Materials Science & Engineering A 632 (2015) 120–126
Contents lists available at ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
Rapid communication
Variation of work hardening rate by oxygen contents in pure titanium alloy Duck-soo Kang a,b, Kwang-jin Lee a,n, Eui-pyo Kwon a, Toshihiro Tsuchiyama b,c, Setsuo Takaki b,c a Convergence Components & Agricultural Machinery Application Center, Korea Institute of Industrial Technology (KITECH), 838-11, Palbok-dong, Deokjin-gu, Jeonju-City 561-202, South Korea b Department of Materials Science and Engineering, Kyushu University, 744 Motooka, Nishi-ku, Fukuoka 819-0395, Japan c International Institute for Carbon Neutral Energy Research (WPI-I2CNER), Kyushu University, 744 Mooto-oka, Nishi-ku, Fukuoka 819-0395, Japan
art ic l e i nf o
a b s t r a c t
Article history: Received 13 January 2015 Accepted 26 February 2015 Available online 5 March 2015
Pure titanium–oxygen alloys with different oxygen contents were tensile-tested to investigate the effect of oxygen on work hardening rate and deformation behavior. Yield and ultimate tensile strengths markedly increased with increasing oxygen contents, although the elongations were decreased. Work hardening rate was also enhanced with increasing oxygen contents resulting in increase in the uniform elongation. The improved work hardening rate was ascribed to transition of primary deformation mode from twin deformation to dislocation slip by oxygen addition. When twin deformation is suppressed by oxygen addition, however, the 〈c þa〉 dislocation must function as a substitute for twinning to permit the homogeneous plastic deformation. It contributed that the improved work hardening rate without deformation twinning is thought to be a restriction of dislocation slips to a certain special plane by oxygen addition. & 2015 Elsevier B.V. All rights reserved.
Keywords: Oxygen Work hardening rate Deformation behavior Pure titanium alloy
1. Introduction Interstitial elements such as carbon, nitrogen and oxygen are wellknown solid-solution strengtheners in titanium alloys. Especially, oxygen is widely used in commercial pure Ti and Ti alloys to obtain good mechanical properties [1]. For example, while grade 5 Ti–6Al–4V alloy with high oxygen content of 2000 ppm exhibits superior mechanical properties, grade 1 commercially pure Ti (CP Ti) has the lowest ultimate tensile strength among Ti alloys due to the lowest oxygen concentration. In particular, oxygen contents need to be carefully controlled to obtain optimal balance between strength and ductility [2]. It is well known that the hcp-structure CP Ti has an axial ratio of c/a ¼1.587, which is strongly plastically anisotropic. Furthermore, the c/a ratio increases with oxygen concentrations. The mechanical response of hcp materials is strongly dependent on the combination of active deformation modes, which is affected by c/a ratio [3]. The most common slip modes, in the order of ease of operation, are the f1010g, f1011g and {0001} planes, with 〈1120〉 as the slip direction, that constitute a total of four independent slip systems [4]. However, five independent slip systems are necessary n
Corresponding author. Fax: þ82 632103700. E-mail address:
[email protected] (K.-j. Lee).
http://dx.doi.org/10.1016/j.msea.2015.02.074 0921-5093/& 2015 Elsevier B.V. All rights reserved.
for the polycrystalline material to undergo general homogeneous plastic deformation. Thus, twin deformations are operated in order to maintain the deformation compatibility [5]. Hence, twin and slip deformations occur concomitantly during plastic deformation [6]. Oh et al. [7] studied the effect of oxygen content on the mechanical properties and lattice parameter of Ti–6Al–4V and CP Ti alloy. If the change in ultimate tensile strength is compared between these two alloys, it is clear that both exhibit a similar gradient with respect to oxygen concentration. Also, the lattice parameter c/a ratio increased with oxygen concentration at a rate dependent on the alloy. It is thought that the change in lattice strain with respect to the interstitial concentration affects the mechanical properties, i.e. as both the a and c lattice parameters increase with increasing oxygen concentration, they begin to limit the number of slip planes and improve the strength, but at the same time limit the ductility of the alloy. Although the mechanical properties and deformation behavior in hcp α-Ti alloys with various oxygen contents have been widely reported [8–10], precise influence of oxygen on the work hardening rate is not yet fully investigated. In this study, various oxygen content was added to a pure Ti alloy to clarify the effect of oxygen on the work hardening rate. In particular, the variation of the work hardening rate by oxygen concentrations was discussed in terms of lattice parameter and deformed microstructures.
D.-s. Kang et al. / Materials Science & Engineering A 632 (2015) 120–126
2. Experimental procedure
Table 1 Chemical composition of the α single-phase Ti–O alloys used in this study (mass%).
Table 1 shows the chemical compositions of the alloys used. All alloys containing oxygen (0.082–0.268 mass%) were prepared by a crucible levitation melting furnace. The oxygen contents were controlled by the appropriate addition of TiO2 during melting. The α single-phase Ti–O alloys were hot-rolled in α and β single-phase regions, and after annealing the α single-phase region, were coldrolled to a size of 800 300 5 (mm). After that, these cold-rolled specimens were solution-treated for 3.6 ks at 873 K, which is in α single-phase region, and then water quenched. A schematic illustration of the heat treatment procedure for the α single phase is given in Fig. 1. The oxygen contents in all of the alloys were measured using oxygen–nitrogen analyzer (EMGA-620 W, Horiba Japan). Microstructures were observed with optical microscope (OM) and scanning electron microscope (SEM) for the specimens finely polished and chemically etched by using a mixture of 5% hydrofluoric acid solution. The SEM (VE9800, KEYENCE Japan) observation was conducted at 15 kV. The transmission electron microscope (TEM) was also used for detailed microstructure observation. Thin foil specimens for TEM observation (∅3 mm) were prepared by twin-jet electro polishing in a chemical solution of 60% methanol, 35% butanol and 5% perchloric acid at 243 K. The TEM observation was carried out with JEM-2010EX, JEOL Japan operated at 200 kV. Also, the detailed microstructure and element distribution were investigated by using a scanning transmission electron microscope (STEM) with energy-dispersive X-ray spectroscopy (EDS). Thin foil specimens for STEM with EDS were manufactured in same method with TEM. The STEM with EDS was operated at an accelerating voltage of 200 kV (JEM-ARM 200F, JEOL JAPAN). The X-ray diffraction (XRD: RINT RAPID, RIGAKU) analysis was performed for identifying the constituent phases and measuring their volume fraction and lattice parameters, by using Co Kα radiation at an accelerating voltage of 40 kV and a current of 30 mA. The specimens for XRD were chemically polished with a water solution containing 60% hydrogen peroxide and 10% hydrofluoric acid. The hcp structural lattice parameters were evaluated using Eq. (1) with the experimentally observed sin2 θ [11]: sin 2 θ
λ2 2 λ2 2 2 ðh þ hk þk Þ 2 ðl Þ ¼ D sin 2 2θ 2 3a 4c
121
ð1Þ
where λ is the wave length, hkl is the Miller indices of the Bragg plane, a and c are the lattice parameters of the hcp structure, and D is the displacement. The mechanical properties of the prepared alloys were evaluated by Vickers hardness tests, nano-indentation test, and tensile tests at room temperature. The Vickers hardness (MICROHARDNESS, AKASI Japan) and the nano-indentation hardness (ENT-1100A, Elionix) were measured under a 300 gf load and a 2 mN load, respectively.
3. Results 3.1. Variation in the microstructure of α single-phase Ti alloy with oxygen content Fig. 2 shows optical micrographs of Ti–0.082, 0.132, 0.197, 0.218 and 0.268 mass% O alloys (hereafter, the mass% is omitted), which were subjected to solution treatment at 873 K for 3.6 ks. It is evident from this that all of these alloys consist of equiaxed grains with an average grain size of about 28–36 μm in diameter. The XRD patterns obtained from the solution treated Ti–O alloys within a 2-theta angle range of 44–481 as shown in Fig. 3
Alloy
mass%
Ti–0.082O Ti–0.132O Ti–0.197O Ti–0.218O Ti–0.268O
L
Hot rolling
O
N
H
C
0.082 0.132 0.197 0.218 0.268
0.009 0.001 0.001 0.001 0.001
0.007 0.006 0.005 0.006 0.006
0.004 0.003 0.005 0.004 0.004
Hot rolling
Anealing
Cold Recrystallization rolling 873K, 3.6ks
β α+ β α
Fig. 1. Schematic illustration of the heat treatment procedure for the α singlephase T–O alloys used in this study.
indicate that the phase obtained by solution treatment at 873 K for 3.6 ks is an α phase, with no other precipitates detected. It is observed that the texture is varied depending on oxygen contents. The Ti–0.268O alloy exhibits obvious texture because their strongest peaks are (0002)α peak rather than (1011)α peak, which is normally the strongest peak of α phase. Since the intensity of (0002)α peak is larger than the theoretical diffraction strength ratio (I1010:I0002:I1011 ¼ 25:30:100), it could be known that texture is developed based on 〈0001〉ND direction during rolling. It is well known that texture in single α phase, which is called as B (basal)texture, is greatly influenced by alloying elements such as Al and O [12]. This B-texture easily occurred in which the cross rolling is conducted. However, it is unclear why the texture is developed in rolled Ti–O alloy containing large amount of oxygen. It can also be seen that there is a shift in the diffraction peaks toward a lower angle with increasing oxygen content, which indicates an increase in the lattice parameter. One should note that this shift to lower angles is significantly larger for the (0002)α peak when compared to the peak shifts of the ð1011Þα peak. This means that an octahedral site is favorable to accommodate an oxygen atom as an interstitial, and thus preferentially affected to change the lattice parameter of (0002)α peak [13]. Given this, the lattice parameters were calculated using not only the (0002)α and ð1011Þα peaks, but also the ð1010Þα, ð1012Þα, ð1120Þα, ð1013Þα, ð2020Þα, ð1122Þα, ð2021Þα, (0004)α, ð2022Þα and ð1014Þα peaks. The variation in lattice parameter as a function of oxygen content is given in Fig. 4, which indicates that the lattice parameters rise with increasing oxygen content. It is therefore evident that oxygen addition expands the crystal lattice of the α phase. However, if we look at the change in lattice parameter in more detail, then it is apparent that the a axis varies constantly with increasing oxygen content. Meanwhile, in the Ti–0.082O and Ti–0.132O alloys, the c axis exhibits a slight gradual slope that results in a slight increase in the c/a ratio. In contrast, there is a sharp increase in the c/a ratio of the Ti–0.197, 0.218 and 0.268O alloys that is brought about by an increase in the c axis. It is well known that the ductility of hcp-group metals is strongly related to their c/a ratio, as this determines their slip modes [10]. Consequently, the relationship between c/a ratio and ductility will be discussed.
122
D.-s. Kang et al. / Materials Science & Engineering A 632 (2015) 120–126
10μm Fig. 2. Optical micrographs of solution-treated Ti–O alloys at 873 K for 3.6 ks: (a) Ti–0.082 mass% O, (b) Ti–0.132 mass% O, (c) Ti–0.197 mass% O, (d) Ti–0.218 mass% O and (d) Ti–0.268 mass% O alloys.
c/a
1.5885 1.5875
(d)
(c)
0.4688 0.4687
(b)
0.4686 0.4685 0.4684 0.4683 0.2951
a (nm)
Intensity (a. u. )
(e)
c (nm)
Lattice parameter (nm) and axial ratio
—
α (1011)
α (0002)
1.5895
0.2950 0.2949
(a)
44
0
0.05
0.1
0.15
0.2
0.25
0.3
Oxygen contents (mass%)
45
46
47
48
2 theta (degree) Fig. 3. X-ray diffraction patterns of solution-treated Ti–O alloys at 873 K for 3.6 ks: (a) Ti–0.082 mass% O, (b) Ti–0.132 mass% O, (c) Ti–0.197 mass% O, (d) Ti–0.218 mass % O and (d) Ti–0.268 mass% O alloys.
Fig. 4. Change in lattice parameters and c/a ratio of solution-treated Ti–O alloys as a function of oxygen contents.
3.2. Mechanical properties of α single-phase Ti–O alloys Nominal stress-curves of solution-treated Ti–O alloys are shown in Fig. 5, wherein there is an apparent change in the stress–strain
D.-s. Kang et al. / Materials Science & Engineering A 632 (2015) 120–126
123
Oxygen contents (at.%)
0.6 0.002
Ti-0.268 mass%O Ti-0.218 mass%O
0.01
0.6 Tensile strength Yield strength
Ti-0.186 mass%O
0.3 Ti-0.082 mass%O
0.2
Ti-0.132 mass%O
0.1 0.0 0.0
Tensile strength, σTS/GPa
0.4 Yield strength, σ0.2/GPa
Nominal stress, σn/GPa
0.5
0.5 ΔσTS [GPa]= 0.43[O]
1/2
0.4
0.3 Δσ0.2 [GPa]= 0.25[O]
0.1
0.2 Nominal strain, εn
0.3
Fig. 5. Nominal stress–strain curves of solution-treated Ti–O alloys.
2.4
Δσ TS ½GPa ¼ 0:43½O1=2 ð½O : atomic fractionÞ
ð3Þ
The work hardening rate and true stress of solution treated Ti–O alloys are shown in Fig. 7 as a function of true strain, from which it is clearly evident that the work-hardening rate tends to increase with increasing oxygen concentration. Here, there is a need to emphasize that the Ti–0.268O alloy exhibits the highest work hardening rate at the high strain region of all the Ti–O alloys tested, meaning that plastic instability is suppressed at high strain by adding oxygen. Increase in work hardening rate contributes to the enhancement of uniform elongation as in Fig. 7.
4. Discussion In order to investigate the deformation microstructure, TEM images were obtained from 5.0% deformed alloys and are shown in Fig. 8. Plate-like features identified as deformation twinning are observed in all of these alloys, with a twin 1 μm in width observed in the Ti–0.082O alloy. With increasing oxygen content, the deformation twinning becomes more refined, and their volume fraction is reduced. In other words, an increase in oxygen content suppresses twinning and promotes slip deformation. Twin deformation is affected by several factors, which include grain size,
Vickers hardness, Hv (GPa)
2.2 2.0 1.8 1.6
1.4
1.0 0.05
0.15
0.10
0.25
0.20
0.30
Oxygen contents (mass%) Fig. 6. Change in (a) yield and tensile strengths, and (b) Vickers hardness as a function of oxygen contents in solution treated Ti–O alloys.
1.0
Ti-0.268mass %O
0.8 True stress, σT/GPa
ð2Þ
Vickers hardness
1.2
Work hardening rate, (dσt/dε t/GPa)
Δσ 0:2 ½GPa ¼ 0:25½O1=2
([O]: atomic fraction)
0.2
0.4
behavior with increasing oxygen concentration. All materials exhibit continuous yielding behavior, with the yield strength of the Ti– 0.082O alloy around 0.28 GPa; however, with an increase in oxygen content, the yield strength is gradually increased to a maximum of 0.40 GPa at 0.268O. It is possible that this oxygen could form interstitial solid solution and intermetallic compounds with the Ti alloys, which have higher strength [14]; but as mentioned above in Fig. 3, no such intermetallic compounds were confirmed. It would therefore seem that oxygen addition leads to solid solution strengthening by asymmetric lattice distortion [15], as shown in Fig. 4. Moreover, even though there is some decrease in ductility between the Ti–0.132O and 0.186O alloys, this is only a gradual change. Fig. 6(a) shows the yield and tensile strengths as a function of oxygen content and the variation in Vickers hardness given in Fig. 6(b). The oxygen addition contributes to an increase in not only the yield and tensile strengths, but also hardness. Furthermore, given that the reinforced strength is directly proportional to oxygen atomic fraction per square root, the strengthening rates can be roughly expressed as follows:.
1/2
Ti-0.218mass %O
0.6
Ti-0.198mass %O
0.4
Ti-0.132mass %O Ti-0.082mass %O
0.2 0
0
0.05
0.10
0.15
0.20
True strain, εT Fig. 7. True stress (σT) and work hardening rate (dσt/dɛt) as a function of true strain (ɛT) in solution treated Ti–O alloys.
composition, and deformation rate, and is therefore suppressed by grain size refinement, increased oxygen content or a decrease in deformation rate [6]. In this instance, the effect of oxygen is considered to be only related to the deformation behavior of Ti– O alloys. Namely, the principal deformation mechanism is changed from twinning to slip with increasing oxygen content. This suppression of twinning by oxygen is considered to be due to the following two factors: Firstly, the formation of twinning is strongly affected by variation in the c/a lattice parameter [10], with Sasano et al. [16] reporting that twinning rarely occurs in Ti alloys containing various alloying elements due to an increase in shear strain that accompanies a growing c/a. In this study, it is clear that
124
D.-s. Kang et al. / Materials Science & Engineering A 632 (2015) 120–126
μ Fig. 8. Transmission electron micrographs of (a) Ti–0.082 mass% O, (b) Ti–0.197 mass% O and (c) Ti–0.218 mass% O alloys tensile-deformed by 5%.
twinning is suppressed by increasing the value of c/a with oxygen content, as shown in Fig. 4. As for the second reason, atomic shuffles are considered to be an important factor in producing a twinned structure. This means that twin deformation in hcp metals without oxygen occurs by atoms moving to a twinned position by a shuffling of octahedral interstitial sites, a process that is inhibited by oxygen occupying octahedral interstitial sites and creating a resistance to lattice shear [17]. Plastic deformation in hcp structures has long been known to be a complex of both slip and twinning mechanisms [6]. In the schematic illustration of the slip and twin systems in hcp structures shown in Fig. 9, the three most common slip modes are the f1010g, f1011g, and {0001} planes, with 〈1120〉 as the slip direction in Fig. 9(a). These three glide planes, combined with the slip direction in the basal plane, make up four independent slip systems in total [4]. However, five independent slip systems are necessary for a polycrystalline material to be able to homogeneously plastically deform. As a result, twinning systems or 〈cþa〉 pyramidal slip shown in Fig. 9(b) and (c) are necessary to maintain deformation compatibility, with these two deformation mechanisms occurring simultaneously [5]. When twin deformation is
suppressed by oxygen addition, however, the 〈c þa〉 dislocation must function as a substitute for twinning to permit the homogeneous plastic deformation. Generally, it is well known that the principal slip plane for 〈c þa〉 dislocation is f1101g in Ti alloys. The 〈cþ a〉 dislocation slips for f1101g are indeed observed in Ti– 0.082O and –0.218O alloys as shown in Fig. 10. It can be seen that the 〈cþ a〉 dislocation density is increased with increasing oxygen concentration. This is because, as mentioned earlier, 〈c þa〉 dislocation acts as a substitute instead of twin deformation for homogeneous plastic deformation in high oxygen concentration alloy. It is well known that the twin deformation in steel contributes to increasing its uniform ductility and work hardening rate due to a increase in the dynamic Hall–Petch effect [18,19]. However, in spite of the decreased twinning the work hardening rate of Ti–O alloys was improved as shown in Fig. 7. To understand this contradiction, it is necessary to consider the work hardening mechanism associated with dislocation motion in Ti–O alloys. A possible reason for the improved work hardening rate without deformation twinning is therefore thought to be a restriction of dislocation slips to a certain special plane by oxygen
D.-s. Kang et al. / Materials Science & Engineering A 632 (2015) 120–126
(1012)
125
[1126]
[1011]
(1011) (1010)
[1210]
(0001)
(1122)
[1210]
(1121) [1123]
[1123]
[1123]
(1122)
(1011)
Fig. 9. Schematic illustration of the (a) slip, (b) twin and (c) 〈cþ a〉 deformation systems in a hcp structure [4,5].
< c+a >
< c+a >
g:0002
g:0002
200 nm
Fig. 10. TEM micrographs of 5% deformed (a) Ti–0.082O and (b) Ti–0.218O alloys.
addition [20]. Fig. 11 indicates the schematic illustration of deformation microstructure in Ti–O alloys illustrated as basis of Fig. 8. According to the relation between dislocation and tensile strain, the equation is established as follows: ρ ¼ 2ε=bx
ð4Þ
where ρ is the dislocation density, ɛ is the tensile strain, b is the
Burgers vector and x is the distance between dislocations. By oxygen addition, the spacing (x) between the 〈cþa〉 dislocation is near to them. In other words, reducing the distance between them leads to the high dislocation density. This should, in turn, contribute to an increase in the work hardening rate and uniform elongation because this would form a large-range stress field, which generates a back stress against the trailing moving dislocations. Meanwhile, the high
126
D.-s. Kang et al. / Materials Science & Engineering A 632 (2015) 120–126
Twin
x
< c+a > dislocation Dislocation
x
Oxygen addition
(3). Local elongation was markedly reduced by adding oxygen as the high dislocation density produced a concentration of strain.
Acknowledgment This study was supported by the Grant-in-Aid for Scientific Research (C) No. 23560840 (2011–2013) from Japan Society for the Promotion of Science and the International Institute for Carbon Neutral Energy Research (WPI-I2CNER), sponsored by the Japanese Ministry of Education, Culture, Sports, Science and Technology.
Low dislocation density
High dislocation density
Fig. 11. Schematic illustration of the deformation microstructure in Ti–O alloy.
dislocations density may also produce a concentration of strain, thereby reducing local elongation and potentially causing fracture in the vicinity of the grain boundaries.
5. Conclusion To provide a better understanding of the effects of oxygen on the deformation behavior of Ti alloys, α single-phase alloys were produced with different oxygen contents. The results obtained from these alloys can be summarized as follows: (1). The yield and tensile strengths markedly increased with increasing oxygen contents. (2). The work hardening rate and uniform elongation increased by adding oxygen due to occurrence of 〈c þa〉 slip and the limiting of dislocation slips to a certain special plane. This corresponds to significant dislocations density, which generates a back stress against the trailing dislocation movement.
References [1] G. Lütjering, J. Williams, Titanium, Springer, Berlin Heidelberg, New York (2007) 89–93. [2] M. Wasz, F. Brotzen, R. McLellan, A. Grin, Int. Mater. Rev. 41 (1996) 1–12. [3] A. Serra, D. Bacon, R. Pond, Metall. Mater. Trans. A 33 (2002) 809–812. [4] J. Christian, S. Mahajan, Prog. Mater. Sci. 39 (1995) 1–157. [5] A. Santhanam, R. Reed-Hill, Metall. Trans. A 2 (1971) 2619–2622. [6] H. Conrad, Prog. Mater. Sci. 26 (1981) 123–404. [7] J. Oh, B. Lee, S. Cho, S. Lee, G. Choi, J. Lim, Metall. Mater. Int. 17 (2011) 733–736. [8] D. Chichili, K. Ramesh, K. Hemker, Acta Mater. 46 (1998) 1025–1043. [9] M. Brandes, M. Baughman, M. Mills, J. Williams, Mater. Sci. Eng. A 551 (2012) 13–18. [10] P. Kwasniak, M. Muzyk, H. Garbacz, K.J. Kurzydlowski, Mater. Sci. Eng. A 590 (2014) 74–79. [11] A.I. Kahveci, G. Welsch, Scr. Metall. 20 (1986) 1287–1290. [12] The Japan Society for Technology of Plasticity, Fundamentals of Titanium and Its Working, Corona Publishing Co. Ltd. (2008) 25–26. [13] S. Yamaguchi, M. Koiwa, M. Hirabayashi, J. Phys. Soc. Jpn. 21 (1966) 2096. [14] J.P. Bars, E. Etchessahar, J. Debuigne, J. Less-Common Met. 52 (1997) 51–76. [15] Q. Wei, L. Wang, Y. Fu, J. Qin, W. Lu, D. Zhang, Mater. Des. 32 (2011) 2934–2939. [16] H. Sasano, H. Kimura, J. Jpn. Inst. Met. 41 (1977) 933–939. [17] Z. Wyatt, W. Joost, D. Zhu, S. Ankem, Int. J. Plast. 39 (2012) 119–131. [18] K. Renard, P. Jacques, Mater. Sci. Eng. A 542 (2012) 8–14. [19] H. Idrissi, K. Renard, L. Ryelandt, D. Schryvers, P. Jacques, Acta Mater. 58 (2010) 2464–2476. [20] M. Yoo, S. Agnew, J. Morris, K. Ho, Mater. Sci. Eng. A 319 (2001) 87–92.