Slashdot Mirror

According to ASML's reports last summer, the EUV source outputs 250 W. Are you talking about electrical power?

Right. I confess to a bit of conflation there, to bring home the full scary picture of an EUV stepper. Hardly your granddaddy's gestetner. Semiwiki says 300 watts of EUV is desirable for high volume. According to Wikipedia the input power rounds up to a megawatt.That doesn't directly heat the pellicle of course, but it does require a small creek's worth of cooling. The pellicle is relatively late in the optical path so only a fraction of the 300 watts goes through it (twice) but it's still a huge problem for a membrane just tens of nm thick.

Right, the pellicle protects the mask. Everybody with EUV plans (just TSMC, Samsung and Intel now, by my count) is going to start introducing EUV without pellicles, but only for larger features like contacts and vias that are widely separated so that dirt accumulating on the mask is unlikely to create chip defects. For a full EUV process as is absolutely required to go beyond 7nm, not using pellicles will prohibitively shorten the mask life. So they are absolutely required. The difficulty with EUV is, all matter becomes opaque, except for extremely thin membranes. Something like 16nm I think. You can see how that might be a bit tricky to fabricate and work with.

This article gives good analysis of the latest nodes:
https://www.semiwiki.com/forum...

Samsung's 10nm is about 10% denser than Intel's 14nm

Only none of what you said is true.

They haven't blown their advantage, though it's certainly shrunk, and they will continue to hold it through the "10nm" node where Intel's process is actually below the standard 10nm and TSMC et al's most certainly is not.

The table here: https://www.semiwiki.com/forum...
give a breakdown of the different foundries nodes.
As you can see, TSMC's 10nm is about 15% denser than Intel's 14nm, however density isn't the only factor. Performance-wise I would say Intel's 14nm is going to be better for a server chip, because it's specifically tuned for high performance computing, while TSMC's nodes are tuned for low power mobile SoC's

I doubt that.
TSMC 20nm will be ready for GPUs a lot sooner than their 16nm process. The only reason there are no 20nm GPUs yet is because the initial ramp was fully booked out by Apple.
Meanwhile, a comparison of Apple's 20nm A8 density versus 14nm Core M, indicates Intel's 14nm may not have such a density advantage as they claim: https://www.semiwiki.com/forum...

An interesting article here discribes the horrendiously difficult challenges that face EUV:
https://www.semiwiki.com/forum...

Patents are a part of it, but they're minuscule compared to the capital requirements. Semiconductor manufacture isn't a basement hobbyist game; it's the absolute cutting edge of technology, and the people who make the machines that make the chips are creating custom, precision hardware for a very small customer base. Commercial-scale semiconductor plants run about $1 billion minimum, for a 10k-30k wafers per month "minifab" and can run up to $8-10 billion for a "gigafab" churning out 80k-100k wafers per month.

Read more at SemiWiki.