Materials Science and Engineering A 494 (2008) 132–138

Contents lists available at ScienceDirect

Materials Science and Engineering A

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

The elevated-temperature fatigue behavior of boron-modifiedTi–6Al–4V(wt.%) castings

W. Chen, C.J. Boehlert ∗

Department of Chemical Engineering and Materials Science, Michigan State University, East Lansing, MI 48824-1226, USA

a r t i c l e i n f o

Article history:Received 30 January 2008Received in revised form 2 April 2008Accepted 3 April 2008

Keywords:

a b s t r a c t

This work investigated the effect of nominal boron additions of 0.1 and 1.0 wt.% on the elevated-temperature (455 ◦C) fatigue deformation behavior of Ti–6Al–4V(wt.%) castings for maximum appliedstresses between 250 and 450 MPa (R = 0.1 and 5 Hz). Boron additions resulted in a dramatic refinementof the as-cast grain size, and larger boron additions resulted in larger titanium-boride (TiB) phase volumepercents. The boron-containing alloys exhibited longer average fatigue lives than those for Ti–6Al–4V,

FatigueMicrostructureTB

which was suggested to be related to the reduced as-cast grain size and the addition of strong and stiff TiBphase. The Ti–6Al–4V–0.1B alloy exhibited the longest average fatigue lives. The TiB phase cracked during

nd thase cr

1

bdtwRaa([mai(satadfitpc

v

tamfftba(Baa

2

2

m

0d

itaniumoron

the fatigue experiments aEach alloy exhibited �-ph

. Introduction

Conventional titanium (Ti) alloys containing small additions oforon (B) have been gaining considerable interest in recent yearsue to their attractive mechanical properties including high roomemperature (RT) specific strength and Young’s modulus (E) alongith reasonable elongation-to-failure (εf) [1–8]. Table 1 lists theT tensile properties of as-cast Ti–6Al–4V–xB(wt.%)1 (0 ≤ x ≤ 1)lloys [7,8]. The significant increase in strength and E in Ti–Blloys arise from the strong and stiff titanium-boride (TiB) phaseE ∼ 371–482 GPa) that precipitates in situ during solidification2,5]. TiB reinforcement in Ti is advantageous over other reinforce-

ents as there exists a crystallographic arrangement between Tind TiB resulting in a clean interface without significant reactiv-ty [9]. Due to the similarity of the thermal expansion coefficients8.2 × 10−6 ◦C−1 for Ti and 6. 2 × 10−6 ◦C−1 for TiB), processing ofuch alloys is less challenging than for other ceramic-reinforcedlloys. In addition, recent work has shown that the addition ofrace amounts (∼0.1 wt.%) of B to Ti–6Al–4V(wt.%) decreases thes-cast grain size by approximately an order of magnitude [3]. Thisrastic reduction in the as-cast grain size leads to significant bene-

ts including increasing yield strength while reducing or avoidingime spent on expensive and energy-intensive thermomechanicalrocessing. Small B additions do not result in significant densityhanges, thus the higher strength and stiffness of Ti–B alloys pro-

∗ Corresponding author. Tel.: +1 517 353 3703; fax: +1 517 432 1105.E-mail address: boehlert@egr.msu.edu (C.J. Boehlert).

1 All alloy compositions are given in weight percent.

seFc2f

cm

921-5093/$ – see front matter © 2008 Elsevier B.V. All rights reserved.oi:10.1016/j.msea.2008.04.004

is resulted in a decreasing Young’s modulus with increased cycle number.acking and environmentally assisted surface edge cracking.

© 2008 Elsevier B.V. All rights reserved.

ides important increases in specific strength and stiffness.Although Ti–B alloys are currently used for various ambient-

emperature commercial applications (e.g. exhaust valves ofutomotive engines) [1], understanding the mechanisms of defor-ation and fracture is necessary to consider these materials

or fracture-critical applications (e.g. aerospace) especially underatigue loading conditions. To date no studies of the elevated-emperature fatigue behavior of cast Ti–6Al–4V–xB alloys haveeen performed. This study focused on the deformation evolutionnd mechanisms in Ti–6Al–4V–xB during elevated-temperature455 ◦C) fatigue. The main objectives were to access the effects of

additions on the fatigue behavior of Ti–6Al–4V and to evaluatend understand the fatigue deformation behavior of Ti–6Al–4V–xBlloys as a function of stress level and B concentration.

. Experimental

.1. Materials and processing

Table 2 lists the compositions of the alloys, which wereeasured through inductively coupled plasma optical emission

pectroscopy and inert gas fluorescence. Castings of 70 mm diam-ter and 500 mm length were produced for each composition atlowserve Corporation, Dayton, OH via induction skull melting. Allastings were hot isostatically pressed at 900 ◦C and 100 MPa for

h. No post-processing heat treatments were performed prior to

atigue testing.The as-processed materials were prepared for imaging using

onventional metallography techniques and scanning electronicroscopy (SEM) was used to examine the as-cast grain size, the

W. Chen, C.J. Boehlert / Materials Science and Engineering A 494 (2008) 132–138 133

Table 1Room temperature properties of Ti–6Al–4V–xB castings

Alloy YS (MPa) UTS (MPa) RT εf (%) E (GPa)

Ti–6Al–4V [7] 834 900 6 107TTTT

�gI(Fl�

2

cwfifbasttdmidic5sfstamctows3ta

3

3

ft

Table 3TiB and � phase volume percents (Vp) and �- and �-lath widths (�)

Alloy Vp TiBa Vp �a �� (�m) �� (�m)

Ti–6Al–4V as-cast 0 12.8 ± 2.6 3.8 ± 1.1 0.6 ± 0.1TT

isTluwraaiTbrB

3

i4itNTtts

iciTlacuttpopa

TC

A

TTT

i–6Al–4V–0.05B [7] 879 944 8.5 110i–6Al–4V–0.1B [7] 900 966 7.3 116i–6Al–4V–0.4B [7] 913 973 5.2 126i–6Al–4V–1B [8] 975 1045 2.5 146

- and �-lath widths, the phase distributions and their morpholo-ies. Phase volume percents and lath widths were determined usingmageJ image analysis software on several backscatter electronBSE) SEM photomicrographs acquired using either a CamScan44FEield Emission SEM or a Cambridge Stereoscan 250 Mk 2 SEM. Theath widths were measured for over 300 laths for each of the � and

phases, for each alloy composition, via the line-intercept method.

.2. Fatigue experiments

Dogbone-shaped specimens were machined via electrical dis-harge machining (EDM). Recast layers, which formed during EDM,ere removed through silicon carbide paper grinding to a 600-gritnish. Open-air, tension–tension fatigue experiments were per-

ormed on a horizontal servohydraulic frame previously describedy Hartman and Russ [10]. The test temperature was 455 ◦C and thepplied maximum stresses ranged between 250 and 450 MPa with atress ratio of 0.1. Strain was monitored using an alumina-rod high-emperature extensometer, with a 12 mm gage length, attachedo the gage section of the specimen. The strain-life behavior wasocumented throughout the experiments and both the Young’sodulus and hysteresis behavior were tracked at certain cycles dur-

ng the experiments. In such cases, approximately 20 stress–strainata points were acquired throughout the desired cycle. The spec-

mens were locally heated using a Barber Coleman’s temperatureontroller and two sets of heating banks, located approximatelymm above and below the sample, each containing four evenly

paced quartz lamps. Specimen temperatures were monitored byour chromel–alumel type-K thermocouples located within thepecimen’s gage section. Targeted test temperatures were main-ained within ±5 ◦C. The test specimens were soaked at 455 ◦C fort least 20 min prior to applying load in order to minimize the ther-al stresses. Runout was determined to be greater than 1,000,000

ycles and the test frequency was 5 Hz. For the fractured samples,he furnace was turned off immediately upon fracture to preventxidation of the exposed surfaces. A minimum of three experimentsere performed for each alloy at the following maximum applied

tress levels: 350, 400, and 450 MPa. Maximum stresses of 250 and00 MPa were also evaluated. The fractured surfaces and gage sec-ions of the deformed samples, for both samples taken to failurend runout samples, were evaluated using SEM.

. Results

.1. Microstructure

Table 3 lists the average volume percents of the TiB and � phasesor each alloy along with the average �- and �-lath widths. Pho-omicrographs of the microstructures for each alloy are illustrated

wi0ow

able 2ompositions of the studied alloys

lloy Ti (wt.%) Al (wt.%) V (wt.%) B (wt.%)

i–6Al–4V Balance 6.6 4.1 0i–6Al–4V–0.1B Balance 6.0 4.0 0.1i–6Al–4V–1B Balance 5.9 4.0 0.98

i–6Al–4V–0.1B as-cast 0.6 ± 0.3 14.4 ± 2.3 4.2 ± 0.5 0.8 ± 0.1i–6Al–4V–1B as-cast 5.8 ± 1.8 15.5 ± 2.0 3.6 ± 0.8 0.7 ± 0.2

a The �-phase was the balance in the microstructure.

n Fig. 1. The addition of 0.1B to Ti–6Al–4V refined the as-cast grainize (referred to as the prior-� grain size) from 1700 to 200 �m.he photomicrographs in Fig. 1 show that higher B concentrationsead to higher TiB phase (darkest phase) volume percents. The vol-me percents of TiB in the Ti–6Al–4V–0.1B and Ti–6Al–4V–1B alloysere 0.6 and 5.8, respectively. The volume percents of the �-phase

emained relatively constant (13 < Vp � < 16) for each alloy. The 0.1Bddition reduced the thickness of the grain boundary �-phase frompproximately 10 �m to approximately 2 �m, which is expected tomprove the workability of this alloy system and increase the εf [6].he TiB phase formed a necklace structure around the prior-� grainoundaries in the Ti–6Al–4V–0.1B alloy, see Fig. 1b, which helpedestrict grain growth. The average �-lath width varied slightly withconcentration and was always between 3.6 and 4.2 �m.

.2. Fatigue behavior

The fatigue S–N curves obtained for the as-cast alloys are shownn Fig. 2. For the maximum stress levels of 350 MPa, 400 MPa, and50 MPa, the three Ti–6Al–4V–0.1B specimens tested each exhib-

ted longer lives than the specimens of the other alloys. Comparinghe Ti–6Al–4V–1B and Ti–6Al–4V alloys, there was overlap in thef ranges exhibited at a given maximum stress level, though thei–6Al–4V–1B alloy exhibited greater average Nf values. It is notedhat the S–N behavior for the Ti–6Al–4V alloy was quite similar tohat for a cast-then-forged Ti–6Al–4V alloy, with an �-lath widthize between 1 and 2 �m, tested at 450 ◦C [11].

Hysteresis plots of selected samples during defined cycles arellustrated in Fig. 3. The E tended to decrease with increasingycle number indicating that damage was occurring in the spec-mens. The damage was observed to be in the form of bothiB-phase cracking and �-phase cracking which will be shownater. An E value decrease of 14% was the maximum decrease for

Ti–6Al–4V–0.1B sample tested at �max = 400 MPa (E = 88 GPa atycle 100 and E = 76 GPa at cycle 900,000; Nf = 990,554). However,sually only an E decrease of 5% was exhibited for the samples prioro fracture. For all of the runout samples, an E value decrease of lesshan 5% was observed. No phase instability was observed on theost-fatigue gage sections even after more than two million cyclesf fatigue exposure at 455 ◦C. In such cases the measured volumeercents of the TiB, �, and � phases were similar both before andfter the experiments.

For the B-containing alloys, TiB-phase cracking was evident

ithin the deformed gage sections (see Fig. 4b and c). This behav-

or was similar to that for a Ti–6Al–4V/TiB composite containing.5 wt.%B [4]. The �-phase also exhibited a significant amountf cracking, see Fig. 4a, while cracks were only rarely observedithin the fine �-phase. For each alloy, cracks were observed along

Fe (ppm) O (ppm) N (ppm) C (ppm) H (ppm)

2300 1800 100 200 801300 1500 100 200 501300 1800 200 200 90

134 W. Chen, C.J. Boehlert / Materials Science and Engineering A 494 (2008) 132–138

F b) Ti–T

tdpoi

ilmbrdrostdc

wTtfToTt

4

4

ig. 1. Low- and high-magnification backscattered electron SEM images of the (a andhe TiB phase is black, the �-phase is gray, and the �-phase is white.

he edges of the samples and grew perpendicular to the loadingirection (see Fig. 5). Thus the environment affected the fatigueerformance and it is felt that the �-case on the surface, due to thexidation, assisted this surface cracking. The final failure was likelynitiated by one of these cracks which grew into the sample.

SEM photomicrographs of fracture surfaces for each alloy arellustrated in Figs. 6 and 7. Fracture initiation sites were typicallyocated at surface locations (see Fig. 6a and b). In some cases,

ultiple surface crack initiation sites were evident. Fig. 6a andillustrates the crack initiation, crack propagation, and overload

egions for a Ti–6Al–4V specimen. The overload regions containeductile dimples (see for example Figs. 6b and 7c). The crack growthegions of the sample consumed a large extent of the thickness

f the samples (see Fig. 8). Fig. 8 illustrates the environmentalusceptibility of the alloys as the colored region on the frac-ure surface indicated the crack growth region which oxidizeduring the experiment. The shiny metallic regions outside theolored region in Fig. 8 represent the overload regions/surfaces

teTt

6Al–4V, (c and d) Ti–6Al–4V–0.1B, and (e and f) Ti–6Al–4V–1B alloy microstructures.

hich were not oxidized. The fracture surface appearance of thei–6Al–4V–1B specimens (Fig. 7b) was somewhat different thanhat for the Ti–6Al–4V and Ti–6Al–4V–0.1B specimens as largeraceted regions were evident and fewer dimples were evident.he fracture surface of Ti–6Al–4V–0.1B indicated a more tortu-us crack path than that for Ti–6Al–4V–1B. In some cases, thei–6Al–4V–1B fracture surfaces exhibited cleavage, indicating brit-le fracture.

. Discussion

.1. Microstructure

The �-lath width is controlled by the thermodynamics duringhe � → � phase transformation. The TiB and � phases form in theutectic reaction [6]. When the B concentration is low, such as ini–6Al–4V–0.1B, the TiB phase primarily forms a necklace struc-ure around the prior-� grains (see Fig. 1b) as has been described

W. Chen, C.J. Boehlert / Materials Science an

Fm

pTbftw

iTmewmsmiaffow�i�w

4

The improved average fatigue lives of the B-modified alloyscompared with Ti–6Al–4V is believed to be due to the microstruc-

F4

ig. 2. Maximum applied stress versus cycles to failure curves where runout speci-ens are marked with an arrow.

reviously [3]. When the B concentration is higher, such as ini–6Al–4V–1B, the TiB phase forms not only at the prior-� grainoundaries, but also inside the grains (see Fig. 1c). The TiB phase

acilitates the nucleation of more � laths [12]. Thus it is expectedhat larger TiB volume percents would result in a reduced �-lathidth during the � → � phase transformation.

tbt

ig. 3. Stress versus strain plots at the marked cycle numbers for (a) Ti–6Al–4V tested at50 MPa, and (c) Ti–6Al–4V–1B tested at a maximum stress of 300 MPa.

d Engineering A 494 (2008) 132–138 135

Sen et al. [7] studied the fatigue crack growth behav-or of as-cast Ti–6Al–4V, Ti–6Al–4V–0.05B, Ti–6Al–4V–0.1B, andi–6Al–4V–0.4B cast alloys. The Ti–6Al–4V and Ti–6Al–4V–0.1Baterials of their study were cut from the same ingot as those

xamined in the current study. They observed decreasing �-lathidths for the alloys containing greater B contents. The �-widtheasurement (6.8 ± 1.8 �m) for their as-cast Ti–6Al–4V alloy was

ignificantly larger than that measured in the current study. Theeasured �-lath width of the Ti–6Al–4V–0.1B (4.2 ± 1.2 �m) alloy

n their work was almost identical to that measured for the samelloy in the current study. Their measurement of the �-lath widthor a Ti–6Al–4V–0.4B (3.5 ± 1.1 �m) alloy was very close to thatound for the Ti–6Al–4V–1B alloy of the current study. The trendbserved in both studies was that little decrease in the �-lathidths occurs for B contents greater than 0.1B. Overall, neither the-lath nor �-lath widths changed dramatically for the alloys exam-

ned in the current study, although a slight decrease in the average-lath width was observed for the Ti–6Al–4V–1B alloy comparedith the Ti–6Al–0.1B alloy.

.2. Fatigue behavior

ural differences. Ti–6Al–4V exhibits superior high-cycle smoothar fatigue lives when the slip length is small [13–16]. Reducinghe grain size from 12 to 2 �m in equiaxed Ti–6Al–4V microstruc-

a maximum stress of 450 MPa, (b) Ti–6Al–4V–0.1B tested at a maximum stress of

136 W. Chen, C.J. Boehlert / Materials Science and Engineering A 494 (2008) 132–138

F 44,15e mumm -modi

tt0tcaswlaihv

sTotewrεTTw

tTvs

bilTl f6Al–4V–1B alloy may explain why the Nf values in the HCF regimewere closer to those for Ti–6Al–4V–0.1B. The load sharing by theTiB phase was expected to be responsible for the greater averagefatigue lives of Ti–6Al–4V–1B compared with those for Ti–6Al–4V.

ig. 4. SEM BSE photomicrographs of a (a) Ti–6Al–4V specimen which failed afterxhibited, (b) Ti–6Al–4V–0.1B specimen which failed after 990,554 cycles at a maxiaximum stress of 350 MPa. Cracking within the TiB phase was exhibited for the B

ures increased the room-temperature fatigue strength from 560o 720 MPa [17,18] and reducing the �-lath width from 10 to.5 �m in lamellar microstructures increased the Ti–6Al–4V room-emperature fatigue strength from 480 to 675 MPa [17]. In theurrent study, the B additions reduced the as-cast grain size bylmost an order of magnitude. This grain refinement could haveignificantly enhanced the fatigue resistance. In addition, the TiBhiskers were expected to have effectively decreased the slip

ength (and impede dislocation motion) compared to the Ti–6Al–4Vlloy and thereby increased the fatigue resistance. The small changen �-lath width observed in the current study was not expected toave had as prominent of an effect on the fatigue lives since it onlyaried to a small degree (less than 0.7 �m).

The 0.1B addition resulted in a significant as-cast grainize reduction from 1.7 mm for Ti–6Al–4V to 200 �m fori–6Al–4V–0.1B. It also resulted in increased εf compared to thether alloys studied (see Table 1). For maximum stresses greaterhan 400 MPa, most of the fatigue lives presented here were consid-red to be in the low-cycle-fatigue (LCF) regime (<100,000 cycles)here the greater the alloy’s εf and damage tolerance tends to cor-

espond to greater fatigue lives [19]. Thus the significantly greaterf value of the Ti–6Al–4V–0.1B alloy (8.5%) compared with that fori–6Al–4V–1B (2.5%) is expected to have enhanced its fatigue life.his may explain why the average fatigue lives of Ti–6Al–4V–0.1Bere longer than those for Ti–6Al–4V–1B.

Comparing the S–N behavior of the B-containing alloys,he Ti–6Al–4V–0.1B alloy significantly outperformed thei–6Al–4V–1B alloy at high stresses (LCF regime), while the Nfalues were closer at lower applied stresses. For maximum appliedtresses less than 400 MPa, most of the Nf values were considered to

Fs

3 cycles at a maximum stress of 350 MPa. Cracking within the �-phase laths wasstress of 400 MPa. (c) Ti–6Al–4V–1B specimen which failed after 32,718 cycles at afied alloys in (b) and (c).

e in the high-cycle-fatigue (HCF) regime (>100,000 cycles) wherencreased tensile strength tends to correspond to greater fatigueives and greater fatigue endurance limits or thresholds [19]. Thei–6Al–4V–1B alloy exhibited the greatest tensile strength and theowest ε value (see Table 1). Thus the increased strength of the Ti–

ig. 5. SEM low-magnification BSE photomicrographs of the edges for a Ti–6Al–4Vpecimen which failed after 44,153 cycles at a maximum stress of 350 MPa.

W. Chen, C.J. Boehlert / Materials Science and Engineering A 494 (2008) 132–138 137

Fig. 6. SEM secondary electron photomicrographs of the fracture surface for a Ti–6Al–4V specimen which failed after 28,848 cycles at a maximum stress of 350 MPa. Thecrack initiation sites are indicated with arrows. The overload region in (b) exhibited ductile dimples.

Fig. 7. SEM secondary electron photomicrographs of the fracture surface for a (a) Ti–6Al–4V–0.1B specimen which failed after 990,554 cycles at a maximum stress of 400 MPa,and for a (b and c) Ti–6Al–4V–1B specimen which failed after 32,718 cycles at a maximum stress of 350 MPa. The overload region in (c) exhibited ductile dimples.

138 W. Chen, C.J. Boehlert / Materials Science an

Fcm

StascwsrT

5

wbddcfahTplcT

rte

A

FTFsaL

R

[

[

[

[

[

[[

[

ig. 8. Low-magnification digital image of the oxidized crack growth region (darkolored) for a Ti–6Al–4V–1B specimen which fractured after 25,555 cycles for aaximum applied stress of 400 MPa. The ruler scale is in millimeters.

ince the strength and E of the TiB phase are significantly higherhan those for the � and � phases, much of the load placed on thelloy is supported by the TiB phase. As a consequence of the loadharing by the TiB phase, stress relaxation in the � and � phasesan occur. The highly loaded TiB phase can break at locationshere stresses are maximum and where flaws accumulated. Load

haring was also explained to be the reason for the increased creepesistance of B-modified Ti–6Al–4V cast alloys compared withi–6Al–4V castings [20].

. Summary and conclusions

Cast Ti–6Al–4V, Ti–6Al–4V–0.1B, and Ti–6Al–4V–1B alloysere fatigue tested at 455 ◦C and maximum applied stressesetween 250 and 450 MPa (R = 0.1). Boron additions dramaticallyecreased the as-cast grain size, however, the �- and �-lath widthsid not change significantly with respect to B content. The B-ontaining alloys exhibited longer average fatigue lives than thoseor Ti–6Al–4V, and the Ti–6Al–4V–0.1B alloy exhibited the longestverage fatigue lives. This was explained to be a result of theigher εf value of Ti–6Al–4V–0.1B compared with the other alloys.

he enhanced fatigue resistance of the Ti–6Al–4V–1B alloy com-ared with Ti–6Al–4V was explained to be a consequence of the

oad-sharing mechanism by the strong and stiff TiB phase that pre-ipitated as well as its presence as a barrier to dislocation motion.hus, in spite of the observed TiB-phase cracking, small B additions

[[

[

d Engineering A 494 (2008) 132–138

esulted in improved fatigue resistance. Overall, this work suggestshat Ti–6Al–4V–xB alloys may serve as adequate replacements forlevated-temperature fatigue-driven applications of Ti–6Al–4V.

cknowledgments

This work was partially sponsored by the National Scienceoundation (DMR-0533954). The authors are grateful to Dr. S.amirisakandala (FMW Composites, Inc.) and Dr. D.B. Miracle (Airorce Research Laboratory) for donating the material used in thistudy as well as their helpful technical support and guidance. Theuthors are also grateful to Ms. Sara Longanbach and Mr. Jeromeeboeuf of Michigan State University for their technical support.

eferences

[1] T. Saito, J. Met. 56 (2004) 33–36.[2] S. Gorsse, D.B. Miracle, Acta Mater. 51 (2003) 2427–2442.[3] S. Tamirisakandala, R.B. Bhat, J.S. Tiley, D.B. Miracle, Scripta Mater. 53 (2005)

1421–1426.[4] W.O. Soboyejo, W. Shen, T.S. Srivatsan, Mech. Mater. 36 (2004) 141–159.[5] K.B. Panda, K.S. Ravi Chandran, Acta Mater. 54 (2006) 1641–1657.[6] S. Tamirisakandala, D.B. Miracle, R. Srinivasan, J.S. Gunasekera, Adv. Mater. Pro-

cess. 12 (2006) 41–43.[7] I. Sen, S. Tamirisakandala, D.B. Miracle, U. Ramamurty, Acta Metall. Mater. 55

(15) (2007) 4983–4993.[8] S. Tamirisakandala, R. Bhat, D. McEldowney, J. Tiley, D.B. Miracle, Presentation

Given at the Workshop on Titanium Alloys Modified with Boron, Dayton, OH,October 11–13, 2005.

[9] D.X. Li, D.H. Ping, Y.X. Lu, H.Q. Ye, Mater. Lett. 16 (1993) 322–326.10] G.A. Hartman, S.M. Russ, in: W.S. Johnson (Ed.), Metal Matrix Composites: Test-

ing, Analysis and Failure Modes, American Society for Testing and Materials,Philadelphia, PA, 1989, pp. 43–53.

11] R.K. Nalla, I. Altenberger, U. Noster, G.I. Liu, B. Scholtes, R.O. Ritchie, Mater. Sci.Eng. A 355 (2003) 216–230.

12] Y. Komizo, H. Terasaki, M. Ikeda, F. Nishino, K. Toudai, in: M. Niinomi, S. Akiyama,M. Hagiwara, M. Ikeda, K. Maruyama (Eds.), Ti 2007 Science and Technology,vol. 1, 2007, pp. 741–745.

13] M. Peters, A. Gysler, G. Lutjering, in: H. Kimura, O. Izumi (Eds.), Titanium ’80,Science and Technology, vol. 1, TMS-AIME, 1981, p. 1777.

14] C.A. Stubbington, 1976 AGARD Conference Proceedings (no. 185), reprinted,Titanium and Titanium Alloys Source Book, ASM, 1982, pp. 140–158.

15] C.A. Stubbington, A.W. Bowen, J. Mater. Sci. 9 (1974) 941.16] J.C. Williams, G. Lutjering, in: H. Kimura, O. Izumi (Eds.), Titanium’80, Science

and Technology vol. 1, TMS-AIME, 1981, p. 671.17] M. Peters, G. Lutjering, P183 of titanium and titanium alloys, Report CS-2933,

Electric Power Research Institute, 1983.

18] W. Trojahn, diploma thesis, Ruhr University Bochum, 1980.19] R.W. Hertzberg, Deformation and Fracture Mechanics of Engineering Materials,

4th edition, John Wiley and Sons, New York, NY, 1996.20] C.J. Boehlert, S. Tamirisakandala, D.B. Miracle, in: M. Niinomi, S. Akiyama, M.

Hagiwara, M. Ikeda, K. Maruyama (Eds.), Ti 2007 Science and Technology, vol.1, 2007, pp. 697–700.