Dielectric Film Influence on Stress-Migration
`
`
`
`
`
`A.lsobe, K Okamura, S.0gura and 0.Mizuno
`
`
`NEC Corporation
`
`
`
`
`
`ll2Z Shimokuzawa, Sagamihara, Kanagawa 229,
`
`
`
`Japan
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`The
`this work is to clarify the relation between dielectric
`purpose of
`
`
`
`
`
`
`
`
`
`film characteristics and stress-migration(SM).
`It has been considered
`that
`
`
`
`
`
`
`
`
`
`caused
`temperature
`the dielectrics deposition
`induced
`stress,
`by
`the
`
`
`
`
`
`
`
`
`
`
`
`
`
`difference of
`the expansion coefficients to Al,
`is
`the main factor for SM.
`
`
`
`
`
`
`
`
`
`
`that
`the
`authors
`found
`diffusion
`into dielectrics,
`However,
`Al
`Al
`
`
`
`
`
`
`
`
`recrystalization and dielectrics flexibility are also significant
`for SM.
`
`
`
`
`
`
`
`
`
`
`
`or Al-Si wirings
`of dielectric films were deposited on Al
`kinds
`3%
`
`
`
`
`
`
`
`
`
`oxidized
`Si wafers.
`Especially,
`sputter-SiO2, whose
`formed
`on
`film
`
`
`
`
`
`
`
`
`
`charactaristics are not
`influenced by deposition temperature, was deposited
`
`
`
`
`
`
`
`
`
`
`
`with following ways -
`(A)RT depo.,
`(B)l$flC depo.,
`(C)2EfiC depo., and (D)RT
`
`
`
`
`
`
`
`
`
`
`depo.
`3BZC
`anneal.
`To evaluate the material
`dependence,
`(E)SiO,
`after
`
`
`
`
`
`
`
`
`
`
`
`
`(F)Si0N and (G)SiN were depositted by PECVD under
`the same conditions except
`
`
`
`
`
`
`
`
`
`
`
`
`
`process gas.
`These wafers were annealed at 403C for
`lflmin,
`held
`then
`at
`
`
`
`
`
`
`
`
`
`
`
`hours
`l75C(N2)
`for
`evaluate
`the
`SM.
`The
`sample
`preparation
`Tflflfl
`to
`
`
`
`
`
`
`
`
`conditions and results are shown in the tables.
`
`
`
`
`
`
`
`
`
`
`
`
`
`the
`and
`Samp|e(A)
`(B)
`show more open failures than (C).
`in spite of
`
`
`
`
`
`
`
`
`
`
`
`
`
`the
`same
`deposition temperature as (A),
`(D)
`shows no failure.
`In case of
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`(E)
`shows many,
`(F)
`shows
`few and (6)
`PECVD samples,
`shows no failure. All
`
`
`
`
`
`
`
`the other samples show no failure.
`of
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`(B) and (C) can be
`the results of sample(A),
`The differences between
`
`
`
`
`
`
`
`
`
`
`
`the
`explained
`Al atom diffusion into dielectricsfll‘ and dielectrics
`by
`
`
`
`
`
`
`
`
`
`
`
`
`
`flexibility.
`in case of
`(A) or
`(B), high tensile stress is generated by the
`
`
`
`
`
`
`
`
`
`
`
`Al diffusion into dielectrics during the annealing after the
`deposition. On
`
`
`
`
`
`
`
`
`
`
`
`
`
`the contrary,
`in case of
`(C), as the diffusion occurs during deposition,
`the
`
`
`
`
`
`
`
`
`
`
`
`
`
`stress is
`lower
`than that of
`(A),
`because the dielectrics is flexible until
`
`
`
`
`
`
`
`
`
`
`
`
`it becomes a certain thickness.
`No failure in sample(D) can be explained as
`
`
`
`
`
`
`
`
`
`
`
`
`
`follows. When Al
`film is heated up,
`the Al stress becomes compressive but,
`
`
`
`
`
`
`
`
`
`
`
`
`over ZUEC,
`is reluxed by hillock formation and Al recrystalization.
`If
`the
`
`
`
`
`
`
`
`
`
`
`reluxation
`wirings
`is
`covered with dielectrics,
`this
`stress
`(only
`by
`
`
`
`
`
`
`
`
`
`
`recrystalization) causes the high tensile stress after cooled down. However,
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`the tensile stress is lower
`than (A),
`because the Al
`case
`in
`of
`(D),
`is
`
`
`
`
`
`
`
`
`
`The
`already
`recrystalized
`the
`annealing
`before
`the
`deposition.
`by
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`differences between the results of
`(F) and (G) can be also explained by
`(E),
`
`
`
`the Al diffusivity.
`
`
`in conclusion,
`
`”depositting
`
`
`
`reluxation.“
`
`
`
`
`
`
`
`
`
`
`
`
`
`the best way of dielectrics formation against SM is,
`
`
`
`
`
`
`
`
`low AI diffusivity dielectrics at
`low temperature after stress
`
`
`
`
`(Reference)
`
`
`
`
`
`
`
`
`
`
`[I] A.Tanikawa and H.0kabayashi,"AEM Observation of passivation films for Al
`
`
`
`
`
`
`
`
`
`
`
`
`Metallization", Ext. abst. of The Jap. Soc. of Appl. Phys. Autumn 1989
`
`P628
`
`
`
`
`
`
`June 12-13, 1990 VMIC Conference
`
`
`
`
`
`TH-0325-1/90/0000-0363 $01.00 C 1990 IEEE
`
`
`
`
`
`SAMSUNG ET AL. EXHIBIT 1048
`Page 1 of 2
`
`

`
`
`
`
`
`Dielectric Film Influence on Stress-Migration
`
`
`
`
`
`A.lsobe, K.Ukamura, S.0gura and 0.mizuno
`
`
`
`Tables
`
`
`
`
`
`
`Typical data of dielectrics deposition conditions and
`
`
`
`
`
`
`
`
`open failures after 7000H held in 1751? (NZ).
`
`
`
`
`
`
`
`
`
`I-
`
`
`
`
`Experimental® Pure Al
`(l'l=0.8um)
`
`
`
`
`
`
`
`In
`
`
`
`
`
`
`- In
`
`
`-In
`
`
`
`
`
`
`-wjjln
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`n
`
`
`
`
`
`
`an ammun-
`
`
`
`*1:Annealed by furnace
`
`
`
`
`
`
`
`*2:T is for tensile stress, others are compressive
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`jwI
`
`
`
`
`
`Experimentaléfi Al-1%Si
`
`
`
`
`
`
`
`(W=U.8L£n1)
`
`
`
`
`
`
`
`
`
`
`
`(%§I
`
`
`<T1me>
`<T1me>
`EMPa)
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`31
`
`
`
`
`
`
`-311
`
`
`
`
`
`
`-11
`
`
`
`
`
`
`
`!E—Wu
`
`
`
`
`
`
`
`—EK
`
`
`
`
`
`
`
`
`%—E
`
`
`
`
`
`
`
`jflflj
`
`
`
`
`
`*3:The same deposition conditions as (A)
`
`
`*42No bias
`
`*5!Set
`
`
`
`EfE
`
`
`
`
`
`
`
`
`
`
`
`is RT but wafer
`
`
`
`temp.
`
`
`
`
`
`temp. becomes 300*v350Y3 during depo.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`No failure in polyimide and (no bias) S-Si02 can be explained by film flexi-
`
`
`bility.
`
`364
`
`
`
`SAMSUNG ET AL. EXHIBIT 1048
`Page 2 of 2