Nikon said to be delaying EUV lithography development

[Idontcare]

Nikon said to be delaying EUV lithography development

Lithography specialist Nikon is understood to have put on hold development of some of its EUV tools, which could give ASML a huge early advantage in commercial deployments of the emerging chip making technology.

According to Richard Windsor, industry specialist at securities house Nomura (London) this is somewhat surprising as this technology was loudly touted earlier this year. "Nikon at that time also discussed a 2012 shipment date but clearly that is now uncertain."

Windsor notes that by contrast ASML's EUV program looks "very much on track" and the Veldhoven, Holland-based company "aims to ship beta tools to customers in the second half of 2010."

He suggests that if the Japanese tools developer does not restart investments in EUV soon, "ASML's virtual monopoly may well become a real one. We estimate this lead to be at 18 months but clearly that will now meaningfully increase."

http://www.eetimes.com/news/se...l;?articleID=217500293

 

[magreen]

Hey idc, what does this mean to us? I've never heard of Nikon or ASML. And who's slated to use this euv lith technique? And at what process node? And what companies would use their fabs?

I've heard of tsmc, tfc, intel, but this stuff is beyond me. Thanks!

 

[Idontcare]

(Nikon and ASML are equipment providers, they make the lithography tools that are necessary for everyone to make their ICs, be it CPU's or memory or analog chips, etc)

EUV is the next generation of litho needed to make chips at 16nm node and beyond.

A delay in EUV means a potential delay in 16nm, or more likely the deployment of a more costly/expensive alternative litho approach.

(for example Intel's 45nm uses dry litho because immersion litho was simply not ready for manufacturing environment when Intel took their 45nm to production...in Intel's case the dry litho approach required multiple-pass patterning, increasing the time it takes to manufacture the IC and increasing its cost)

In this case ASML say they are still on track with EUV, but being a sole provider of the equipment doesn't bode well from the cost viewpoint so again for you and mean it signals that the cost of an EUV processed wafer could be higher than it would have otherwise been were Nikon able to field a competitive EUV tool when ASML does.

If ASML is the only EUV supplier at the time, then EUV could be viewed as the "i7" of its time (figuratively speaking of course)...available for anyone who wants the best performance, but the cost to "step-up" to the equipment is going to be all the more painful without the competition around.

 

[aigomorla]

why is it the moment i read lith, i knew it had to be from you IDC?

I think i even learned what a lith was from you directly.

 

[taltamir]

+1 to that aigomorla... idontcare is just a fountain of knowledge on this subject.

 

[magreen]

Cool, thanks for the info. :thumbsup:

I have nothing else meaningful to say

 

[Idontcare]

Originally posted by: aigomorla
why is it the moment i read lith, i knew it had to be from you IDC?

I think i even learned what a lith was from you directly.

I can only apologize for having infected you...:laugh:

It is true I am passionately interested in the subject of process technology. It's actually a long story and I'll spare everyone the details, but I find this side of the technology backstory to be very dynamic and exciting.

I'm sure there is an equally exciting and ongoing backstory to the design and layout of these CPU's and GPU's as well, but I have little to no experience there so I can't comment on most of it. But there are those that can and do if you know how to listen to what they are saying, and sometimes more importantly you have to notice what they aren't saying.

It'll never be as obvious as someone saying "I can't tell you what the L1$ latency is for Westmere, but I can tell you it isn't 1, 2, 4, 5, or 6 clock cycles..." but the tea leaves can be sifted if you know whose tea leaves are worth sifting.

 

[pm]

So my question is, do we need EUV for 16nm?

Based on my recollection, people have been saying that EUV is coming since 180nm. Once the prospect of 157nm DUV disappeared, EUV became the next big thing. But I don't think anyone would have thought we could do 22nm without EUV a decade ago, and yet it's almost here and EUV is still on the horizon.

Could the industry do 16nm as 193nm immersion with a (as-yet-unknown) high-NA fluid and some form of double-patterning or something combination of techniques?

(lest anyone think that I'm asking some form of tricky leading question, I'm not. I'm working in CPU testing now, and I haven't followed the litho scene in literally a decade)

 

[Idontcare]

Originally posted by: pm
So my question is, do we need EUV for 16nm?

Based on my recollection, people have been saying that EUV is coming since 180nm. Once the prospect of 157nm DUV disappeared, EUV became the next big thing. But I don't think anyone would have thought we could do 22nm without EUV a decade ago, and yet it's almost here and EUV is still on the horizon.

Could the industry do 16nm as 193nm immersion with a (as-yet-unknown) high-NA fluid and some form of double-patterning or something combination of techniques?

(lest anyone think that I'm asking some form of tricky leading question, I'm not. I'm working in CPU testing now, and I haven't followed the lith scene in literally a decade)

You are thinking of high-RI fluids.

Here is a pretty decent article discussing the potential and limitations of extending immersion litho to higher RI fluids: http://spie.org/x20167.xml?ArticleID=x20167

And yes it certainly can be extended to 16nm, and beyond, its all just a question of cost and risk. 193nm immersion litho displaced 157nm dry litho based on expectations of timetable risks and costs associated with the CaF2 source materials needed for 157nm litho.

The slipping node intersection point for EUV is not too unlike the continual pushback on incorporating ULK dielectrics into the BEOL, or HKMG into the FEOL. There is a certain eventuality to it as nothing else will solve the fundamental problem at hand.

So do we need EUV for 16nm? No, no more than we need HKMG to make 16nm chips. But EUV is expected to make for cheaper production of 16nm chips versus the estimated costs associated with high-RI immersion fluids combined with double patterning and computational litho. (this assumes EUV costs per wafer pass continue to decrease per projections and expectations...something that this news from Nikon puts in jeopardy)

 

[daw123]

When I first read your IDC's post, I must say that I didn?t have a clue what he was talking about (put that down to my complete ignorance on the subject ).

But a quick search of Wikipedia using Google I read this then applied it to this (specifically the first picture on the right) then this and finally this and I now understand the very basics of what IDC's post is talking about.

Thank you for introducing the un-initiated, IDC; it looks like a very interesting subject.

IDC, do you have any links for slightly more in depth information on this subject (but not too in depth that it will baffle a layperson, such as myself).

Back on topic, if only AMSL is able to produce EUV tools capable of manufacturing future processors because the competition are lagging behind in the technology, then it will ultimately have a financial impact on the end-consumer.

Look at AMD vs. Intel as an example; Intel can command higher prices for their high-end tech because AMD is unable to sell equivalent performing chips (competition doesn?t lower the prices).

P.S. I included the first bit of my post for the benefit of those who were as clueless on this subject as myself.

 

[Idontcare]

daw that is quite the rabbit hole you've plunged into For being a self-professed "uninitiated" you seem to betray a certain uncanny ability to accidentally wander thru the right manner of links nonetheless :laugh:

For starters on your request, take a gander at the wiki material on EUV. It's not overly technical but does a decent job of touching on a lot of relative facets that compromise the overall gem.

Basically you can't google for EUV or Extreme Ultraviolet and find a link that would do a disservice to you. Nothing will be overly technical as all the technical stuff is hidden away in published journals requiring membership to view the articles anyways. But there is plenty of "trade journal" articles out there on the web to satiate anyone's appetite to know more about EUV.

Here's some recent proof of capability for EUV: IMEC shows record 22nm SRAM cell density using EUV

 

[daw123]

Originally posted by: Idontcare
daw that is quite the rabbit hole you've plunged into For being a self-professed "uninitiated" you seem to betray a certain uncanny ability to accidentally wander thru the right manner of links nonetheless :laugh:

For starters on your request, take a gander at the wiki material on EUV. It's not overly technical but does a decent job of touching on a lot of relative facets that compromise the overall gem.

Basically you can't google for EUV or Extreme Ultraviolet and find a link that would do a disservice to you. Nothing will be overly technical as all the technical stuff is hidden away in published journals requiring membership to view the articles anyways. But there is plenty of "trade journal" articles out there on the web to satiate anyone's appetite to know more about EUV.

Here's some recent proof of capability for EUV: IMEC shows record 22nm SRAM cell density using EUV

Thanks for the links IDC, I'll read them tomorrow.

P. S. I'll confess that there were a few hit-and-miss website attempts, before I thought about using Wikipedia.

 

[Nemesis 1]

Intel just Bought into ASMLis company . I wonder what customer is first?

 

[Idontcare]

Plenty of Reasons to Be Confident About EUV Lithography

ASML presented process data at 28 nm half-pitch for 1:1 lines and spaces (L/S) with <10% critical dimension uniformity (CDU). Presenters also showed a large process window (>200 nm for 28 nm L/S) for EUVL compared with ArF immersion, which is why the industry continues to stand behind EUVL.

ASML, which dominates this particular scanner market, will happily sell you two 193 nm immersion scanners for double pattering instead of one EUVL scanner. But its own CoO calculations, shown in the keynote talk, concluded that end users will realize significant cost savings from EUVL compared with immersion/double patterning at 22 nm. This should put an end to the debate on whether the cost of EUVL will be competitive with that of double patterning.

In another keynote talk, AMD showed its process development results for the 22 nm node (although AMD in 2008 showed process development results for the 45 nm node, its researchers and partners in the IBM Alliance skipped over 32 nm because they expect EUVL implementation to occur beyond 22 nm).

http://www.semiconductor.net/a....html?industryid=47299

daw123, pm, this article is perhaps one of the best ones I've read in the public domain regarding the progress, performance, cost analysis, and other careabouts (LER, resists, source power, etc) for EUV.

Its a definite read, will bring you up to speed and traverse about 10yrs of prior history on EUV in the span of a few minutes of your time.

 

[daw123]

Thanks for link, IDC.

Sorry I haven't been able to read the article, but our ISP was down all day yesterday.

I'll read it now and post any questions I have.

 

[Idontcare]

Originally posted by: daw123
Thanks for link, IDC.

Sorry I haven't been able to read the article, but our ISP was down all day yesterday.

I'll read it now and post any questions I have.

You are most welcome, and this isn't a class, there won't be a quiz at the end, no need to apologize, take all the time you want. I guarantee you'll fully understand EUV long before it goes into production :laugh:

 

[PCTC2]

Hey IDC. Have you read anything about Cymer? I haven't really heard anything about Cymer Inc lately or their ventures into EUV. I know they currently use 193nm DUV immersion litho for their product venture into 45nm.

(Former Electrical Engineering major. I miss my old micro-processor lab to make 0.18um chips).

 

[Idontcare]

Nikon: EUV program alive and well

Responding to an analyst, Japan's Nikon Corp. insists that its extreme ultraviolet (EUV) lithography program is still alive and well.

There is also no delay with its EUV program, according to officials from Nikon (Tokyo). As reported, lithography specialist Nikon was understood to have put on hold development of some of its EUV tools, which could give ASML Holding NV a huge early advantage in commercial deployments of the emerging chip-making technology, according to Richard Windsor, industry specialist at securities house Nomura (London).

http://www.eetimes.com/news/se...l;?articleID=217500676

Nikon calls shens on analyst :laugh:

I believe Nikon. The analyst needs the publicity to make money, Nikon doesn't need publicity, their customers already know if the project was mothballed or not.

I wonder if in a case like this if the shareholders of Nikon can sue to recover lost equity if this analyst report can be shown to have adversely effected market price?

Originally posted by: PCTC2
Hey IDC. Have you read anything about Cymer? I haven't really heard anything about Cymer Inc lately or their ventures into EUV. I know they currently use 193nm DUV immersion litho for their product venture into 45nm.

(Former Electrical Engineering major. I miss my old micro-processor lab to make 0.18um chips).

Yeah at 193nm we had Cymer light sources on our litho equip.

Cymer is critical to the success of EUV as the power of the beam determines wafer throughput which determines cost.

http://www.semiconductor.net/article/CA6614099.html

Power and lifetime are critical metrics of success for the EUV source, and cymer is the key player in solving those issues.

 

[PCTC2]

Thanks IDC. I'm looking forward to see their end-of-calendar year results with their tin-deployment and debris-management systems for their EUV laser. There's quite a few job openings, even for full-time engineers over interns, and I'd definitely want to jump on that train if I had the correct job experience.