Scripta Materialia 138 (2017) 66–70

Contents lists available at ScienceDirect

Scripta Materialia

journal homepage: www.elsevier.com/locate/scriptamat

Regular Article

Mechanical properties of fully martensite microstructure in Ti-6Al-4V

alloy transformed from refined beta grains obtained by rapid heat
treatment (RHT)
Yan Chong a,⁎, Tilak Bhattacharjee a,b, Jangho Yi a, Akinobu Shibata a,b, Nobuhiro Tsuji a,b
a
Department of Materials Science and Engineering, Kyoto University, Kyoto, Japan
b
Element Strategy Initiative for Structural Materials (ESISM), Kyoto University, Kyoto, Japan

a r t i c l e i n f o a b s t r a c t

Article history: Fully martensite microstructures transformed from refined β grains were obtained in a Ti-6Al-4V alloy by rapid
Received 6 April 2017 heat treatment (RHT) into a single β phase region followed by water quenching. Ultrafine-grained equiaxed α +
Received in revised form 25 May 2017 β microstructure obtained by hot compression of martensite was used as the starting microstructure for subse-
Accepted 25 May 2017
quent RHT. By increasing the heating rate from 10 °C/s to 50 °C/s, the resulting β grain size was refined from 40 μm
to 8 μm, which was the smallest size of β reported so far. Both strength and ductility increased with decreasing the
Keywords:
β grain size, which was attributed to the reduced micro-crack lengths by fractography analysis.
Ti-6Al-4V © 2017 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
Martensite
β grain size
Rapid heat treatment

Titanium alloys especially α + β titanium alloys have been widely explore the limit of β-grain size refinement that can be achieved via
used in different industrial applications due to their unique combination RHT method by using UFG equiaxed α + β microstructure as the initial
of attractive properties like high strength to weight ratio, exceptional microstructure. Mechanical properties of fully martensite microstruc-
corrosion resistance as well as good biocompatibility [1]. Ti-6Al-4V is tures transformed from the fine α + β grains are also discussed.
the most widely used α + β titanium alloy which has been reported Chemical composition of the Ti-6Al-4V alloy used in the present
to account for more than 50% of the worldwide titanium tonnage [1, study was Ti-6.29Al-4.35V-0.155O-0.225Fe (in mass%). The β transus
2]. Martensite microstructure of Ti-6Al-4V alloy obtained by solution temperature of this alloy was estimated to be ~970 °C. Cylindrical spec-
treatment in a single β phase region followed by water quenching is imens with 8 mm in diameter and 12 mm in height were cut from the
comprised of fine α′-lamellae having various orientations (variants) as-received billet and solution treated at 1050 °C for 1 h in vacuum
within coarse prior β grains. The grain size of β is usually hundreds of followed by water quenching to get fully martensite microstructure
microns or even several millimeters [3], because of rapid grain growth shown in Fig. 1(a). The average α′-lamellae thickness in the martensite
in BCC structure of β phase. Compared with a lamellar microstructure microstructure was 1 μm. The specimens with the martensite micro-
obtained by solution treatment and slow cooling, the martensite micro- structure were uniaxially hot compressed on a thermomechanical pro-
structure generally possesses higher strength but lower ductility. cessing simulator (Thermecmastor-Z) at the temperature of 700 °C
Previous investigations [4–9] have shown that by using rapid heat and strain rate of 0.01 s−1 to a true strain of 0.8. The specimens were
treatment (RHT) into the single β phase region (at a heating rate in heated to the deformation temperature at a heating rate of 10 °C ∙s−1
the order of tens or even hundreds degrees per second), a unique micro- and held for 20 min for homogenization before the hot compression.
structure comprising relatively fine β-grains with fully transformed In the thermomechanical processing simulator, the temperature of the
intragranular α-lamellae can be obtained. This microstructure shows specimen was controlled by induction heating, so that rapid heating de-
much better balance of strength and ductility as well as good fatigue scribed later could be also carried out. After the hot compression, the
performance, compared with coarse-grained counterparts [7]. Howev- samples were immediately reheated to a temperature in the single β
er, the smallest β-grain size obtained by the RHT method so far is still phase region (995 °C) at a heating rate of 10 °C ∙ s−1, 30 °C ∙ s−1 or 50
30–50 μm because of the relatively coarse (~ 5 μm) equiaxed used as °C∙s−1 and held for 5 s before water quenching to room temperature.
the initial microstructure. The main purpose of the present study is to After the thermomechanical process, the samples were cut in half
along a diameter of the cylindrical specimens. The sections parallel to
⁎ Corresponding author. the compression axis and including the center of the samples were
E-mail address: [email protected] (Y. Chong). then ground and polished in standard metallographical methods.

http://dx.doi.org/10.1016/j.scriptamat.2017.05.038
1359-6462/© 2017 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
 Y. Chong et al. / Scripta Materialia 138 (2017) 66–70 67

Fig. 1. (a) Initial martensite microstructure transformed from coarse β grains. An EBSD IPF map. (b) IPF map of α phase before the hot deformation. The specimen for (b) and (c) was heated
to 700 °C at a heating rate of 10 °C∙s−1, held for 20 min, and then water-cooled without deformation. (c) IPF map of β phase before the hot deformation. (d) IPF map of α phase after the hot
deformation. (e) IPF map of β phase after the hot deformation. The colors in the IPF maps (b)–(e) indicate crystallographic orientation parallel to the compression direction, according to
the key stereographic triangles inserted.

Electrical polishing in a solution of 10% perchloric acid and 90% metha- After the tensile test, captured images were analyzed using Vic-2D com-
nol at −30 °C was conducted. Microstructural observations were car- mercial software to determine the strain in the micro-tensile specimens
ried out at the center of the samples on the sections described above. through tracking the positions of markers. For each condition, at least
Electron backscattered diffraction (EBSD) investigations were conduct- two specimens were tested and the average values were used for eval-
ed using TSL system attached to a field emission gun SEM (JEOL 7100F) uating strength and ductility.
operated at an accelerating voltage of 15 kV. The collected data were an- The initial martensite microstructure shown in Fig. 1(a) had a large β
alyzed using OIM software. Micro-tensile specimens with a gauge grain size which was measured to be ~ 200 μm by a line intersection
length of 2 mm, width of 1.2 mm and thickness of 0.5 mm were cut method on optical micrographs. Microstructure of the sample water
from the center of the cylindrical specimens compressed. The tensile di- quenched before the hot compression was characterized. It is found in
rection was perpendicular to the compression axis. We have confirmed the IPF maps (Fig. 1(b) and (c)) that during the holding period, the mar-
in previous studies that the micro-tensile specimen could give reliable tensite microstructure was tempered and a small amount of β phase
stress-strain date equivalent to those obtained from standard-sized formed, having the same orientation within the same prior β grain
specimen of the same material [10]. Tensile tests were conducted on (Fig. 1(b) and (c)). During the hot compression at relatively low tem-
Shimadzu AG-X plus system at an initial strain rate of 8.3 × 10−4 s−1 perature (700 °C), the tempered α′-lamellae and a small amount of β
at room temperature. The tensile strains of the micro-tensile specimens were supposed to change into equiaxed α grains via dynamic recrystal-
were precisely measured by a digital image correlation (DIC) method. lization (probably via discontinuous dynamic recrystallization [11,12]),
Surfaces of the micro-tensile specimens were sprayed with white and which could be confirmed by various orientations of both α and β grains
black contrast particles which acted as markers for the DIC analysis. shown in Fig.1 (d) and (e). Microstructure of the sample immediately
During the tensile test, the images of the micro-tensile specimens water quenched after the hot compression was characterized. Inverse
were captured by a CCD camera at a speed of 10 frames per second. pole figure (IPF) maps indicating crystallographic orientation parallel
68 Y. Chong et al. / Scripta Materialia 138 (2017) 66–70

to the compression direction of the α and β phases are shown in

Fig. 1(d) and (e), respectively. An equiaxed microstructure with ultrafine
grain sizes (α grain ~ 500 nm, β grain ~ 200 nm) was obtained as expect-
ed. The volume fraction of retained β phase was about 4.0%. The β grains
with the ultrafine grain size would act as nuclei for α to β phase transfor-
mation in the subsequent rapid heat treatment.
The microstructures after RHT at different heating rates followed by
water quenching are shown in Fig. 2. Fully martensite microstructures
were obtained in all the specimens. The average grain sizes of prior β
were measured to be 40 μm, 15 μm and 8 μm for the heating rates of
10 °C∙s−1, 30 °C∙s−1 and 50 °C∙s−1, respectively. The β grain sizes ob-
tained by the RHT method were all much smaller than that of the initial
martensite microstructure. The β grain refinement was finer in the
specimens heat-treated at higher heating rate. The grain size of 8 μm ob-
tained in the specimen with the heating rate of 50 °C/s is the finest body
centered cubic (BCC) β grain size reported so far for Ti-6Al-4V alloy.
With the refinement of the β grain size, the average length of α′-lamel-
Fig. 3. Engineering stress-strain curves of the fully martensite microstructures with
lae formed within the β grain was significantly reduced. However, it
different β grain sizes (Dβ). Positions and values of the tensile strength and total
should also be noted that there existed a size gradient of β grains elongation are indicated for each curve.
along the radial direction of the cylindrical specimen possibly due to
higher temperature near the surface part in RHT by induction heating.
The distribution of the β grain size from the center to the surface in ductility increased with decreasing the β grain size. As the average β
the specimen is plotted for each heating rate condition in a supplemen- grain size decreased from 200 μm to 8 μm, the tensile strength increased
tary material (Fig. A). It was clearly shown that the β grain size was larg- from 1047 MPa to 1298 MPa while the total elongation also increased
er near the surface than that in the center regions. This phenomenon from 9.4% to 19.8%. This result suggested that by refining the β grain
indirectly affected the pinning effect for the grain growth of β grains size, the strength and ductility of fully martensite microstructures can
by remaining α grains during heating and holding processes, as α grains be simultaneously increased. The increase of strength was expected to
near the surface dissolved earlier. The area with uniform β grain sizes in be attributed to a grain refinement strengthening, since the α′-lamellae
the center decreased with the increase of the heating rate. width were reduced with the grain refinement of β. On the other hand,
In order to evaluate the effect of β grain size on mechanical proper- because the uniform elongation remained unchanged with the
ties of the fully martensite microstructures, micro-tensile specimens refinement of β grain size, as is indicated by the arrows in Fig. 3, the in-
were cut from the area with uniform β grain sizes and then tensile test- crease of total elongation was mostly attributed to the increase of post-
ed at room temperature. The engineering stress-strain curves of the uniform elongation.
specimens with different β grain sizes are shown in Fig. 3. All the curves In the previous studies [13], it is found that the martensite micro-
showed high yield strength but limited work-hardening followed by structure of Ti-6Al-4V alloy followed the quasi-cleavage fracture behav-
large post-uniform elongation, which was typical for martensite micro- ior. Therefore, the microstructures near the tensile fractured surfaces
structures of Ti-6Al-4V alloy [13]. Surprisingly, both strength and were observed in the specimens with different β grain sizes, with a

Fig. 2. EBSD IPF + IQ maps (a–c) indicating crystallographic orientations parallel to the compression direction as well as corresponding BSE images (d–f) of the fully martensite
microstructures after RHT and water quenching. (a) and (d) heating rate of 10 °C∙s−1; (b) and (e) 30 °C∙s−1; (c) and (f)50 °C∙s−1.
 Y. Chong et al. / Scripta Materialia 138 (2017) 66–70 69

special attention to the size and distribution of micro-cracks. The EBSD could be obtained when the heating rate was 50 °C∙s−1, which was the
image quality (IQ) maps of specimens with the largest and smallest β smallest β grain size reported by now. With the refinement of the β
grain sizes, together with the plot showing the evolution of average grain size, both strength and ductility of the final martensite specimens
crack length with β grain size are shown in Fig. 4, in which the micro- increased. The tensile strength and total elongation of the sample with
cracks are indicated as dark regions. Regardless of the β grain size, the minimum β grain size of 8 μm reached 1298 MPa and 19.8%. The
micro-cracks were found to form, and the cracks propagated along α′- total elongation was twice of that with a coarse β grain size (200 μm).
lamellae. In order to statistically measure the average length of the The improvement in total elongation was attributed to the reduced
micro-cracks, at least five areas like Fig.4 were observed for each speci- micro-crack length associated with the decrease in the length of α′-la-
men. The average crack lengths in the specimens with mean β grain size mellae. The present result would open a possibility for obtaining the
of 200 μm, 40 μm, 15 μm, and 8 μm were 80 μm, 30 μm, 20 μm, and 10 full martensite microstructures transformed from fine β grain sizes in
μm, respectively. It was quite clear that with the refinement of β grain Ti-6Al-4V alloy, which process higher strength and ductility comparable
size, the average micro-crack length decreased. This was due to the to the bimodal microstructures of the alloy. Superior creep resistance
fact that the propagation of micro-cracks along one α′-lamellae got nature of the martensite microstructures could be expected at the
blocked when meeting another α′-lamellae with different orientations. same time.
Therefore, in the specimen with smaller α′-lamellae length, the chance Supplementary data to this article can be found online at http://dx.
of micro-cracks getting blocked significantly increased and subsequent- doi.org/10.1016/j.scriptamat.2017.05.038.
ly the average crack propagation length was reduced, which was
believed to be the main reason for the increased post-uniform elonga- Acknowledgement
tion with the refinement of the β grain size.
To summarize, rapid heat treatments to a temperature in the single β The authors were financially supported by the Cross-Ministerial
phase region followed by water quenching were carried out on the UFG Strategic Innovation Promotion Program on Structural Materials for In-
equiaxed microstructure of Ti-6Al-4V alloy formed by hot-deformation. novation (SIP-SMI), and the Elements Strategy Initiative for Structural
Full martensite microstructures with prior β grain sizes as small as 8 μm Materials (ESISM). The supports are gratefully appreciated.

Fig. 4. EBSD IQ maps near fracture surfaces of the tensile fractured specimens with the average β grain sizes (Dβ) of 200 μm (a) and 8 μm (b). (c) The plot of average crack length against the
averageβ grain size with error bas include.
70 Y. Chong et al. / Scripta Materialia 138 (2017) 66–70

References [8] P.E. Markovsky, Key Eng. Mater. 436 (2010) 185–194.
[9] P.E. Markovsky, S.L. Semiatin, J. Mater. Process. Technol. 210 (2010) 518–528.
[1] G. Lutjering, J.C. Williams, Titanium, second ed. Springer, Berlin, 2005. [10] L. Zhao, N. Park, Y. Tian, A. Shibata, N. Tsuji, Sci. Rep. 6 (2016) 39127.
[2] D. Banerjee, J.C. Williams, Acta Mater. 61 (2013) 844–879. [11] H. Matsumoto, B. Liu, S.-H. Lee, Y.P. Li, Y. Ono, A. Chiba, Metall. Mater. Trans. A 44A
[3] G. Lutjering, Mater. Sci. Eng. A243 (1998) 32–45. (2013) 3245–3260.
[4] S.L. Semiatin, J.C. Soper, I.M. Sukonmik, Acta Mater. 44 (1996) 1979–1986. [12] Q. Chao, P.D. Hodgson, H. Beladi, Metall. Mater. Trans. A 45A (2014) 2659–2671.
[5] O.M. Ivasishin, R.V. Teliovich, Mater. Sci. Eng. A263 (1999) 142–154. [13] H. Matsumoto, H. Yoneda, K. Sato, S. Kurosu, E. Maire, D. Fabreque, T.J. Konno, A.
[6] O.M. Ivasishin, G. Lutjering, Mater. Sci. Eng. A168 (1993) 23–28. Chiba, Mater. Sci. Eng. A528 (2011) 1512–1520.
[7] P.E. Markovsky, S.L. Semiatin, Mater. Sci. Eng. A528 (2011) 3079–3089.