Excerpted from the latest "Electronic News/VLSI Report" published by SEMI Newsletters:
2. CORNELL SCIENTISTS CLAIM ALTERNATIVE TO STOICHIOMETRY IN
ADDRESSING LATTICE MISMATCH WHEN GROWING GALLIUM ARSENIDE
Cornell University scientists led by Yu-Hwa Lo, a professor of
electrical engineering at Cornell, claim that they have come up
with an alternative to stoichiometry to address the dilemma of
lattice mismatch when growing some of the more obscure
semiconductor compounds such as gallium arsenide, gallium
phosphide, indium phosphide and silicon carbide.
Dan Rose, president of materials research firm Rose
Associates, called the Cornell research"underwhelming" and said
he could not find cause to assume that "there is any great
breakthrough in semiconductor technology, as some press repor
ts seem to suggest."
In the scientific journal, Applied Physics Letters, an article
by Dr. Lo; Felix E. Ejeckam, a doctoral student; and Shanthi
Subramanian, a former student now with the Exxon Research and
Engineering Co., describes the new technique. An intermediate
thin layer is bound to a substrate of growing semiconductor
crystals at a slight angle. Thus the crystal orientation absorbs
the stress caused when lattices do not match. Dr. Lo is reported
to have stated, "You put the new semiconductor on top of the thin
layer, and as stress energy builds up, it goes down into the thin
layer below and does not stay in the film on top to form defects.
The bottom layer sacrifices itself for the top layer." Hong Q.
Hou, senior member of the technical staff at Sandia National
Laboratories, who collaborated with Dr. Lo in the research,
inferred that the new technique will not lead to a replacement
for relatively inexpensive silicon. "It will still be very much
more economical to use a standard substrate, but this technique
seems promising for specific applications that need a unique
lattice, such as might be necessary in infrared wafer detectors
and lasers for optical fiber communications." Mr. Hou claims the
process allows a higher degree of flexibility than might be found
in stoichiometry because an engineer can "create an over layer of
a lattice constant anywhere."
Suggestions were made in press reports last week that the new
technique could lead to the development of a universal substrate,
on which pure, single crystals of exotic semiconductor compounds
could be grown. The skeptical Mr. Rose stated simply, "The
findings don't lead to a universal substrate," and Mr. Hau said
the technique is "universal" only in terms of its flexibility --
that a constant of any degree might be created.
The technique is only worthwhile to semiconductor
manufacturers if it is economical. Currently, it's expected that
the new technique will approach the high cost of stoichiometry,
which utilizes molecular-beam epitaxy and a process of at
omic ratios to address the lattice mismatch issue. The new
technique, said Mr. Hau, utilizes a wafer-bonding technique and
furnaces. |
