7
October 1999 Ž . Materials Letters 41 1999 9–15 www.elsevier.comrlocatermatlet HREM studies on the microstructure of severely cold-rolled TiNi alloy after reverse martensitic transformation Y.F. Zheng a,b, ) , J.X. Zhang a , L.C. Zhao a , H.Q. Ye b a School of Materials Science and Engineering, Harbin Institute of Technology, P.O. Box 433, Harbin 150001, China b Laboratory of Atomic Imaging of Solids, Institute of Metal Research, Chinese Academy of Sciences, Shenyang 110015, China Received 2 December 1998; accepted 10 April 1999 Abstract The microstructure of Ti — 49.8 at.% Ni alloy, which was cold rolled to about 30% reduction in thickness in its martensitic condition and subsequently heated up to 2008C for half an hour, has been studied by high resolution electron w x w x microscopy. The newly-formed parent phase has an orientation relationship with the residual martensite: 111 ´ 110 , P M Ž . Ž . 101 6.58 away from 001 . The parent phase-martensite interface is not smooth and well coherent. The boundary P M between two subgrains of the parent phase is not straight and perfectly coherent, with partial dislocation observed at the 4 4 interface. Inside some parent phase grains, thin plate-like 114 and spear-like 112 twin-related parent phase variant pairs 4 are observed. The 114 twinning boundary is relatively straight, but with two or three atomic-height blurred layers existing near the interface. q 1999 Elsevier Science B.V. All rights reserved. Keywords: Titanium nickel alloy; Reverse martensitic transformation; Microstructure; High resolution electron microscopy; Cold rolling 1. Introduction Among many shape memory alloys, TiNi alloys are most frequently used for practical applications of both shape memory effect and pseudoelasticity. Re- cently, the effect of cold rolling on the martensitic transformation and microstructure of near equiatomic w x TiNi alloy have been intensively studied 1–3 through TEM, IF, DSC and XRD. Tadaki and Way- wx man 1 observed the aligned lamellar structure with ) Corresponding author. School of Materials Science and Engi- neering, Harbin Institute of Technology, P.O. Box 433, Harbin 150001, China. Tel.: q86-451-6412163; fax: q86-451-6413922; E-mail: [email protected] fine segments along different directions inside each lamellae in the martensite in TiNi alloy cold rolled wx up to about 30% reduction in thickness. Lin et al. 2 described the microstructure as mainly variant- crossed and cross-hatched morphologies in the 40% thickness reduced TiNi alloy specimen. The phe- nomenon of mechanical-induced martensite stabiliza- wx tion was observed in the cold rolling martensite 2 . Both deformed martensite structures and deformed- induced defects are considered pertinent to the wx w x martensite stabilization 2 . Many studies 4,5 have been concerned with the influence of annealing treat- ment right after the cold rolling on the shape mem- ory effect and pseudoelasticity characteristics of TiNi alloy. Except for the experimental result that the recrystallization would occur at temperatures G 00167-577Xr99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved. Ž . PII: S0167-577X 99 00095-6

HREM studies on the microstructure of severely cold-rolled

  • Upload
    others

  • View
    1

  • Download
    0

Embed Size (px)

Citation preview

Page 1: HREM studies on the microstructure of severely cold-rolled

October 1999

Ž .Materials Letters 41 1999 9–15www.elsevier.comrlocatermatlet

HREM studies on the microstructure of severely cold-rolled TiNialloy after reverse martensitic transformation

Y.F. Zheng a,b,), J.X. Zhang a, L.C. Zhao a, H.Q. Ye b

a School of Materials Science and Engineering, Harbin Institute of Technology, P.O. Box 433, Harbin 150001, Chinab Laboratory of Atomic Imaging of Solids, Institute of Metal Research, Chinese Academy of Sciences, Shenyang 110015, China

Received 2 December 1998; accepted 10 April 1999

Abstract

The microstructure of Ti — 49.8 at.% Ni alloy, which was cold rolled to about 30% reduction in thickness in itsmartensitic condition and subsequently heated up to 2008C for half an hour, has been studied by high resolution electron

w x w xmicroscopy. The newly-formed parent phase has an orientation relationship with the residual martensite: 111 ´ 110 ,P MŽ . Ž .101 6.58 away from 001 . The parent phase-martensite interface is not smooth and well coherent. The boundaryP M

between two subgrains of the parent phase is not straight and perfectly coherent, with partial dislocation observed at the� 4 � 4interface. Inside some parent phase grains, thin plate-like 114 and spear-like 112 twin-related parent phase variant pairs

� 4are observed. The 114 twinning boundary is relatively straight, but with two or three atomic-height blurred layers existingnear the interface. q 1999 Elsevier Science B.V. All rights reserved.

Keywords: Titanium nickel alloy; Reverse martensitic transformation; Microstructure; High resolution electron microscopy; Cold rolling

1. Introduction

Among many shape memory alloys, TiNi alloysare most frequently used for practical applications ofboth shape memory effect and pseudoelasticity. Re-cently, the effect of cold rolling on the martensitictransformation and microstructure of near equiatomic

w xTiNi alloy have been intensively studied 1–3through TEM, IF, DSC and XRD. Tadaki and Way-

w xman 1 observed the aligned lamellar structure with

) Corresponding author. School of Materials Science and Engi-neering, Harbin Institute of Technology, P.O. Box 433, Harbin150001, China. Tel.: q86-451-6412163; fax: q86-451-6413922;E-mail: [email protected]

fine segments along different directions inside eachlamellae in the martensite in TiNi alloy cold rolled

w xup to about 30% reduction in thickness. Lin et al. 2described the microstructure as mainly variant-crossed and cross-hatched morphologies in the 40%thickness reduced TiNi alloy specimen. The phe-nomenon of mechanical-induced martensite stabiliza-

w xtion was observed in the cold rolling martensite 2 .Both deformed martensite structures and deformed-induced defects are considered pertinent to the

w x w xmartensite stabilization 2 . Many studies 4,5 havebeen concerned with the influence of annealing treat-ment right after the cold rolling on the shape mem-ory effect and pseudoelasticity characteristics of TiNialloy. Except for the experimental result that therecrystallization would occur at temperatures G

00167-577Xr99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved.Ž .PII: S0167-577X 99 00095-6

Page 2: HREM studies on the microstructure of severely cold-rolled

( )Y.F. Zheng et al.rMaterials Letters 41 1999 9–1510

6008C, however, there has been no systematic inves-tigation of the microstructure of cold-worked marten-sitic TiNi alloys after annealing treatment at different

temperatures, especially that subjected to a reversemartensitic transformation at temperature below3008C.

Fig. 1. The microstructural condition that the parent phase and the retained martensite coexisted in the 30% thickness reduced specimenŽ . Ž . Ž .after heated to 2008C; a . HREM image; b, c corresponding EDPs achieved by FFT method from the areas marked by P and M in a ,

w x w xrespectively. Electron beam´ 111 ´ 110 .P M

Page 3: HREM studies on the microstructure of severely cold-rolled

( )Y.F. Zheng et al.rMaterials Letters 41 1999 9–15 11

In the present study, the severely cold rolledspecimens are selected to investigate the microstruc-

tural aspects inside it after the first reverse marten-sitic transformation of B19X

™B2. The size of mar

Ž . Ž . Ž .Fig. 2. The boundary between two subgrains in the wholly austenised area. a HREM image; b corresponding EDPs to a , Electronw xbeam´ 111 .P

Page 4: HREM studies on the microstructure of severely cold-rolled

( )Y.F. Zheng et al.rMaterials Letters 41 1999 9–1512

tensite colony with a single orientation in violentlydeformed martensite had been found to be on a

w xnanometer scale 3,6 , which made the conventionalTEM analysis complicated and difficult. Much highermagnification is desirable to differentiate one tinycertain oriented microstructure from another. There-fore, high resolution electron microscope is em-ployed by the present authors to facilitate the identi-fication of the derivation of diffraction spots, at thesame time to obtain the detailed interfacial structureinformation.

2. Experimental procedure

A Ti — 49.8 at.% Ni alloy was prepared byconsumable arc-melting under an Ar atmosphere in awater-cooled copper crucible. The electrode was acompact of 99.92 wt.% sponge Ti and 99.95 wt.%electrolytic Ni plate. The ingot was hot swaged androlled at 8508C to strips of approximately 5 mmthickness. After annealing the samples at 8508C for1 h followed by cooling with the furnace, the trans-formation temperatures Mf, Ms, As and Af weremeasured by electrical resistance vs. temperaturemeasurement to be y158C, 188C, 428C and 608C,respectively. The specimens were quenched into theliquid nitrogen to achieve full martensite structure atthe beginning, then cold-rolled at room temperatureto the extent of 30% reduction in thickness. Subse-quently, the specimen was annealed at 2008C for halfan hour. The foils for HREM observations weremechanically polished to 50 mm and ion thinned.HREM observations were performed by a JEOL-2000EX II electron microscope operated at 200 kVusing a top-entry type double-tilt specimen stagewith angular ranges of "108.

3. Results and discussion

HREM observation shows that the majority of themicrostructure in the specimen are parent phases,

with a small amount of retained martensites coex-isted. This microstructural condition could be easilyunderstood to be the result of martensite stabiliza-tion. The finishing temperature of reverse martensitictransformation for 31% thickness reduced equiatomicTiNi specimen was reported to be about 2318C for

w x Ž .the first cycle 7 . Fig. 1 a shows the high resolutionimage of the dual phases microstructure. The corre-sponding diffraction pattern derived from the areas P

Ž .and M in Fig. 1 a by FFT method are illustrated inŽ . Ž .Fig. 1 b and c , respectively. The orientation rela-

Ž .tionship o.r. between the parent phase and themartensite can be written simply as follows:

w x w x111 ´ 110P M

101 6.58 from 001Ž . Ž .P M

This o.r. established by electron diffraction pat-w xterns is close to that reported by Otsuka et al. 8

between the resulting thermal martensite and theparent phase in TiNi binary alloy. This suggests thatthe crystallographic reversibility of martensitic trans-formation had not been violently deteriorated by thesevere cold rolling deformation, which is believed tobe the origin of shape memory effect. It could benoticed that the parent phase-martensite interface is

Ž .not smooth and well-coherent in Fig. 1 a , on theupper right side of which these two phases interlaceeach other, exhibiting broad strain field contrast.

Some degree of recovery occurs in the cold-worked specimen after annealing at low temperature.

Ž .Fig. 2 a shows the HREM image of the boundarybetween two parent phase subgrains, with the corre-

Ž .sponding EDPs shown in Fig. 2 b . The boundary isnot very straight and perfectly coherent, with a rela-tively strong strain contrast observed near it. A

Ž .partial dislocation can be noticed in Fig. 2 a , asmarked by a ‘‘H ’’ symbol. The difference of theorientation between the two parent phase grains is

Ž .about 4–58, as directly measured from the Fig. 2 aŽ .or b , indicative of subgrains rather than a random,

recrystallized, fine grain structure.

Ž . Ž .Fig. 3. Twinning variants inside individual parent phase grains. a Typical morphology of parent phase variants; b bright field image of aŽ . Ž . w x Ž .single spear-like parent phase variant pair; c corresponding EDPs taken from the variant pair ArB in b , Electron beam´ 110 ; dP

Ž . w x Ž . � 4corresponding EDPs taken from the variants A, B and C in b , Electron beam´ 110 ; e HREM image of the 114 twinning boundaryPŽ .between variants A and B in b .

Page 5: HREM studies on the microstructure of severely cold-rolled

( )Y.F. Zheng et al.rMaterials Letters 41 1999 9–15 13

Page 6: HREM studies on the microstructure of severely cold-rolled

( )Y.F. Zheng et al.rMaterials Letters 41 1999 9–1514

Some parent phase variants could be found insideŽ .individual grains. Fig. 3 a shows the typical spear-

like morphology of parent phase variants. In Fig.Ž .3 b , the thin plate variants coupling A and B are

� 4found to be 114 twin related, as clearly indicatedŽ .by the corresponding EDPs shown in Fig. 3 c ,

whereas the spear-like variants pair A and C could� 4be indexed to be 112 twin related, as obviously

indicated by the corresponding EDPs shown in Fig.Ž .3 d . The variants B and C could also regarded as

� 4114 twin related variant coupling, since their con-Ž . Ž .junction planes are 114 and 114 , which areB C

Ž .from the same family in the B2 structure. Fig. 3 eillustrates the HREM image of the ArB intervariantboundary. It is relatively straight, with two or threeatomic layers height blurred regions existing near theinterface, like a midrib. The incident electron beam

w xis parallel to 110 zone axis in Fig. 3, from whichB2

direction the schematic diagram of the relative atom-Ž .istic orientations between 114 twin-related variantsw xwas given by Goo et al. 9 . As discussed by Goo,

Ž .twinning on 114 plane involves atoms undergoingw xa shear strain of 0.707 in the 221 direction. In order

to preserve the B2 structure, Goo’s model involvesŽ .half the atoms, on alternating 110 planes, shuffling

w x0.5a in the 001 direction. Clearly, two types of0Ž .atomic matching might exist at the 114 twinning

boundary. These special atomic configurations attwinning boundary might cause the appearance ofthis midrib featured interfacial structure.

� 4 � 4In the previous study, the 114 and 112 twinswere only found to be formed by deformation of theB2 parent phase without the occurrence of stressinduced martensitic transformation, and believed tobe mechanical twinnings in Ti Ni Fe and50 47 3

w xTi Ni alloys 9,10 . For the present specimen49 51

annealed above the recrystallization temperature, ourrecent study show that no twinning substructure isobserved inside the parent phase grain at room tem-

w xperature. According to Koike et al. 3 , the disloca-tion density had been estimated to be about 1–5=

1013rcm2 in the local area in the 30% reduced TiNialloy. These dislocation configurations must haveimportant influence on the microstructural aspect inthe specimen with different post annealing treatment.It can be suggested that for the present low tempera-ture annealed specimen some dislocations would dieout during the reverse martensitic transformation

while others would retain in the newly-formed parentphase. Might these residual irreversible defects dur-ing the reverse transformation from the severe cold-rolled martensite to parent phase be responsible tothe formation of the above several kinds of parentphase twinning variants. They may help accommo-date the twin formation and the associated shuffle,rather than being a hindrance, if the Burger vector ofthe slip dislocations are in the same directions as theshuffle to realize mechanical twinning. Once theseresidual defects are eliminated by the recovery andrecrystallization annealing treatment at high tempera-ture, the twinning parent phase microstructure disap-pears.

4. Conclusions

Ž .1 The newly-formed parent phase has an orien-tation relationship with the residual martensite:w x w x Ž . Ž .111 ´ 110 , 101 6.58 away from 001 . TheP M P M

parent phase-martensite interface is not smooth andwell-coherent.

Ž .2 The boundary between two subgrains of theparent phase is not straight and perfectly coherent,with partial dislocation observed at the interface.

Ž . � 43 Inside some parent phase grains, 114 and� 4112 twin-related variant pairs are observed. The� 4114 twinning boundary is relatively straight, withtwo or three atomic-height blurred layers existingnear the interface. The residual irreversible defectsduring the reverse transformation from the severecold-rolled martensite to parent phase might be re-sponsible to the formation of these twinning variants.

Acknowledgements

Project 59471029 supported by the National Natu-ral Science Foundation of China.

References

w x Ž .1 T. Tadaki, C.M. Wayman, Scripta Metall. 14 1980 911.w x2 H.C. Lin, S.K. Wu, T.S. Chou, H.P. Kao, Acta Metall.

Ž .Mater. 39 1991 2069.

Page 7: HREM studies on the microstructure of severely cold-rolled

( )Y.F. Zheng et al.rMaterials Letters 41 1999 9–15 15

w x Ž .3 J. Koike, D.M. Parkin, M. Nastasi, J. Mater. Res. 5 19901414.

w x Ž .4 Y. Liu, P.G. McCormick, Acta Metall. Mater. 38 19901321.

w x5 S. Miyazaki, T. Imai, Y. Igo, K. Otsuka, Metall. Trans. A 17Ž .1986 115.

w x6 J.C. Ewert, I. Bohm, R. Peter, F. Haider, Acta Mater. 45Ž .1997 2197.

w x Ž .7 H.C. Lin, S.K. Wu, Acta Metall. Mater. 42 1994 1623.w x8 K. Otsuka, T. Sawamura, K. Shimizu, Phys. Status Solidi A

Ž .5 1971 457.w x9 E. Goo, T. Duerig, K. Melton, R. Sinclair, Acta Metall. 33

Ž .1985 1725.w x10 W.J. Moberly, J.L. Proft, T.W. Duerig, R. Sinclair, Acta

Ž .Metall. Mater. 38 1990 2601.