Globularization Behavior of ELI Grade Ti 6Al 4V Alloy during Non Isothermal Multi Step Forging

Globularization Behavior of ELI Grade Ti-6Al-4V Alloy

during Non-Isothermal Multi-Step Forging

Jin Young Kim

, Kyung-Tae Park

, In Ok Shim

and Soon Hyung Hong

The 4thTechnology Research Center, Agency for Defense Development, Daejeon 305-600, Korea

2_{Division of Advanced Materials Science and Engineering, Hanbat National University, Daejeon 305-719, Korea}

3_{Department of Materials Science & Engineering, Korea Advanced Institute of Science and Technology, Daejeon 305-701, Korea}

The globularization behavior of ELI grade Ti-6Al-4V alloy having different initial lamellar structures during multi-step forging under non-isothermal condition was investigated. The samples with either a thin or thick lamellar structure, which were produced fromannealing followed by different cooling conditions, were upset and stretched repeatedly at forging start temperatures of 940_{C and 900C, and}

subsequently air cooled. The microstructural changes after non-isothermal multi-step forging were analyzed with respect to globularization of lamellae. After multi-step forging at 940_{C, the initially thin lamellar structure changed to homogeneous equiaxedglobules, but elongated}

globules with high aspect ratio were obtained from the initially thick lamellar structure. By forging at 900_{C, the initially thin lamellar structure}

was changed to a mixed structure of both equiaxed and elongatedglobules, while severely deformed lamellae were preserved inside distorted colonies with little globularization for the initially thick lamellar structure. Globularization of the samples by forging at 900_{C with both initially}

thin and thick lamellar structures was less effective compared to that by forging at 940_{C. A quantitative microstructural analysis revealed that,}

in addition to the initial microstructure and forging start temperature, globularization of lamellae was significantly affected by the instantaneous microstructural development by thermal fluctuation during non-isothermal multi-step forging.

[doi:10.2320/matertrans.MER2007224]

(Received September 25, 2007; Accepted November 15, 2007; Published December 25, 2007)

Keywords: titanium-6aluminum-4vanadium Alloy, multi-step forging, globularization, lamellar structure, thermal cycle

1. Introduction

þTi-6Al-4V alloy with a lamellar structure generally

exhibits the superior fatigue properties and creep resistance to the alloy with an equiaxed or a bimodal structure. But due to the poor ductility at room temperature of this lamellar structure, the equiaxed or bimodal structure is preferred in

most industrial applications.1–4)_{Therefore, the conversion of}

a lamellar structure to an equiaxed or a bimodal structure is one of the key steps in the semi-product fabrication of the alloy. This conversion process, called globularization, is known to occur by dynamic/static recrystallization during

deformation and post heat treatment below thetransus.5–8)

The globularization behavior of Ti-6Al-4V alloy with a lamellar structure is known to be closely related to the initial

microstructural features such as the prior grain size,

colony size andlamellar thickness. In addition, it is affected

by the processing parameters such as the forging temperature, strain, strain rate, and deformation mode. Accordingly, extensive and intensive investigations have been performed previously on the globularization behavior of the lamellar Ti

alloys focusing on the effects of the microstructures9–12)_and

processing parameters.13–15) _{However, most of the previous}

studies were conducted under the isothermal deformation condition, and it is rare to find a systematic one carried out under the non-isothermal deformation condition. In most of industrial practices, hot working for globularization of the Ti alloy billet is carried out under the non-isothermal condition — that is, the billet temperature concurrently decreases during deformation, and several reheating steps are employed to restore the initial temperature. Under such temperature fluctuation, the globularization behavior of a lamellar structure is expected to be different from that under the ideal isothermal deformation condition.

The objective of the present research is to investigate the effects of the initial lamellar thickness and forging start temperature on the globularization behavior of Ti-6Al-4V alloy during non-isothermal multi-step forging, particularly focusing on the instantaneous microstructural change under thermal fluctuation. For this purpose, the samples having different lamellar thickness were prepared and forged at different temperatures, and then their microstructures after forging were analyzed. In addition, in order to find the effect of thermal fluctuation during non-isothermal forging, the forged microstructures were compared with those of the samples subjected to the identical thermal cycle without multi-step forging.

2. Experimental Procedures

2.1 Material

Mill annealed Ti-6Al-4V alloy of ELI (Extra Low

Interstitial) grade16) was used as the starting material. Its

chemical composition in wt.% was 5.96 Al, 3.88 V, 0.10 Fe, 0.001 C, 0.097 O, 0.006 N, 0.0076 H with the balance being

Ti and its beta transus was975_{C. The samples of 40 mm}

diameter and 60 mm height were annealed at 990_{C for}

30 min. After this annealing, either air cooling (AC) or

furnace cooling (FC) was employed to produce a thin or thick lamellar structure, respectively. In the case of AC, the cooling rate measured at the center of the sample was

100_{C/min, while in FC, the samples were cooled to 700C}

with a cooling rate of 1_{C/min in the furnace and then air}

cooled to room temperature. The microstructures of the AC and FC samples are shown in Fig. 1. Microstructure of the

AC sample (Fig. 1(a)) consisted of thinlamellae of1mm

in thickness and thick grain boundary layer. For the FC

boundaryincreased due to the slow cooling rate, and the

average thickness oflamellae was8mm.

2.2 Non-isothermal multi-step forging

The AC and FC samples were subjected to multi-step forging under the non-isothermal condition. At first, a sample

was heated at the forging start temperature, i.e. 900C or

940C, for 30 min prior to deformation, and then upset to

40 mm height. After reheating to the forging start temper-ature for 10 min to compensate for tempertemper-ature drop during upsetting, the upset sample was stretched to 36 mm thickness by side pressing. The side pressing was repeated 6 times by rotating a sample properly so that the initial height of sample (60 mm) was restored. In this case, intermediate reheating was employed after every two side pressing operation. The changes of the sample shape during upsetting and side pressing are shown in Fig. 2(a), and the initial and final shapes of the sample are compared in Fig. 2(b). A series of one upsetting and six side pressing operations was repeated 3 times. The accumulated macroscopic true strain of the

sample was2:4after the completion of multi-step forging.

All forging was carried out between the two flat dies heated at

350C with a ram speed of 5 mm/s. After final forging, the

sample was air cooled without additional heat treatment. In

addition, in order to examine the microstructural evolution due to intermediate reheating, the samples were reheated to the forging start temperature for 30 min, and then either water quenched or air cooled several times without forging.

The microstructures were observed at the center of the longitudinal plane of the forged samples by optical mi-croscopy and scanning electron mimi-croscopy. The size and

volume fraction ofglobules were measured and analyzed

quantitatively by an image analyzer.

3. Results

3.1 Microstructures after multi-step forging 3.1.1 940_{C forging}

The microstructures of the AC and FC samples

multi-forged at 940_{C are shown in Fig. 3. The microstructure of}

the AC sample (initially thin lamellar structure) exhibited the

bimodal structure of equiaxed particles and interparticle

lamellae (Fig. 3(a)). By contrast, elongatedparticles with

high aspect ratio (Fig. 3(c)) were observed in the FC sample (initially thick lamellar structure) instead of equiaxed ones. It

is noticed that the thickness of elongatedparticles of the FC

sample was nearly the same as the diameter of equiaxed

particles in the AC sample. In addition, while equiaxed

αα

β

Fig. 1 SEM micrographs of the present Ti-6Al-4V alloy afterannealing at 990_{C for 30 min. (a) air cooled (AC) sample (b) furnace}

cooled (FC) sample.

after upsetting

(b) (a)

before forging after forging

the 2nd side pressing the 4th side pressing the 6th side pressing reheating

repeat 3 times

reheating reheating

particles were evenly distributed in the AC sample

(Fig. 3(b)), the traces of the lamellar colonies and thick

layers at the prior grain boundaries remained even after

multi-step forging in the FC sample (Fig. 3(d)).

3.1.2 900_{C forging}

Figure 4 shows the optical micrographs of the AC and FC

samples multi-step forged at 900_{C. The microstructure of}

the AC sample exhibited a mixture of equiaxed particles

and elongatedlamellae (Fig. 4(a)). At a low magnification

(Fig. 4(b)), the traces of smallcolonies of50mmin size

could be identified. In the FC sample, the severely deformed

but unbroken lamellae were preserved inside distorted

colonies after forging as in Fig. 4(c) and (d), in which both

lamellae and colonies were more distorted than those in the

FC sample multi-step forged at 940_{C (Fig. 3(d)).}

3.2 Quantitative comparison of multi-step forged mi-crostructures

In this section, the characteristics of the final micro-structures of non-isothermal multi-step forged samples are analyzed quantitatively. The values of the microstructural

features of the AC and FC samples after forging at 940_{C and}

900_{C are summarized and compared in Table 1.}

3.2.1 Aspect ratio

The distributions of the aspect ratio (L=D;LandDare the

longitudinal and short transverse dimensions of the particle,

(a)

(b)

(c)

(d)

equiaxed

αα

α+β

elongated

α

Fig. 3 Optical micrographs after non-isothermal multi-step forging with the forging start temperature of 940_{C: (a) and (b) the AC}

sample, (c) and (d) the FC sample.

(a)

(b)

(c)

(d)

Fig. 4 Optical micrographs after non-isothermal multi-step forging with the forging start temperature of 900_{C: (a) and (b) the AC}

respectively) of particles after forging are compared in Fig. 5. At the same forging start temperature, it is obvious

that the thin lamellar structure (i.e.the AC sample) was more

effective for globularization than the thick lamellar structure

(i.e. the FC sample). For instance, at 940_{C, the maximum}

frequency peak of the AC sample appeared at L=D<2,

which indicates full globularization, but that of the FC

sample appeared atL=D<3(Fig. 5(a)) with broad

distribu-tion. The same trend was also found at 900_{C (Fig. 5(b)).}

However, by decreasing the forging start temperature from

940_{C to 900C, the fraction of particles with L=D<2}

decreased from 92% to 39% for the AC sample, and from 17% to 5% for the FC sample. This result shows that a decrease in the forging start temperature suppresses globula-rization in both AC and FC samples.

particles, and the thickness was measured for elongated or

lamellar ones. After multi-step forging at 940_{C (Fig. 6(a)),}

the diameter of equiaxed particles of the AC sample

(Fig. 3(a)) was nearly the same as the thickness of elongated

particles of the FC sample (Fig. 3(c)), though these samples

had very different initial lamellar thickness as shown in

Fig. 1. After multi-step forging at 900_{C (Fig. 6(b)), the}

particle size of the AC sample less increased (from1mmto

4mm) as compared with that forged at 940_{C showing a}

significant increase (from 1mmto 7mm), while a small

difference was observed in the FC sample.

3.3 Effect of intermediate reheating during non-isother-mal multi-step forging

Thermal fluctuation as well as plastic deformation during non-isothermal multi-step forging could affect the final microstructure of the forged sample. In order to examine the sole effect of intermediate reheating and subsequent cooling during non-isothermal multi-step forging on glob-ularization, the AC and FC samples were reheated to the

forging start temperature (i.e.940_{C or 900C) for 30 min,}

and then either water quenched or air cooled repeatedly without forging. By water quenching, the shape and volume

fraction of lamellae at the forging temperature can be

preserved, while the matrix phase is transformed to

martensite. Therefore, information on the instantaneous microstructural state just prior to forging operation can be

morphology equiaxed mixed (equiaxed and elongated) elongated deformed lamellar

average size of and aspect ratio

L: 10.9mm D: 7.4mm L=D: 1.5

L: 9.6mm D: 4.3mm L=D: 2.3

L: 22.5mm D: 8.4mm L=D: 2.6

L: 23.7mm D: 7.3mm L=D: 3.2

colony trace No Yes Yes Yes

volume fraction of globularized particles

(L=D<2)

92% 39% 17% 5%

1 2 3 4 5 6 7 8

0 20 40 60 80 100

940°C AC sample FC sample

F

raction (%)

Aspect Ratio, L/D

1 2 3 4 5 6 7 8

0 20 40 60 80 100

900°C AC sample FC sample

F

raction (%)

Aspect Ratio, L/D

(a) (b)

Fig. 5 Histograms showing the distribution of the aspect ratio ofparticles in the AC and FC samples after multi-step forging at (a) 940_{C, and (b) 900C.}

(a) (b)

2 3 4 5 6 7 8 9 10 11 12 13 14

0 10 20 30 40 50

940°C AC sample FC sample

F

raction (%)

α Particle Size, D/µm

2 3 4 5 6 7 8 9 10 11 12 13 14

0 10 20 30 40 50

900°C AC sample FC sample

F

raction (%)

α Particle Size, D/µm

Fig. 6 Histograms showing the distribution of theparticle size in the AC and FC samples after multi-step forging at (a) 940_{C, and}

obtained. Similarly, by air cooling, the microstructural change during air cooling after multi-step forging can be inferred.

3.3.1 AC sample

The microstructures of the AC sample subjected to

repetitive thermal cycle without forging at 940_{C and}

900_{C are shown in Figs. 7 and 8, respectively. A}

compar-ison of Figs. 7(a) and 7(c) reveals that lamellae became

thicker and shorter, and their volume fraction decreased by

reheating at 940_{C. Subsequent air cooling after reheating}

increased thelamellar thickness as shown in Figs. 7(b) and

7(d). The same trend was observed in the thermal cycle at

900_{C (Fig. 8), but less significant.}

3.3.2 FC sample

Figures 9 and 10 represent the microstructures of the FC

sample thermally cycled at 940_{C and 900C, respectively. In}

the case of the FC sample, both thickness and volume

fraction oflamellae decreased during reheating compared

to the microstructure of the initial FC sample, Fig. 1(b). But, similar to the AC sample, they slightly increased again after subsequent air cooling: for example, a comparison between Figs. 9(c) and 9(d). It is also noticed that a fine lamellar

structure existed inside the individuallamellae which was

transformed from thephase,i.e.the area ‘A’ in Figs. 9 and

10. This fine lamellar structure, which was not observed in the initial state of Fig. 1(b), resulted from the faster cooling rate than that of the furnace cooling.

The variation of thelamellar thickness and its volume

fraction with the number of thermal cycle is summarized in Figs. 11 and 12, respectively. It is of interest to note that the

increase of the lamellar thickness of the AC sample at

940C was the most significant such that it reached that of the

FC sample after 5 thermal cycles at the same temperature.

The volume fraction oflamellae was more affected by the

temperature and the initial lamellar thickness rather than the number of thermal cycle. As shown in Fig. 12, after the rapid

decrease by the first thermal cycle, the volume fraction of

lamellae remained nearly unchanged with further thermal cycle at all conditions. But, the initial decreases in both AC

and FC samples by 940_{C thermal cycle were more}

significant than those by 900_{C thermal cycle and, at the}

same temperature, the dissolution ofin the FC samples was

significant compared to that in the AC samples.

4. Discussion

4.1 Globularization under non-isothermal condition

In the preceding sections, the globularization behavior of Ti-6Al-4V alloy subjected to non-isothermal multi-step forging was analyzed. According to the globularization

mechanism proposed by Margolin and Cohen6)and Weisset

al.,7) at first - grain boundaries were formed across

lamellae during deformation by dynamic recrystallization.

Then the phase penetrates into these - boundaries to

reduce the interface energy. Finally, equiaxedparticles are

formed through severance and/or partition of lamellae by

complete penetration of thephase. For recrystallization of

thephase, sufficient strain should be imposed onlamellae

beforepenetration, and the amount of strain required varies

with the lamellar thickness. It is known that the phase of

bcc structure is softer and more deformable than thephase

of hcp structure at conventional forging temperatures ofþ

titanium alloys.17,18) Accordingly, the amount of strain

accumulated onlamellae associated with their

recrystalli-zation depends on the thickness and volume fraction of

lamellae.

It is worth mentioning that, under non-isothermal

con-dition, the thickness and volume fraction of lamellae

(a)

(b)

(c)

(d)

αα

Martensite (

β

)

Fig. 7 Microsturctural changes of the AC sample during thermal cycle (reheating and cooling at 940_{C) without multi-step forging:}

change continuously, and the amount of stain imposed on them varies due to diffusional relaxation at high temper-atures. This makes it difficult to analyze the separate effect of deformation and thermal fluctuation on globularization. However, only in terms of thermal fluctuation, the micro-structures formed at each forging step in the present non-isothermal forging are expected to exhibit the similar trend of microstructural evolution to that described in Section 3.3.

That is, the microstructural changes during cooling and reheating for the next forging step of either deformed lamellae or recrystallized globules formed by one forging step can be deduced from Figs. 11 and 12. In the next, the globularization behavior observed in this study is discussed in association with instantaneous microstructural develop-ment in the deformation process considering the thermal fluctuation effect.

(c)

(d)

Fig. 8 Microsturctural changes of the AC sample during thermal cycle (reheating and cooling at 900_{C) without multi-step forging:}

(a) the first reheating and water quenching, (b) the first reheating and air cooling, (c) the fifth reheating and water quenching, and (d) the fifth reheating and air cooling.

(a)

(b)

(c)

(d)

A

αα

Martensite (

β

)

A

Fig. 9 Microsturctural changes of the FC sample during thermal cycle (reheating and cooling at 940_{C) without multi-step forging: (a) the}

4.2 Globularization of the AC and FC samples 4.2.1 AC samples

From a comparison between Fig. 3(a) (940_{C) and}

Fig. 4(a) (900_{C), it is obvious that globularization at the}

higher forging start temperature was more effective than

that at the lower forging start temperature, but the

lamellar thickness of the latter was smaller than the

globule diameter of the former. The previous result regarding the effect of the forging temperature on the globularization behavior of the same alloy under the

isothermal compression8,13)showed that as the temperature

decreased, deformation of lamellae became non-uniform. A similar argument can be applied to the present case. In addition to the lower forging temperature, non-uniform deformation would be more prevalent due to concurrent

cooling of the sample during non-isothermal forging, resulting in less globularization in the AC sample forged

at 900_{C. As shown in Fig. 11, the lamellar thickness of}

the sample thermally cycled at 940_{C increased faster than}

that of the sample thermally cycled at 900_{C with the}

number of thermal cycle. This observation can be explained by the followings. The lamella thickening is more

signifi-cant at the end of lamella (i.e.lamella node) due to the fault

migration.19,20)_{Dynamic recystallization at 940C was more}

active than at 900_{C due to high temperature, resulting in}

more recrystallized particles with the low aspect ratio

which were formed through severance and/or partition of

lamellae by complete penetration of the phase. Then, the

lamella thickening in the recrystallized particles with the

low aspect ratio resulted in the more equiaxed globules

(a)

(b)

(c)

(d)

A

A

Fig. 10 Microsturctural changes of the FC sample during thermal cycle (reheating and cooling at 900_{C) without multi-step forging:}

(a) the first reheating and water quenching, (b) the first reheating and air cooling, (c) the fifth reheating and water quenching, and (d) the fifth reheating and air cooling.

0 2 4 6 8 10

Number of Thermal Cycle, n

5 3

FC 940°C FC 900°C

AC 940°C

α

Lamellar

Thic

kness (

µ

m)

AC 900°C

1 0

solid : during heating open : after cooling

Fig. 11 The variation of thelamellar thickness in the AC and FC samples during thermal cycle without multi-step forging.

0 20 40 60 80 100

Number of Thermal Cycle, n

5 3

FC 940°C FC 900°C

AC 940°C

F

raction of

α

Lamellae (%)

AC 900°C

1 0

with a relatively large diameter. By contrast, at 900_C, lamellae with the high aspect ratio were formed due to less recrystallization and non-uniform deformation, and the lamella thickening in these lamellae was to be limited at the end of each lamella. This interpretation of thermal effect

can be extended to explain the fact thatglobule diameter

of the sample forged at 940_{C (Fig. 3(a)) was larger than}

the lamellar thickness of the sample forged at 900C

(Fig. 4(a)).

4.2.2 FC samples

Globularization of the FC sample forged at 940_{C resulted}

in the elongatedparticles with high aspect ratio (Fig. 3(c)),

while the microstructure of the FC sample forged at 900_C

(Fig. 4(c)) showed the distorted colony structure consisting

of severely deformed lamellae, indicating less effective

globularization in the latter. As listed in Table 1, the

lamellar thickness change in the FC sample was not sensitive to the forging temperature and this result was in agreement with that observed in the thermally cycled sample (Fig. 11). Figure 4(c) reveals that there are many irregular grooves

(indicated by arrows) on the boundaries of deformed

lamellae at 900C forging, indicating the incomplete

glob-ularization. As aforementioned, globularization oflamellae

is known to occur by penetration of thephase through such

grooves called ‘cusp’ during dynamic and/or static

recrys-tallization processes.9,21) _{In the case of non-isothermal}

deformation at 900_{C, penetration becomes difficult}

compared to that at 940_{C due to not only the lower}

temperature but also concurrent cooling of the sample during

forging under non-isothermal condition.22) _{In addition,}

although the sample was reheated to forging temperature, the time was not sufficient for globularization. Eventually,

unlike forging at 940_{C,lamellae are not globularized, but}

only deformed at 900_{C forging.}

4.3 Comparison of globularization behavior of the AC and FC samples

The instantaneous microstructures just prior to forging can be inferred from the water quenched samples after reheating.

Their characteristics shown from Fig. 7 to Fig. 12 can be schematically summarized as Fig. 13. At the same forging temperatures, the instantaneous microstructures of the FC

sample are manifested by thicker and lamellae but the

smaller volume fraction of thephase compared to the AC

sample. Such lamellar configuration of the FC sample just prior to forging is not favorable for globularization because

thicklayers are likely to deform preferentially than thick

lamellae, and therefore only a small portion of strain will be

accumulated on lamellae, resulting in insufficient plastic

deformation of lamellae for recrystallization. Eventually,

the final microstructure of the FC sample after non-isothermal multi-step forging exhibits the traces of deformed

colonies and elongatedparticles as shown in Fig. 3(c) and

Fig. 3(d). In the same point of view, the AC sample, which

has both thin and lamellar structure in addition to the

higher volume fraction of thephase, is thought to be more

favorable for the recrystallization than the FC sample, resulting in the more homogeneous distribution of fully

globularizedparticles as shown in Fig. 3(a) and Fig. 3(b).

As mentioned earlier, the diameter of theglobules of the

AC sample was comparable to the thickness of elongated

particles of the FC sample after forging at 940C in spite of

very different initial lamellar thickness. This can be ex-plained in terms of the microstructural development due to thermal fluctuation. As shown in Fig. 11, the thickness of initial thin lamellae of the AC sample increased continuously with increasing the number of thermal cycle, and became nearly the same as that of the FC sample after the 5 thermal

cycles at 940_{C. The similar result of lamella thickening}

during isothermal compression of very thin lath structure

was reported by Shell and Semiatin.12)_{In their experiment,}

the thickness oflaths increased with the heating

temper-ature and holding time. It is worth mentioning that the

thickness change of lamellae of the AC sample during

cooling was larger than that during reheating as seen in

Fig. 11, indicating that the coarsening ofglobules can be

enhanced by repetition of cooling during non-isothermal multi-step forging.

f_α= 90%, t_α= 1µm

α β

sample

FC sample

f_α= 70%, t_α= 1µm f_α= 50%, t_α= 1µm

f_α= 90%, t_α= 8µm f_α= 55%, t_α= 3.5µm f_α= 35%, t_α= 3µm

5. Summary

(1) The globularization behavior of ELI grade Ti-6Al-4V alloy with either initially thin or thick lamellar structure during non-isothermal multi-step forging was

investi-gated. It was found that globularization of lamellae

was significantly affected by the instantaneous micro-structural state developed by thermal fluctuation during non-isothermal multi-step forging.

(2) After multi-step forging at a start temperature of 940_C,

the alloy with initially thin lamellar structure exhibited

homogeneous distribution of equiaxedglobules, but

elongated globules with high aspect ratio was

dominant in the alloy with initially thick lamellar structure.

(3) After the forging at 900C, the alloy with initially thin

lamellar structure exhibited a mixed structure of

equiaxedparticles and elongatedlamellae. On the

contrary, the microstructure of the alloy with initially thick lamellar structure consisted of severely deformed lamellae inside the distorted colonies with little glob-ularization. In case of the present non-isothermal multi-step forging, the more effective globularization occur-red at higher forging start temperature.

(4) To obtain equiaxed microstructure by non-isothermal multi-step forging, a thin lamellar structure is to be used as a starting microstructure. A fully globularized but coarsened structure was obtained at higher forging start

temperature. The coarsening of globules was

sup-pressed, but globularization became less effective at the same time with decreasing the forging start temper-ature. In this case, the higher deformation and/or post annealing heat treatment should be applied to acquire

finer equiaxedglobules.

REFERENCES

1) G. W. Kuhlman:Microstructure/Property Relationships in Titanium Aluminides and Alloys, ed. by Y.-W. Kim and R. R. Boyer, (TMS, Warrendale, Pennsylvania, 1991) pp. 465–491.

2) G. Lu¨tjering, J. Albrecht and A. Gysler: Titanium ’92 Science and

Technology, Proc. of the 7th Int. Conf. on Titanium, ed. by F. H. Froes and I. Caplan, (TMS, Warrendale, Pennsylvania, 1993) pp. 1635–1646. 3) R. R. Boyer and G. Lu¨tjering:Advances in the Science and Technology of Titanium Alloy Processing, ed. by I. Weiss, R. Srinivasan, P. J. Bania, D. Eylon and S. L. Semiatin, (TMS, Warrendale, Pennsylvania, 1996) pp. 349–367.

4) J. R. Wood and P. A. Russo:Advances in the Science and Technology of Titanium Alloy Processing, ed. by I. Weiss, R. Srinivasan, P. J. Bania, D. Eylon and S. L. Semiatin, (TMS, Warrendale, Pennsylvania, 1996) pp. 421–433.

5) M. Peters and G. Lu¨etjerung: Titanium ’80 Science and Technology, Proc. of the 4th Int. Conf. on Titanium, ed. by H. Kimura and O. Izumi, (AIME, Warrendale, Pennsylvania, 1980) pp. 925–935.

6) H. Malgolin and P. Cohen: Titanium ’80 Science and Technology, Proc. of the 4th Int. Conf. on Titanium, ed. by H. Kimura and O. Izumi, (AIME, Warrendale, Pennsylvania, 1980) pp. 1555–1561.

7) I. Weiss, G. E. Welsch and F. H. Froes: Titanium Science and Technology, Proc. of the 5th Int. Conf. on Titanium, ed. by G. Lu¨tjering, U. Zwicker and W. Bunk, (DGM, Germany, 1985) pp. 1503– 1510.

8) S. L. Semiatin, V. Seetharaman and I. Weiss: Advances in the Science and Technology of Titanium Alloy Processing, ed. by I. Weiss, R. Srinivasan, P. J. Bania, D. Eylon and S. L. Semiatin, (TMS, Warrendale, Pennsylvania, 1996) pp. 3–73.

9) I. Weiss, F. H. Froes, D. Eylon and G. E. Welsch: Metall. Trans. A17A (1986) 1935–1947.

10) M. Burn, N. Anoshkin and G. Shakhanova: Mater. Sci. & Eng. A243 (1998) 77–81.

11) P. Ari-Gur and S. L. Semiatin: Mater. Sci. & Eng. A257(1998) 118– 127.

12) E. B. Shell and S. L. Semiatin: Metall. Mater. Trans. A30A(1999) 3219–3229.

13) S. L. Semiatin, J. F. Thomas and P. Dadras: Metall. Trans. A14A (1983) 2363–2374.

14) S. L. Semiatin, V. Seetharaman and I. Weiss: Mater. Sci. & Eng. A263 (1999) 257–271.

15) R. Ding, Z. X. Guo and A. Wilsom: Mater. Sci. & Eng. A327(2002) 233–245.

16) AMS 4930F (2000.6), Aerospace Material Specification, Titanium Alloy Bars, Wire, Forgings, and Rings: Ti-6Al-4V, Extra Low Interstitial Annealed.

17) S. L. Semiatin and T. R. Bieler: Acta Mater.49(2001) 3565–3573. 18) I. Weiss and S. L. Semiatin: Mater. Sci. & Eng. A243(1998) 46–65. 19) H. E. Cline: Acta Metall.19(1971) 481–490.

20) L. Y. Lin, T. H. Courtney and K. M. Ralls: Acta Metall.25(1977) 99– 106.

21) N. Stefansson and S. L. Semiatin: Metall. Mater. Trans. A34A(2003) 691–698.