Hic­cups for the Next Big Thing in chip­mak­ing

Chip­mak­ers await ma­chines that make faster tran­sis­tors “We may never see a pay­back on the in­vest­ment and time”

Bloomberg Businessweek (Europe) - - NEWS - The bot­tom line In­tel, Sam­sung, and TSMC may start to see signs that their $1.6 bil­lion in­vest­ment in ASML is pay­ing off.

As chip­mak­ing ad­vances have run up against the lim­its of physics, threat­en­ing to push up man­u­fac­tur­ing costs, the in­dus­try’s lead­ers have placed a great deal of faith in a Dutch com­pany called ASML. For most of the past decade, ASML has been promis­ing that its new tech­nique for cre­at­ing tran­sis­tors would al­low chips to keep get­ting slim­mer and more pow­er­ful at the rates we’re used to. In 2012, In­tel, Sam­sung, and TSMC took the un­prece­dented step of in­vest­ing about $1.6 bil­lion in ASML to speed its re­search and paid close to $5 bil­lion for 23 per­cent of the com­pany.

ASML ex­pects to ship as many as seven new ma­chines this year so chip­mak­ers can start test­ing the tech­nol­ogy, known as ex­treme ul­tra­vi­o­let lithog­ra­phy (EUV). Chief Ex­ec­u­tive Of­fi­cer Peter Wen­nink says most of his cus­tomers ex­pect to in­cor­po­rate EUV by 2019, and he’s prep­ping for or­ders within the next year. “The in­dus­try needs EUV,” he says.

So far, though, the tech­nol­ogy for EUV is prov­ing less ef­fi­cient than once thought and pos­si­bly un­able to de­liver the re­turns in­vestors hoped for. Even al­ter­na­tives that would make smart­phone and PC parts costlier have started to look more at­trac­tive, says Pa­trick Ho, an an­a­lyst at in­vest­ment bank Stifel Ni­co­laus. “The in­dus­try made their bet sev­eral years ago that EUV would be the next gen­er­a­tion,” he says. “It’s been very dis­ap­point­ing.” TSMC de­clined to com­ment for this story. Sam­sung didn’t re­spond to a re­quest for com­ment.

Lithog­ra­phy is the process of us­ing con­cen­trated rays of light to burn lines into lay­ers of ma­te­ri­als de­posited on sil­i­con, a cru­cial step in cre­at­ing tran­sis­tors—and a nat­u­ral choke point for en­gi­neers work­ing at a near-atomic level. Th­ese days, com­pa­nies are try­ing to etch lines smaller than the wave­length of the light used to do the work, hence the push into ex­treme ul­tra­vi­o­let beams with shorter wave­lengths. “With­out EUV … your eco­nom­ics are worse,” says We­ston Twigg, an an­a­lyst at Pa­cific Crest Se­cu­ri­ties. “If EUV is not ready, things get a lot harder.”

The prob­lem is that beams with shorter wave­lengths use a lot of en­ergy, and ASML’s ma­chines re­quire sub­stan­tial down­time be­cause EUV dir­ties the mir­rors used in the process. Ac­cord­ing to its pub­lic state­ments, ASML aims to cut the EUV ma­chine’s re­quired down­time, from the cur­rent 25 per­cent to 30 per­cent, to 20 per­cent by yearend. Not ex­actly what you want to hear when the price tag (even with­out the re­search and de­vel­op­ment fund­ing) runs to eight fig­ures, mak­ing it the most ex­pen­sive de­vice in the plant.

Stifel Ni­co­laus’s Ho says ASML, one of Europe’s few tech­nol­ogy pow­er­houses, has to prove it can keep to its sched­ule. Its lat­est, that is. In 2007, for­mer CEO Eric Meurice said EUV ma­chines would be cost-ef­fec­tive for chip­mak­ers by 2012. Says Pa­cific Crest’s Twigg: “It’s prob­a­bly the most ad­vanced sci­en­tific re­search pro­gram in the world. Yet the pro­gram is still be­hind.”

In Fe­bru­ary, TSMC co-CEO Mark Liu told in­vestors that his com­pany has backup plans. That month, In­tel’s di­rec­tor of lithog­ra­phy strate­gic sourc­ing, Jan­ice Golda, wrote in a com­pany blog post that the ques­tion with EUV is when, not if. But, she added, “the road to EUV lithog­ra­phy pro­duc­tion is a long one.” Chip­mak­ers tend to in­cor­po­rate man­u­fac­tur­ing ad­vances in two- to three­year cy­cles, so if EUV isn’t ready this year, ASML’s next big chance would be closer to 2020.

For now, the most ob­vi­ous way to get smaller lines is to use cur­rent lithog­ra­phy tech­niques a greater num­ber of times on each chip. The big chip­mak­ers have been loath to do that be­cause it takes longer, al­ways top of mind in a $10 bil­lion fac­tory that will be ob­so­lete within five years. Yet the com­plex na­ture of EUV de­vel­op­ment should be the big­ger con­cern, says Robert Maire, pres­i­dent of Semi­con­duc­tor Ad­vi­sors. “There are so many things that can go wrong,” he says. “We may never see a pay­back on the in­vest­ment and time put into EUV.” Elco van Gronin­gen and Ian King

ASML’s ex­treme ul­tra­vi­o­let lithog­ra­phy ma­chine

Newspapers in English

Newspapers from Bahrain

© PressReader. All rights reserved.