HT

Ia

b

a

ARRA

KTETR

1

iiwrebT˛tbbmrcFiw

0d

Materials Science and Engineering A 527 (2010) 6157–6165

Contents lists available at ScienceDirect

Materials Science and Engineering A

journa l homepage: www.e lsev ier .com/ locate /msea

igh temperature deformation processing maps for boron modifiedi–6Al–4V alloys

ndrani Sena,∗, Ravi Sankar Kottadab, U. Ramamurtya

Department of Materials Engineering, Indian Institute of Science, Bangalore 560012, IndiaDepartment of Metallurgical and Materials Engineering, Indian Institute of Technology-Madras, Chennai 600036, India

r t i c l e i n f o

rticle history:eceived 27 January 2010eceived in revised form 24 April 2010ccepted 16 June 2010

eywords:itanium alloyslectron microscopy

a b s t r a c t

The alloy, Ti–6Al–4V is an ˛ + ˇ Ti alloy that has large prior ˇ grain size (∼2 mm) in the as cast state.Minor addition of B (about 0.1 wt.%) to it refines the grain size significantly as well as produces in-situTiB needles. The role played by these microstructural modifications on high temperature deformationprocessing maps of B-modified Ti64 alloys is examined in this paper. Power dissipation efficiency andinstability maps have been generated within the temperature range of 750–1000 ◦C and strain rate rangeof 10−3–10+1 s−1. Various deformation mechanisms, which operate in different temperature–strain rateregimes, were identified with the aid of the maps and complementary microstructural analysis of thedeformed specimens. Results indicate four distinct deformation domains within the range of experimen-

◦ −3 −2 −1

hermomechanical processingecrystallization

tal conditions examined, with the combination of 900–1000 C and 10 –10 s being the optimumfor hot working. In that zone, dynamic globularization of ˛ laths is the principle deformation mecha-nism. The marked reduction in the prior ˇ grain size, achieved with the addition of B, does not appearto alter this domain markedly. The other domains, with negative values of instability parameter, showundesirable microstructural features such as extensive kinking/bending of ˛ laths and breaking of ˇ lathsfor Ti64–0.0B as well as generation of voids and cracks in the matrix and TiB needles in the B-modified

alloys.

. Introduction and background

Ti–6Al–4V (also referred as Ti64), an ˛ + ˇ titanium alloy is anmportant engineering alloy that is extensively used particularlyn aerospace industry. This is due to its low density combined

ith high strength and toughness as well as outstanding corrosionesistance. An additional benefit associated with Ti alloys, in gen-ral, is that their properties are relatively temperature-insensitiveetween cryogenic temperature and ∼500 ◦C. In the as-cast state,i64 exhibits the classical Widmanstätten microstructure of (hcp)and (bcc) ˇ phases. However, Ti alloys – and Ti64 is no excep-

ion – typically suffer from large prior ˇ grain size, which tendse in the order of a few mm. Therefore it becomes necessary toreak this coarse microstructure down, through several thermo-echanical steps. Typically, this involves upset forging in the ˇ

egime, i.e. above 1000 ◦C. This not only adds considerably to the

ost of the final product, but also brings in additional complexities.or example, the oxide layer that forms on the surface during forg-ng has to be machined out at each step, causing loss of material as

ell as adding to the manufacturing cost, as it otherwise could get

∗ Corresponding author. Fax: +91 80 2360 0472.E-mail address: indrani.indrup@gmail.com (I. Sen).

921-5093/$ – see front matter © 2010 Elsevier B.V. All rights reserved.oi:10.1016/j.msea.2010.06.044

© 2010 Elsevier B.V. All rights reserved.

included in the material leading to low fatigue performance. Thus,the necessity to break the coarse as-cast structure makes the fin-ished Ti alloy products considerably expensive vis-à-vis competingalloys.

Recently, it was discovered that the addition of B in minoramounts (within the hypo-eutectic range) to Ti64 reduces the grainsize significantly (by an order of magnitude) [1,2]. This circum-vents the need for processing steps such as ˇ upset forging andhence makes Ti alloys relatively more affordable. As a result, therehas been considerable interest in understanding the mechanicalbehavior of these alloys. It has been shown that the microstruc-tural refinement leads to anomalous increase in elastic modulus,moderate enhancement in yield and ultimate strengths, and signif-icant benefit in terms of the unnotched fatigue performance of theTi64 [3–7].

Although the addition of B leads to a markedly reduced grainsize, it does not completely eliminate the need for some thermo-mechanical processing steps subsequent to casting of the alloy.These are necessary for, at least, two reasons. The first is to impart

the desired shape to the alloy and the second is to close any as-castporosity, which is otherwise detrimental to the fatigue perfor-mance of the alloy. Therefore, it is essential to determine the hotworking conditions, i.e., optimum temperature and strain rate com-bination, for deformation that will yield required microstructures

6 d Engi

wt

emztTpticatcpomamrdaodmo

n[

�

wepw

�

Tcmabtcd

m

T

J

WdtJ

�

TarF

158 I. Sen et al. / Materials Science an

ithout significant defects in them. The identification of these,hrough the processing maps approach, is the objective of this work.

Hot working of ˛/ˇ Ti alloys is a widely studied area with consid-rable literature available on it. The principle dynamic deformationechanism that leads to an optimum condition in the working

one of Ti64 alloys with lamellar structures is the spherodiza-ion/globularization of ˛ laths [8–10]. Since the addition of B toi64 reduces its grain size and also introduces stiff and hard TiBarticles into the microstructure, a pertinent question to ask ishe following: “Do these microstructural modifications, which aremparted through the addition of B to Ti64, alter the optimum pro-essing conditions?” An additional, and related, question is “whatre the differences in micro-mechanisms during high tempera-ure deformation that occur at different temperature–strain rateombinations?” Answers to these questions are sought throughrocessing maps, which are generated based on the resultsbtained by conducting high temperature compression experi-ents on a series of Ti64–B alloys with the test temperature and

pplied strain rate as variables. These are coupled with detailedicrostructural studies of the deformed specimens for corrobo-

ating the inferred deformation mechanisms operating in variousomains. The stress–strain responses and detailed analyses of themre reported in a companion paper [11]. Since this paper’s focus isn processing maps, the scientific basis for their construction isescribed briefly below. Interested reader can find detailed infor-ation about them in the reviews by Prasad et al. [12,13] who

riginally came up with the concept of processing maps.When a material is deformed at a strain rate, ε̇, its instanta-

eous response in terms of true stress, �, will follow the power law13–15]:

= K · ε̇m (1)

here K is a constant and m is the strain rate sensitivity. The totalnergy spent on the work piece, given by the product �. ε̇, can beartitioned into that dissipated as heat (G-content) and the onehich leads to change in the microstructure (J-co content) as:

· ε̇ =∫ ε̇

0

� dε̇ +∫ �

0

ε̇ d� (2)

he first integral on the right-hand side of Eq. (2) gives the G-ontent whereas the second gives the J-co content. Practically,ajor portion of the energy input to the work piece is dissipated

s the rise in temperature i.e. thermal energy which is representedy the G-content. J-co content, which represents the microstruc-ural changes that occur during deformation, is relatively small inomparison to the G-content. The partitioning between the two isecided by m, which is defined as:

= dJ

dG= �log �

�log ε̇(3)

he value of m varies between 0 and 1. From Eqs. (1)–(3), we get:

= m · � · ε̇

m + 1(4)

hen m = 0, J = 0 whereas J = G for m = 1 [13]. For a linear powerissipater, m = 1 and J ∼ Jmax [13]. The efficiency of power dissipa-ion, �, for microstructural changes can be obtained by normalizingwith Jmax as:

= J

J= 2m

m + 1(5)

max

hus, the constitutive equation, which captures the intrinsic char-cteristics of the work piece, will dictate the path it will follow inesponse to the applied strain. This path is a strong function of ε̇.or example, application of higher ε̇ may lead to internal crack-

neering A 527 (2010) 6157–6165

ing whereas the same material may undergo dynamic recovery atlower ε̇.

From the true stress, �– true strain, � curves generated at var-ious T– ε̇ combinations, m can be calculated by assuming that thework-piece is a non-linear energy dissipater. Consequently, � canbe computed for a fixed value of � by using eqn. (5), which in turncan be utilized to construct iso-efficiency contour maps in the 2-Dspace of T and ε̇ (� maps). These contour maps are utilized to iden-tify the different deformation mechanisms prevailing over differentcombinations of T and ε̇. For instance, the dynamic recrystalliza-tion (DRX) regime, which is important from the workability pointof view, typically exhibits the highest � over a broad range withwidely spaced contours indicating small gradation to the efficiencyhill. The width of the DRX regime obtained from the processingmaps will dictate the limits, in terms of T and ε̇, for the industrialhot working schedule. However, a region with very high � value(>90) in the efficiency maps is also usually associated with voidformation at hard particles or can be an indication of superplasticregime as well. The former is also characterized by a rapid rise in �with the decrease in temperature. Therefore, the information fromprocessing maps alone is insufficient. Complementary microstruc-tural characterization thus plays a vital role in understanding of thedeformation mechanisms associated with a specific regime.

An instability criterion, which is based on the maximum rate ofentropy production, gives the following equation [8,13,16]:

�(ε̇) = ∂ ln(m/m + 1)∂ ln ε̇

+ m < 0 (6)

where �(ε̇) is a dimensionless parameter that represents the insta-bility in the system. Similar to � maps, � contour maps can alsobe generated in the T– ε̇ space for a fixed value of �. The region ofinstability is indicated by a negative � value. Typical signatures ofinstability are shear band formation, flow localization, formationof cavities, breaking of particles etc. [8–10,13,17–19]. The � and �maps together can then be utilised to identify the domains that aresuitable for hot working as well as the ones that give rise to unde-sirable microstructures and hence should be strictly avoided duringindustrial practice. This approach has been successfully utilized tocorrectly predict hot working conditions for a variety of metals andalloys including Ti alloys [8,10,13].

2. Experiments

Five as-cast Ti–6Al–4V-xB alloys, where x = 0.0, 0.04, 0.09, 0.30and 0.55 wt.%, were examined in this study (referred as Ti64–xBin the remainder of the text). Uniaxial compression tests wereconducted at temperatures, (T) of 750, 800, 850, 900, 950 and1000 ◦C. This range of T has been chosen for the following rea-son. The reported ˇ transus temperature for commercial grade Ti64alloy, which contains about 1500–2000 ppm of oxygen, is ∼995 ◦C[19]. The high temperature bcc ˇ phase exhibits large ductility andhence is easy to process. Therefore, the present work is restrictedto the two phase ˛ + ˇ regime. At each T, constant true strain rates(ε̇) of 10−3, 10−2, 10−1, 100, 10+1 s−1 were employed and truestrains � of 0.7 were achieved at each T– ε̇ combination. Furtherdetails of the materials, their initial microstructures, stress–strainresponses, and details of analyses of the latter can be found in Ref.[11].

3. Results

3.1. Stress–strain curves

All the �–ε plots show that softening followed by a peak � areexistent at lower ε̇ of testing (Fig. 2a of [11]), whereas at higherε̇ the flow curves exhibit oscillations (Figs. 2b and c of [11]). The

d Engi

e�[

fovttow[

tI2id[

3

treεpdWw�wwtiae

patmafc

4

tdteod

I. Sen et al. / Materials Science an

ffect of B addition on the flow stress behavior is illustrated in–ε plots at a particular T and ε̇ of 750 ◦C and 10−2 s−1 (Fig. 3 of

11]).The values of m were computed for each alloy and at dif-

erent temperatures from the slopes of the logarithmic plotsf �– ε̇ [20]. ε value of 0.5 is used for the purpose. The malues were found to vary in two distinct temperature ranges,ermed as ‘regime’. Regime is therefore defined as the range ofemperature over which a distinct deformation mechanism isperative. Regime I corresponds to T = 750–850 ◦C with m ∼ 0.17,hereas regime II is associated with T = 900–1000 ◦C and m ∼ 0.22

11].Activation energy, Q, has also been found to vary in two

emperature ranges, regime I (750–850 ◦C) as well as regimeI (900–1000 ◦C). The values of Q in these two regimes are60–380 and 500–640 kJ/mole, respectively [11]. Such values

ndicate that the rate controlling deformation mechanisms areislocation glide in regime I [10,21,22] and DRX in regime II23].

.2. Processing maps

Efficiency maps for all the five grades of alloys (Ti64–0.0Bo Ti64–0.55B) generated at ε = 0.5 are displayed in Figs. 1a–5a,espectively. These maps show peaks of �, which signify differ-nt deformation mechanisms, at different combinations of T and

˙ . Instability maps for all the compositions generated at � = 0.5 arelotted in Figs. 1b–5b. A domain in these maps is defined as the win-ow over which a change in both the values of � and � are evident.ithin the individual domains several subdomains are presenthich have subtle variations in terms of values and contours ofand �. However, microstructural features in all these subdomainsithin a particular domain are similar. Different domains alongith the subdomains, represented by A and B are indexed and illus-

rated on both the efficiency and instability maps in Figs. 1–5. Fornstance, the second subdomain of the third domain is representeds III-B. Representative micrographs obtained for all the alloys inach domain (I–IV) are presented in Figs. 6–9.

A summary of the information that has been extracted from therocessing maps is presented in Table 1. It lists the various domainsnd the subdomains within them, operative T and ε̇ ranges forhese domains and subdomains, corresponding � and � values, key

icrostructural features observed and broad deformation mech-nisms for all the compositions examined in this work. It is clearrom Table 1 and Figs. 1–5, that the processing maps can be broadlylassified into four domains, which are listed below.

Domain I: T = 750–850 ◦C and ε̇ = 10−3 to 10− s−1; high �, negative�.Domain II: T = 900–1000 ◦C and ε̇ = 10−3 to 10−2 s-1; high �, broadcontours with a positive value of �.Domain III: T = 750–850 ◦C and ε̇ = 100 to 10+1 s−1; low �, (often)negative values of �.Domain IV: T = 900–1000 ◦C and ε̇ = 100 to 10+1 s−1; low �, (often)negative values of �.

. Discussion

As mentioned in Section 3.1, the analysis of m, Q as well ashe Zener-Hollomon parameter, Z, obtained from the stress–strain

ata of all the alloys, indicate to two temperature regimes wherehe deformation mechanisms are distinctly different [11]. How-ver, the � and � maps help further delineate the domains in termsf T and ε̇ combinations. As mentioned in Section 3, four distinctomains, I–IV have been identified from the efficiency and instabil-

neering A 527 (2010) 6157–6165 6159

ity maps for all the alloy compositions irrespective of the amountof B addition. In addition, different prominent subdomains in termsof variation of � and � are present within each domain. These arediscussed below.

4.1. Domain I (T ∼ 750–850 ◦C and ε̇ ∼ 10−3 to 10−1 s−1)

For the B-free alloy, domain I is associated with two distinctsubdomains. The first one, denoted as I-A is at T ∼750–850 ◦C andε̇ = 10−3 to 10−2 s−1, with � ∼ 40–96% (Fig. 1a) and a positive � valueof ∼0.4–0.9 (Fig. 1b). Such high efficiency, which was observed onlyin the B-free alloy, has been found to increase sharply with increasein ε̇ and decrease in T (narrow contours). This typically denotes theoccurrence of voids [13]. Sharp bending of laths as well as breakingor discontinuities in the ˇ laths are the principle microstructuralfeatures observed in specimens that were deformed within thisregime (Fig. 6a). A second subdomain I-B operates at T ∼750–800 ◦Cand ε̇ ∼ 10−2 to 10−1 s−1. It exhibits a wide variation in �, between12 and 44%, and a negative � ranging between −0.3 and −0.7. Thisunstable region corresponds to lath bending as the principle defor-mation mechanism (Fig. 6a).

For the B-modified alloys, domain I is associated with � vary-ing between 24 and 40% (Figs. 2a–4a) except for Ti64–0.55B whichhas lower values of �, varying between 16 and 32% (Fig. 5a). The� values for all the alloy compositions in this domain are about−0.1 (Figs. 2b–5b) indicating instability. Microstructural featuresassociated with this region include cracking of the Ti64 matrix aswell as the TiB needles along with bending of laths, breaking of lathboundaries and generation of voids in the matrix where TiB parti-cles are present (Figs. 6b and c). Strain incompatibility between thematrix and the TiB particles is a possible source for the void for-mation. Due to the applied stress, the hard particles do not deformsignificantly but the matrix surrounding the particle can undergoextensive plastic deformation leading to strain concentration nearthe particles. This in turn leads to separation of the matrix fromthe particle or particle cracking, which in turn could aid in prema-ture fracture of the component [12]. However, as T increases, theremust be lower work hardening of the matrix due to recovery, andalso a concomitant decrease in the strength of the particles. Con-sequently, the stress concentration at particle and matrix interfacewill be less at higher temperature and the rate of void/cavity for-mation will be reduced. Presence of TiB needles therefore increasesthe probability of instability in this domain.

4.2. Domain II (T ∼ 900–1000 ◦C and ε̇ ∼ 10−3 to 10−2 s−1)

Domain II is characterised by a peak efficiency of 40–56% and apositive � value of 0.2, for all the alloy compositions. Microstruc-tural characteristics in this domain are similar for the B-free as wellas B-modified alloys. These include globularised ˛ phase with thick-ness and aspect ratios being 8 mm and 1.25 respectively (Figs. 7a–c).Another interesting feature of this domain is the generation offiner alternate ˛ (thickness ∼0.5 �m) and ˇ laths (Fig. 7a). To showthat such microstructural features emerge from the beginning ofthe deformation, a specimen of Ti64–0.09B alloy was deformedto ε = 0.3 at T = 900 ◦C and ε̇ = 10−3 to 10−2 s−1. It was then evalu-ated for its microstructure. It is evident from the micrographs [11]that globularization as well as existence of finer laths is prominenteven at such lower strain. Microstructural observations thus sug-gest that recrystallization under dynamic condition is prevalent inthis domain (II).

From the observed differences in microstructural features ofdomains I and II, it is apparent that different deformation mecha-nisms are active in them, although both the domains exhibit similarpower dissipation efficiencies. However, the steepness of the hills,as reflected by the broadness of the contours, is different. The tem-

6160I.Sen

etal./M

aterialsScience

andEngineering

A527 (2010) 6157–6165

Table 1List of domains and subdomains within the processing window of temperature, T varying between 750 and 1000 ◦C and strain rate, ε̇ varying between 10−3 and 10+1 s−1. Domains are differentiated based on the values of powerdissipation efficiency, � and instability parameter, �, complemented with microstructural observations.

Composition (wt.%B) Domain Sub domain T range (◦C) ε̇ range (s−1) � range (%) � range (fraction) Key microstructural features* Broad mechanism

0.0 I A 750–850 10−3 to 10−2 40–96 0.4 to 0.9 Sharp BL and BRL InstabilityB 750–800 10−2 to 10−1 48–12 −0.3 to −0.7

II A 900–1000 10−3 to 10−2 40–48 0.2 GL and FL with width of � lath∼0.5 �m DRXIII A 750–800 5 × 100 to 10+1 32–40 1.0 LDB, BL and � BRL Instability

B 825–875 100 to 10+1 12 −0.2IV A 950–1000 100 to 10+1 12–28 −0.1 to −0.5 BL and BRL Instability

0.04 I A 750–850 10−3 to 5 × 10−3 24–32 0 to −0.2 BL, BRL and MC wherever TiB are present InstabilityII A 900–1000 10−3 to 10−2 40–52 0.2 GL and FL DRXIII A 750–825 5 × 10−1 to10+1 4–16 −0.2 to −1.0 MC, BL and TC InstabilityIV A 875–950 100 to 10+1 20–24 0 to 0.1 MC and TC Instability

0.09 I A 750–825 10−3 to 10−2 32–40 0 to −0.1 BL, BRL, MC and TC InstabilityII A 875–1000 10−3 to 5 × 10−2 36–52 0.2 GL and FL. Smaller extent of MC and TC are also apparent DRXIII A 750–800 5 × 100 to 10+1 20–28 0.5 to 1.0 MC, TC, BL. Cracking on the specimen at an

angle of 45◦ to the loading axis is observedInstability

B 800–900 100 to 10+1 12–20 −0.1 to −1.0IV A 900–950 100 to 10+1 24 0.1 to 0.3 BL, MC, TC and FL Instability

B 950–1000 5 × 100 to 10+1 24 0.0 to −0.2

0.30 I B 750–800 10−3 to 10−1 20–40 −0.1 MC InstabilityII A 950–1000 10−3 to 5 × 10−3 44–52 0.2 GL and FL DRX

B 975–1000 5x10−3to10−1 24–40 0.0 to −0.1III A 750–800 100 to 10+1 16 0.1 to 0.5 Extensive MC and

TC, BLInstability

B 800–850 100 to 10+1 16–20 −0.1IV A 900–950 5 × 100 to 10+1 20–24 −0.1 MC, GL Instability

B 975–1000 100 to 10+1 28–36 0.2 to 0.5

0.55 I A 750–800 5 × 10−3 to 10−1 32–16 −0.1 Extensive MC, TC and BL InstabilityII A 900–1000 10−3 to 10−2 40–56 0.1 to 0.2 GL and FL. Somewhat MC DRXIII A 750–800 100 to 10+1 16–20 0.1 to 0.5 Extensive MC InstabilityIV A 925–975 100 to 10+1 24 0.1 to 0.2 MC, TC Instability

B 975–1000 100 to 10+1 16–20 −0.1 to −0.2

* Sharp bending of laths (BL), breaking of laths (BRL), Globularised � laths (GL), finer laths (FL), Localized deformation band (LDB), matrix cracking (MC), TiB needle cracking (TC).

I. Sen et al. / Materials Science and Engineering A 527 (2010) 6157–6165 6161

F fig. ae bdom

ptTDo(Sgbipholoarb

ig. 1. (a) Efficiency and (b) instability maps for Ti64–0.0B. The contour numbers infficiency level whereas that in (b) indicates instability. Significant domains and su

erature range for domain II corresponds to regime II [11], thereforehis domain has Q varying from 500 to 640 kJ/mole, and m ∼ 0.22.hese values are similar to the values that are usually obtained forRX in Ti alloys [20,23,24]. Microstructural observations carriedut in this work also suggest that DRX is prevalent in this domainII) with a significant change in lamellar to globular morphology.emiatin et al. [9] have reported that the main driving force forlobularization was the formation of high energy, intense shearands within individual ˛ lamellae. As a result, the ˛/˛ interfaces

n contact with the ˇ lamellae give rise to surface tension drivenenetration of ˛ lamellae by the ˇ phase. Seshacharyulu et al. [10]ave suggested the following mechanism for this globularizationf ˛ laths. According to them, the shear part of the applied sheareads to shearing of the laths. As expected, the colonies favourably

riented for shear will participate in the process and will cause thedjacent colonies (which were initially unfavourably oriented) tootate and shear. In the process, dislocations of both the signs wille generated along the line of shear. Cross slip will lead to annihila-

Fig. 2. (a) Efficiency and (b) insta

represent percent efficiency of power dissipation as well as the color bar indicatesains have been indicated.

tion of opposite signed dislocations at intersecting slip planes. Thusa group of dislocations with similar sign will be left behind whichwill nucleate an interface along the line of shear. In order to mini-mize the surface energy, these interfaces will migrate by diffusionand form globular structures. Since this entire process consists oftime related nucleation and migration, it is referred to as a dynamicrecrystallization process.

Scrutiny of the �-maps reveals minor changes in the tem-perature (∼25 ◦C) for peak efficiency of power dissipation in theB-modified Ti64. We consider this ‘minor’ for the following two rea-sons. (a) The temperature interval that was used for conducting theexperiments, 50 ◦C, is much larger than the relative shift observed.Therefore, paying emphasis on the relative changes would be incor-rect. (b) During industrial practice, controlling temperatures to a

high degree is not possible. Since this paper’s goal is to aid in suchpractice, such small temperature changes are ignored.

Critical observations of the map reveal that the values of �exceed 40% in the temperature range of 900–1000 ◦C. Microstruc-

bility maps for Ti64–0.04B.

6162 I. Sen et al. / Materials Science and Engineering A 527 (2010) 6157–6165

) insta

tttDcnvDioa

4

veB

Fig. 3. (a) Efficiency and (b

ural observations also support the occurrence of DRX in the aboveemperature range for low strain rates. These therefore suggest thathe optimum hot working zone or the domain for the occurrence ofRX does not alter with B addition to Ti64. In other words, signifi-ant refinement of prior ˇ grains with B addition to Ti64 has little oro effect on the DRX mechanism active in domain II. Similar obser-ations were also made by Prasad et al. [25]. This is because, theRX mechanism acts in globularizing the initially acicular ˛ laths

rrespective of the size of the prior ˇ grains. Thus no prominent rolef B addition could be found in the optimum working zone of Ti64lloys.

.3. Domain III (T ∼ 750–850 ◦C and ε̇ = 100–10+1 s−1)

The third domain has two distinct subdomains in terms ofariation in � and �. The first one, represented as III-A, has mod-rately higher values of � ∼ 16–28% and a positive � ∼ 0.1–1.0 for-modified alloys. A domain of high � of 40% (however, a steep

Fig. 4. (a) Efficiency and (b) insta

bility maps for Ti64–0.09B.

efficiency hill as shown by narrow contours) and positive � existsfor Ti64–0.0B in the temperature range of 750–850 ◦C and a highε̇ range of 5 × 100 to 10+1 s−1. Microstructural analyses of spec-imens tested at 750 and 800 ◦C and at 10+1 s−1 exhibit localiseddeformation bands (indicated by dashed lines in Fig. 8a) along withsignificant bending of the ˛ laths. All these microstructural featuressignify instability. Thus, although this domain is associated with apositive � value, it should not be confused with the DRX domain.Moreover, even if the value of � seems to have similar magnitudein both domains II and III the broadness in the contours in thesetwo domains are different (II being higher than III).

The second subdomain, III-B has � ∼ 4–20% and negative� ∼ −0.1 to −1.0. However, both III-A and III-B are associated withmicrostructural signatures of instability. These features include lath

bending for the B-free alloy (Fig. 8b) and along with that, generationof voids in the matrix and TiB needle cracking for the B-modifiedalloys (Fig. 8c). Also, microscopy of a partially deformed (ε = 0.3)Ti64–0.09B specimen tested at a strain rate of 100 s−1 at T = 850 ◦C

bility maps for Ti64–0.30B.

I. Sen et al. / Materials Science and Engineering A 527 (2010) 6157–6165 6163

Fig. 5. (a) Efficiency and (b) instability maps for Ti64–0.55B.

F he latf ws sig

eo

4

ti−it

Fs

ig. 6. Representative microstructures obtained in domain I. (a) shows bending of tormation of voids in the matrix in subdomain B of the Ti64–0.04B alloy and (c) sho

xhibits cracking of ˛ laths and TiB needles, besides the occurrencef voids at the origins of TiB needles.

.4. Domain IV (T ∼ 900–1000 ◦C and ε̇ = 100–10+1 s−1)

In spite of having similar � values, the distinct contours for both

he domains III and IV can easily be identified. The fourth domains associated with narrow � values of 12–28% and a negative � of0.1 to −0.2. As expected this domain is also associated with sim-

lar microstructural features as those of the unstable regimes. Forhe B-modified alloys, this domain exhibits clustering of TiB nee-

ig. 7. Representative microstructures obtained in domain II. (a) shows formation of globhow the evidences of globularization of laths for Ti64–0.04B and Ti64–0.55B at subdoma

hs in subdomain B in the B-free alloy. (b) reveals bending of the laths as well as thenature of instability at subdomain A for Ti64–0.55B.

dles. Considerable stress generated due to this clustering leads tocracking of the hard TiB needles, along with the generation of voidsat the matrix. Bending of ˛ laths and presence of shear bands arethe other important phenomena observed in this domain (Figs. 9band c).

Thus, amongst all the four domains as differentiated on the basis

of variation of � and � on the contours of individual regimes; domainII is associated with dynamic deformation mechanism and hencecan be considered as the optimum working zone. B additions toTi64 alloys do not alter the span of DRX regime significantly. All theother domains having minute to large variations of � and negative �,

ular ˛ laths and finer acicular ˛/ˇ laths at subdomain A for Ti64–0.04B. (b) and (c)in A.

6164 I. Sen et al. / Materials Science and Engineering A 527 (2010) 6157–6165

Fig. 8. Representative microstructures obtained in domain III. Existence of localized deformation bands (marked by dashed lines) as well as kinking of laths is evident from(a) and (b) respectively for Ti64–0.0B alloy. (c) shows breaking of TiB needles and formation of voids for Ti64–0.30B.

F and (bd

hio

5

iih7oeatdItodoasdvnwk

wbdt

[

[[

ig. 9. Representative microstructures obtained in domain IV for (a) Ti64–0.30Bomains are also evident in IV as well.

owever signify instability and should be avoided during process-ng. The presence of TiB needles further enhances the probabilityf instability in those domains.

. Conclusions

Hot deformation behavior of Ti64 alloys modified with vary-ng wt.% of B (starting from 0.0 up to 0.55 wt.%) has been studiedn detail. Isoefficiency and instability parameter contour mapsave been constructed within the ˛ + ˇ temperature range of50–1000 ◦C and strain rate range of 10−3–10+1 s−1. The variationsf the values and spans of � and � as obtained from the maps,xhibit presence of four distinct domains. Domain I (T ∼ 750–850 ◦Cnd ε̇ ∼ 10−3 to 10−1 s−1), domain II (T ∼ 900–1000 ◦C and ε̇ ∼ 10−3

o 10−2 s−1), domain III (T ∼ 750–850 ◦C and ε̇ = 100–10+1 s−1) andomain IV (T = 900–1000 ◦C and ε̇ = 100–10+1 s−1). For the domain

I, �, �, Q (Regime II) as well as the microstructural observa-ions reveal that dynamic recrystallization/globularization (DRX)f ˛ laths is the principle operating deformation mechanism. Thisomain is the optimum working regime for the alloys. The additionf B has not been found to alter the span of this domain consider-bly. This is because, addition of B to Ti64 refines the prior ˇ grainize significantly which has not been found to control/affect theynamic globularization of the ˛ laths. The other domains with �arying over a range of values and negative values of � exhibit sig-atures of instabilities such as ˛ lath bending in the Ti64–0.0B asell as generation of voids in the matrix, cracking of the TiB needles,

inking of laths etc. in the B-modified alloys.

Thus the significant refinement in prior ˇ grain size that occurs

ith B addition has minor effect in controlling the processingehavior. In other words, the domains suitable for working, i.e.ynamic globularization regime and the domains to be avoided, i.e.he unstable regime do not alter significantly for the B-modified

[

[[[

) Ti64–0.55B. The signatures of instabilities as mentioned for the other unstable

Ti64 alloys. However the generation of TiB needles with B addi-tion has an adverse effect in terms of strain incompatibility withthe matrix which leads to the cracking of both the needles and thematrix.

Acknowledgements

This work at IISc was funded by The Boeing Company through aResearch Partnership. We thank Dr. K.K. Sankaran and Mr. R. Talwarof Boeing for their support and technical discussion. We acknowl-edge the help of Mr. S. Sasidhara for his technical assistance withthe experiments reported in this paper.

References

[1] S. Tamirisakandala, R.B. Bhat, J.S. Tiley, D.B. Miracle, Scripta Mater. 53 (2005)1421.

[2] M.J. Bermingham, S.D. McDonald, K. Nogita, D.H. St. John, M.S. Dargusch, ScriptaMater. 59 (2008) 538.

[3] S. Tamirisakandala, R.B. Bhat, J.S. Tiley, D.B. Miracle, J. Mater. Eng. Perform. 14(2005) 741.

[4] W. Chen, C.J. Boehlert, Mater. Sci. Eng. A 494 (2008) 132.[5] I. Sen, S. Tamirisakandala, D.B. Miracle, U. Ramamurty, Acta Mater. 55 (2007)

4983.[6] I. Sen, Ph.D. Thesis, Indian Institute of Science, 2010.[7] I. Sen, U. Ramamurty, Scripta Mater. 62 (2010) 37.[8] Y.V.R.K. Prasad, T. Seshacharyulu, Mater. Sci. Eng. A 243 (1998) 82.[9] S.L. Semiatin, V. Seetharaman, I. Weiss, Mater. Sci. Eng. A 263 (1999) 257.10] T. Seshacharyulu, S.C. Medeiros, J.T. Morgan, J.C. Malas, W.G. Frazier, Y.V.R.K.

Prasad, Scripta. Mater. 41 (1999) 283.11] I. Sen, U. Ramamurty, Met. Trans. A, in press, doi:10.1007/s11661-010-0352-x.12] Y.V.R.K. Prasad, Indian J. Tech. 28 (1990) 435.

13] Y.V.R.K. Prasad, S. Sasidhara, Hot Working Guide, ASM International, 1997, p.

5.14] M.Q. Li, W.F. Zhang, Mater. Sci. Eng. A 502 (2009) 32.15] C. Poletti, F. Warchomicka, H.P. Degischer, Mater. Sci. Eng. A 527 (2010) 1109.16] T. Seshacharyulu, S.C. Medeiros, W.G. Frazier, Y.V.R.K. Prasad, Mater. Lett. 47

(2001) 133.

d Engi

[[[

[

[

[22] M. Doner, H. Conrad, Metall. Trans. 4 (1973) 2809.[23] Y.Y. Zong, D.B. Shan, M. Xu, Y. Lu, J. Mater. Process. Technol. 209 (2009)

I. Sen et al. / Materials Science an

17] R.M. Miller, T.R. Bieler, S.L. Semiatin, Scripta Mater. 40 (1999) 1387.18] S.L. Semiatin, T.R. Bieler, Acta Mater. 49 (2001) 3565.

19] T. Seshacharayulu, S.C. Medeiros, J.T. Morgan, J.C. Malas, W.G. Frazier, Y.V.R.K.

Prasad, Mater. Sci. Eng. A 279 (2000) 289.20] S.L. Semiatin, K.A. Lark, D.R. Barker, V. Seetharaman, B. Marquardt, Metall.

Mater. Trans. A 23 (1992) 295.21] T. Seshacharyulu, S.C. Medeiros, W.G. Frazier, Y.V.R.K. Prasad, Mater. Sci. Eng.

A 325 (2002) 112.

[[

neering A 527 (2010) 6157–6165 6165

1988.24] T. Sheppard, J. Norley, Mater. Sci. Technol. 4 (1988) 903.25] Y.V.R.K. Prasad, T. Seshacharyulu, S.C. Medeiros, W.G. Fraizer, J. Mater. Eng.

Perform. 9 (2000) 153.