A TEM STUDY OF DIFFUSION-INDUCED GRAIN BOUNDARY MIGRATION IN Ni-Cu DIFFUSION COUPLES

(1)

HAL Id: jpa-00228078

https://hal.archives-ouvertes.fr/jpa-00228078

Submitted on 1 Jan 1988

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

A TEM STUDY OF DIFFUSION-INDUCED GRAIN BOUNDARY MIGRATION IN Ni-Cu DIFFUSION

COUPLES

D. Liu, W. Miller, K. Aust

To cite this version:

D. Liu, W. Miller, K. Aust. A TEM STUDY OF DIFFUSION-INDUCED GRAIN BOUNDARY

MIGRATION IN Ni-Cu DIFFUSION COUPLES. Journal de Physique Colloques, 1988, 49 (C5),

pp.C5-635-C5-640. �10.1051/jphyscol:1988581�. �jpa-00228078�

(2)

A TEM STUDY OF DIFFUSION-INDUCED GRAIN BOUNDARY MIGRATION IN Ni-Cu DIFFUSION COUPLES

D. LIU, W.A. MILLER a n d K.T. AUST

D e p a r t m e n t of M e t a l l u r g y a n d M a t e r i a l s S c i e n c e , U n i v e r s i t y o f T o r o n t o , T o r o n t o M5S 1A4, C a n a d a

ABSTRACT

A study was conducted on defect s t r u c t u r e s , o r i e n t a t i o n r e l a t i o n s h i p s , and compositional p r o f i l e s a t DIGMboundaries i n NiXu d i f f u s i o n couples. TEM revealed d i s l o - cations a t the i n i t i a l g r a i n boundary positions of the DIGM zones i n the N i substrate.

The misorientation between t h e DIGM zones and the m t r i x (across the dislocation w a l l ) was determined by electron d i f f r a c t i o n and found t o be l e s s than 0.5'. Cu p r o f i l e s across the DIGM zones were obtained by TEMIEDS analysis. The f o m t i o n of t h e dislocation wall i s discussed i n terms of the misorientation and l a t t i c e misfit between the DIGM zones end the m t r i x .

INTRODUCTION

Dif fusion-Induced Grain Boundary Migration (DIGM) i s a well recognized phenomenon i n vhich s o l u t e diffusion along grain boundaries gives r i s e t o motion of the boundaries (1). Two d i f f e r e n t mechanisms have been proposed t o explain DIGM. One of these,the coherency s t r a i n model ( 2 ) , s t a t e s that s t r e s s e s due t o compositional inhomogeneity i n the v i c i n i t y of a boundary during diffusion provide adriving force f o r DIGM. The other ( 3 ) ( 4 ) i s based upon grain boundary dislocation climb caused by unequal boundary dif f u s i v i t i e s of solute and solvent a t m s . However, n e i t h e r model can explain a l l of the DIGM-related observations recorded i n the l i t e r a t u r e . Although mny

theoretical and experimental papers on DIGM have been published, only a few have involved TEM studies such a s the work on bulk sarrples of Fe-Zn(2), Cu-Zn(5) ( 6 ) , Al-Zn(7) and A1-4.7% Cu a l l o y ( 8 ) . The s c a r c i t y of TEM data i n the l i t e r a t u r e i s probably due t o d i f f i c u l t i e s i n preparing TEM f o i l s , since the DIGM zone occurs within a few microns of the o r i g i n a l i n t e r f a c e between the two phases ( 9 ) . The purpose of t h i s paper i s t o report our i n i t i a l TEM studies of DIGM i n the N i

substrate of Cu-Ni diffusion couples a f t e r annealing.

EXPER IWNTAL

High p u r i t y N i b a r s (99.99%) were reduced t o a thickness of 250 m by rmltiple-pass cold r o l l i n g . The N i s t r i p s were annealed a t 9 0 0 ~ ~ f o r 24h i n A r . The recrysta- l l i z e d N i s t r i p s were then electropolished i n an e l e c t r o l y t e containing 90% methanol and 1% perchloric acid a t -20°c t o reduce the thickness t o approximtely 150um. The electropolished N i s t r i p s were then electrodeposited with Cu i n a solution consisting of copper sulphate and sulphuric acid. A Cu l a y e r , 15 urn t h i c k , was coated on each s i d e of the N i s u b s t r a t e t o form a Cu-Ni-Cu sandwich. The d i f f u s i o n couples so pro- duced were sealed i n pyrex capsules f i l l e d with p u r i f i e d A r and annealed a t 6 1 5 ' ~ for 26h. Disks of 3 mn i n diameter were then spark-cut from the annealed Ni<u s t r i p s . In order t o study the DIGM region adjacent t o the original Ni/Cu i n t e r f a c e , the following perforation method was used. F i r s t , the d i s k s were protected on one sur- face by a t h i n PTFE f o i l and jet electropolished i n an e l e c t r o l y t e consisting of 400 m l a c e t i c a c i d , 300 m l phosphoric a c i d , 200 m l n i t r i c acid and 100 m l d i s t i l l e d water a t -15O~ f o r 2 min. A s a r e s u l t , one l a y e r of Cu was dissolved and the N i substrate was thinned t o about 20 urn. The d i s k s were cleaned i n d i s t i l l e d water and methanol and then j e t polished on both sides using the conditions described previously, u n t i l perforation occurred. The second stage of jet polishing requires about 5 t o 10 sec.

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1988581

(3)

C5-636

JOURNAL DE PHYSIQUE

The TEM microstmcture was studied with a HITACHI H-800, 200 k v 'TEM which was

equipped f o r EDS microanalysis. X-ray microanalysis of Cu p r o f i l e s across DIGM zones was performed using a probe s i z e of 50 nm. The take-of f angle f o r the X-rays was 69.5O. The EDS processing was carried out using a Apple

II:

conputer and Dapple soft- ware. Convergent beam d i f f r a c t i o n p a t t e r n s with a spot s i z e of 0.1 urn were recorded i n the a r e a where EDS a n a l y s i s was t o be performed in order t o measure the f o i l thickness. The f o i l thickness was found t o be l e s s than the mximun a l l m b l e l i m i t using the C l i f f-Lorimer approach without an absorption correction. The mximm allowable limit f o r N i was calculated t o be2

l f i ~ J (

according t o Tixier and Philibert

(10).

RESULTS

In the present study, TEM observations were m d e i n the N i substrate i n which f i f t e e n boundaries were observed t o migrate e i t h e r i n a single d i r e c t i o n o r i n opposinq d i r e c t i o n s by DIGM. Fig.1 shows a representative boundary which has moved by DIGM i n one d i r e c t i o n . The i n i t i a l g r a i n boundary position i s mrked by a dislocation wall.

The Cu concentrations along AB,CD and EF in Fig.1 were obtained using EDS. These composition p r o f i l e s gave similar p a t t e r n s i l l u s t r a t e d i n Fig.2, i n which the zone swept by DIGM i s seen t o be Cu-rich.

Fig.1 ( l e f t ) : A segment of grain boundary showing the i n i t i a l position XY and the f i n a l position PQ a f t e r migration by DIGM. Note the dislocation wall a t XY.

I

^% ^Cu"'²²₂₀

I 8 16 14 12 l o

-

4 2 0

Fig.2 ( r i g h t ) : Cu concentration p r o f i l e s along AB in Fig.1 showing a Cu rich region i n the DIGM zone. Cu p r o f i l e s along

CD

and EF i n Figl. shoved a similar p a t t e r n .

W T R I X MATRIX

- - - - - -

-

B

-4-C. A -.---

- -

0 I 2 3 0 6

D6TAMCE I x m l

(4)

t r a l area of the DIGM band shown in Fig.1. The dislocations a t the original grain boundary position were tangled, and some of them extended i n t o e i t h e r the DIGM zone o r the m t r i x . The i m g e was checked in different g vectors; the m j o r i t y of the dislocations s t i l l remained v i s i b l e , indicating that the Burger's vector of the dislocations i s not unique. Misorientations between the DIGM zones and the m t r i x across the dislocation w l l s were measured using Kikuchi l i n e analysis a s shown i n Fig.b(a) and ( b ) . Three different beam directions were chosen a t positions DP1 and DP2 a s shown i n Fig.3.- The measured misorientations were 0. lb: O.lloand 0.18~corresponding t o the [002], [I141 and ~ 2 2 3 1 beam directions respectively. The misorienta-

tion measurements were a l s o carried out a t three different positions ( AB,CD and EF ) a s indicated in Fig.1, using a [I141 beam direction. The r e s u l t s showed misorientations of 0.32: 0.28'and 0.46'at AB,CD and EF, respectively.

Fig.3 ( l e f t ) : Dark f i e l d i m g e of the central part of the DIGM band in Fig. 1, showing the dislocation wall a t the original boundary position, grain boundary dislocations and ledges a t the migrated grain boundary.

Fig.h(a) and ( b ) (right 1: A p a i r of Kikuchi/spot patterns taken a t DP1 ( m t r i x ) and DP2 (DIGM zone) i n Fig.3 using a [I141 beam direction. Fig.b(a) was recorded by overlapping the Kikuchi and spot centres. Fig.4(b) shows a s m l l deviation between the two centres, corresponding t o a 0.11 misorientation between the DIGM zone and the the m t r i x .

A typical bi-directional migration of DIGM i s shown i n Fig

.5.

I n i t i a l l y , the boundary migrated from position A X t o position EFC, leaving behind a dislocation wall a t the original boundary position. Then, the part

FC

together with boundary

CD

moved backwards t o form an S-shaped DIGM zone.

:ig.5: The original boundary position i s shown a s AECD. I n i t i a l l y , a boundary segment A X moved by DIGM and

stopped a t position

EFC.

Then, segment FC migrated backwards, c d i n i n g with the unrnigrat ed segment CD and leaving a dislocation wll a t FCD.

(5)

C5-638 JOURNAL DE PHYSIQUE

Cu concentration p r o f i l e s along FQ and XY in Fig.5 a r e plotted in Fig.6(a) and (b) respectively. A peak concentration (22 wt% Cu) occurred just behind the dislocation wall i n the second DIGM zone ( l i n e XY ) which i s higher than the value of 15.5wt% Cu

in the f i r s t DIGM zone ( l i n e

PQ).

The measured misorientations across the dislocation wall i n the f i r s t DIGM zone

(PQ)

a r e 0.21' and 0.18' using beam directions of

[ i l 4 ] and 10021 respectively

.

^Fig.⁷shows relatively unif o m arrays of dislocations a t an original boundary position i n another DIGM region. Cu concentrations across the wall were found t o be 4.5wt% i n the m t r i x and 19.5wt% i n the DIGM zone. The misfit parameter was-estimted t o be 0.4%, and the misorientation across the wall was found t o 0.2' i n a [I141 beamdirection.

O 1 2 1 1 5 6 7 1 0 ~ m

DISTANCE ( ~ m )

CU%

-

_MATRIX

2 2 L-

20 -

I

14

-

Fig.G(a) ( l e f t 1: Cu concentration p r o f i l e measured by EDS along the l i n e FQ i n the f i r s t DIGM zone show in Fig.5. Note the peak ( 15.5wt% Cu ) a t the dislocation wall i n the DlGM zone.

Fig.6(b) ( r i g h t ) : Cu concentration p r o f i l e measured by EDS along the l i n e XY in the second DIGM zone i n Fig 5, showing a higher peak ( 22.5wt% Cu ) a t the dislocation wall i n c c q a r i s o n with that ( 15.5wt% Cu ) i n Fig.6(a).

Fig.7:A dislocation u a l l a t the original boundary position in a DIGhl zone.

The dislocation arrays have a misfit of 0.4% and a misorientation of 0 . 2 ~ between the DIG.4 zone and the m t r i x .

DISCUSSION

Dislocation w a l k a t the original grain boundary position have been observed i n several DIGM systems (21, ( 5 ) , ( 9 ) . H i l l e r t and Purdy ( 2 ) reported that the disloca- w a l l contained a misfit array i n the Fe-Zn system. However. Grovenor (11) argued that the dislocations constituted a low angle boundary in the Au-Cu thin film sanples, Recently, Hackney et a1 (5) pointed out that there was a smll misorientation across the dislocation -11 i n the Cu-Zn system, although no experimental measurements were reported. In the present study, the steep Cu concentration gradient a t the original grain boundary positions and the s m l l misorientations ( l e s s than 0.5O ) across the dislocation walls indicate that these d i s l o u t i o n arrays appear t o have both l a t t i c e misfit and m i s o r i e n t a t i a characteristics.

The presence of the dislocation wall a t the original boundary position m y be rationalized i n terms of H i l l e r t ' s coherency s t r a i n model on the basis of s t r e s s relaxation leading t o the i n i t i a t i o n of boundary migration. According t o H i l l e r t

,

the coherency s t r a i n energy can be e s t i m t e d a s follows (8) :

(6)

the bulk concentration of the solute, and

X2

i s the concentration a t the grain boundary. For Ni,

E

⁼2.07~11)~hpa(12) and

V

= 0.3 (13), with the measured corrpositions

X2=0.184 ( 19.5wt% )

, XI=

^0.042 ⁽ ^{4.5wt% Cu} ⁾

,

and the calculated value of

q

⁼^0.4%

i n Fig.7, Eq( 1 ) y i e l d s a coherency energy of about 1 x lo5 j /m3. The curvature of the DIGM zone shown i n Fig.7 i s 5 x107m-I. I t has been reported that the required driving force f o r DIGM i s

lo6

t o 107 j/m3 with boundary energies of the order of 1 j /m2 and observed curvatures of

lo6

t o 107 ni1(14), about an order of mgnitude larger than the value indicated by the present r e s u l t s . Hackney e t a1 ( 5 ) e s t i m t e d the coherency

s t r a i n energy i n the Cu-Zn system using a different equation:Eq2>6uhere

E

a n d q a r e a s defined previously,

8

i s the boundary energy, and /( i s the boundary curvature.

They a l s o found a value of coherency s t r a i n energy of one order of magnitude too l w f o r t h e observed curvature. However, they pointed out that the coherency s t r a i n energy calculation result does not invalidate the coherency s t r a i n theory, and suggested that an additional driving force may be present.

Cu concentration steps across the original grain boundary positions shown i n Fig.6(a) and ( b ) resulted i n a 0.287 coherency s t r a i n a t the i n i t i a l l y migrated DIGM segment of the boundary and a 0.46% coherency s t r a i n a t the second. The l a r g e r coherency s t r a i n ( 0.46% ) m y not be necessary t o provide a larger coherency s t r a i n energy, since the driving force would a l s o depend on the orientation dependence of the e l a s t i c modulus (15). The DIGM d i r e c t ions i n the neighbouring grain were not determined, so that the exact e l a s t i c moduli i n the two directions a r e unknown. However, i t i s clearly shown that the direction of i n i t i a l movement ( from P t o Q i n Fig.5 ) corresponds t o the s m l l e r coherency s t r a i n ( 0.287 )

.

The phenanenon of the same boundary moving i n both the forward and backward directions i s d i f f i c u l t t o explain by the dislocation c l i h model. This model predicts that t h e structure of the boundary determines the direction of migration. A reversal i n the direction of migration would require change in the dominant set of grain boundary dislocations a t various positions of the same boundary.

This paper reports on the f i r s t TEM study of DIGM in a Cu-Ni bulk system. The grain boundaries i n the N i substrate were observed t o move e i t h e r i n a single direction o r in opposing directions by DIGM. The i n i t i a l boundary positions a r e outlined by dislocation u a l l s which have both misfit ( 0.2 t o 0.4& ) and misorientation

(

-=

^O.SO⁾characteristics. The coherency s t r a i n energy ( driving force ) was estimated a s 1 x l o 5 j/m3, *ich i s one order of mgnitude s m l l e r than that required f o r the DIGM process. I t is suggested that an additional driving force f o r the i n i t i a t i o n of the DIGM process m y be required.

This work uas supported by the Natural Science and Engineering Research Council of Canada. One of the authors wishes t o acknowledge several useful discussions with Professor G.C. Weatherly during the course of t h i s work.

REFERENCES :

(1) P.G. S h m and G. Meyrick " Diffusion i n So1ids:Recent Developments" Edited by M.A. Dayananda and G.E. k r c h , A I E , ( 1985 ) p. 261

(2) M. H i l l e r t and G.R. Purdy, Acta Met. 26, (19781, 333 ( 3 ) R.W. Balluffi and J .W. Cahn, Acta Met. 29, (19811, 493 ( 4 ) D.A. Smith and A.H. King, Phil. k g . A44, (19811, 333

(5)

S.A. Hackney, F.S. Biancaniello, D.N. Yoon and C.A. Handwerker. Scripta Met.

20, (1986)

,

93?

( 6 ) S.A. Hackney, Scripta Met. 20. (1986). 1385 . -

( 7 ) K. Tashiro %d G.R-. Purdy, Scripta ~ e t

.

17, (19831, 455

(8) J .R. Michael and D.B. Williams, " Interface Migration and Control of Microstruc- ture" ( Proc. Conf .), D e t r o i t , Michigan, USA, 17-21 Sept. 1984, American Society f o r Metals, Metals Park, Ohio 44073, USA, 1986. p . 12

(7)

C5-640 JOURNAL

DE

PHYSIQUE

(9) F. J

.A. den Broeder and S. Nakahara, Scripta Met.

17, (19831, 399

(10)

R. Tixier and J . P h i l i b e r t , (

1969

) , Proc. 5th Int. Ccmf. on X-ray Optics and Microanalysis, eds. G.&llenstendt and K.H.Gaukler, Springeryerlag, Berlin,p.

180 (11)

C.R.M. Grovenor, Acta Met.

33, (1985) , 579

(12)

C.R. Barrett

, W.D.

Nix and A.S. T e t e l m " The Principles of Engineering k t e r i a l s " Prentice-Hall,

11973), p.540

(13)

ibid. p.

197 (14)

J

.W.

Cahn, J .D.Pan and R.W. Balluffi, Scripta Met.

13, (19791, 503

(15)

D.N. Yoon,

J

.W.Cahn, C.A. Handweker, J .E. Blendell and Y . J

.

^Baik,^"^Interface

Migraticm and Control of hiicrostructure

"

( Proc. Conf

.

⁾Detroit, Michigan, USA,

17-

21

Sept.

1984,

American Society for Metals, Metals Park, Ohio

44073,

USA,