Journal of Materials Science & Technology, 2020, 49(0): 56-69 DOI: 10.1016/j.jmst.2020.02.026

Research Article

Microstructural evolution and FCC twinning behavior during hot deformation of high temperature titanium alloy Ti65

Zhixin Zhanga,c, Jiangkun Fan,a,b,*, Bin Tanga,b, Hongchao Koua,b, Jian Wangc, Xin Wangc, Shiying Wangd, Qingjiang Wange, Zhiyong Chene, Jinshan Li,a,b,*

a State Key Laboratory of Solidification Processing, Northwestern Polytechnical University, Xi’an, Shaanxi, 710072, China

b National & Local Joint Engineering Research Center for Precision Thermoforming Technology of Advanced Metal Materials, Xi’an, Shaanxi, 710072, China

c Baoti Group Ltd., Baoji, Shaanxi, 721014, China

d School of Materials Science and Engineering, Changzhou University, Changzhou, Jiangsu, 213164, China

e Institute of Metal Research, Chinese Academy of Sciences, Shenyang, Liaoning 110016, China

Corresponding authors: * E-mail addresses:jkfan@nwpu.edu.cn(J. Fan),ljsh@nwpu.edu.cn(J. Li).

Received: 2019-11-18   Accepted: 2020-02-28   Online: 2020-07-15

Abstract

Although the development of titanium alloys with working temperatures above 600 ℃ faces enormous difficulties and challenges, the related research has not stopped. In the present work, detailed analyses on microstructure evolution and hot deformation behavior of a new temperature resistant 650 ℃ titanium alloy Ti65 were investigated from micrometer scale to nanometer scale. The results revealed that lamellar α grains gradually fragmentized and spheroidized during the α + β phase region compression and the orientation of the c-axis of α grains gradually aligned to radial directions, forming two high Schmid factors (SFs) value texture eventually with the increase of strain to 0.7. Moreover, there were some strengthening characters in the α + β phase region such as lenticular αs and nano silicide (TiZr)6Si3. In the β phase region, fine equiaxed dynamic recrystallized (DRX) β grains were formed. Besides, the variant selection of α´ martensite followed Burgers orientation relationship during the compression process. The main deformation mechanisms of the α + β phase region were dislocation slip and orientation dependent spheroidization. Whereas, the deformation process in the β phase region was controlled by β grain DRX. Interestingly, many nano scale FCC twins were generated at the interface of α´ lath during deforming in the β phase region, which was firstly observed in Ti65 alloy.

Keywords: High temperature titanium alloy ; Hot deformation ; Microstructure evolution ; Texture ; FCC twin

PDF (11202KB) Metadata Metrics Related articles Export EndNote| Ris| Bibtex  Favorite

Cite this article

Zhixin Zhang, Jiangkun Fan, Bin Tang, Hongchao Kou, Jian Wang, Xin Wang, Shiying Wang, Qingjiang Wang, Zhiyong Chen, Jinshan Li. Microstructural evolution and FCC twinning behavior during hot deformation of high temperature titanium alloy Ti65. Journal of Materials Science & Technology[J], 2020, 49(0): 56-69 DOI:10.1016/j.jmst.2020.02.026

1. Introduction

Near α titanium alloy is the titanium alloy with relatively large amounts of α stabilizer and low concentrations of β stabilizer [1]. Such alloy generally has excellent creep resistance property at high temperature [2]. The typical near α titanium alloy, such as IMI834 [3] and Ti60 [4], exhibits high tensile strength and remarkable creep resistance at 600 ℃ [5]. Ti65 (Ti-Al-Sn-Zr-Mo-Si-Nb-Ta-W titanium alloy) is a newly developed near α titanium alloy to be utilized as disks, blades or blisks in aero-engine at 600 ℃~650 ℃ [6], which is completely unattainable in all previous high temperature titanium alloys. Compared to Ti60 alloy, more molybdenum element is added to Ti65 to improve its tensile strength and comprehensive mechanical properties [7,8]. More silicium element is dissolved in the matrix to obtain the best creep resistance [9]. Wolfram element, a newly added element, can refine grains and improve high temperature tensile strength [10].

Near α titanium alloys always compose of multiple alloying elements, along with high strength [11,12] and relatively undesirable high temperature processibility [13,14], so investigation about deformation behavior and microstructure development of near α titanium alloys is a hotspot. The flow behavior of IMI834 and Ti60 alloy were modeled by Min Zhou [15] and Wenwen Peng [16,17]. It suggested that the flow softening can be caused by deformation heating, dynamic recovery (DRV), DRX, kinked α platelets, flow localization and free-surface cracking. For the hardening process, strain hardening and precipitation hardening are dominated in near α titanium alloy. The role of hardening mechanism and the influence of phase stability on strain hardening were clarified by Guohua Zhao [18], he indicated that the flow stress increase can be attributed to the internal stress fields induced by kinematic hardening. The strain-hardening ability is influenced by microstructures (lamellar, bi-lamellar, equiaxed and bimodal microstructures) [19] and α-phase orientation [20]. The strain-hardening ability of the ‘dual-phased’ microstructures and basal and prismatic texture was much better. K V Sai Srimadh [21] noticed that silicide along α/β interfaces and ordered Ti3Al precipitates would strengthen the α matrix during hot deformation process in IMI834 alloy. Besides, Jingyuan Shen [22] proposed a competitive DRX mechanism between α grain (continuous dynamic recrystallization) and β grain (discontinuous dynamic recrystallization and continuous dynamic recrystallization) in Ti-4Al-3V-2Mo-2Fe alloy during high temperature tensile deformation process. Ke Wang [23] reported that the dominant mechanism in the β phase varied from DRV to DRX with increasing temperature, and the great deformation made β phase experience DRX in dominant at 940 ℃, which could refine the β grain in TC8 alloy. Moreover, Zibo Zhao et al. [24] found that two main texture components were acquired after cooling in compressed Ti60 alloy and Wenyuan Li et al. [25] studied the anisotropy of mechanical properties at 600 ℃ in Ti60 alloy. In a word, most of the investigation above mainly focuses on stress-strain curves and microstructure evolution of the titanium alloy with target service temperature no more than 600 °C. The study of the deformation texture of high temperature titanium alloy materials is not systematic, and as a newly designed temperature resistant 650 ℃ titanium alloy, the related work of Ti65 alloy on microstructure evolution during hot deformation is limited. The research on microstructure, texture evolution and precipitation behavior of Ti65 alloy is necessary.

In the present study, isothermal compression tests were carried out at various processes to clarify precipitation behavior and deformation mechanism during hot deformation process. Detailed characterization of FCC twins and semi-in-situ observation of the microstructure evolution were conducted from micrometer scale to nanometer scale. These results would be expected to figure out some new insights into a deep understanding of the precipitation behavior and deformation mechanism of the new high temperature alloy Ti65, which could provide some guide to the optimization of industrial thermal process such as single-phase region ingot cogging or two-phase region finish rolling and make some progress on the basic accumulation of this new titanium alloy.

2. Materials and research methodology

The chemical composition of Ti65 alloy is Ti-5.8Al-4.0Sn-3.5Zr-1.0Ta-0.3Nb- 0.5Mo-0.5Si-0.8 W (wt.%). The as received Ti65 alloy was supplied in form of a billet which went through the ingot breakdown process. After primary working in the single β field, the ingot was further worked in the α + β phase region. The cylindrical specimens for hot compression tests in this study were machined from the billet with 8 mm in diameter and 12 mm in height, and the compression axis of specimens was parallel to the original deformation direction of the billet. The initial microstructure, Fig. l(a) and (b), is composed of acicular martensite α´ with a continuous lamellar of about 2 μm in width bestrewing in the coarse β grains, which is was obtained by solution treatment at 1070 ℃ (above the β transus temperature ~1035 ℃) for 30 min followed by water quenching. The texture distribution of initial Ti65 sample is shown in Fig. 1(c). There are several texture peaks with weak strength.

Fig. 1.   (a) Initial microstructure, (b) EBSD map (IPF) and (c) three dimensional ODF map of Ti65 alloy.


Uniaxial compression tests were carried out in the temperature range of 930~1140 ℃ with 30 ℃ intervals and strain rate range of 0.001~10 s-1 with 0.7 true strain on a Gleeble-3800 simulator. Specimens were heated to the target test temperature with a heating rate of 10 ℃/s, and then held for 5 min to ensure uniform and stable temperature in Gleeble chamber. Thermocouples were welded in the middle surface of the specimen to measure the actual temperature. After compression, the specimen was cooled in the water immediately to preserve the deformation microstructure of the high temperature region, as shown in Fig. 2(a). To reduce die friction and get uniform deformation, the top and bottom surface of the specimen were coated with graphite powder, and a foil of tantalum was placed between the face of the specimen and anvils. In order to trace metallographic and texture revolution, three parallel compression tests at the strain of 0.04, 0.3 and 0.7 were conducted with fixed temperature and strain rate to realize semi-in-situ observation.

Fig. 2.   (a) The heating and deformation schedule for thermal compression process and (b) schematic representation of compression direction.


Deformed specimens were sectioned parallel to the compression direction and prepared for metallographic examination with an Olympus / PMG3 optical microscope. Texture was measured in a section containing the compression direction (CD) and radial directions (RDs) of the deformed samples as shown in Fig. 2(b). Specimen and crystal coordinate system were chosen with OX=<10$\bar{1}$0>, OY=<$\bar{1}$2$\bar{1}$0> and OZ=<0001> for the HCP α-phase, as shown in Fig. 2(b). The EBSD specimen were prepared by metallographically polishing the plane of compression direction (CD) and radial directions (RDs), and finally electro-polishing the plane using the electrolyte of 5 % perchloric, 35 % butanol, and 60 % methanol at 0 ℃ and 28 V. Micro-texture was measured on a Gemini-SEM-500 equipped with the Channel 5 system (HKL Technology, Denmark). EBSD maps were acquired with different step sizes, according to the required map resolution. Typically, a step size of 1 μm was used to collect data over a large area covering millimeter long macro-zones. A smaller step size of 0.2 μm was used for a detailed microstructure analysis. The EBSD analysis software (HKL-Tango) was set to index only α phase to construct inverse pole figure (IPF) maps, pole figure (PF) maps, orientation distribution function (ODF) maps, SF histogram and KAM maps, mainly because few β phase could be preserved at room temperature in Ti65 alloy. Microscopic deformation features were characterized by bright field (BF) and dark field (DF) TEM, HRTEM and EDS using Tecnai F30 field emission transmission electron microscope at a voltage of 300 kV. TEM specimen were prepared by cutting 0.3 mm thick slice from deformed samples, then grinding them to 50 μm thick to punch out 3 mm diameter wafers, and finally thinning to finished samples by a twin-jet machine using the electrolyte of 20 % perchloric acid in methanol at 5 ℃ and 35 V.

3. Results and discussion

3.1. Microstructure and texture evolution during hot compression

3.1.1. Deformation in the α + β phase field

Fig. 3 is the stress-strain curve and the corresponding microstructure of Ti65 alloy deformed at 960 ℃ (below Tβ) and 0.1 s-1. The flow stress reaches to a peak at a critical strain and then decreases with further deformation, showing a flow softening behavior. When the strain reaches to 0.04, the peak stress is obtained as 171 MPa. When the deformation is completed (strain = 0.7), the flow stress has dropped to 78 MPa. Besides, the α colonies undergo a dramatic deformation, so α colonies gradually fragmentizes and spheroidizes, eventually transforming into equiaxed α grains.

Fig. 3.

Fig. 3.   (a) Flow stress-strain curve with the corresponding microstructure of Ti65 alloy deformed at 960 ℃ and 0.1 s-1 with different true strain: (b) 0.04; (c) 0.4; (d) 0.7.


The original β boundary can be observed at the temperature of 960 ℃ with a small strain (0.04) and lamellar α colonies with are bestrewed in the coarse initial β grains (Fig. 3(b) and Fig. 4(a)). The lamellar α colonies are kinked and woven under external force, the thickness of lamellar is slightly reduced, and α colonies take spheroidization response locally with strain increasing to 0.4 (Figs. 3(c) and 4 (c)). The region of spheroidized α grains expands and the average size of α grains is reduced further with the strain increasing to 0.7 (Figs. 3(d) and 4 (e)). The distortion and spheroidization response of α colonies are the dominate transformation of Ti65 alloy under the coupling effect of thermal and stress action at 960 ℃ from 0.04 to 0.7 strain, which may make contribution to the deformation softening process.

Fig. 4.

Fig. 4.   EBSD inverse pole figures (IPF) and the corresponding pole figures (PF) of Ti65 alloy deformed at 960 ℃ and 0.1 s-1 with different strain: (a) (b) 0.04; (c) (d) 0.4; (e) (f) 0.7.


The microstructure retains the original morphology of the initial β quenching treatment at the temperature of 960 ℃ with 0.04 true strain, but the initial α´ martensites in β grains have transformed into lamellar α colonies [26]. IPF shows the crystal direction parallel to CD keeps random and PF shows there are no intensive peak zones. It means the orientation of α colonies nearly keeps random, that is, there is no obvious preferred orientation at 0.04 strain, as illustrated in Fig. 4(a) and (b).

Fig. 4(c) shows that the proportion of the spheroidized α grains in the microstructure has reached to 55 % when the strain increased to 0.4. The number of grains colored green and blue with the crystal directions [$\bar{1}$ 2 $\bar{1}$ 0] and [01 $\bar{1}$ 0] parallel to CD increases, that is, the grains gradually rotate to the crystal directions with c-axis aligned to RDs. The number of grains colored red with [0001] aligned to CD decreases. Fig. 4(d) illustrates that there are A~D four peak zones, which means some preferred orientation appears. A texture component, with highest strength, corresponds to the c-axis aligned to 75~90° away from CD towards RDs, whose c-axis is approximately parallel to RDs. A texture is colored as green and blue on IPF. B texture component corresponds to the c-axis aligned to 49~70° away from CD towards RDs. C texture component corresponds to the c-axis aligned to CD. D texture component has the same c-axis orientation as B zone.

Most of spheroidized grains are with the orientation of [$\bar{1}$2$\bar{1}$0] aligned to CD and have an average grain size of 1.13 μm at 0.7 true strain, as shown in Fig. 4(e). Fig. 4(f) illustrates that there are E and F two peak zones in the pole figures, which have the orientation with [0001] parallel to RDs. It means that almost all α grains rotate to a same direction with the increase of strain under the external forces, and finally forms the deformation texture with the c-axis aligned closer to RDs, and other texture components disappear gradually.

Fig. 5 shows the evolution of Euler angles and texture by Euler space and ODF maps. Euler space exhibits the main texture components with strength more than 5. Random texture generally has a large spread in the Euler space and ODF maps as illustrated in Fig. 5(a) and (b). With strain increasing, some identified texture appears in Euler space on φ = 90°, as illustrated in Fig. 5(d) and (g). The miller indices of the identified texture can be calculated by Eq. (1) [27] or some low index texture can be determined by comparing the tested ODF and the standard ODF sections [28]. There are many texture peaks in Fig. 5(b) and (c) ODF, which means the 0.04 strain can not force lamellae to form intense deformed texture. This is consistent with the related analysis results of Fig. 4(a) and (b).

Fig. 5.

Fig. 5.   Three dimensional ODF (Euler space) and constant φ2 = 0°/30° ODF maps of Ti65 alloy deformed at 960 ℃ and 0.1 s-1 with different true strain: (a) (b) (c) 0.04; (d) (e) (f) 0.4; (g) (h) (i) 0.7.


For the texture evolution with a strain of 0.4, the preferred orientation distribution is affected by the fragmentation and spheroidization behavior obviously. Some identified textures form, but many original textures at 0.04 strain have disappeared. The main texture (Fig. 5(e) ODF (φ2 = 0°)) can be calculated as ($\bar{1}$ 2 $\bar{1}$ 5)[1 $\bar{2}$ 11], ($\bar{1}$ 2 $\bar{1}$ 5)//OXY plane and [1 $\bar{2}$ 11]//OX direction, with Euler angles {90° 32° 0°} at 0.4 strain. Moreover, there is a weak texture forming at the Euler angles {82° 43° 30°} in Fig. 5(f).

The grains spheroidize and rotate more drastically from the strain of 0.4 to 0.7. The texture with the deformation train of 0.7 is enhanced and the ODF peak has some degree rotation compared to that of strain of 0.4. The texture in Fig. 5(h) is estimated as ($\bar{1}$ 2 $\bar{1}$ 2)[4 $\bar{5}$ 16] with Euler angles {78° 53° 0°}, and the texture in Fig. 5(i) is calculated as (02 $\bar{2}$ 3)[0 $\bar{1}$ 11] with Euler angles {80° 41° 30°}. The intensity of texture on ODF (φ2 = 0°) weakens and the intensity of texture on ODF (φ2 = 30°) strengthens with the strain increasing from 0.4 to 0.7.

The onset of plastic deformation occurs when the shear stress acting on the incipient slip plane and in the slip direction reaches a critical value. The magnitude of the shear stress under a certain external force is determined by SF. SF evolution could speculate the possibility of the activation of different slip systems during straining. The average SF keeps a relatively small values (0.31~0.38) with all slip systems in Fig. 6(a), which suggests the crystal orientations at early stage of deformation (strain of 0.04) are unfavorable for activation of Basal < a>, Prismatic < a> and Pyramidal < a> slip systems. The SF of Prismatic < a> and Pyramidal < a> slip systems increase to 0.35 and 0.43 at the true strain of 0.4 (Fig. 6(b)), and then changes to 0.42 and 0.41 respectively at the true strain of 0.7 (Fig. 6(c)). The increasement of SF is due to the formation of the texture ($\bar{1}$ 2 $\bar{1}$ 5)[12- 11], ($\bar{1}$ 2 $\bar{1}$ 2)[45- 16] and (02 2- 3)[0 $\bar{1}$ 11], whose SF value will be calculated below.

Fig. 6.

Fig. 6.   Schmid factors (SF) histogram of α phase with Basal < a>, Prismatic < a> and Pyramidal < a> slip systems for Ti65 alloy deformed at 960 ℃ and 0.1 s-1 with different true strain: (a) 0.04; (b) 0.4; (c) 0.7.


SFs can be calculated by the equation SF = cosφ·cosλ [29], where φ is the angle between the specimen axis and the normal to the slip plane and λ the angle between the load axis and the slip direction. The SFs for different texture components forming at 0.4 and 0.6 strain are calculated in Table 1. It indicates that the SF of Basal < a> for ($\bar{1}$ 2 $\bar{1}$ 5)[1 $\bar{2}$ 11] is 0, so the average SF of Basal < a> in Fig. 6(b) decreases at strain of 0.4. The SFs of Prismatic < a> and Pyramidal < a> for ($\bar{1}$ 2 $\bar{1}$ 5)[1 $\bar{2}$ 11] are 0.5 and 0.4, so the average SF of these slip systems in Fig. 6(b) keeps a large value. It means the slip system of Basal < a> is restricted, and Prismatic and Pyramidal slip systems are highly active. That is to say, the grains are inclined to rotate to the soft orientation with Prismatic < a> and Pyramidal < a> slip systems during the hot deformation process. This situation has intensified after deformation at strain of 0.7. The average SF of Prismatic < a> continuously increase to 0.42 and the average SF of Pyramidal < a> keeps a high value 0.41 as shown in Fig. 6(c). This would be due to the formation of ($\bar{1}$ 2 $\bar{1}$ 2)[4 $\bar{5}$ 16] and (02 $\bar{2}$ 3)[0 $\bar{1}$ 11] texture, whose SF are listed in Table 1. The increase of the average SF makes it easy to activate the slip process, leading to the decrease of flow stress and the softening process.

Table 1   SFs of Basal, Prismatic and Pyramidal slip systems for different texture components formed at 0.4 and 0.6 strain.

Slip systemTexture component
($\bar{1}$2$\bar{1}$5)[1$\bar{2}$11]($\bar{1}$2$\bar{1}$2)[4$\bar{5}$16](02$\bar{2}$3)[0$\bar{1}$11]
Basal < a> (0001)[11$\bar{2}$0]00.130.04
Prismatic < a> (1$\bar{1}$00)[11$\bar{2}$0]0.50.470.35
Pyramidal < a> (1$\bar{1}$01)[11$\bar{2}$0]0.400.310.26

New window| CSV


The Kernel Average Misorientation (KAM) map can be obtained by calculating the orientation differences between adjacent points on EBSD map. It shows the average local misorentation below the subgrain angle and can be used to represent the dislocation density and locate deformed regions [30]. The area of yellow and red regions, high KAM region, expands gradually from Fig. 7(a) to Fig. 7(c). It reveals that the deformation degree and dislocation density increase with the strain increasing from 0.04 to 0.7. Fig. 7(d) histogram illustrates the increment of average KAM from 0.04 to 0.4 strain is higher than that from 0.4 to 0.7 strain. It means the initial deformation makes more contribution to the increase of dislocation density. Besides, the increase of KAM value demonstrates the process of deformation hardening. This hardening process may be related with the crystallographic orientation. It can be found that the grains with specific orientation have high KAM value by comparing the IPF (Fig. 4) and KAM maps (Fig. 7). It is because the SF plays a significant role in dislocation increment and its distribution, and the grains with high SF results in high KAM [31].

Fig. 7.

Fig. 7.   KAM maps of Ti65 alloy deformed at 960℃and 0.1 s-1 with different true strain (a) 0.04, (b) 0.4, (c) 0.7 and (d) KAM histogram.


Grain boundary misorientation can reveal the coordinated deformation of grain boundaries and the transformation of substructure. Fig. 8(a) shows some LABs form under the action of stress with 0.04 strain. The variant orientation relationships of 10.53°, 60°, 60.8°, 63.2° and 90° are maintained between the α variants which are inherently in a Burgers orientation relationship with the β matrix [32,33], as illustrated in Fig. 8(d). Much more LABs are distributed in both Fig. 8(b) and (c) with large strain. Furthermore, the α misorientation angle curve no more keeps the peaks at 10.53°, 60°, 60.8°, 63.2° and 90° as shown in Fig. 8(d). That is, the Burgers orientation relationship maintained by the partial α phase and the β matrix may has been destroyed during the deformation process due to the grain spheroidization response and rotation [34]. Fig.8 (e) illustrates the grain size distribution of deformed Ti65 samples. The sample of 960 ℃ with a small strain (0.04) shows the lamellar α colonies with average size of 3.31 μm are bestrewed in the coarse initial β grains. The lamellar α colonies are kinked and woven under external force, the thickness of lamellar is slightly reduced to 1.43 μm, and α colonies take spheroidization response locally with strain increasing to 0.4 (Fig. 8(b) and (e)). The region of spheroidized α grains expands and the average size of α grains is further reduced to 1.13 μm with the strain increasing to 0.7 (Fig. 8(c) and (e)).

Fig. 8.

Fig. 8.   Grain boundary contrast maps for Ti65 alloy deformed at 960 ℃ and 0.1 s-1 with different true strain (a) 0.04, (b) 0.4, (c) 0.7 and (d) statistical diagram of misorientation angle, (e) the α grain size histogram. The high-angle boundaries (HABs) with misorientation over 15 deg and the low-angle boundaries (LABs) with misorientation at 2 to 15 deg are depicted as black lines and red lines, respectively.


In summary, the process of hardening and softening under hot deformation is controlled by microstructure, texture, dislocation density (KAM) in the α + β phase field. The KAM gradually increases with the increasing of strain, leading to deformation hardening. The orientation dependent spheroidization of lamellar α with the c-axis aligned to RDs direction gradually forms soft texture with high SF value. The orientation dependent spheroidization makes contribution to deformation softening. The dominant slip systems of plastic deformation are Prismatic < a> and Pyramidal < a> slip systems. The main deformation mechanisms in the α + β phase field are dislocation slip and orientation dependent spheroidization [35,36].

3.1.2. Deformation in the β phase field

The true stress-strain curve of the single β phase region (1080℃, above Tβ) is slightly different from that in the α + β phase region (Fig. 3a). The stress reaches a peak (52 MPa.) at 0.08 strain and follows by the near steady state subsequently (Fig. 9a). When the deformation is completed (strain = 0.7), the flow stress has dropped to 41 MPa. The dynamic softening and work hardening during deformation process are balanced with each other in the steady state. As the strain increases, the initial coarse β grains are elongated and become fibrous. Martensite α´ is densely distributed in all deformed microstructures.

Fig. 9.

Fig. 9.   (a) Flow stress-strain curve with the corresponding microstructure of Ti65 alloy deformed at 1080 ℃ and 0.1 s-1 with different true strain: (b) 0.04; (c) 0.4; (d) 0.7.


The deformed microstructure with 0.04 true strain consists of almost no deformed coarse equiaxed β grains with acicular α´ martensite in β grains, as illustrated in Fig. 9(b). The elongated fiber is perpendicular to the CD. The β grain boundaries become serrate with the strain increasing to 0.4 (Fig. 9(c)). Moreover, some fine equiaxed DRX β grains form around the deformed and elongated β grains, which has an average grain size of about 11 μm. With the strain increasing to 0.7, β grains are further elongated to fibrous morphology and much more DRX β grains with an average grain size of about 15 μm can be observed easily (Fig. 9(d)). Therefore, dynamic recrystallization is the main feature of microstructure evolution of Ti65 alloy deformed at 1080 ℃, which contributes to the softening process of flow stress [37,38].

There is no obvious dominate deformation texture existing in Fig. 10. IPF (a) (c) (e) shows that there is no obvious preferred orientation at all the deformed microstructures. The α´ lath in each β grain have a number of different orientations. Fig. 10(b) (d) (f) PF shows that the peak zones are basically consistent which has the c-axis aligned to about 60° away from CD towards RDs or the c-axis aligned to RDs at different strains. At 0.04 strain, the strength of texture is relatively low because the martensitic transformation can decrease the strength of the texture [39]. With the strain increasing, the texture of c-axis aligned to RDs directions strengthens and the texture of c-axis aligned to about 60° away weakens.

Fig. 10.

Fig. 10.   IPF and the corresponding PF of Ti65 alloy deformed at 1080℃ and 0.1s-1 with different true strain: (a) (b) 0.04; (c) (d) 0.4; (e) (f) 0.7.


Fig. 11 shows the evolution of Euler angles and texture by Euler space and constant ODF maps. Both Euler space and ODF maps show a large spread and multiple peaks. It implies that there is not obvious preferred orientation during deforming at 1080 ℃, which is consistent with the result in Fig. 10. The main texture of the strain of 0.4 (Fig. 11(f)) and 0.7(Fig. 11(h)) can be calculated as the texture (033¯1)[0 $\bar{1}$16] with Euler angles {85° 79° 30°} and the texture (3¯63¯ 2)[1 $\bar{2}$ 19] with Euler angles {90° 78° 0°} respectively. In addition, there are many other components of the texture, which is different from the case of deformation in the α + β phase field (Fig. 5).

Fig. 11.

Fig. 11.   Three dimensional ODF (Euler space) and constant φ2 = 0°/30° ODF maps of Ti65 alloy deformed at 1080 ℃ and 0.1 s-1 with different true strain: (a) (b) (c) 0.04; (d) (e) (f) 0.4; (g) (h) (i) 0.7.


Influenced by the near random texture at 1080℃, the average SFs is kept within a small fluctuation range (0.27~0.40) and has small changes with all slip systems at all strain (Fig. 12). It reconfirms that some random texture forms and there is no obviously soft texture (high SF texture) existing. What needs to be pointed out is that the SF of Basal < a> decrease with the increasing strain to 0.7. It suggests the slip system of Basal < a> is restricted, and Prismatic and Pyramidal slip systems are inclined to activate.

Fig. 12.

Fig. 12.   Schmid factors histogram of α phase with Basal < a>, Prismatic < a> and Pyramidal < a> slip systems for Ti65 alloy deformed at 1080 ℃ and 0.1 s-1 with different true strain: (a) 0.04; (b) 0.4; (c) 0.7.


Fig. 13 shows that high KAM regions expand gradually with the deformation strain increasing. The high KAM regions colored yellow are almost concentrated around the β grain boundaries. The KAM histogram (Fig. 13(d)) demonstrates the average KAM rises from 0.51 to 0.61 with the increasing strain. The process of deformation hardening aroused by KAM rising in the β phase field is not as obvious as that in the α + β phase field.

Fig. 13.

Fig. 13.   KAM maps for Ti65 alloy deformed at 1080 ℃ with different true strain (a) 0.04, (b) 0.4, (c) 0.7 and (d) KAM histogram.


Grain boundary maps indicate that there are many LABs distributing near the β grain boundaries (Fig. 14(a)-(c)). The content of LABs gradually increases from 26 % to 31 % with the strain increasing, which makes contribution to deformation hardening. Besides, the nucleation of DRX grains is initiated when a critical dislocation density is reached during hot deformation, nuclei for DRX typically appears by subgrain growth or subgrain rotation at the grain boundaries, as shown in Fig. 14(b) and (c). DRX is a process of dislocation annihilation and a process of softening [35]. The areas highlighted in Fig. 14(b) and (c), some fine equiaxed grains around the elongated β grain boundary, are detected as recrystallized grains colored in blue, as shown in Fig. 14(e) and (f). As the deformed samples are cooled in the water immediately, the microstructure deformed at 1080 ℃ can be preserved. Therefore, the grains in highlighted areas can be confirmed to be dynamic recrystallized grains (DRX) forming in the deformation process of 1080 ℃ and 0.1 s-1. Besides, the DRX fraction increases from 12.06 % to 16.28 % with the strain rising from 0.4 (Fig. 1(e)) to 0.7 (Fig. 1(f)). The DRX behavior during deformation in the β phase field is also reported in some other near α alloy [13].

Fig. 14.

Fig. 14.   Grain boundary contrast maps with true strain (a) 0.04, (b) 0.4, (c) 0.7 and recrystallized maps with true strain (e) 0.4, (f) 0.7 for Ti65 alloy deformed at 1080 ℃and 0.1 s-1 and (d) statistical diagram of misorientation angle.


The martensitic transformation occurs during the quenching process of the alloy after high temperature deformation in the single β region [40]. These martensites maintain a Burgers orientation relationship with the β matrix (Fig. 14(d)). It can be seen that the misorientations of all variants of martensites (10.53°, 60°, 60.8°, 63.2° and 90°) are consistent with the literatures [33] at all strains. It illustrutes that the crystallographic orientation relationship between martensite and β phase matrix is not affected by hot deformation.

In conclusion, the character of the morphology of α´ lath and texture changes significantly with the variation of strain in the β phase region. Whereas, the martensite α$\acute{v}$ariant selection and the orientation relationship of α$\acute{v}$ariants have not been affected, and the Burgers relationship is still maintained with the β phase. Dislocation proliferation and DRX respectively cause work hardening and dynamic softening during the deformation, which is the essential reason of the steady state of the true stress-strain curve in the β phase field.

3.2. Precipitates and FCC twinning during hot compression

3.2.1. Precipitation behavior and deformation mechanism

As discussion above, the stress-strain curves of different deformation conditions are quite different (Figs. 3(a) and 9 (a)). This difference can be reflected not only in the deformation microstructure and texture, but also in the dynamic phase transitions. Therefore, in order to more thoroughly discuss and understand the hot deformation behavior and deformation mechanism of Ti65 alloy, it is necessary to carry out more observation and characterization analysis of the deformed structure on the nanometer scale.

Fig. 15 describes the microscopic characteristics of the microstructure deformed in the α + β and β phase region, respectively. It can be seen that the microstructure contains many equiaxed α grains (Fig. 15(a)). The selected area electron diffraction (SAED) pattern at the gap of α grains indicated that Ti65 alloy composes of full α phase with HCP structure at 960 ℃ and a small number of dislocations. The lenticular nano secondary α grains (αs) precipitate and fill in the gap of equiaxed α grains (Fig. 15(b)). Moreover, a particle phase is dispersed in the primary α (αp) matrix. The SAED pattern (Fig. 15(b)) and EDS (Fig. 15(c)) could confirm that the particle precipitation is silicide (TiZr)6Si3 with HCP structure, which is consistent with the investigation in Ti-1100-6Zr alloy [41]. All results in the α + β phase region could demonstrate that the main strengthening mechanisms are boundary strengthening, αs strengthening and nano-silicide strengthening as modeled in Fig. 16, and the main deformation mechanisms are dislocation slip and orientation dependent spheroidization which has been established in 3.1.1.

Fig. 15.

Fig. 15.   (a) (b) TEM images of Ti65 alloy deformed at 960 ℃ (0.1 s-1) with 0.7 strain and (c) corresponding EDS results of silicide in (b). (d) (e) TEM images deformed at 1080 ℃ (0.1 s-1) with 0.7 strain and (f) the SAED pattern of FCC twins highlighted in (e). The SAED patterns corresponding to the TEM images are embedded in the picture respectively.


Fig. 16.

Fig. 16.   The schematic diagram of the precipitation process during the hot deformation of Ti65 alloy.


In comparison, the deformation characters in the β phase region are obviously different with the α + β phase region. It can be seen from Fig. 15(d) that a large number of α´ lath structures fill the entire image. Moreover, there is a large amount of precipitated phase in the gap of the α´ lath (Fig. 15(e)). A further analysis is conducted on the second phase by SAED, as shown in Fig. 15(f), and indicates that there is a FCC twinning behavior widely existing in the interface. It suggests that the main deformation mechanisms are FCC twinning and β grain DRX (See Section 3.1.2 for details). The FCC twinning and β grain DRX behavior are modeled in Fig. 16.

The difference of deformation mechanism in the α + β and single β phase regions could be further explained by calculating the deformation activation energy. The peak stress of hot compression flow curves from 0.001 s-1 to 10 s-1 in the α + β phase region and β phase region has been tested to determine the activation energy.

The constitutive equation used in this research is a kinetic equation as shown in Eq. (2) [42] which is used to derive activation energy. The activation energy Q can be determined based on Eq. (3) Q=R⋅p⋅q [43]. Therefore, the value of Q = 686.67 kJ/mol in the α + β phase region and Q = 202.64 kJ/mol in the β phase region are obtained by the slope p of the plot of lnε˙ against lnsinhασ and the slope q of the plot of lnsinhασ against 1/T, as shown in Fig. 17.

$\dot{ε}=A[sinh(ασ)]^{n}exp(\frac{-Q}{RT})$
$Q=R\{\frac{∂ln\dot{ε}}{∂ln[sinh(ασ)]}\}_{T}\{\frac{∂ln[sinh(ασ)]}{∂(1/T)}\}_{\dot{ε}}$

Fig. 17.

Fig. 17.   (a) and (b) Plots of ln(strain rate ε˙) vs. lnsinhασ for various deformation temperatures and (c) (d) 1/T vs. lnsinhασ for various strain rates.


The hot deformation process is a thermal activation process. Activation energy Q depends essentially on the melting point of the material and the diffusion mechanism (e.g. direct interstitial or vacancy mechanism). Therefore, different titanium alloys have different activation energy values. Similar to other near α titanium alloys [44], the deformation activation energy of Ti65 alloy in the β phase region is calculated to be 203 kJ/mol, which is close to the Ti activation energy for self-diffusion in β phase region, i.e. 153 kJ/mol [45]. It suggests that the slip and twining process may be controlled by diffusion mechanism in the β phase region. The deformation activation energy of Ti65 alloy in the α + β phase region is 686.67 kJ/mol which is much larger than that in the β phase region. The high activation energy means more energy is needed to start deformation. This is in accord with the microscopic deformation strengthening features of nano silicide, lenticular αs and refined αp grains as show in Fig. 15(a and b) and Fig. 16, which could hinder dislocation motion and increase deformation resistance, leading to the increase of activation energy.

3.2.2. Deformation induced FCC twinning behavior

FCC twins as a special deformation mechanism in titanium alloy has attracted much attention. Fig. 18(a) illustrates that FCC twins in DF TEM image exist at the interface of α´ lath and spread over the whole deformed microstructure. HRTEM observation is performed on the twin region as shown in Fig. 18(b) and the twin boundaries (TB1, TB2, TB3 and TB4) are marked. It is found that the twin boundary is arranged in parallel at the same interface. Fig. 18(c) is obtained by FFT transformation in the red frame region, and the FCC twin could be confirmed again. The zone axis is [011] and the twinning plane is identified as ($\bar{1}$1$\bar{1}$). The IFFT image in red frame region is constructed in Fig. 10(d). It illustrates that there are no other defects in the twins and the corresponding TB1 and TB2 twin planes are marked in IFFT image. The value of lattice parameter a of FCC phase calculated by measuring the ($\bar{1}$1$\bar{1}$) crystal plane spacing d (d = 0.249 nm, $d=a/\sqrt{h^{2}+k^{2}+l^{2}}$) is 0.4321 nm, as shown in Fig. 18(d).

Fig. 18.

Fig. 18.   FCC twins in Ti65 alloy deformed in the β phase field: (a) DF TEM image and its corresponding magnified BF TEM image is inset; (b) HRTEM image of FCC twins; (c) Fast Fourier-filtered (FFT) pattern and (d) Inverse fast Fourier-filtered (IFFT) image transformed from red frame in (b).


A further comparison of the precipitation conditions and lattice parameters of FCC phase in different titanium alloys is summarized in Table 2. It suggests that the FCC phase could be formed by severe plastic deformation and some specific heat treatments, and the lattice parameter a of FCC structures is close in different titanium alloys. In conclusion, the nano FCC twins exist generally in the gap of α´ lath, which is a main deformation product in the β phase region of Ti65 alloy. Such a FCC twining behavior aroused by high temperature deformation in titanium alloy has not been reported before.

Table 2   FCC phase in different titanium alloys.

AlloyGenerating conditionLattice parameter a(nm)
CP-Ti [46]Cryogenic channel-die compression0.4302
Ti-6Al-4 V [47]High energy shot peening0.4158
Ti-20Zr-6.5Al-4 V [48]Solution treatment at 950℃0.4385
Ti65Compression in the β phase region0.4321

New window| CSV


4. Conclusions

Based on a systematic analysis of microstructure evolution and hot deformation behavior of a new high temperature titanium alloy Ti65 in the α + β phase field and β phase field, the following conclusions can be drawn.

(1)Lamellar α colonies spheroidized and the orientation of its c-axis aligned to RDs direction forming two high SF value textures ($\bar{1}$ 2 $\bar{1}$ 2)[45- 16] and (02 2- 3)[0 1- 11] with the increase of strain during the deformation in the α + β phase region. There were some hardening phases precipitating such as the lenticular αs in interface and nano-silicide (TiZr)6Si3 with HCP structure.

(2)The α$\acute{v}$ariants were not affected by straining and following the crystallographic orientation relationship of Burgers all the time during the deformation in the single β region.

(3)The main deformation mechanisms were dislocation slip and orientation dependent spheroidization in the α + β phase region. The deformation in the β phase region was controlled by β grain DRX and nano scale FCC twins were generated at the lath interface. The activation energy of Ti65 alloy in the α + β phase region and β phase region were 686.67 kJ/mol and 202.64 kJ/mol respectively.

(4)The nano FCC twin formed numerously at the interface of α´ lath during the hot compression process. It was a main deformation product of Ti65 alloy in the β phase region. The twinning plane is ($\bar{1}$1$\bar{1}$) and the lattice parameter a of the FCC twin is 0.4321 nm.

Acknowledgments

The authors gratefully acknowledge the Major State Research Development Program of China (No. 2016YFB0701305), the National Natural Science Foundation of China (No. 51801156) and the Natural Science Basic Research Plan in Shaanxi Province of China (Nos. 2018JQ5035 and 2019JM-584) for the financial support to this work.

Reference

Christoph Leyens, Manfred Peters, Darmstadt, 2004, pp. 23-25.

[Cited within: 2]

J. Matthew, Titanium: A Technical Guide, 2000, pp. 18.

[Cited within: 2]

R.W. Evans, R.J. Hull, B. Wilshire, J. Mater. Process. Technol. 56 (1-4) (1996) 492-501.

URL     [Cited within: 1]

Shouyong Wei, Wei-min Shi, Ding-chun Wang, Chin. J. Nonferrous Met. 20(S1) (2010) 801-806.

[Cited within: 3]

I. Weiss, S. Semiatin, Mater. Sci. Eng. A 263 (2) (1999) 243-256.

URL     [Cited within: 3]

Q.J. Wang, J.R. Liu, R. Yang, J. Aeronaut. Mater. 34 (4) (2014) 1-26.

[Cited within: 1]

J.Q. Qi, Y. Chang, Y.Z. He, Mater. Des. 99 (2016) 421-426.

URL     [Cited within: 1]

Wei Shouyong, Shi Weimin, Guo Jialin, J. Chin. J. Nonferrous Met. 23 (1) (2013) 121-124.

[Cited within: 1]

H. Kestler, H. Muahrabi, H. Remner, Titanium Sci. Technol. (1995) 1171-1178.

[Cited within: 1]

Michael Y. Mendoza, Peyman Samimi, David A. Brice, et al., Metall. Mater. Trans. A (48A) (2017) 3594-3605.

[Cited within: 1]

W.J. Zhanga, X.V. Songa, sx. Hui, et al., Mater. Sci. Eng. A 595 (2014) 159-164.

DOI      URL     [Cited within: 3]

R.W. Evans, R.J. Hull, B. Wilshire, J. Mater. Process. Technol. 56 (1996) 492-501.

DOI      URL     [Cited within: 1]

Wenwen Peng, Weidong Zeng, Qingjiang Wang, et al., Mater. Sci. Eng. A571 (2013) 116-122.

[Cited within: 2]

Feng Sun, Jinshan Li, Hongchao Kou, Mater. Sci. Eng. A 626 (2015) 247-253.

DOI      URL     [Cited within: 1]

Min Zhou, Mater. Sci. Eng. A 245 (1998) 29-38.

DOI      URL     [Cited within: 1]

Wenwen Peng, Weidong Zeng, Qingjiang Wang, Mater. Des. 51 (2013) 95-104.

URL     [Cited within: 3]

Wenwen Peng, Weidong Zeng, Qingjiang Wang, Mater. Sci. Eng. A 593 (2014) 16-23.

URL     [Cited within: 1]

Guohua Zhao, Xin Xu, David Dye, et al., Acta Mater. 183 (2020) 155-164.

DOI      URL     [Cited within: 1]

Yan Chong, Guanyu Deng, Jangho Yi, et al., J. Alloys. Compd. 811 (2019), 152040.

URL     [Cited within: 1]

P. Ahmadian, S.M. Abbasi, M. Morakabati, Mater. Today Commun. 13 (2017) 332-345.

URL     [Cited within: 1]

K.V. Sai Srinadh, Nidhi Singh, V Singh, Bull. Mater. Sci. 30 (2007) 595-600.

DOI      URL     [Cited within: 1]

Jingyuan Shen, Yu Sun, Yongquan Ning, et al., Mater. Charact. 153 (2019) 304-317.

DOI      URL     [Cited within: 1]

Wang Ke, Miaoquan Li, Qing Liua, Mater. Charact. 120 (2016) 115-123.

DOI      URL     [Cited within: 1]

Z.B. Zhao, Q.J. Wang, J.R. Liu, et al., Metallurg. Mater. Trans. A (2018) 1-10.

[Cited within: 1]

Wenyuan Li, Zhiyong Chen, Jianrong Liu, et al., Mater. Sci. Eng. A 688 (2017) 322-329.

URL     [Cited within: 1]

Sun Feng, Li Jinshan, Kou Hongchao, et al., Rare Met. Mater. Eng. 44 (4) (2015) 0848-0853.

URL     [Cited within: 1]

O. Engler, V. Randle,

Introduction to Texture Analysis: Macrotexture, Microtexture, and Orientation Mapping

CRC Press Taylor, New York, 2009, pp. 21-27.

[Cited within: 1]

Y.N. Wang, J.C. Huang, Mater. Chem. Phys. 81 (2003) 11-26.

DOI      URL     [Cited within: 1]

Richard W. Hertzberg, Richard P. Vinci, Jason L. Hertzberg,

Deformation and Fracture Mechanics of Engineering Materials

WILEY-VCH Verlag GmbH & Co, KGaA, 1995, pp. 7-8.

URL     [Cited within: 1]

M.G. Jiang, C. Xu, H. Yan, et al., Acta Mater. 157 (2018) 53-71.

URL     [Cited within: 1]

Eralp Demir, Dierk Raabe, Nader Zaafarani, et al., Acta Mater. 57 (2009) 559-569.

URL     [Cited within: 1]

S. Birosca, F. Di Gioacchino, S. Stekovic, et al., Acta Mater. 74 (2014) 110-124.

URL     [Cited within: 1]

Ehsan Farabi a, Peter D.Hodgson a, Gregory S. Rohrer, Acta Mater. 154 (2018) 147-160.

DOI      URL     [Cited within: 2]

Jiangkun Fan, Jinshan Li, Yudong Zhang, et al. Adv. Eng. Mater. 19 (2017).

DOI      URL     PMID      [Cited within: 1]

Hongwei Li, Xinxin Sun, He Yang, Int. J. Plast. 87 (2016) 154-180.

DOI      URL     [Cited within: 2]

Shibayan Roy, Satyam Suwas, Scr. Mater. 154 (2018) 1-7.

DOI      URL     [Cited within: 1]

E. Ghasemi, A. Zarei-Hanzaki, E. Farabi, et al., J. Alloys. Compd. 695 (2017) 1706-1718.

DOI      URL     [Cited within: 1]

C.K. Yan, A.H. Feng, S.J. Qu, et al., Acta Mater. 154 (2018) 311-324.

DOI      URL     [Cited within: 1]

Shibayan Roy, Satyam Suwas, Acta Mater. 134 (2017) 283-301.

URL     [Cited within: 1]

Hossein Beladi, Qi Chao S. Gregory, Acta Mater. 80 (2014), 478-48.

URL     [Cited within: 1]

Binguo Fu, Hongwei Wang, Chunming Zou, et al., Mater. Charact. 99 (2015) 17-24.

URL     [Cited within: 1]

H.J. McQueen, N.D. Ryan, Mater. Sci. Eng. A 322 (2002) 43-63.

DOI      URL     [Cited within: 1]

K. Wang, M.Q. Li, Mater. Sci. Eng. A 600 (2014) 122-128.

URL     [Cited within: 1]

P. Wanjara, M. Jahazi, H. Monajati, S. Yue, J.-P. Immarigeon, Mater. Sci. Eng. A 396 (2005) 50-60.

URL     [Cited within: 1]

F. Warchomickaa, C. Poletti, Mater. Sci. Eng. A 528 (2011) 8277-8285.

URL     [Cited within: 1]

D.H. Hong, T.W. Lee, S.H. Lim, et al., Scr. Mater. 69 (2013) 405-408.

URL     [Cited within: 1]

Y.G. Liu, M.Q. Li, H.J. Liu, Scr. Mater. 119 (2016) 5-8.

URL     [Cited within: 1]

R. Jing, C.Y. Liu, M.Z. Ma, J. Alloys. Compd. 552 (2013) 202-207.

URL     [Cited within: 1]

/