当前位置：文档库 › Stress-induced phase transformation and detwinning in NiTi polycrystalline shape memory alloy tubes-

Stress-induced phase transformation and detwinning in NiTi polycrystalline shape memory alloy tubes-

Stress-induced phase transformation and detwinning in NiTi polycrystalline shape memory alloy tubes

K.L.Ng,Q.P.Sun

Department of Mechanical Engineering,The Hong Kong University of Science and Technology,Clear Water Bay,

Kowloon,Hong Kong SAR,China

Abstract

Deformation behavior associated with initial austenite (A),rhombohedral (R)and martensite (M)phase structures was studied in polycrystalline NiTi shape memory alloy tubes by tensile testing at di?erent temperatures.The nominal stress–strain curves of the tubes from room temperature (23°C)to 70°C were recorded.The deformation of NiTi tubes with initial structure of R-phase proceeded via R !M martensitic transformation,while the deformation of NiTi tubes with initial structure of M-phase proceeded via martensitic detwinning.It was found that the R !M martensitic type transformation was realized,at the macroscopic level,by nucleation and growth of an inclined cylindrical band,while the detwinning process of the tube was macroscopically homogeneous.Further,two-stage yielding,which is associated with austenite to rhombohedral (A !R)and R !M phase transformations,was observed in the stress–strain curves of NiTi tubes in a certain testing temperature range.With a further increase in temperature,the shape of the nucleated band remained cylindrical until 60°C (>A f )when the shape of the initial band suddenly became helical which was well observed in the superelastic microtubing.ó2005Elsevier Ltd.All rights reserved.

Keywords:NiTi polycrystalline;Shape memory alloy;Martensitic transformation;Detwinning;Martensitic band;R-phase

1.Introduction

NiTi polycrystalline shape memory alloy (SMA)microtubes are increasingly used in medical surgery and human implants (see Pelton and Duerig,2003).

The applications make use of the NiTi SMAs ?bio-compatibility,large recoverable deformation,good fatigue life and outstanding superelastic and shape memory properties around body temperature or over other temperature ranges.They have been successfully used to manufacture medical devices in recent years.One of the most interesting applica-tions for NiTi tubing is in extremely ?ne instru-ments.For instance,the 1mm diameter grasper is

*Corresponding author.Tel.:+852********;fax:+852********.

E-mail address:meqpsun@ust.hk (Q.P.

Sun).

Mechanics of Materials 38(2006)

41–56

composed of a very thin-walled NiTi tube with a NiTi wire inside(Duerig et al.,1997).This combi-nation not only enables it to bent around radii of less than3cm without kinking,but also allows opening and closing of the distal grasper jaws with-out increasing resistance.Stainless steel or other metallic instruments would kink and be destroyed by even very slight mishandling,while this NiTi de-vice will continue to operate smoothly even after being bent around tortuous paths.Systematic investigations are crucial for understanding and further modeling the thermo-mechanical behavior of the material in the device design.In the past dec-ade some fundamental mechanics studies on bulk NiTi wires and strips have been performed and re-ported(for example,see Leo et al.,1993;Shaw and Kyriakides,1995,1997,1998;Tse and Sun,2000). Due to the di?culties in testing the long and thin microtubes and the cost of the testing facility,only limited experimental research has been conducted by several groups(Li and Sun,2000,2002;Sun and Li,2002;Berg,1997;Helm and Haupt,2001; McNaney et al.,2003).Even for tubes the experi-mental investigations were so far limited to super-elastic NiTi tubes which experienced reversible phase transformation between austenite and mar-tensite(A!M and M!A)under loading.Due to the special geometric shape of the tubes,signi?-cant di?erences in the transformation band mor-phology between the tube and those in the wire and strip can be expected.For superelastic NiTi tubing,a helical shaped martensitic transformation band was observed during phase transformation. The discovery brought up several important issues in the fundamental understanding and modeling of the phase transition in polycrystalline NiTi and therefore it is worth further investigation.One of the issues is about the competition between the interfacial energy of the A/M transformation front and the strain energy of the tube in determining the ?nal morphology of the macroscopic martensite band during the deformation process.Such macro-scopic A/M interfacial energy manifested itself in the deformation process and may play a very important role in the morphology evolution of the band.This is di?erent from that in a strip where a straight inclined band forms through the whole cross-section.The observation of tube deformation which involves other deformation processes such as phase transformation associated with the rhombo-hedral(R)to martensite phase(R!M)and mar-tensitic detwinning(i.e.,M!M)process is still not available in the literature.

The objective of this paper is to examine and report the deformation behavior of the polycrys-talline NiTi tubes during A!M,R!M,A! R!M and M!M processes under uniaxial ten-sion.It focuses on the stress–strain response of the material and the corresponding surface morphol-ogy during the above processes.The main purpose of this preliminary study is,through investigation, to provide a quantitative base in developing a con-stitutive model for this type of material.

2.Experimental procedures

2.1.Material and sample preparation

The material used(from Shape Memory Appli-cation,Inc.)was a commercial binary polycrystal-line NiTi alloy and was received in tube form with a dark oxidized surface layer.The nominal compo-sition is54.1at.%Ni.The grain size is about50–100nm by TEM as shown in Fig.1(see Ng, 2002).The original inner and outer diameters of the tubes were1.1mm and2.15mm respectively. The tubes were cut into pieces of110mm long for the test.Two types of samples were prepared. The original dark oxidized surface layer which was then coated with ethanol and rosin was kept on type I sample.A brittle and transparent layer (about20l m in thickness)of rosin would be formed after the volatilization of the ethanol. Microcracks would be formed due to the signi?-cant increase of strain during martensitic transfor-mation.The color of the rosin layer would change from transparent to white due to the change of re?ectivity of the microcracked coating layer. Thus,the transformed and non-transformed re-gions could be observed very easily.Type II sam-ple was chemically etched into‘‘dog-bone’’shape and mechanically polished by?ne grained sand papers(?nished with50nm aluminum oxide sand-paper)after etching.The?nal inner and outer diameters(the etched section)of Type II sample

42K.L.Ng,Q.P.Sun/Mechanics of Materials38(2006)41–56

were 1.1mm and 1.87mm (with tolerance of 0.01mm)respectively (Fig.2).

X-ray ?uorescence spectrometer (XRF)and Di?erential Scanning Calorimeter (DSC)were used to measure the composition and the phase transformation temperatures of the NiTi tube respectively (see Table 1).It is noticed that the ini-tial phase structure of the sample at room temper-ature depends on the heating/cooling history of

the sample as shown in Fig.3.Therefore,by cool-ing down from a higher temperature and heating up from a lower temperature,specimens with ini-tial stress-free room temperature R-phase and room temperature M-phase were prepared respec-tively for testing.2.2.Experimental set-up

Uniaxial tensile tests were performed on both the Universal Test Machine (UTM)and a small loading frame from room temperature to 70°C through two specially designed water chambers (Fig.4)in order to study the mechanical behavior of these NiTi tubes at di?erent temperatures.On the UTM machine the load and displacement were measured by a load cell and cross-head movement respectively.In the small loading frame,the load was measured by the load cell and displacement was controlled by a stepping motor through the gear system.Both were under displacement con-trolled loading condition.All tests were performed at a relatively low loading rate of 0.2mm/min (nominal strain rate of 5.6·10à5/s)in order to minimize the self-heating e?ect caused by the transformation latent heat (Leo et al.,1993;Shaw and Kyriakides,1995).In addition,higher loading rates of 0.5mm/min (nominal strain rate of 1.4·10à4/s),1.0mm/min (nominal strain rate of 2.8·10à4/s)and 2mm/min (nominal strain rate of 5.6·10à4/s)were also used to examine the loading rate e?ect on the test

results.

Fig.1.TEM image of the grains of the polycrystalline NiTi tube

(115kX).

Fig.2.Dimensions of ‘‘dog-bone’’shaped type II specimen.

K.L.Ng,Q.P.Sun /Mechanics of Materials 38(2006)41–5643

In the test by the UTM machine,the specimens were clamped at two ends by two specially de-signed screw-clamping blocks which were con-nected to the loading grips of the test machine through two double-hinge connectors(Fig.5(a) used in room temperature test only).This clamp-ing system could transfer the concentration force from the grip into a uniformly distributed shear stress at the end of the tube,therefore producing a uniform tensile stress in the measurement section of the tube.The clamping length of the tube was 20mm at each end,and the overall length of the specimen was110mm with an etched length of 60mm.In the test by the small loading frame, the specimen clamping method is similar to that in the UTM machine and is shown in Fig.5(b).

Surface morphology observations of the tube surface during loading were conducted by a CCD camera with well adjusted lighting(Fig.6).

3.Results and discussion

We?rst give a general description of the main deformation features of the tube in Section3.1,special aspects such as the band nucleation and temperature e?ect are given and discussed in Sections3.2and3.3,respectively.

3.1.Main deformation features of NiTi tube with initial R and M phases

Fig.7shows the typical stress–strain curve of the‘‘dog-bone’’shaped specimen(Type II)with initial state of R-phase at room temperature under a loading rate of0.2mm/min.The stress–strain curve could be divided into three regions.Region I represents the macroscopic homogeneous elastic deformation of the R-phase.Once the stress reached the critical value(about98.05MPa),the deformation became inhomogeneous through the formation of a macroscopic deformation(or trans-formation)band at the middle portion of the tube, which was accompanied by a very small amount of stress decrease(from98.05MPa to97MPa).The band grew under further loading while the stress was maintained at almost a constant value until the band grew all over the sample and the whole material within the etched section was fully trans-formed into the martensitic phase(see discussion

Table1

Composition(at.%)and transformation temperatures of the NiTi tube

Ni Ti R s R f M s M f A s A f 54.1%45.9%52.1°C28.2°C 5.5°Cà32.7°C23.3°C57.1°

in a recent paper by Brinson et al.,2004).The onset of the stress drop has been generally recog-nized as a signal for band formation,and the stress plateau (Region II)demonstrates the growth of the nucleated macroscopic band via visible propaga-tion of the band front (R–M front).After all the material was fully transformed,the stress increased monotonically corresponding to the elastic defor-mation and further detwinning of M-phase (Re-gion III).The subsequent unloading response was almost linear elastic and the remaining trans-formation strain was about 4.2%which could be fully recovered by heating the specimen to A f (57.1°C).

Fig.8shows the typical stress–strain curve of the ‘‘dog-bone’’shaped specimen (Type II)with initial state of twinned M-phase at room tempera-ture under a loading rate of 0.2mm/min.This stress–strain curve could be roughly divided into two regions.When compared with the regions

Fig.4.Water chamber used in (a)UTM,(b)small loading frame.

K.L.Ng,Q.P.Sun /Mechanics of Materials 38(2006)41–5645

and II in R !M transition in Fig.7there is much less distinction in the detwinning process in Fig.8.Region I in Fig.8demonstrates the elastic defor-mation of twinned martensite.With increasing stress (above 60MPa),the stress–strain curve gradually bent and its slope gradually changed into that of region II where the martensitic detwin-ning process dominated the deformation.Further loading led to further detwinning and elastic defor-mation of the twinned martensite (Liu,2002),this whole process occurs gradually instead of in one single step.Another contrast to the deformation of R !M process is that no macroscopic defor-mation band was found in the tube.Macroscopi-cally,the stress–strain curve with considerable strain hardening shown in Fig.8gives a homoge-neous deformation over the whole etched section.The response during unloading is almost linear elastic and the residual strain due to detwinning is about 3.4%which could be fully recovered by heating the specimen up to A f (57.1°C).

The e?ect of the loading rate on the stress–strain curves and deformation of tubes are shown in Figs.9and 10.For the specimen which was ini-tially in the R-phase,a ?at stress plateau was ob-served at the loading rate of 0.2mm/min.This

Fig.5.(a)Screw-clamping block with two-pin connector used in UTM.(b)Screw-clamping blocks used in small loading frame.

46K.L.Ng,Q.P.Sun /Mechanics of Materials 38(2006)41–56

because the required stress for the almost quasi-static and isothermal movement of the R/M inter-faces was almost constant during R !M trans-formation.However,such stress plateau did

not

Fig.6.A CCD camera system used for capturing the surface morphology of the tube under tensile test.

K.L.Ng,Q.P.Sun /Mechanics of Materials 38(2006)41–5647

exist in the cases of higher loading rates and the stress–strain curves maintained a positive slope during transformation.This means the required stress for the growth of band via interface migration in these cases kept increasing due to the latent heating e?ect.Surface morphology observation revealed that the higher the loading rate,the more bands formed sequentially with loading(see Fig.11).On the other hand,for the specimen(type II specimen)which was in the M-phase initially,the stress–strain curves under di?erent loading rates almost coincided.This re-veals that the deformation due to martensitic de-twinning was not a?ected in this loading range. We should notice that this may not be the case if the loading rate is very high(such as under an impact load)therefore further investigation at higher loading rates is crucial for studying the loading rate e?ect.

48K.L.Ng,Q.P.Sun/Mechanics of Materials38(2006)41–56

3.2.Observation of macroscopic band formation and propagation at room temperature under a loading rate of 0.2mm/min

The evolution of the R !M martensitic bands (Fig.12)was recorded by observing the tube sur-face morphology.The starting structure of the specimen was R-phase and the specimen de-formed homogeneously (Fig.12(a))before mar-tensitic phase transformation (stress plateau).When the stress reached the critical level (peak stress)an inclined cylindrical martensitic band (Fig.12(b))appeared and a small nominal stress decrease was observed in the stress–strain curve.The band width of this initial martensitic band was about 1.2mm and the inclination angle of the interface is about 55°to the loading axis.The growth process of this martensitic band is shown in a series of photos of the tube (without surface coating)in Fig.12(f)where the initial in-clined cylindrical band gradually evolved into a cylindrical tube with the two fronts perpendicular to the loading axis.Fig.12(b)–(e)shows the pho-tos of the surface image of the tube with brittle coating where the inclined M band can be clearly seen and is in agreement with the surface mor-phology in Fig.12(f).Also the second band formed during loading is shown in Fig.12(d).The martensitic band maintained a cylindrical shape in the subsequent growth process under continued loading.During the heating process the reverse transition made the band gradually shrink back into an inclined band and ?nally

disappeared.

Fig.11.Variation of macroscopic band nucleation sites with loading rate:(a)0.2mm/min,(b)0.5mm/min,(c)1.0mm/min and (d)2.0mm/min.

K.L.Ng,Q.P.Sun /Mechanics of Materials 38(2006)41–5649

3.3.Mechanical response of NiTi tubes at di?erent temperatures

Uniaxial tensile tests were performed at 10dif-ferent temperatures from 23to 70°C using the water chamber to study the e?ect of temperature on material ?s response.The measured nominal stress–strain curves are summarized in Fig.13and are classi?ed into three groups according to the initial phase structures.Each stress–strain curve is divided into three regions (similar to Fig.7)for convenient discussion.The main re-sults of the observation are summarized as follows.

In the temperature range of M s (5.5°C)

the

Fig.12.Images of the tube surface with brittle coating,showing the martensitic band formation and growth in NiTi tube.(a)Before phase transformation,(b)formation of the ?rst inclined cylindrical martensitic band and (c)the growth of the martensitic band.(d)The growth of the ?rst and the second cylindrical martensitic bands.(e)Two cylindrical martensitic bands merged into one band.(f)Photo series of the tube surface without coating,showing band morphology evolution (focusing on one band only).

50K.L.Ng,Q.P.Sun /Mechanics of Materials 38(2006)41–56

initial phase of the specimen by cooling from a higher temperature.The stress–strain curves at 23°C and28°C(Fig.13)are in this range.Firstly, region I represents the elastic deformation of R-phase.When the stress reached the critical level, a small stress decrease was observed and this was the starting point of the stress-induced R!M transformation.A stress plateau(Region II)was formed during the growth of the martensitic bands.Two cylindrical martensitic bands were nucleated(loading rate0.2mm/min)within this temperature range.Region III was the elastic deformation and further detwinning of M-phase when the whole specimen was fully transformed from R to M phase.

In the temperature range of R f(28.2°C)< T

At the temperatures T>R s(52.1°C),the ini-tial structure of the specimen was A-phase.The stress–strain curves at56°C,61°C,65°C and 70°C(Fig.13)are within this temperature range. The specimen exhibited SME at56°C(

K.L.Ng,Q.P.Sun/Mechanics of Materials38(2006)41–5651

deviation from linearity represented the A!R phase transition and the second stage R!M phase transition started with a load drop. However,this two-stage yielding phenomenon gradually merged into one-stage A!M phase transition when the testing temperature was fur-ther increased to61°C,65°C and70°C (>A f=57.1°C)at which the specimen exhibited superelasticity(refer to the phase diagram in Fig.15).As shown in Fig.16(b),region I consists of linear elastic deformation of A-phase only(no obvious deviation from linearity).The stress-

52K.L.Ng,Q.P.Sun/Mechanics of Materials38(2006)41–56

K.L.Ng,Q.P.Sun/Mechanics of Materials38(2006)41–5653

induced A !M transition process (Region II)started with a sharp load drop signaling the forma-tion of band and followed by the single band growth at almost constant stress plateau.During unloading,the M-phase was transformed back to A-phase by the reverse movement of the band front on a lower stress plateau.Special attention was paid to the band morphology variation with temperatures.Below 60°C the nucleated macro-scopic bands remained cylindrical in shape.As the temperature was raised to above 60°C the band shape suddenly changed to helical as shown in Fig.17(a)and (b)at 60°C and 65°C respec-tively.This type of helical band was well observed in previous researches on superelastic tubing (Sun and Li,2002;Li and Sun,2002).There also seems to be a correlation between the initial band volume and the amount of stress drop.The stress drop increased almost monotonically with temperature.

A phase diagram in temperature and stress space can be drawn by plotting the transformation stress versus temperature as shown in Fig.15.The required stresses are represented by solid squares for the A !R transition,by solid circles for R !M and by open circles for the A !M transi-

tions.They all increased with temperature and generally satisfy the Clausius–Clapeyron relation-ships for each type of phase transition (except when there was a small deviation at low tempera-tures,i.e.,T =34°C and 40°C in this case).Moreover,the slope of the Clausius–Clapeyron relationship was steeper for the A !R transition than for the R !M and A !M transitions,which are responsible for the observed two-stage and one-stage yielding (transition)phenomena.

4.Conclusions

To investigate the stress-induced phase trans-formation and detwinning processes in polycrys-talline NiTi tubes under tension,displacement controlled uniaxial tensile tests on the tubes with di?erent initial phase structures were performed.Depending on the pre-testing heating/cooling his-tory and the test temperature,di?erent initial phases of the material were obtained and four di?erent types of deformation process (A !M,R !M,A !R !M and M !M)were realized by the loading.The key experimental ?ndings of this preliminary research are listed as

follows:

Fig.17.Helical band formation in the tube at high temperature (a)60°C and (b)65°C.

54K.L.Ng,Q.P.Sun /Mechanics of Materials 38(2006)41–56