Cai Msea 2009

Materials Science and Engineering A 501 (2009) 166181

Contents lists available at ScienceDirect

Materials Science and Engineering A

journa l homepage: www.e lsev ier .co

Evoluti ssetemper

S. Caia, M . Ola Dept. of Mech ton, Cb NRC, Canadiac ISIS, Rutherfo

a r t i c l

Article history:Received 18 AuReceived in re23 SeptemberAccepted 8 Oc

Keywords:InterphaseIntergranular strainNeutron diffractionAsymmetric yieldingStrength differentialZr2.5Nb

testsong aranuln diffpropeionading oduri

tension and compression on texture. The combination of the thermal residual stress and the asymmetricCRSS in the c axis gives the {0002} grain family a higher strength in compression than in tension.

2008 Elsevier B.V. All rights reserved.

1. Introdu

Due toand strongproperties,material fopropertieshave beenlished to sithat the irrdependentual stresseschedule. Tconsists ofstrain and aof 90% hcbcc -Zr, wing, -Zr pmetastableture of prestage and

CorrespoE-mail ad

0921-5093/$doi:10.1016/j.ction

its good mechanical strength, high corrosion resistancecreep resistance, combined with excellent neutronicZr2.5wt.% Nb has been selected as the pressure tuber CANDU reactors since the 1970s [1]. Its in-reactorsuch as hydride cracking, irradiation growth and creepof interest for decades and models have been estab-mulate the irradiation behavior [26]. It has been foundadiation behavior of Zr2.5Nb pressure tubes is stronglyon the texture, microstructure and intergranular resid-s [710], which are controlled by the manufacturinghe general fabrication route of CANDU pressure tubesextrusion at 800 C, air cooling, cold draw to 2030%utoclaving at 400 C for 24h. After extrusion, it consists

p -Zr, which has Nb content

S. Cai et al. / Materials Science and Engineering A 501 (2009) 166181 167

twinning [19,20]; twinning occurs preferentially in coarse grainsand produces a sharp texture change during deformation, whileprismatic, basal and pyramidal c+ a slip are responsible for thetexture evolution in ne grain structures. Cheadle et al. [21] testedZr2.5Nb in tension along a specimen direction with a high con-centration of basal plane normals, but did not observe any texturechange associated with twinning for plastic strains less than 10%.Kims study [22] of an annealed Zr2.5Nb pressure tube material,which has an equiaxed grain structure with size of 10m, showsthat deformation twinning occurred only after strains greater than5%. Twinning is thus not likely to be operating in the early stage ofdeformation where strain is less than 5% [23]. Despite the lack oftwinning, Zr2.5Nb pressure tube material yields anisotropically,and attempts to model the anisotropic mechanical behavior by apower law constitutive model or a polycrystalline model [23,24]have not been very successful.

Part of the reason for the lack progress in modeling the behav-ior of this material is that much of the previous work focusedon the irradiation behavior such as creep and growth, while theshort term deformation mechanisms of Zr2.5Nb such as therelative contribution of each phase, the interactions between dif-ferent deformation systems and the evolution of interphase andintergranular stressstrain during deformation has not been thor-oughly studied. Even the studies of irradiation induceddeformationwere restricted to pressure tube material with a limited range ofmicrostructures and texture, where the basal plane normals aredistributed in the radial and transverse plane and, mostly, concen-trated in the transverse direction [25]. The models developed wereaccurate ovand could nfrom that othe understture to irraddeformatioa material tnature hasto date havvolume fraccrystalline [error within

Fig. 1. Microstructure of hot rolled Zr2.5Nb showing elongated -phase grains(light) surrounded by -phase (dark).

the experimental data. However, recent studies [27] have suggestedthat the overall properties of Zr2.5Nbmay be highly dependent onthe properties and the distribution (both geometric and texture) ofthe -phase. Neglecting the -phase is thus likely to introduce sig-nicant errors in modeling the deformation response of the overallmaterial an

xturdelstructpaphot

nsioal plitorhe loollecmecn the

Fig. 2. Texture -phase, the lower part shows the selected pole gures of the -phase. TheND is in the ceer only a narrow range of textures and microstructureot be accurately extrapolated to textures quite differentf these pressure tubes. This has signicantly restrictedanding of the relationship of texture and microstruc-iation growth and creep. To the authors knowledge, a

n model which can be used and applied directly to suchaking into account the anisotropy and the dual phasenot yet been constructed. Indeed, most of the studiese ignored the -phase because of its relatively smalltion and have treated thematerial as single phase poly-3,4,23,26] with the aim of accommodating any relatedthe t of the parameters of the single phase models to

new tethe momicros

Thisior of aBoth teprincipto monalong tbeen cmationadditio

of hot rolled Zr2.5Nb plate. The upper part shows the selected pole gures of thenter of the pole gure, the RD is vertical and TD is horizontal.d in particular in extrapolating model predictions toes and microstructures. This is one likely reason whyin [3] and [4] are accurate only for certain texture andure combinations.er presents our recent study on the deformation behav-rolled Zr2.5Nb plate material at room temperature.

n and compression tests were carried out along threeate directions, while neutron diffraction was employedthe evolution of interphase and intergranular strainsading and two Poisson directions. A large data set hasted to help us to understand the fundamental defor-hanism such as slip and twinning of this material. Indata basewill be used to evaluate and improve current

168 S. Cai et al. / Materials Science and Engineering A 501 (2009) 166181

modeling approaches. We present here the experimental resultsof the mechanical behavior in both macroscopic and microscopicscales from which the texture dependence and the inuence of the-phase can be shown.

2. Materials and experimental method

The composition of the experimental material is 2.5wt.% Nb,1200ppm O, 950ppm Fe, 110ppm C and balance Zr. Ingotswere forged to 112mm thick at 1065 C and then hot rolled to56mm thick at about 700 C followed by a furnace cool. Themicrostructure at room temperature is shown in Fig. 1. The pre-dominant -grains have a plate-like shape with length and width30m and thickness 3m, while the -phase is distributedmore or less continuously between the -grains. The O is expectedto be segregated to the -phase while the Fe and Nb are expectedto be mostly concentrated in the -phase [28].

Texture measurement was carried on High-Pressure-PreferredOrientation (HIPPO) at Lujan Jr. Neutron Scattering Center, LosAlamos National Laboratory and calculated based on the gener-alized spherical-harmonics description of the effect of texture onreection intensities in aRietveld renementprogram inGSAS [29].Fig. 2 shows the texture of thismaterial at room temperature. In the-phase thethe 10 1 0of basal pla(fR, fT and fthe -phaseTexture waResearch Uratory, on aalthough afacilities, theach pole Rietveld rethe nomina

Dog-bon6mm6and cylindrfrom the rto the rolliplate normdifferent sastate withintests were c

and lattice strain evolution in the axial and transverse directionswas measured intermittently during testing by neutron diffraction.Samples are dened based on the loading direction and the diffrac-tion collecting direction, e.g. a NDRD sample means that theloading direction is the plate normal direction and the diffractionvector is along the rolling direction.

Tensile testswere carried out on the L3 beam line at theNationalResearchUniversal reactorat theAECLChalkRiver Laboratory (CRL).The (331) diffraction peak from a germanium single crystal wasused as the incident beam. Thewavelengthwas2.046. Thebeamsizewas dened to select the center area of the sample gauge. Sam-ples were mounted in an Applied Test System screw-driven loadframe equipped with a load cell and hydraulic grips. A series ofincreasing uniaxial tensile loadswere applied along the axial direc-tion and the deformation strain was measured by an extensometermounted on the sample. At each selected stress level, the detectorand the sample table were rotated together such that the diffrac-tion peaks from plane families of interest could be collected in thedesired specimen direction. By rotating the load frame out of thescattering plane, both axial (parallel to the sample axis) and trans-verse (normal to the sampleaxis) lattice strainsweremeasured. The(10 1 0), (1 0 1 1), (1 0 1 2), (1 0 2 0) and (0002) diffraction peaks ofthe-phase and (110), (2 00) and (211) diffraction peaks of the-

eret in sn conelaxaed to

ate acom

lsedaxi

r baneam,ctionoreeasinirectredm hsca

contandstrai.

rectioc axis is mostly orientated towards TD and ND, whilea axis is concentrated in RD. The resolved fractionsne normal in rolling, transverse and normal directionN) are 0.27, 0.39 and 0.34 [7]. The 100 direction ofis slightly aligned along the three principle directions.

s also measured at the E3 spectrometer at the Nationalniversal (NRU) reactor at the AECL Chalk River Labo-dditional samples cut from the plate. It is noted thatgenerally consistent texture are obtained at these twoere is a small difference in the intensity maxima forgure. The volume fraction of the-phase obtained fromnement is 12%, consistent with that expected froml Nb content.e tensile samples (with cross-section areamm for the RD and TD, 15mm4mm for the ND)ical compressive samples (9mm20mm) were cutolled plate with the longitudinal direction parallelng direction (RD), transverse direction (TD) and theal direction (ND). FE analysis demonstrated that themple geometries had the same (i.e. uniaxial) stressthe measurement volume. Tension and compression

arried out along these three principle plate directions

phase wried outo straiature rdeformelimin

TheISIS puloadingdetectodent bin dire[30]. Mof incraxial dmonitowas 8mvided astrainstrainsStressperiod

Fig. 3. Macroscopic mechanical behavior in the three principal plate dicollected. In the elastic region, the experiment was car-tress control, while in the plastic region itwas switchedtrol. To handle the stress drop caused by room temper-tion observed in the plastic regime, samples were rstthe desired strain and then backed off 0.025% strain to

ny stress relaxation during neutron data acquisition.pression tests were carried out on ENGIN-X at theneutron facility, Rutherford Appleton Laboratory. Thes is horizontal and at 45 to the incident beam. Twoksare setuphorizontally andat angles90 to the inci-allowing simultaneous measurement of lattice strainss both parallel and perpendicular to the applied loaddetails of the instrument can be found in [31]. A seriesg uniaxial compressive loads were applied along theion to produce a nal true strain of 10%. Strain wason the samples using a clip gauge. The incident beamigh and 4mm wide, the radial collimators in use pro-

ttered beam aperture of 4mm. Experiment was underrol, each time the sample was deformed to desiredheld for 30 s to allow relaxation before data acquisition.ns were the average values during the data acquisition

ns during (a) tension and (b) compression.


Table 1Mechanical properties in the three plate directions during tension and compression. Results are based on the average of 23 samples in each plate direction.

RD TD ND

Tension Comp. Tension Comp. Tension Comp.

E (GPa) 90.5 7 84.1 6.5 101 5.6 105.5 10.3 91 11.5 10014.60.2 (MPa) 370 8.1 345 12.9 395 7 420 7 385 7 422.510.6

The mechanical responses of individual grain families were eas-ily determined by single peak tting, while different techniqueswere used to estimate the average phase behavior based on thecharacteristics of the beam used. For tensile tests carried out at CRLwhere monochromatic beam is used, the average phase strain wascalculated using the Method B weighting equations discussed in[32].

For compression tests carried out at ISIS where time-of-ightis employed, the average phase strains are commonly determinedby a multiphase Rietveld renement for materials with a weak tomedium texture [33,34]. Due to the weak texture of this material,Rietveld renement is employed to determine the average phasestrains of the-and-phases. The advantage of using this approachto determine the average phase response is that it can obtain rea-

Fig. 4. Phase(100200)the -phase. Dresponse during tension (left) (a) RD/TD, (c) TD/RD, (e) ND/RD and compression (right106 for the -phase and (200500)106 for the -phase. For compression data, theashed lines correspond to the yield of the - and -phases.) (b) RD/TD, (d) TD/RD, (f) ND/RD. For tension data, the errors areerrors are (1050)106 for the -phase and (50200)106 for


Fig. 5. Responloading directiDashed lines c

sonably acccompared tgrain familiphase respoof the Zr inweak, thusticable for aaverage phawhile for thby:

ave = fc +

where c anf is the resoses of individual grain orientations in the -phase of RD samples during tension (left) aon, the others show the results along the other two transverse directions. The errors are orrespond to the yield of the - and -phases.

urate results in a relatively shorter data collection timeo that needed for strain measurement of individuales. This is especially useful in determining the averagense of the Zr in Zr2.5Nb. Since the volume fractionZr2.5Nb is very small, the diffraction peaks are very

the peak average method [32] was found to be imprac-nalysis of the Zr in Zr2.5Nb. For the bcc -phase, these strain was obtained directly from the lattice strain,e hcp -phase, the average phase strain was calculated

(1 f )a (1)

d a are the lattice strains of c and a axes respectively,lved fraction of basal plane normals in the direction

of interest [cussed in [(2a + c)/3code [35]. Tand thus dstresses.

3. Experim

3.1. Macros

Macroscpression arestrengths (0nd compression (right). Fig. 5(a) and (b) shows the results along the(100200)106 for tension and (10100)106 for compression.

7], which is equivalent to the weighting approach dis-32]. For a random or weak texture, this simplies to[34]. Rietveld renementwasperformedusing theGSAShe strains reported here are relative to the initial state,o not take into account any initially existing residual

ental results

copic mechanical response

opic mechanical responses during tension and com-shown in Fig. 3. The average Youngs moduli and yield.2%) in the three plate directions are listed in Table 1.


Observations can be summarized as:

(a) All of the three samples have a similar Youngs modulus90100GPa during tension and compression.

(b) An obvious strength differential (SD) between tension andcompression can be seen in Fig. 3. The yield strength of theTD and ND samples are higher in compression than in ten-sion, while the strength in the RD is very close in tension andcompression.

(c) The texture dependence of strength is weaker in tension thanin compression, i.e. the strength differential in the three platedirections is much smaller in tension than that observed incompression. In compression, the TD and ND, which havethe larger portion of c axis, have higher yield strengths(0.2 =420MPa) than RD (350MPa), while in tension all thethree directions have a similar yield strength (370395MPa).

(d) The elasto-plastic transition is gradual and smooth during com-pression, while this transition period is much shorter andsharper during tensile testing.

3.2. Average phase response

Fig. 4 shows the evolution of average phase elastic strains duringtension and compression along the three principle plate directions,as determined by neutron diffraction. For reasons of space, eachplot only shows the responses along the axial loading direction andthe results along one of the two Poissons directions. Qualitativelysimilar behavior is obtained in the other Poissons direction. Onlythe elastic strains relative to the starting point are plotted in Fig. 4,thus showing the development of internal strains caused by defor-mation. In both tension and compression, the elastic anisotropy isvery small in all the three plate directions. The Youngs modulus of

Fig. 6. Respon ft) an-phases.ses of individual grain orientations in the -phase of TD samples during tension (le d compression (right). Dashed lines represent the yield of the - and


the -phase is smaller than that of the -phase, hence more stressis taken up by the -phase in the elastic region, which may cause itto yield at lower applied stress than it would without the -phasepresent. The yield point of the -phase is represented by the rstinection point in the stressstrain curve and is indicated by thedashed lines in Fig. 4a. It is seen that during tension in RD the -phase yields at an applied stress 320MPa, which is close to theyield strength of the RD sample. Once the-phase yields, its elasticstressstrain curve shifts towards compression relative to its lin-ear elastic response, while the elastic strain of the -phase shiftsto the opposite direction, indicating that with increasing appliedstress, more and more stress is taken by the -phase, i.e. thereis a load transfer from the -phase to the -phase. The -phasenally yields at an applied stress of about 400MPa showing a sec-

ond inection in the stressstrain curve. More stress is then sharedby the-phasewhose lattice strain shifts back towards tension. Thetransverse data show a qualitatively similar trend. Similar load par-titioning between the - and -phases during tension is observedin other samples (Fig. 4c and e) and during compression (Fig. 4b, dand f).

3.3. Response of individual grain orientations

3.3.1. Evolution of intergranular strains in the -phaseThe evolution of intergranular strains in the-phase during ten-

sion and compression is plotted in Figs. 57. The lines all start fromzero strain, thus thermal residual strains are excluded and guresonly show the evolution of lattice strains caused by deformation,

Fig. 7. Respon ft) an-phases.ses of individual grain orientations in the -phase of ND samples during tension (le d compression (right). Dashed lines represent the yield of the - and


relative to the starting state. Only selected stressstrain curves forindividual grain families are plotted due to space limitations. Theobservations can be summarized as:

(a) In the elastic regime, there is a small level of elastic anisotropyduring tension and compression along all the loading andtransverse directions, however, an obvious plastic anisotropyis observed once plastic deformation occurs.

(b) During tension, the -phase shows a small plastic anisotropyin the axial direction (Figs. 5a, 6a, 7a) but a larger plasticanisotropy in the Poisson directions (Figs. 5c, 5e, 6c, 6e, 7c, 7e).However during compression the-phase exhibits a large plas-tic anisotropy in the loading direction (Figs. 5b, 6b, 7b) but

a relatively small plastic anisotropy in the Poisson directions(Figs. 5d, 5f, 6d, 6f, 7d, 7f). This feature is perhaps more clearlyseen in Fig. 8, where the lattice strain is plotted instead againstthemacroscopic true strain to highlight the strain anisotropy inthe plastic region. Similar behavior has been observed in singlephase Zircaloy-2 [3638].

(c) During compression along the axial direction, the {1010}grain family yields rst, dened as the rst inection of itsstressstrain curve, at a lattice strain about 3000106 and itsstressstrain curve shifts backwards relative to its elastic line.Load is transferred to other orientations whose lattice strainsthus increase by larger steps for a given applied stress incre-ment; the 0002 is the plastically hardest orientation and thus

Fig. 8. Lattice (righTD; (c) tests alstrain plotted against macroscopic true strain during tension (left) and compressionong ND. Diffraction data from only one transverse direction is shown.t) for the -phase diffraction peaks. (a) Tests along RD; (b) tests along


yields last at a lattice strain of about 8000106. In betweenthe yield of the {1010} and the {0002} grain families, the{10 1 1}, {10 1 2} and {10 1 3} grain families yield as the appliedload progressively increases. For grain families such as {10 1 1},{10 1 2}, {10 1 3} and {0002}, the designation of yield is basedon the observation of strains parallel to the applied load curvingback towards their elastic lines.

(d) This sequence in yielding is not really observable during tensiontests (e.g. Fig. 6a). All the selected grain orientations yield at asimilar lattice strain level ((30004000)106). The {10 1 3}grain family appears to actually be the last one to yield.

3.3.2. Evolution of intergranular strains in the -phaseThe evolution of lattice strains of the -grains along the loading

direction is plotted in Fig. 9. Generally, the tension and compression

directions show a similar mechanical behavior, i.e. a small elasticanisotropy in the elastic region and an obvious plastic anisotropyin the plastic regime. In the elastic region, the {200} has a smallerelastic modulus while {110} and {211} have the same value aseach other. It is also shown that the elastic anisotropy inND is largerthan that in the other two plate directions. In the plastic region,the strains of the -grains are strongly affected by the responseof the -phase, i.e. the lattice strains of all the selected orienta-tions of the -phase increase at larger steps once {10 1 0} grainfamily of the-phase yields. These grains thus take more and morestresses from the -phase as the strain increases, and nally yieldat a stress close to the yield point of the {0002} grain family ofthe -phase. Among the selected grain orientations, {200} is thelast one to yield. Compared with the -phase, the deformationof the -phase is more symmetric. All the selected grain orien-

Fig. 9. Respon ons d(200500) e yieldses of individual grain orientations of the -phase along three plate directi106 for tension and (50300)106 for compression. Dashed lines represent thuring tension (left) and compression (right). The uncertainty isof the - and -phases.


Fig. 10. Poisso

tations yielcompressio10,0001yields at lacompressio

Similar sin the Poissintergranul{200} andthe axial in

4. Discussi

Slight ditial (pre-defThismayben strains of individual grain orientations of the -phase during tension (left) and compres

d at a similar lattice strain level during tension andn. The {110} and {211} both yield at lattice strain06 during tension and compression, while {200}ttice strain 15,000106 during both tension andn.train development of the -grain families can be seenon direction (Fig. 10). However, in the plastic region, thear Poissons strains (e.g. the strain difference between{110} grain families) are much smaller compared totergranular strains.

on

fferences were observed in the pole gures of the ini-ormation)materialmeasured atHIPPOandChalk River.due to the fact that theyweregeneratedby thedifferent

techniquesferent typesthe plate caspecied romonotectoifront part ofsome of thafter coolinintensity inHowever, ifthe back ofbelow the mmined by thmodes. Thufrom the pheach samplesion (right). Dashed lines represent the yield of the - and -phases.

that are used for texture measurement at these two dif-of facility. It is also possibly due to texture variation in

used by inhomogeneities in the hot rolling process. Thelling temperature is700 C,which is slightly above thed temperature of this material. During hot rolling, thethe platewould be in theZr andZr two phase region;

e deformed -phase then transforms to the -phaseg to room temperature, which produced the {0002}RD based on the Burgers relationship [39] (see Fig. 2).the temperature dropped quickly during hot rolling,the plate would then be rolled at temperatures justonotectoid temperature. Texture would then be deter-e combination of rolling strain and active deformations a lower {0002} intensity in RD would be expectedase transformation after cooling. Since the texture ofwas notmeasured before deformation, the exact start-


ing texture of each sample is not known. However the texture ofeach sample after deformation has been measured, and can beused, to some extent, to extrapolate the initial sample texture. Fur-ther, since this experiment and thediscussionbelow focusesmostlyon relative intensity changes for a given sample during the in situdeformation which can be determined directly from the mea-sured diffraction spectra any small texture variations betweensamples should not cause signicant trouble in interpretation ofexperimental data in terms of active deformation modes, althoughabsolute starting intensities may vary slightly from sample to sam-ple.

The similar Youngs modulus obtained in the three plate direc-tions indicate aweak texture dependence of elastic stiffness, whichcan be attributed to a combination of the weak texture and thestrong constraints both between the - and -phases and amongindividual grains.

Fig. 3 shows that the strength differential between the ten-sion and compression is largest in TD and smallest in RD, whichhave the highest and lowest basal plane normal concentrationsrespectively, indicating that this strength differential is primarilydetermined by the distribution of the c axes. As explained in[21,24], since both pyramidal c+ a slip and twinning are muchmore difcult to activate than other slip systems and the c-axisof the -Zr in this material is mostly concentrated along the TDand ND, higher strengths are expected along these two directions.However, the different texture dependence of strength during ten-sion and compression implies that other factors such as the grainshape, the thermal residual stresses and/or twinning may take partof the responsibility for the strength anisotropy and the strength

differential observed between tension and compression. Thermalresidual stresses have been shown to cause the strength differencein tension and compression in single phase Zircaloy [4042]. Twin-ning is another potential source for the strength asymmetry due tothe strength difference between tensile and compressive twinning,and/or between twinning and dislocation slip [24,37,38]. Lowdenand Hutchinson have observed a similar strength asymmetry inTi6Al4V [43] and they claimed that the strength differential is infact caused by the different hydrostatic level and the sense of shearbetween tension and compression that is experienced by disloca-tions, leading to 11 2 3 c+ a slip being easier in tension than incompression, andhence ahigher yield strength in compression. Thepossible contributions and their inuence in this material will bediscussed later.

Contrary to [1517], where in the explanation of texture devel-opment during cold rolling the -grains are treated as rigid bodiesduring deformation and the accommodation of plastic strain ismainly by the -phase, the phase strains shown in Fig. 4 givedirect evidence that in this material the -phase yields at a lowerapplied stress during room temperature deformation than doesthe -phase. From Fig. 4, it can be seen that the lattice strain atyield is about 4000106 for the -phase and about 8000106for the -phase respectively. Provided the Youngs modulus forthe - and -phases are 100GPa and 60GPa [44], the phaseyield strengths are estimated as 400MPa and 480MPa respec-tively. Thus, despite the fact that the {0002} grain family in the-phase is the strongest among all the grain families in this dualphase system, the average -phase is actually stronger than the -phase. This brings the explanation proposed in [1517] regarding

d (c)Fig. 11. Intensity change during tension of RD sample, (a) loading direction, (b) an Poissons directions. Errors are in the range of 316%.


the effect of the -phase on the texture development in Zr2.5Nbinto question. The high strength of the -phase could be due toits very ne grain size, and/or to preferential segregation of alloy-ing elements. It is also in contrast to two phase titanium alloyssuch as Ti6Al4V, where the -phase is usually considered to besofter.

Given that the single crystal elastic constants for 10 1 0 and0002 crystal orientations are 100GPa and 125GPa, the strainslisted inSection3.3.1 (c) roughly correspond to tensile yield stresses350MPa and 440MPa and compressive yield stresses300MPaand1000MPa for {10 1 0} and {0002} grain families respectively,showing a signicant asymmetrical yielding along the 0002orientation under tension and compression. This observation pro-vides the explanation to the macroscopic behaviors described inSection 3.1. Since during compression, the {0002} grain fam-ily can be seen to be plastically harder than other orientations(Figs. 5b, 6b, 7b), the yield stress of the material is mainly deter-mined by the distribution of the c axis and thus shows a strongtexture dependence. However, during tension all the grain orien-tations yield at a similar stress level (Figs. 5a, 6a, 7a), thus thematerial strength is less affected by the grain orientations, implyinga small texturedependenceof strength.All three sampleshavecloseyield strengths in tension and a similar short elasto-plastic transi-tion period (Fig. 3a). The asymmetric yielding also causes differentbehaviors in the Poisson directions. Since the Poisson directions ofa tensile loaded sample are under compression, the {0002} grainfamily in the Poisson directions behaves much more strongly thanother directions and causes a relative larger plastic anisotropy. Sim-ilarly, along the Poisson directions of a compression sample, all

the grains have a similar strength and thus produce a relativelysmaller plastic anisotropy compared to that in the loading direc-tion.

5. Asymmetry of yielding

Two widely accepted sources for the asymmetric yielding in theZr are thermal residual stresses [4042] and twinning [24,37,38].The Zr is thermally anisotropic with the thermal expansion coef-cient of the c axis is 10.3106 K1, almost double the valueof the a axis, 5.8106 K1 [45]. After cooling from the stressfree temperature at 900K, the {10 1 0} grain family will be undercompression while the {0002} grain family is under tension, thusthe {10 1 0} grain family will reach the yield surface at a higherapplied stress in tension but a lower applied stress in compression,while the {0002} grain familywill be stronger in compression thatin tension. Another potential contribution to asymmetric yieldingis that due to the texture and the limited number of slip modesavailable, different deformation systems are activated duringtension and compression [24,46]. (1 0 1 2) 10 1 1 tensile twinninghas been considered as easier than c+ a pyramidal slip at roomtemperature and thus would occur when there is a tensile stressalong the c axis or compression along the a axis [47]. Xu [38]showed that tensile twinning plays a major role in the deformationand texture evolution of textured Zircaloy-2. One interestingobservation from [38] is the sharp kick back of the {0002} grainfamily strain in the loading direction after compression along thedirections with fewer basal plane normals. This strain kick backwas attributed to tensile twinning in the Poissons direction which

d (c)Fig. 12. Intensity change during tension of TD sample, (a) loading direction, (b) an Poissons directions. Errors are in the range of 314%.


rotated theaxial compstrain. Thisincrement idrop in theobservationRD sampleoccurrencetensile twindecreases twhich henctwinning b{10 1 0} graverify whettively lowerto its compr{10 1 0}, {0the three ptexture variin differenttion peak idetermineddeformatiosity increasthe tensiondeformatioThe {0002instead like

The interolled Zr2Fig. 13. Intensity change during tension of ND sample, (a) loading direction, (b) and (c)

crystal lattice and generated new {0002} grains in theression direction to reduce the average axial {0002}explanation is supported by a large {0002} intensityn the loading direction with a corresponding intensity{0002} intensity in the Poissons direction. A similarof a strain kick back is obtained in the compressed

here (see Fig. 5b), which may at rst sight suggest theof tensile twinning in our material. (1 0 1 2) 10 1 1ning causes a 90 rotation of the c axis and thushe {0002} intensity in the tensile loading direction,e allows the conrmation of the presence of tensiley analyzing the intensity change of the {0002} andin families in the two perpendicular detector banks. Toher or not tensile twinning is responsible for the rela-tensile strength of the {0002} grain family comparedessive strength, the evolution of peak intensities of the002} and {10 2 0} grain families during tension alonglate directions are plotted in Figs. 1113. Due to theation in the plate (i.e. texture may be slightly differentsamples), the relationship between measured diffrac-ntensity and texture coefcient for each sample wasbased on the measurement made on each sample after

n. It can be seen that during tension the {0002} inten-ed rather thandecreased in the tensile axis, especially inTD (Fig. 12a), indicating that twinning is not the majornmode during tension along the c axis of thismaterial.} intensity increase in the tensile loading direction isly caused by pyramidal {10 1 1} 11 2 3 slip [38].nsity change during compression along RD of the hot.5Nb plate is plotted in Fig. 14. It is shown that the

{0002} intand decreasgest that tenthe {0002canbe seenthe incremeby Xu [38],played a mi{0002} intexplained bthe {0002also seen intensile loadsample (Figdiscussedbtigation weWe concludlikely causeto the dislo

Based otwinning, aZr alloys, cexperimentmaterial anyielding of tsion.

HutchinexplanationIn their studcompressedPoissons directions. Errors are in the range of 410%.

ensity increased in the compressive loading directioned in the Poissons direction (e.g. TD), whichmight sug-sile twinning in the Poisson direction is responsible for

} strain kick back in the compressed RD. However itthat the initial {0002} intensity inRD is veryweak, andnt in intensity is very small compared to that observedindicating that if tensile twinning has occurred, it onlynor role in the deformation of this material. The smallensity increment in the compressed RD could also bey activity of basal a slip [38]. It is worth noting that} strain kick back observed in the compressed RD istensile loading tests such as in the axial direction of theed TD sample (Fig. 6a) and to a lesser extent in the RD. 5a), which are unlikely to have twinning occurring, asefore. In addition, despite extensive SEMandTEMinves-were unable to nd microscopic evidence of twinning.e therefore that the {0002} strain kick back is mored by other stress relaxation mechanisms which relatecation evolution occurring during deformation.n the above discussions, it is evident that althoughs one of the important deformation modes in someannot be totally excluded from the results of this, it is considered a minor deformation mode in thisd is unlikely to be responsible for the asymmetriche {0002} grain family between tension and compres-

son and co-workers [43,48] have given another possiblefor the asymmetry in deformation along the c axis.ies of textured Ti6Al4V alloys, they found that whenalong the 0002 direction, the 11 2 3 dislocations


F ) and

move on thnot cross slfor pyramidAssociatedwdirections opression, thdependentier in tensiohas a similais reasonabthe asymmout by Chris

For theTD (Fig. 9ations, while(Fig. 9f), indbe attributegrain shapeand {211}{200} graiand thus ya higher TaSimilar obset al. [50].

Fig. 15 shbothphasescontrast toKonly observbroadeningdeformatioig. 14. Intensity change during compression of RD sample, (a) loading direction, (be {10 1 1} planes with considerable cross slip, but doip during tension. They attributed this to a higher CRSSal c+ a slip during compression than during tension.ith thedilatationnormal to the slip plane anddifferent

f the dislocation movement during tension and com-ey proposed that the slip of 11 2 3 c+ a dislocation ison the stress state and the shear direction, and it is eas-n than in compression. Bearing in mind that Zr2.5Nbr microstructure and crystal structure to Ti6Al4V, itle to believe that Hutchinsons theory may also explainetric yielding observed in zirconium alloys, as pointedtodoulou [49].-phase, the small elastic anisotropy in the RD andd) implies a strong constraint along these two direc-a relatively larger elastic anisotropy along the ND

icates a relatively weak constraint in ND, which mayd to the character of the microstructure such as the, geometry and phase distribution. Since the {110}grain families have a larger elastic modulus than then family, they bear more stress in the elastic regimeield at a lower applied stress even though they haveylor factor than {200} for (110) 111 slip system.ervations in bcc ferrite have been explained by Oliver

ows the changes in diffraction peak width observed induringdeformation, parallel to the loadingdirection. Inumar et al. [17],where signicant peakbroadeningwased in the -phase peaks and not in the -phase, peakwas seen in all diffraction peaks in both phases duringn with broadening increasing with deformation strain.

While therewidth, it seein contrast tdiffraction wsurface of ththe {10 1 2}tributed tothere is a trcomponentduring tenssamples. Peby heterogestrain varyitors [51]. Ifa test, andassumed toasymmetricevidence focompressio

From Figphaseat yiemodulus isthis materiis comparab100% Zr min [44] waoxygen andin this hoting of niobsize.(c) Poissons directions. Errors are in the range of 16%.are many possible contributions to an increasing peakms likely that plastic strain is occurring in both phases,o Kumars rigid-phasemodel. In Kumars study, X-rayas used, which limited the data acquisition to near thee sample, and the comparison was made only between peak and the {200} peak which may have con-

the observation. Fig. 15 also shows that in the -phaseend of peak broadening increasing with increasing cduring compression but that this trend is rather vagueion. Similarphenomenawereobserved in theTDandNDak broadening in the loading directions could be causedneous dislocation distribution, strain distributions (i.e.ngwith position), grain size and other instrumental fac-the effect of instrumental factors is unchanging duringthe strain distribution and grain size evolution can bebe similar during tension and compression, then thepeak broadening shown in Fig. 15may provide anotherr asymmetric dislocation evolution during tension andn.. 4, it can be seen that the lattice strain of the -ld is about (800010,000)106. Assuming theYoungs60GPa [44], the yield strength of the -phase in

al can be roughly estimated as 500600MPa, whichle to the macroscopic yield strength obtained for theaterial in [44]. The high strength of the 100% Zr

s attributed to the solution strengthening from theniobium content, while the high -phase strengthrolled Zr2.5Nb is likely caused by the strengthen-

ium content with a contribution from the small grain


Fig. 15. Peak b compis less than 5%

6. Conclus

Both inout for texDeformatiobeen studilar strains abeen investhat:

1. A strengstrengthence of tin tensio

2. A signicfound dustrengththe strenstrength

3. The evolmined btexture dduring tmaterialof the {0in the loaduring te

4. The therpyramidable reasofound forroadening in the loading direction of the RD sample, (a)-grains, (b)-grains duringfor alpha peaks and less than 10% for beta peaks.ions

situ tension and compression tests have been carriedtured Zr2.5Nb plate material at room temperature.n along all the three principle plate directions hased and the evolution of interphase and intergranu-long all the loading and the Poissons directions has

tigated through neutron diffraction. It is concluded

th differential and different texture dependence ofwas found during tension and compression. The inu-exture on strength is much larger in compression thann.ant asymmetric yielding of the {0002} grain family isring tension and compression, where the compressionis much higher than the tensile strength. This causesgth differential and different texture dependence ofobserved in this material.ution of interphase and intergranular strain was deter-y the relative phase properties, crystal properties andistribution. The average phase behaviors are similar

ension and compression, where the -phase in thisis stronger than the -phase. The asymmetric yielding002} grain family results in a larger intergranular strainding direction during compression than that obtainednsion.mal residual stress and different CRSS for the c+ al slip during tension and compression are two possi-ns for this asymmetric yielding, and little evidence istwinning as contributing to this asymmetry.

Acknowled

The authKhan for TEAlamos NatThis work isionMetalsMaterials awere carrieAlamos Nat(NRU) reacand ISIS puUK.

References

[1] D.G. Hurs1997.

[2] R.A. Holt,[3] N. Christo

Trans. A 3[4] C.N. Tome[5] E.E. Ibrah[6] R.A. Holt,[7] R.A. Holt,[8] M. Grift[9] A.R. Caus

[10] R.A. Holt,[11] R.A. Holt,[12] R.A. Holt,[13] R.A. Holt,[14] B.A. Chea[15] K. Kapoor

Technol. 2[16] M.K.Kum

J. Nucl. Mression, (c)-grains and (d)-grains during tension. The uncertaintygements

orswish to acknowledge Drs. Igor Yakobtsov and AbdulM observations of this material, and Dr. Sven Vogel (Losional Lab.) for texture measurements on the material.s sponsored by NSERC, COG, OPG and Nu-Tech Preci-under the Industrial Research Chair Program inNucleart Queens University. Neutron diffraction experimentsd out at the Los Alamos Neutron Scattering Center, Losional Laboratory, USA, the National Research Universaltor at the AECL Chalk River Laboratory (CRL), Canadalsed neutron facility, Rutherford Appleton Laboratory,

t, Canada Enters the Nuclear Age, McGill-Queens University press,

N. Christodoulou, A.R. Causey, J. Mater. Nucl. 317 (2003) 256.doulou, P.A. Turner, C.N. Tome, C.K. Chow, R.J. Klassen, Metall. Mater.3 (2002) 1103., N. Christodoulou, Philos. Mag. 80 (2000) 1407.im, J. Nucl. Mater. 118 (1983) 260.J. Nucl. Mater. 159 (1988) 310.E.F. Ibrahim, Acta Metall. Mater 27 (1979) 1319.hs, N. Christodoulou, S.A. Donohue, J. ASTM Int. (2005) 2.ey, C.H. Woo, R.A. Holt, J. Nucl. Mater. 159 (1988) 225.J. Nucl. Mater. 90 (1980) 193.J. Nucl. Mater. 372 (2008) 182.S.A. Aldridge, J. Nucl. Mater. 135 (1985) 246.P. Zhao, J. Nucl. Mater. 335 (2004) 520.dle, C.E. Ells, J. Electrochem. Technol. 4 (1966) 329., D. Lahiri, S.V.R. Rao, T. Sanyal, N. Saibaba, B.P. Kashyap, Mater. Sci.0 (2004) 1281.ar, C.Vanitha, I. Samajdar,G.K.Dey,R. Tewari,D. Srivastava, S. Banerjee,ater. 335 (2004) 48.


[17] M.K. Kumar, I. Samajdar, N. Venkatramani, G.K. Dey, R. Tewari, D. Srivastava, S.Banerjee, Acta Mater. 51 (2003) 625.

[18] S. Cai, M.R. Daymond, R.A. Holt, E.C. Oliver, Adv. Mater. Res. 1517 (2007) 615.[19] A. Salinas-Rodriguez, Acta Metall. Mater. 43 (1995) 485.[20] A. Salinas-Rodriguez, J.H. Root, Texture Microstruct. 1418 (1991) 1239.[21] B.A. Cheadle, C.E. Ells, W. Evans, J. Nucl. Mater. 23 (1967) 199.[22] S. Kim, Metall. Mater. Trans. 37A (2006) 59.[23] N. Christodoulou, P.A. Turner, E.T.C. Ho, C.K. Chow, M. Resta Levi, Metall. Mater.

Trans. A 31A (2000) 409.[24] Z.H.A. Kassam, Z. Wang, E.T.C. Ho, Mater. Sci. Eng. 158 (1992) 185.[25] P. Gangli, J. Root, R. Fong, Can. Metall. Q 34 (1995) 211.[26] R.A. Lebensohn, C.N. Tome, Acta Metall. Mater. 41 (1993) 2611.[27] D.X. Du, C.H. Woo, Comput. Mater. Sci. 23 (2002) 260.[28] R. Choubey, S.A. Aldridge, J.R. Theaker, C.D. Cann, C.E. Coleman, Proceedings of

the 11th International Symposium on Zirconium in the Nuclear Industry ASTMSTP, vol. 1295, 1996, p. 657.

[29] R.B. Von Dreele, J. Appl. Crystallogr. 30 (1997) 517.[30] M.R. Daymond, H.G. Priesmeyer, Acta Mater. 50 (2002) 1613.[31] J.R. Santisteban, M.R. Daymond, L. Edwards, J.A. James, J. Appl. Crystallogr. 39

(2006) 812.[32] M.R. Daymond, J. Appl. Phys. 96 (2004) 4263.[33] M.R. Daymond,M.A.M. Bourke, R.B. VonDreele, B. Clausen, T. Lorentzen, J. Appl.

Phys. 82 (1997) 1554.

[34] M.R. Daymond, M.A.M. Bourke, R.B. Von Dreele, J. Appl. Phys. 85 (1999) 739.[35] A.C. Larson, R.B. Von Dreele, Report No. LAUR 86-748, Los Alamos National

Laboratory, 1994.[36] J.W.L. Pang, T.M. Holden, P.A. Turner, T.E. Mason, Acta Mater. 47 (1999) 373.[37] E.C. Oliver,M.R. Daymond, P.J.Withers,Mater. Sci. Forum495497 (2005) 1553.[38] F. Xu, Ph.D. Thesis, Queens University, 2007.[39] W.G. Burger, Physica 1 (1938) 561.[40] S.R. MacEwen, C. Tome, J. Faber, Acta Metall. 37 (1989) 979.[41] P.A. Turner, C.N. Tome, Acta Metal. Mater. 42 (1994) 4143.[42] P.A. Turner, N. Christodoulou, C.N. Tome, Int. J. Plast. 11 (1995) 251.[43] M.A.W. Lowden, W.B. Hutchinson, Metall. Trans. 6 (1975) 441.[44] S. Cai, M.R. Daymond, R.A. Holt, E.C. Oliver, J. Nucl. Mater., submitted for publi-

cation.[45] C.N. Tom, N. Christodoulou, P.A. Turner, M.A. Miller, C.H. Woo, J. Root, T.M.

Holden, J. Nucl. Mater. 227 (1996) 237.[46] S.R. MacEwen, C.E. Ells, O.T. Woo, J. Nucl. Mater. 101 (1981) 336.[47] C.N. Tome, P.J. Maudlin, R.A. Lebensohn, G.C. Kaschner, Acta Mater. 49 (2001)

3085.[48] I.P. Jones, W.B. Hutchinson, Acta Metall. 29 (1981) 951.[49] N. Christodoulou, Acta Metall. 37 (1989) 529.[50] E.C. Oliver, M.R. Daymond, P.J. Withers, Acta Mater. 52 (2004) 1937.[51] B.D. Cullity, Element ofX-raysDiffraction, Secondedition, Addison-WesleyPub-

lishing Company, Inc., 1978.

Evolution of interphase and intergranular stresses in Zr-2.5Nb during room temperature deformationIntroductionMaterials and experimental methodExperimental resultsMacroscopic mechanical responseAverage phase responseResponse of individual grain orientationsEvolution of intergranular strains in the alpha-phaseEvolution of intergranular strains in the beta-phase

DiscussionAsymmetry of yieldingConclusionsAcknowledgementsReferences

Cai Msea 2009

Documents

Transcript of Cai Msea 2009