The camera in your phone collects light on silicon and
translates that information into digital bits. One of the reasons those cameras
and phones continue to improve is that researchers are developing new materials
that absorb more light, use less power, and are less expensive to produce.
Now, University of Wisconsin-Madison materials science and
engineering researchers have introduced innovations that could make possible a
wide range of new crystalline materials. Writing in ACS Nano, Deborah Paskiewicz and Boy Tanto, along with Donald
Savage and Erwin W. Mueller, describe a new approach for using thin sheets of
semiconductor known as nanomembranes.
Controlled stretching of these membranes via epitaxy allows
the team to fabricate fully elastically relaxed silicon germanium nanomembranes
for use as growth substrates for new materials. The team grew defect-free
silicon germanium layers with any desired germanium concentration on silicon
substrates and then released the silicon germanium layers from the rigid
silicon, allowing them to relax completely as free-standing nanomaterials. The
silicon germanium film is then transferred to a new host and bonded there. From
this stage, a defect-free bulk silicon germanium crystal can be grown
(something not possible with current technology), or the silicon germanium
membrane can be used as a unique substrate to grow other materials.
Epitaxy, growth that controls the arrangement of atoms in
thin layers on a substrate, is the fundamental technology underlying the
semiconductor industry’s use of these new materials. By combining elements, researchers
can grow materials with unique properties that make possible new kinds of
sensors or high-speed, low-power, efficient advanced electronics. It is the
ability to grow them without detrimental defects that makes these alloys useful
to the semiconductor industry. However, making high-quality crystals that
combine two or more elements faces significant limitations that have vexed
researchers for decades.
“Many materials consisting of more than one element simply
cannot be used. The distances between atoms are not the same,” says Lagally. “When one begins to grow such a layer, the atoms start to interfere with each
other and very soon the material no longer can grow as just one crystal because
it starts to have defects in it. Eventually, it breaks up into small crystals
and becomes polycrystalline, or even cracks.”
In addition to its use in the semiconductor industry,
silicon germanium is important to the nascent field of quantum computing. A
quantum computer makes direct use of quantum mechanical phenomena such as
superposition and entanglement to perform calculations. Current computers are limited
to two states; on and off, or zero and one. With superposition, quantum
computers encode information as quantum bits. These bits represent the varying
states and inner workings of atoms and electrons. By manipulating these
multiple states simultaneously, a large-scale quantum computer, if it can be
built, could be millions of times more powerful than today’s most powerful
classical supercomputer.
UW-Madison Physics Professor Mark Eriksson uses silicon
germanium to make two-dimensional electron gases. “A ‘two-dimensional
electron gas’ is a layer of a semiconductor in which charges are able to move
freely over large distances, in analogy with atoms in a real gas, except
confined to a thin layer and hence two-dimensional. For quantum computing, this
2D electron gas is formed in a strained-silicon layer grown on a silicon
germanium substrate. Electrodes put on top of a structure containing the 2D
electron gas in the strained-silicon layer allow one to move and control single
electrons, turning regions of the quantum well into ‘electron buckets,’ if you
will, that are defined by the electric fields from the top electrodes,”
says Lagally.
A major obstacle to developing a quantum computer is
creating multiple quantum buckets as similar as possible. To make rapid
progress, researchers need low-defect and consistent materials.
“With the silicon germanium substrates we have been using,
the electrostatic fields can be quite uncertain because of the defects in the
substrate,” says Lagally. “We believe our new process can fix that. Because the
substrate material is uniform, without defects, it should bring more
predictability and control to Mark’s efforts.”
Beyond silicon germanium, Lagally says the process should
work for a wide range of exotic materials that cannot be grown in bulk but have
interesting properties. Materials Science and Engineering Associate Professor
Paul Evans develops new ways to probe and apply these materials.
“The thin defect-free substrates that can be produced by
transferring and relaxing these layers present exciting opportunities in the
growth of materials beyond silicon and other traditional semiconductors,” Evans
says. “With this approach, it will be possible to produce defect-free
substrates of materials for which no high-crystalline quality bulk materials
exist. In complex oxides, this can lead to thin substrates that stabilize
specific ferroelectric or dielectric phases. That could lead to better
oscillators, sensors, and optical devices, which are important to the cell
phones, cameras, and computers we use everyday.”
