REGISTER | LOGIN
Breaking News
News & Analysis

TSMC Plans New Fab for 3nm

12/12/2016 00:00 AM EST
10 comments
NO RATINGS
1 saves
More Related Links
View Comments: Newest First | Oldest First | Threaded View
resistion
User Rank
Author
Re: EUV very certain for 5nm
resistion   12/13/2016 1:26:24 PM
Oh but there are...

witeken
User Rank
Freelancer
Re: EUV very certain for 5nm
witeken   12/13/2016 12:00:16 PM
SemiEngineering has a few pieces, but I don't think there are any real showstoppers.

resistion
User Rank
Author
Re: EUV very certain for 5nm
resistion   12/13/2016 10:44:08 AM
ASML doesn't or won't show you everything. In fact, all the key problems are hidden behind subscription but of course not trotted out for the press and/or investors.

witeken
User Rank
Freelancer
Re: EUV very certain for 5nm
witeken   12/13/2016 10:06:01 AM
ASML sums it up at slide 83.

http://seekingalpha.com/article/4018888-asml-investor-day-slides

resistion
User Rank
Author
Re: EUV EOL 0.33NA
resistion   12/13/2016 7:05:19 AM
Actually EOL is foreseen for current 0.33NA EUV tools with new >0.5 NA coming out around 2020.

resistion
User Rank
Author
Re: EUV very certain for 5nm
resistion   12/13/2016 4:54:11 AM
Still plenty of EUV problems not under ASML control. Five areas: power/dose, resist, contamination, mask/pellicle/inspection, off-axis effects. Only the first and last are in ASML's jurisdiction. The other three areas recently realized more serious than thought.

witeken
User Rank
Freelancer
Re: EUV very certain for 5nm
witeken   12/12/2016 9:43:01 PM
NO RATINGS
"You mean EUV very certain for 7nm :). You know TSMC 5nm is only a 7nm.....and 7nm is only a little better than a "real" 10nm with old metrics."

Yes I know that. Or maybe I mean very certain for 29/30nm :).

https://twitter.com/witeken/status/806976146273353728

Gondalf
User Rank
Author
Re: EUV very certain for 5nm
Gondalf   12/12/2016 9:22:18 PM
You mean EUV very certain for 7nm :). You know TSMC 5nm is only a 7nm.....and 7nm is only a little better than a "real" 10nm with old metrics.

Anyway i admire Samsung and Intel, they both are very realistic about what will be the future silicon scaling. Good things are possibles but they will are very expensive.

I am curious about who will be the average TSMC customer at these crazy 5/7nm and 3/5nm nodes. Not to mention the EUV equipment cost will be stellar.

IMO EUV has lost the right window to enter in the market, SA multipatterning is a little less effective (in theory) but it requires far lower investiments 

jasiojasiojasio
User Rank
Author
Re: EUV very certain for 5nm
jasiojasiojasio   12/12/2016 7:25:42 PM
NO RATINGS
Observing how aggressive is Mainland China I would advice TSMC to develop that foundry in USA.

witeken
User Rank
Freelancer
EUV very certain for 5nm
witeken   12/12/2016 5:47:43 PM
NO RATINGS
As I noted in a few tweets yesterday, EUV is on good progress in the last 2 years and is on track to become a viable commercial lithography tool around 2018. (ASML has already demoed 210W, while 250W is necessary for HVM.) Given that TSMC's 5nm will go in risk production in 2019 (so the new fab won't be a 5nm launch fab), that should be no problem.

ASML for its part expects to ship 10 EUVs next years, 20 in 2018 and doubling again in 2019.

https://twitter.com/search?q=%23euvstatus2016

Like Us on Facebook
EE Times on Twitter
EE Times Twitter Feed