SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Politics : Formerly About Applied Materials -- Ignore unavailable to you. Want to Upgrade?


To: michael97123 who wrote (46706)5/14/2001 9:48:20 AM
From: Fred Levine  Read Replies (2) | Respond to of 70976
 
From the NY Times:

Ore., May 11 — In a few
weeks, a 26,000-pound machine will
begin shooting ions into silicon wafers like
lightning bolts, one process in hundreds that
silicon must undergo during its transformation
into a computer chip.

The immense machine, called an implanter,
will be a small piece of the Intel Corporation's newest semiconductor factory here,
scheduled to begin mass production early next year. But today, a dozen technicians,
shrouded in white protective suits from head to foot, swarmed around and inside the
machine, painstakingly piecing it together.

"This is how you spend $7.5 billion in capital," Gerald Marcyk, Intel's director of
component research, said after a rare tour of the new factory, 14 miles west of
Portland. That is how much Intel estimates it will spend this year on capital
expenses, in addition to $4.2 billion in research and development.

Similar scenes of ambitious construction will be played out in semiconductor plants
throughout the world in the new few years, as chip makers prepare for the first
radical change in production techniques since the mid-1990's.

The change in processing technology involves increasing the size of the round silicon
wafers from which chips are culled. Instead of the current standard diameter of 200
millimeters (about 8 inches, the size of a salad plate), chip makers are moving
toward wafers of 300 millimeters (roughly 12 inches). The larger wafers can yield
more than twice as many chips, achieving an economy of scale that Intel says will
save 30 percent in manufacturing costs for each wafer.

Engineers have worked to enlarge wafers since the 1960's, when chips were cut
from pieces of silicon the size of quarters. As the wafer size increases, so does the
difficulty of getting perfect chips from that wafer every time, and so does the ability
of equipment makers to design tools to handle the wafers. As a result, tool makers
can charge tremendous amounts for customized machines.

With the entry of Intel, the world's largest chip maker, into the 300-millimeter
market, the changeover to the new process is sure to accelerate. I.B.M. led the
industry by developing the 200-millimeter wafers in 1994, much as Intel led the
150-millimeter transition a decade before. Each company learned its lesson,
choosing to let others take the development lead that time and remain close behind.
I.B.M. is building a $2.5 billion 300-millimeter plant in East Fishkill, N.Y., that is
expected to be in full production by 2003, a year after Intel's.

"The company that drives the move to a new wafer size ends up paying the bill for
much of the tooling development; then the rest of the industry benefits from it," an
I.B.M. spokesman, William O'Leary, said.

This time, Intel and I.B.M. left the trailblazing role to Semiconductor300, a joint
venture between Motorola and Infineon Technologies,the semiconductor unit of
Siemens. In 1999, the venture. financed by the German government, became the
first to produce working memory chips from 300-millimeter wafers and sold them in
small numbers.

Dr. Ulrich Schumacher, Infineon's president and chief executive, said he welcomed
Intel's entrance into 300-millimeter wafer production because it could be expected
to force more equipment makers to build appropriate tools, driving prices down.
"This is a very important push for equipment suppliers," he said, "because now it is
not only just Infineon and Motorola, but also the mighty Intel."

Today, more than two dozen 300- millimeter plants are under construction around
the world. In the first three months of this year, chip makers undertook
300-millimeter development projects worth $35 billion in construction and
equipment, according to Semiconductor Equipment and Materials International, a
trade group for tool makers.

But only Semiconductor300, the joint venture, and TaiwanSemiconductor have put
300-millimeter-produced chips on the market. With demand slumping in many
segments of the industry, chip makers are gambling that sales will improve by the
time the plants are ready. A factory can cost billions of dollars to maintain, even
when running below full capacity.

Intel has warned that its famously high profit margins will suffer in coming months, as
demand for its semiconductors remains weak. But Intel executives remain bullish on
their long-term potential. The factory here produced a limited number of its first
working microprocessors on the new wafer size in March. The chips were the
world's first to feature the next generation of 0.13-micron circuit lines, which allows
Intel to cram more transistors onto each chip. Intel says the combination of the two
technologies will quadruple production efficiency.

Even while construction continues, the factory here is already preparing for the drive
to full-scale production. As technicians assembled the implanter in one bay, an
automated trolley system in the next bay whisked a wafer across the ceiling, then
lowered it to a tool where workers awaited its arrival. The tool deposited oxides
onto the 300-millimeter wafer, then sent it along to the next step.

"From a long-term perspective," said Mark Edelstone, managing director of
semiconductor research for Morgan Stanley Dean Witter, after touring the factory,
"there's no question now about Intel's core competency. It raises the bar significantly

IMO--all the fabs must follow or fall.

fred