简体中文
हिन्दी
Português
Español
What it takes to bring semiconductor manufacturing back to the US
What it takes to bring semiconductor manufacturing back to the US
Rather than playing catch-up in manufacturing today’s most advanced chips, U-M researchers advocate preparation to build those of tomorrow
Story by Katherine McAlpine | Multimedia by Hans Anderson and Jeremy Marble
The Midwest is well positioned to help revive U.S. semiconductor manufacturing.
With chip development talent from the University of Michigan and major automakers down the road in Detroit, U-M plays a key role in designing the advanced chips needed for increasingly autonomous vehicles and many other applications. Meanwhile, Hemlock Semiconductor is building a new manufacturing facility in Hemlock, MI while Indiana and Ohio are building out chip factories, known as fabs.
Semiconductors have become as essential to modern society as electricity. Without them, we’d be unable to do many tasks we take for granted—operate most modern appliances and vehicles, make purchases without cash and do almost anything with a smartphone or computer.
The COVID-19 pandemic disrupted semiconductor supply chains in 2020, leaving parking lots full of new cars that couldn’t be driven dotting landscapes across the country. The average price of a new auto jumped by nearly $10,000 between late 2020 to the end of 2022—a rise that had previously taken eight years. Even in early 2020, electronics accounted for 40% of a new car’s cost.
Today, semiconductor chip manufacturing has one big choke point: the Taiwan Semiconductor Manufacturing Company is reported to be responsible for 68% of global chip manufacturing, including 90% of the most advanced chips. This dependence on a single company is a global risk, but how to distribute that risk by developing more distributed semiconductor manufacturing is not straightforward, researchers say.
“TSMC is a great company, but it’s a single point of failure. So I think there is no benefit for anybody to have a single company that can produce the most advanced technology node,” said Valeria Bertacco, the Mary Lou Dorf Collegiate Professor of Computer Science and Engineering and vice provost for engaged learning at U-M.
Changing that isn’t so simple though, as the U.S. tries to compete with TSMC’s budget, which includes about $40 billion per year on equipment and research. The U.S. CHIPS Act, signed into law in 2022, provides $52 billion over five years.
However, the rapid advancement of semiconductor technologies means that the equipment is constantly turning over, creating opportunities for the U.S.—which is still heavily engaged in research that drives the next generations of chips. The U.S. has the opportunity to jump back into manufacturing, according to Bertacco. Funding from the CHIPS Act benefits Michigan through Hemlock Semiconductor and helps seed the industry in Ohio and Indiana, with grants to Intel and SK Hynix, respectively. Other large awards went to TSMC, Samsung and Micron.
Historically, the big driver of semiconductor advancement has been miniaturization. “Moore’s Law,” the observation that transistor counts in integrated circuits double every two years, roughly held from the 1970s until the 2010s. However, as engineers went deeper into the nanoscale, shrinking transistors became much harder.
“Moore’s Law really drove a huge explosion in the scale of chip manufacturing and the utility of those chips across all sorts of industries, including the phones that we use every day,” said Gus Evrard, the Arthur W. and Alice R. Burks Collegiate Professor of Physics and professor of astronomy at U-M.
Bertacco pointed out that while 2-nanometer transistors are possible, Apple has decided they are cost-prohibitive, at least for now. The M4 chip uses 3 nanometer transistors. Now, researchers are exploring other approaches: different materials and different computer architectures that could lead to different manufacturing.
“The government has a strong interest in developing leading edge technologies. These technologies don’t all use silicon. They use other semiconductors like silicon carbide, gallium nitride, diamond,” said Becky Peterson, director of U-M’s Lurie Nanofabrication Facility and associate professor of electrical and computer engineering.
“One example of this is high-temperature electronics, where we need electronics that can go inside an aircraft engine or a hypersonic missile, or at the bottom of an oil, gas or geothermal well. Here at the University of Michigan, we’re seeking to advance these novel technologies and to explore the edges of what they can do to develop new capabilities for the U.S. and for the world.”
While the U.S. doesn’t have the manufacturing capability it needs to produce many advanced chips, it does have the research capacity for discovering the most effective next-generation technologies. One of the key facilities is the Lurie Nanofabrication Facility, where U-M researchers and tech companies in the Midwest build chips with experimental materials and architectures.
U-M Lurie Nanofabrication Facility (LNF)
With those designs in hand, the U.S. could get into the chip manufacturing game by building fabs for the next generation of chip technologies, the researchers say.
“The CHIPS Act started in 2022. There is still funding being distributed, but I really hope that there will be another investment down the road. Maybe exactly the same, five years from now, when we start to really have some clarity on which are the subsegments where the United States really has a key advantage and we would like to invest,” Bertacco said. “But if it’s a one-time thing, it will be difficult to build enough momentum that is sustained.”
Related stories
Did you know?
The making of integrated circuits was initially pitched as women’s work. As a result, the highly skilled labor of those who built early advanced chips, including the computers that powered the Apollo missions, has only been recently recognized for extraordinary precision and engineering excellence.
