Elm Exhibit 2137
`Samsung, Micron, SK hynix v. Elm
`IPR2016-00387
`
`

`
`
`
`
`
`,11r1o,:a1 11011 10:11 PM mu 343 am
`
`. NUVELLUS
`
`'
`
`'
`
`Elm!-‘I
`
`Attnrnr-y’: Dqokel No. 008442-028
`(fuxmatly 5l.960.P0.l7d)
`Page 2
`
`1 am cu:-nntly Cl'1i¢f’1'Iub.uicnl Officer of N1:-valluu, a §;oI‘lli.on I hm we
`5.
`sing: 1990.
`'
`'
`.
`6. In 1988 or so. ivuvcuua lnuuétuceu : mm-an Pncvn aeposuiun symn.
`This uyuwm provided for t.l1.I firs: rim: 1 npchanhn III comm! the su-us cf deposited
`films. one RF mum: is high fxequenny. and 1111: other RF neuron in law ftaqunncy.
`
`stress In conxrollad by varying :11: may mtln ohm: hlgh and law Ibquanuy RF-
`suumas. The mail: rule of th: high-fieqlazncy RF is to ganarnte the reactlve in-ecies'a:1d
`
`is Iddid to central thi
`prcwirlu sufficient clock-on and ion dannifisa. Th: lqw
`ion bombaxdment In which ‘I116 IU-blllllfll are uubjwnad. during dspouition.
`Insztnsing ti:
`
`low-nequency power incruses 11:: plasma potential and H1: Imoum of ions following Eh:
`low-fnaquuna.-.y RF finld (<1 MI-I2). Thu xullllfillg low-unargy Ion isnplnntattinn Granting
`
`during deposition causu a chug: in this in|'.ri1:I:ic film siren fzurn tensile to nu:-nprcmivo,
`increases film density, and i1:1uruv:a th: chcmlcal ructiana.
`.
`
`7. Sn-an level meammncms of dielectric film 11;: men spoken of and dapicued
`‘tn '*un.iu- of 1 x 10’ dynuI=n1',.w-Ith 1+" buingnsad In damn 1:-nails anus and -J’
`being usui La demon: compressive II.|:'B6S- Both inlngtr and tlcch-11.‘|1 values may b_= used.
`for cumple;
`
`
`
`E
`
`:
`
`-
`
`'
`
`-
`
`.5 I. 10' dyueslcn‘
`
`-4 :1 10’ dyneucm’
`-3 x 10' aynuicsu‘
`-2 1: 10- dynI:a.'cn1"
`-1 J: 10’. dsrneslnnf
`
`‘
`
`-11.1 x 10* dync5In:n'
`
`0 as 10’ d1'nesIc:n'
`
`0.1 :1 10' dynulem‘
`1 x 10' dyneslcm'
`
`'
`
`2 it 10’ dymsh.-Ira’
`3 x 10‘ dynesI'
`
`I
`
`.
`
`
`
`I
`
`-5
`
`,
`
`_
`
`1
`
`mic:
`
`Elm Exhibit 2137, Page 2
`
`Elm Exhibit 2137, Page 2
`
`

`
`__nr1o«.s'r
`
`‘non 10:15 nu; ans m was
`
`NOVELLUS
`
`I
`
`.
`
`1
`
`'m“
`
`
`
`Attorney‘: Docket No! 003442-928
`I
`(fD_l.'mI=|!1Y 5196U'.POI7d)
`1%.]: 3
`
`8. M‘. Novetlus. I would be approached by qmamcr: having various dlflemm tum
`
`requlnrrnnnltl. Malt ofthele eequixemeen were qufie conventional. Mont cu1ttnrt_'u':r:I
`
`in tensile
`requeslted !ihn.s.haVing II stress of ah-nut. I-1.. Customer! ‘marl mt
`flirnsbeeause Dfflilfl propensity orsuchfllgas no crack. Dielncmc films are case used an
`pnssivatlcn layers. cllcldng at which can mutt in an-ioc'con1a:ninat1un and failure In the
`field.
`
`9.
`
`In 1990, I was appxuached by Glenn Laud? thou: the feasibility of proelucing a
`
`thick dielectric film hull!!! very low tczuflc stress, This was I very unconventional
`requnr. one that. no my knowledge, no cnstcxncr Ind made before.
`
`10. Al: the time Mr. Lcedy read: his request. Nevcllua had completed a
`substantial body of work regmllng sues: centre: or thin man. This war}: suggested in
`
`theory that. by cent:-nliing deposition parameters (principally lite ratio of high frequency
`
`RF energy to low frequency RF energy). flux at-can could be controlled anywhere along I
`line passing through me and mending from about -5 to +3. Atunhed is a page from
`Nevel1us's Process Guide for Silicon Nitride. The figure at the bottom or fit: page shew’:
`the theeretiell curve for street Vexnu peteunl-fie of low-frequency powfl-1'. 362611!!! Wit}!
`
`I.'|8l.I points corrclgonding Infilml Ictullly prcthlced. Infrtctlcc. however, the lowest
`tensile stress rum produced at Nnvcllus Print to Ieceivln; Mr. Iaeedy's request was.
`according to my recnllcctiun. about +1.
`'
`I
`11. Based cnrny uplflernca. Ituponded to Mr. LeedYthI.tI'lhG1.l§htthat
`Novetlua euuld pmduec at film. um would meet his zequirctnents. Under m} direction. an
`engineer then undertuok a series of expefimenu. conducted over a. period of several whys.
`
`eunermretlng that indeed a Ihlflk low stress wniilu dielectric fi!_n: could he produced.
`having I. ntntfl ls-nrelbtebour 0.1:»: less-. murder of magnitude en:-mme Iowutlun
`tensile mm: that had previously been produced at Ncvcllus.
`12.
`I have reviewed U.S. Patent 4.702.936 to Mad: et 11-. From the limited
`
`infonmtion presented in the patent. I helievetltnt it is lmponlblc to any with my degree
`
`or certainty what the stress level at films produced In cceurdance with the tracking: or
`
`ruin-:
`
`Elm Exhibit 2137, Page 3
`
`Elm Exhibit 2137, Page 3
`
`

`
`’i
`
`
`
`_.11/10:91
`
`'
`'
`'
`HUN 10:15 mix «as an ans
`
`‘
`NOVELLUS
`
`v
`
`Blues
`
`Auomey-s Docket No, ousacz-nzs
`(fomlcrly 51960. P0176)
`,
`Page 1
`
`I
`
`ma panenl. would be. For eximpic. no mcntion I: made in Maori: er :1. or the reactor
`Iprfilfilllt. an impcnam plums; Variublu, Tia
`pwclsl Vlfinbles in the Ncwcllut
`process and in produce law mus tcuil: dialectic fan: for Mr. Lady included _
`
`tempcmuzc. p:_eI5'lli'¢. high fiequmny I3 pawn-. low fiequcncy RF power. and flow
`ratc: of vnrimu chemical ec.miitue.n:a (Spcclfimtian. page 21).
`13. Both amulslry aha ncunuaniany In llnpochnt ml: in dmminin; fllzn
`chlllcmrlsllca. Sllnlilritiu in chcmlsiry (or cw:n__ld.cn:1:a.i nhcmisliy) simply uaanui bu
`used to pradict film similarity span: from mnsidnrafiann of activation. The aim:
`chcmiuuy coupled with ditfarcnt activation can produce cntinly different mrultn.
`
`In tin
`
`can of Mad: at al.. thc ‘UV activation process m1pioycd'diffcrs grcatly from the dual
`RF activation przuccls uacd at Novclllil to Dmduce films (or Mr. Lccdy. To draw an
`analogy, the mac ingrcdicnu. flour and watcr my be used In ml: a pizza. or may be
`used no make a cdrc.
`~
`
`14. Bascd on my experience. 1 do believe IMI mm produocd in znccxtlance with
`an leaching: uf Mud: at :1. would have zcnsllcsucns.
`I-Inwcvcr. M.II::lu at .1. provides
`no muchanism for contzulling the mum lcvcl. i.c.. no mechanism akin In varying Eh:
`'
`
`ratioofhighfzcqucnqtcluwf¢‘1"¥=l|€:!¢ncrgyuinthac-seofr11eNovalJusay|tcrn.
`The ability to PIOIIIGB I low tcnsilc um: layer rclics on the ability to control atfcaa with
`In independent mechlnim or "lm'n'h.“ Nu mdnknnh is p:I:.Ir:n.t‘in.Maeds. :1 ll.
`
`:5. Silicon nitride meammuy goes down a: rclnllvcly high levels of trzmile am:
`(c.g., +1}1n accuzdance win: both thezmal depa-man -ms ugh frrqunncy 1-Ecvn.
`'
`Ptixthcrmnrc. other ‘work using aimfla: chmnimy and UV activation as Mzéln ct II. has
`
`reluliacd in flinn hlvinz strnilcs of about _-I-1. Ncic, fin: cnmplc. Ncgcyoahi. e1 31..
`“Residual Sins: of 1 Sl,..N,:I-I Pilmn Prcpucd by Afterglow Plum: Ghanaian! Vapor
`Deposition Techniq-uc", Jpn. J. Jppl. rm, von. 31. pp. L857-1.869 (1992). a copy of
`which is attachcd l:_n1'Ito.
`16. Th»: umdculxnsd declare: fuxlhet. that all statements made hcrcin at his own
`
`all slaiemznia made an infianuation and belief ax: believed to
`knuwlcdgc are ‘true. Ind
`bu tl.'I.1I:: Hind further that than ntztemnnn Well: nude with the kxmwlédge that willful false
`
`am":
`
`-a_
`
`Elm Exhibit 2137, Page 4
`
`Elm Exhibit 2137, Page 4
`
`

`
`3 _
`
`H
`
`__________
`
`. ’ _.__ -
`
`—- -—---—--—--'--- -'
`
`.11/1fl_._v(9'l’
`
`_IlON 10:19 FAI I08_ 34.! 3418-
`
`NDVELLUS
`
`AI:mmIy'n Do-chit No. 008442-032!
`(fillflfirly 5-1960.PD17d?
`Page 5
`
`1
`
`summons Ind tho Iih so made In punishahla by flan at Ixnpr-luamncnt. onboth. under
`Sfltfidn 1001 at’ This 18 of the Unitad sum Code. and that such willful fill! Ittbczuenn
`
`mayjeopanlin: ma valldil-‘Jr of the ippllnfinn. or any pamrgt issuing rh:mn_1.
`
`BM 1; [am]
`
`AIILIIH
`
`mm-
`
`I305.)
`
`'.-.-.7
`
`Elm Exhibit 2137, Page 5
`
`Elm Exhibit 2137, Page 5