US Opens First Major Silicon Carbide Chip Plant (nikkei.com) 29
On Monday, the world's largest plant for making silicon carbide chips was opened in central New York. Nikkei Asia reports: The $1 billion, 63,000 sq. meter fabrication plant, or "fab," will be the first of its kind to make 200mm silicon carbide wafers, according to [North Carolina-based Wolfspeed]. Silicon carbide, or SiC, is an alternative to traditional silicon that is gaining popularity due to its energy efficiency when transferring power, something especially useful in electric vehicle manufacturing. Ahead of the fab opening, Wolfspeed announced a partnership with luxury EV maker Lucid Motors. It also has agreements with General Motors and China's Yutong Group to supply silicon carbide chips for their electric vehicles.
The U.S. share of modern semiconductor manufacturing capacity has declined to 12% from 37% in 1990, according to the Semiconductor Industry Association, a group that has lobbied for the CHIPS Act. This trend, the group says, is largely due to a lack of government investment compared to other nations. [...] In New York, there are hopes that other plants will cluster around the Wolfspeed fab. Wolfspeed received $500 million in construction subsidies from New York as the state seeks to expand its semiconductor manufacturing industry, which has generated nearly $6 billion a year in economic impact and over 34,000 jobs, according to [New York Gov. Kathy Hochul].
The grand opening of a modern manufacturing facility has a special resonance in central New York. The fab stands in the Mohawk Valley, an area along the Erie Canal that was once filled with traditional industry but long ago slid into an economic decline that has defined the region in recent times. So far the Wolfspeed facility has created 265 jobs, with a goal of over 600 new jobs by 2029, according to the company. The site sits directly across from a campus of SUNY Polytechnic Institute, New York's public polytechnic college, and Wolfspeed has given the school $250,000 to help train potential future employees.
The U.S. share of modern semiconductor manufacturing capacity has declined to 12% from 37% in 1990, according to the Semiconductor Industry Association, a group that has lobbied for the CHIPS Act. This trend, the group says, is largely due to a lack of government investment compared to other nations. [...] In New York, there are hopes that other plants will cluster around the Wolfspeed fab. Wolfspeed received $500 million in construction subsidies from New York as the state seeks to expand its semiconductor manufacturing industry, which has generated nearly $6 billion a year in economic impact and over 34,000 jobs, according to [New York Gov. Kathy Hochul].
The grand opening of a modern manufacturing facility has a special resonance in central New York. The fab stands in the Mohawk Valley, an area along the Erie Canal that was once filled with traditional industry but long ago slid into an economic decline that has defined the region in recent times. So far the Wolfspeed facility has created 265 jobs, with a goal of over 600 new jobs by 2029, according to the company. The site sits directly across from a campus of SUNY Polytechnic Institute, New York's public polytechnic college, and Wolfspeed has given the school $250,000 to help train potential future employees.
What is this ambiguity? (Score:3)
Title is:
US Opens First Major Silicon Carbide Chip Plant
TFS Says:
[The] fabrication plant, or "fab," will be the first of its kind to make 200mm silicon carbide wafers [...]
Is this a plant where SiC ingots are grown and cut into SiC waffers to be processed elsewhere?
Is this a plant where SiC waffers from elsewhere are processed (etched) to make actual SiC chips?
Both?
Re:What is this ambiguity? (Score:4, Insightful)
TFA is poorly written and somewhat ambiguous, but it appears to be "both".
The SiC ingots will be grown, sliced into wafers, and fabbed into chips, all in the same facility.
Re: (Score:2)
TFA is poorly written and somewhat ambiguous
You can say that again:
as the state seeks to expand its semiconductor manufacturing industry, which has generated nearly $6 billion a year in economic impact and over 34,000 jobs, according to [New York Gov. Kathy Hochul]. [...] So far the Wolfspeed facility has created 265 jobs, with a goal of over 600 new jobs by 2029, according to the company.
What happened to the other 33,400 jobs?
Nobody leaves zer room!
Re: (Score:3)
as the state seeks to expand its semiconductor manufacturing industry, which has generated nearly $6 billion a year in economic impact and over 34,000 jobs, according to [New York Gov. Kathy Hochul]. [...] So far the Wolfspeed facility has created 265 jobs, with a goal of over 600 new jobs by 2029, according to the company.
What happened to the other 33,400 jobs?
Nobody leaves zer room!
The 34k jobs is for the "semiconductor manufacturing industry"
Which includes many employers other than Wolfspeed.
ex: Global Foundries, STMicroelectronics, Applied Materials, ON Semiconductor, IBM.
Re: (Score:2)
I'm also not a fan of articles that describe the new value, then give the old value at some time in the past. It seems to focus on the wrong value.
The U.S. share of modern semiconductor manufacturing capacity has declined to 12% from 37% in 1990.
The U.S. share of modern semiconductor manufacturing capacity has declined from 37% in 1990 to 12%.
.
For me, the first approach requires a second or third re-read to get the timeline and percentage difference aligned with the way I think about changes over time.
Hmm... (Score:2)
Re: (Score:3, Insightful)
It is a different technology ...
Re:Hmm... (Score:5, Informative)
SiC devices are either large diodes or mosfets, stuff rated for high voltage and high current. They don’t need small lithography.
Re: (Score:3)
This wafer will have multiple chips come from the 200 mm wafer not 22 mm.
As to the part size on these chips, it will depend on where the manufacturing tech comes from. If from S. Korea, then it will be 5 nm. If from IBM, it will be 14 nm.
Re: (Score:3)
I don't think it will be either. As far as I'm aware no one in Asia is producing 5nm feature sizes on silicon carbide.
It's currently popular for very high power devices at high voltages. Currently more or less every device out there is discrete, though I've seen the odd cascoded FET. There's not even an IGBT yet. Tiny features aren't especially useful if you're switching 1200V.
I'd love to see some of the early demos. Maybe a nerdy engined fabbing an original pattern 555 for a low integration test that can r
Re: (Score:2)
So, the Asians are cranking out 5 nm chips and this plant is going to make 22mm wafers? Seems kinds of ass backwards doesn't it?
The 5nm (transistor size) chips you talk about are etched on 300mm waffers. These are 200mm SiC waffers, and who knows how many nm the transistor size will be (TFA does not say)...
No need for government investment (Score:3)
That's communism brotha! Capitalism will work.
Now, about those tax cuts for the rich. . .
Uh no. (Score:2)
The U.S. share of modern semiconductor manufacturing capacity has declined to 12% from 37% in 1990, according to the Semiconductor Industry Association, a group that has lobbied for the CHIPS Act. This trend, the group says, is largely due to a lack of government investment compared to other nations.
No, the real problem is lack of end manufacturing.
We need to make the final parts here.
Re: Uh no. (Score:2)
Re: (Score:1)
Why does the solution to every problem require government investment? Here's an idea, cut out the investment and just have the government take less money in taxes in the first place.
Building factories requires capital. It's in the definition of "capital expense". To expend capital to build anything someone has to own capital to expend it. Capitalism isn't the heartless theory of the poor getting poorer and the wealthy gaining more wealth. Capitalism means people can own stuff, that's it. In a capitali
Re: (Score:2)
Yes, it's a classic libertarian stance, we don't need government, private enterprise will solve everything. Often it doesn't work in practice though. In trying to create wealth for themselves people will take the path of least resistance. It makes sense, why take more risks and make less profit when you can have more? But that will mean that people will create wealth for themselves not in the way that fulfills the greatest need for something but in a way that benefits them the most.
To keep it somewhat relev
Re: (Score:2)
Yes, it's a classic libertarian stance, we don't need government, private enterprise will solve everything
It's also a stance that quickly becomes indistinguishable from fascism.
Re: (Score:1)
you might want to look up fascism
Re: (Score:2)
As was clearly stated in the article;
"The U.S. share of modern semiconductor manufacturing capacity has declined to 12% from 37% in 1990, according to the Semiconductor Industry Association, a group that has lobbied for the CHIPS Act. This trend, the group says, is largely due to a lack of government investment compared to other nations."
Just Great (Score:1)
With this new technology, it can't be long before the big tool companies start putting DRM in my abrasive cutting wheels.
Thermal runaway. (Score:3, Interesting)
I'm sure someone will step in with a correction if this is wrong. I'm going from memory on semiconductor physics from a long time ago.
I recall a big problem with silicon semiconductors was thermal runaway, a process in which conductivity increases with temperature, which then increases current, which increases temperature, until the semiconductor burns up. Gallium arsenide and silicon carbide don't have this thermal runaway problem and so are better suited to high power electronics.
Another feature of silicon carbide is that it melts at higher temperatures, which means it can be put in places like electric motors and gasoline engines when silicon electronics won't hold up. But, and again I may be wrong here, silicon carbide can't hold the small features like silicon can. This means no high speed electronics that can be had with silicon and gallium arsenide. Silicon carbide is great for embedded electronics, but horrible for general purpose computing. Maybe someone will work on the silicon carbide process to make it a practical replacement for silicon CPUs and GPUs.
I don't recall anything about how silicon carbide holds up to radiation, as in how suited this technology is to aerospace applications where radiation would kill most off the shelf silicon electronics. I see this important if the USA is going to keep ahead of China and Russia in the new space race we are currently in right now. Building the factories for aerospace rated electronics will be necessary for any nation that wants to keep the "high ground" in warfare. Not just in a shooting war but a war of ideas, economies, or scientific discovery.
What is not likely to be a matter of national security is maintaining an edge on semiconductor technology for photovoltaic power. Solar PV for power to the grid is not viable. Studies showing the land use, material use, labor, fresh water use, and perhaps more proves solar PV a dead end for power to the grid. When going to space we will likely see more investment in radio-isotope thermo-electric generation (RTG) and nuclear fission. The edge of utility for solar PV is about the orbit of Mars, beyond that point solar power takes more mass than RTG and should RTG materials become more abundant, and/or the technology advance more, then solar PV may prove nonviable for all space exploration. More junk in orbit around Earth may even make solar PV nonviable for communication satellites, the PV panels being weak points for damage from collisions.
We need to see more semiconductor production in the USA. If the solar PV fanatics are right about solar PV solving our energy needs in the future then we need to see solar PV production facilities pop up like dandelions, and that's not happening. There's a limit on how quickly we can mine high grade silicon for semiconductors, and when there's a demand for embedded electronics there's no way solar PV can compete on price for raw materials. Recycling old solar PV is proving difficult as once the silicon is doped for PN junctions it is really hard to undo that for high purity silicon. The most likely place for old PV cells to end up is as metallurgical grade silicon, silicon used to make steel alloys and aluminum alloys.
We need more electronics fabrication inside the USA. Relying on foreign sources of electronics is not good for national security. We need to be able to maintain a secure supply of electronics if the shooting war in Europe gets any bigger. Should demand of electronics get much tighter then I expect one of the first things to feel that crunch is solar PV. If that happens then we need something else to fill in on our energy supply. I'm quite confident on that that will be.
Re: (Score:3)
I don't recall anything about how silicon carbide holds up to radiation, as in how suited this technology is to aerospace applications where radiation would kill most off the shelf silicon electronics.
There already are rad-hard processes such as SoS. They're already being used in places where "off the shelf silicon electronics" was deemed inadequate. Why would a high-power, low-density process like SiC be of any help here? Unless you're landing on Venus, you probably don't need SiC.
The edge of utility for solar PV is about the orbit of Mars, beyond that point solar power takes more mass than RTG a
Setting aside all the other of your parroted rubbish about PV being "a dead end", Juno was generating 486 W when it arrived to the orbit of Jupiter, using a solar array weighing 340 kg. That's around 1.4 W/kg. The MMRTG, curre
Re: (Score:3)
Re: (Score:1)
One California has sufficient nuclear power generating capacity and battery backup to manage a cloudy day where the installed solar power capacity produce next to nothing in energy to the grid then where is the added value in having any solar power generating capacity?
Solar power produces more CO2 per energy produced than nuclear power.
Solar power takes more materials, land, labor, and fresh water than nuclear power.
The only situation where solar power costs less than nuclear power is with grid scale solar
In central NYC? (Score:1)
Was there no cheaper real estate for a 63,000 m^2 facility? Not anywhere?
Re: (Score:2)
Re: (Score:2)
Would have sworn TFS said NYC when I wrote this. Never mind.
Neat. (Score:1)