough computer control, can turn on or off each beam
according to precise timing sequence. The fact that growth is controlled by
computer, endows MBE with the ability to deliver even atom-thick layers, of
abrupt composition, in a reproducible and reliable fashion. This, in turn,
enables bandgap engineering, the use of the material band gap as a degree of
freedom to engineer device properties. In the InP HBT, an emitter with a
band gap greater than that of the base, permits high base doping, without
compromising current gain, by virtue of the reduction of hole current
injection into the emitter effected by the latter’s energy barrier. In the RTD,
a lower band gap region, a potential well, sandwiched between two large
band gap regions, barriers, allows preferential current conduction only when
the energy of conduction electrons coincides with the discrete energy state in
the potential well, thus giving rise to the creation of a current-voltage
characteristic exhibiting negative differential resistance. The fact that the
path length of electron transport through the device is very short, leads to
these devices exhibiting very high speeds of operation, e.g., hundreds of
GHz in the case of the HBT, and close to a THz in the case of the RTD.