Rare Metal Materials and Engineering Volume 42, Issue 7, July 2013 Online English edition of the Chinese language journal

Cite this article as: Rare Metal Materials and Engineering, 2013, 42(7): 1356-1361.

Received date: July 25, 2012 Foundation item: Natural Science Fund of China(51001086); “973 Program”(2011CB605502) Corresponding author: Zhang Tiebang, Ph. D., Associate Professor, State Key Laboratory of Solidification Processing, Northwestern Polytechnical University, Xi’an 710072, P. R. China, Tel: 0086-29-88481764, E-mail:tiebangzhang@nwpu.edu.cn

Copyright © 2013 Northwest Institute for Nonferrous Metal Research. Published by Elsevier BV. All rights reserved.

1356

ARTICLE

Hot Workability and Microstructure Evolution of TiAl Alloy in (α2+γ) Dual-phase Field Deng Zhihai, Li Jinshan, Zhang Tiebang, Hu Rui, Zhong Hong, Chang Hui State Key Laboratory of Solidification Processing, Northwestern Polytechnical University, Xi’an 710072, P. R. China

Abstract: To explore the possibility of near conventional forging of TNM alloys using existing equipment for superalloy, the hot workability of a typical TNM alloy with normal composition of Ti-43Al-4Nb-1Mo-0.1B (at%) was investigated in (α2+γ) dual-phase field. Based on the hot workability maps and microstructure observation, it is found that the hot deformability of this alloy is dropped remarkably when the temperature decreases into (α2+γ) dual-phase field. The samples can be soundly compressed with strain rate of 0.001 s-1 at 1050~1150 oC and cracks are formed when the temperature is 1000 oC. As the temperature increases and/or the strain rate decreases, cracks vanish and more dynamic recrystallized grains are formed. Furthermore, its workability can be promoted when it is packed with 304 stainless steel with a thickness of 1.5 mm and the billets can be adequately deformed at 1050~1150 oC with strain rate of 0.1 s-1, which is favorable to current industrial trials. By comparison, there are not cracks but fewer recrystallized grains when packed-compressed under the same conditions.

Key words: hot workability; microstructure evolution; β-solidified TiAl alloys; packed forging

γ-TiAl based alloys represent an important class of alloys

providing a unique set of physical and mechanical properties that can lead to substantial payoffs in the automotive industry, power plant turbines and aircraft engines[1,2]. However, the major obstacle to the application of γ-TiAl based alloys as structural components is the lack of ductility at room temperature[3]. The absence of plasticity at classical fabrication temperatures is also a source of difficulties in processing them to form a component[3]. Therefore, it is important to improve their workability in order to achieve their industrial applications.

Recently, a series of γ-TiAl based alloys referred to be TNM alloy with the composition of Ti-(42~45)Al-(3~5)Nb- (0.1~2)Mo-(0.1~1)(B, C) (at%) have been explored[3]. These alloys solidify via the β-phase, exhibiting an isotropic, fine grained and texture-free microstructure with modest micro-segregations[3], which is favorable to both ingot breakdown and secondary forming operations. At the same time, these alloys show an adjustable volume fraction of disordered β-phase at elevated temperatures which acts as deformation accommodating phase[4], because the bcc β-lattice provides a sufficient number of independent slip

systems. In previous studies, it has been shown that these alloys exhibit an improved deformability at elevated temperatures and have been processed by isothermal forging[5] and near conventional forging[6]. However, the investigation of high temperature deformation behavior and the near conventional forging trials of TNM alloys were performed in (α+β) phase field according to the published data[3]. And the forging temperature, which is even higher than that of superalloys (always deformed at about 1150 oC[3]), poses a new challenge to the existing forging equipment.

In order to take full advantage of the current equipment widely used for hot working of superalloys, the investigation on the possibility of near conventional forging at a lower temperature is necessary. In this work, the hot workability and microstructure evolution of a TNM alloy with normal composition of Ti-43Al-4Nb-1Mo-0.1B (at%) were investigated in (α2+γ) dual-phase field by isothermal compression tests. The effect of cans on the hot workability of this alloy was also investigated based on the hot workability map and microstructure observations.

1 Experiment

Deng Zhihai et al. / Rare Metal Materials and Engineering, 2013, 42(7): 1356-1361

1357

The alloy with a nominal composition of Ti-43Al-4Nb-1Mo-0.1B (at%) (abbreviated as TNM-1 in the following text) was prepared using a consumable electrode arc melting technique in argon atmosphere for three times in order to reduce composition heterogeneity. The ingot with 140 mm in diameter and 340 mm in height was cast, and then hot isostatically pressed (HIPPed) under argon atmosphere of 140 MPa for 4 h to eliminate casting defects.

Two groups of isothermal compression tests were conducted using Gleeble-1500 system in argon atmosphere. Cylindrical specimens with a diameter of 8 mm and a height of 12 mm were machined by electro-discharge machining and were ground with 1000# abrasive paper after removing the oil in acetone. In the first group test, the naked specimens were isothermally compressed at temperatures varying from 1000 oC to 1150 oC and strain rates from 10-3 to 1 s-1. In the second group, the samples were packed by 304 stainless steel sheets with a thickness of 1.5 mm and isothermally compressed at temperatures of 1050, 1100 and 1150 oC and strain rates of 0.01, 0.05, and 0.1 s-1. In order to reduce the friction force between the pressure indenter and two head faces of specimens, a graphite lubricant was used for the isothermal hot-compression tests. Specimens were heated by induction coils with a heating rate of 5 oC. s-1 and held at the testing temperature for 300 s before the tests were conducted. The specimens were deformed to a total true strain of about 0.8. The true stress-true strain curves were obtained from the load-displacement data, which was recorded by computer automatically. In order to investigate the microstructure evolution during the deformation, the specimens were quenched in water immediately right after the deformation.

The microstructure of TNM-1 alloy both in as-cast and HIPed state and the central zone of the deformed samples was characterized by optical microscope, scanning electron microscope (SEM) with backscattered electron mode (BSE), and X-ray diffractometer (XRD) using Cu Kα radiation. The specimens for optical microscope were cut, ground, polished and then etched in Kroll’s agent (10% HF + 10% HNO3 + 80% H2O). The specimens for transmission electron microscopy (TEM) were first mechanically ground to about 50 μm and then thinned using a twin-jet polishing method in a solution of 21% perchloric acid, 29% n-butyl alcohol and 50% methanol with a voltage of 12.5 V at –25 oC.

2 Results 2.1 Initial microstructure

Fig.1a shows the SEM micrograph taken in BSE mode of TNM-1 alloy in the as-cast state after hot isostatic pressing. The alloy exhibits mainly a uniform near lamellar (NL) structure and the microstructure consists of lamellar (γ+α2)-colonies with a typical size of 40~80 μm, and some

brightly appearing phases located along the colony boundaries. From the TEM image shown in Fig.1b, it can be concluded that the lamellae are uniformly arranged and the average lamellar spacing is about 300 nm. The X-ray diffraction pattern in Fig.1c shows that there is a third phase identified as β phase in present alloy. The microstructural features clearly indicate that solidification proceeds completely via the β-phase[4]. 2.2 Flow behavior and hot workability maps

According to our previous study[7], the true stress-true strain curves of TNM-1 alloy exhibit that the flow stress is promptly increased to a peak at a relatively low strain (<0.1), and then gradual softening occurs with the strain increasing. Furthermore, TNM-1 alloy exhibits a strong dependence of the flow stress on strain rate and deformation temperature and the flow stresses increase with the increase of strain rate and the decrease of deformation temperature. The peak stress, which is a critical parameter for hot working, is a little lower than that of Ti-45Al-(8-9)Nb-(W, B, Y) alloy deformed in the same condition[8]. Therefore, it can lower the deformation resistance and the limited capacity of the hydraulic press.

During the hot deformation, the thermal activation is a critical factor to affect the hot deformability. The dependence of flow stress σ on the strain rate ε and temperature T during high temperature deformation can be described as following[9]:

exp( ) [sinh( )]nQZ ART

ε ασ= = (1)

where Z is Zener-Hollomon parameter, n is the stress exponent, A and α are material constants, and Q is deformation activation energy.

Eq. (1) can be transformed into Eq. (2) by taking ε and T as constant:

ln ln[sinh( )]ln[sinh( )] (1 / )T

Q RT ε

ε ασασ

⎧ ⎫ ⎧ ⎫∂ ∂= ⎨ ⎬ ⎨ ⎬

∂ ∂⎩ ⎭ ⎩ ⎭ (2)

According to the true stress-true strain curves (Fig.2, not all of them are shown here), the slopes of lnε - ln[sinh( )]ασ and ln[sinh( )]ασ -1/T curves can be obtained (Fig.2), then the apparent activation energy Q and the stress exponent n can be measured as 498 kJ/mol and 3.13, respectively.

Fig.3 is the workability maps and deformation results (macroscopic pictures of the deformed samples) of TNM-1 alloy. It can be obtained from Fig.3a that external cracks are found for the samples compressed to a true strain of 0.8 at 1000 oC or with the strain rates of 1 s-1. with the temperature increasing and/or the strain rate decreasing, the workability of present alloy increases and the external cracks get fewer and smaller. It is worth noting that there is no crack for all the samples compressed above 1000 oC with the strain rate of 1×10-3 s-1. However, this strain rate is not

Deng Zhihai et al. / Rare Metal Materials and Engineering, 2013, 42(7): 1356-1361

1358

Fig.1 SEM micrograph taken in BSE mode (a), TEM image of one lamellar colony (b) and XRD pattern of HIPed TNM-1 alloy (c) Fig.2 Relationships for the TNM-1 alloy: (a) ln[sinh( )]ασ - lnε ,

(b) ln[sinh( )]ασ -1/T suitable for industrial trials. As shown in Fig.3b, the workability of TNM-1 alloy is promoted remarkably by packing the specimens with stainless steel cans. After decanned, all the samples, deformed at 1050~1150 oC with strain rates of 0.01~0.1 s-1, appear to be sound and do not have any penetrating cracks, indicating that the alloy has good deformation ability. 2.3 Microstructure evolution

Fig.4 shows the microstructure of the naked samples of TNM-1 alloy deformed at different temperatures from 1000 oC to 1150 oC with strain rates of 0.001, 0.01 and 0.1 s-1. It can be seen from Fig.4 that the temperature and the strain rate both have significant effects on the microstructure. When the sample is compressed at 1000 oC with strain rate of 0.1 s-1, its microstructure is mostly unchanged besides that a few

Fig.3 Hot workability maps of TNM-1 alloy under different

conditions: (a) naked, (b) packed. The inserts of each picture in the right are the corresponding macroscopic pictures

lamellar structure parallel to the compression direction starts to kink (area A in Fig.4a) and some cavities can be observed at the lamellar colony boundaries (area B in Fig.4a). As compression temperatures increase and/or strain rates decrease, cavities at lamellar colony boundaries disappear and the lamellar structure commences to break-up, globalizes and recrystallizes. Moreover, the lamellar colony rotates, stretches and finally forms strip-like structure, which is perpendicular to the compression direction during the deformation (Fig.4b to Fig.4f). It can be concluded that more fine equiaxed recrystallized grains form at lamellar colonies boundaries when the strain rates decrease (Fig.4a, Fig.4c and Fig.4e) and the compression temperatures increase (Fig.4a, Fig.4c and Fig.4e). However, γ lamellae in the lamellar colonies decreases and the characterization of the lamellar structure is indistinct when deformed at 1150 oC with strain rate of 0.01 s-1 (Fig.4e). This is may be attributed to the rise of the temperature of the sample.

1000 1100 1200

10

1

0.1

0.01

1E-3

1E-4

Stra

in R

ate/

s-1 Crack Little Crack Sound a

1000 1100 1200

Temperature/ºC

1

0.1

0.01

1E-3

Stra

in R

ate/

s-1 Crack Little Crack Sound

b

20 40 60 80

2θ/(º)

1500

1200

900

600

300

0

Inte

nsity

/cps

γ α2 β

a b

50 µm 1 µm

c

0.00070 0.00074 0.00078

1/T

2

1

0

–1

–2

ln[s

inh(ασ

)]

b

0.001 s-1 0.01 s-1

0.1 s-1

1 s-1

–7 –6 –5 –4 –3 –2 –1 0 1

lnε

2

1

0

–1

–2

ln[s

inh(ασ

)]

a

1000 ℃ 1050 ℃ 1100 ℃ 1150 ℃

·

Deng Zhihai et al. / Rare Metal Materials and Engineering, 2013, 42(7): 1356-1361

1359

Fig.4 SEM images of the naked TNM-1 alloy deformed at different conditions to a true strain of 0.8: (a) 1000 oC, 0.1s-1, (b) 1100 oC,

0.1 s-1, (c) 1000 oC, 0.01 s-1, (d) 1100 oC, 0.01 s-1, (e) 1000 oC, 0.001 s-1, and (f) 1150 oC, 0.01 s-1

Fig.5 SEM images of the packed TNM-1 alloy deformed at different conditions to a true strain of 0.8: (a) 1050 oC, 0.1s-1, (b) 1100 oC, 0.1 s-1, (c) 1100 oC, 0.01 s-1, and (d) 1150 oC, 0.01 s-1

Fig.5 shows the microstructure of the packed samples of

TNM-1 alloy deformed at different temperatures from 1050 oC to 1150 oC with strain rates of 0.01, 0.05 and 0.1 s-1. It can be seen from Fig.5 that the microstructure of the packed samples almost has the similar feature to that of the naked samples, such as kinking of the lamellar structure and formation of fine and equiaxed recrystallized grains at lamellar colony boundaries. However, when compared with the naked samples, something interesting in the packed samples should not be ignored that the deformation is more uniform and there is no cavity and crack in all the samples under the protection of stainless steel cans. Therefore, a well-selected pack can improve the hot workability of TiAl alloys and get rid of formation of cavity and inner crack during hot work. By comparison, there are fewer recrystallized grains in the packed samples when compressed under the same conditions. It is worth noting that there is almost no lamellar structure when the packed samples are compressed at 1150 oC with strain rate of 0.01 s-1. This is attributed to deforming heat and heat preservation effect of the steel can and it will be discussed in Section 3.2.

3 Discussions 3.1 Hot deformation behavior

The shape of the S-S curves reflects the changes in the microstructure during compression deformation. The flow curve can be divided into two stages: work hardening stage and flow softening stage. And there is a third stage, steady-state flow stage, in some curves. In the work hardening stage, the dislocations density increases significantly as the strain increases, and the dislocations are tangled and piled up, which induces the increase of the flow stress, and a peak stress is reached. The decrease of the flow stress after the peak stress may be owing to dynamic recrystallization (DRX)[3], breakup of the lamellar microstructure[10], increase of temperature of the billet due to plastic deformation[11] and occurrence of flow localization and surface cracking[12]. These assumptions, especially about microstructure evolution, need to be checked carefully.

When the temperature is lower than 1050 oC or the strain rate is higher than 0.1 s-1, the stress increases sharply with the increase of strain and cavities and small cracks formed at lamellar colony boundaries because of stress

a b c

f ed

100 µm

a b c d

100 µm

Deng Zhihai et al. / Rare Metal Materials and Engineering, 2013, 42(7): 1356-1361

1360

concentration resulting from poor mobility of the dislocations. Once small cracks are formed, they spread and connect with each other quickly and transform into penetrating surface cracks finally (Fig.3a). So, the reason for the decrease of the flow stress in this condition is the occurrence of surface cracking. As the temperature increases and the strain rate decreases, the lamellar structure starts to kink and break up, which is also one mechanism of flow softening. However, the dominant mechanism of flow softening during the hot deformation is DRX, which can be estimated from the flow curve shapes and microstructure examination. Due to their lower stacking fault energy (SFE) and poor mobility of the dislocations for TiAl alloys, DRX is one of the restoration processes of γ-TiAl base alloys during high temperature deformation and plays an important role in hot working of γ-TiAl base alloys. Because the microstructure of TNM-1 alloy mainly consists of lamellar structure (Fig.1a), the DRX process depends on lamellar orientation [3]. When the lamellae are inclined at an intermediate angle to the loading axis, the lamellar colonies are easily deformed by shear parallel to the interfacial planes. At the same time, the yield stress is moderately high when the lamellae are parallel to the loading axis. Both of them are favorable to the DRX process. However, when the loading axis is perpendicular to lamellar the yield stress is the highest and the stability of the lamellar structure is the highest. Therefore, there are many lamellar colonies with lamellae perpendicular to the compression axis, which may restrict the hot workability of TNM-1 alloy.

To further understand the hot deformation behavior of TNM-1 alloy, the activation energy Q was measured as 498 kJ/mol based on the S-S curves. The value of Q can indicate the hot deformability of materials. The measured value of Q is a little larger than 417 kJ/mol for Ti-45.5Al-2Nb-2Cr[13], 449 kJ/mol for Ti-46Al-2W[14], 448.6 kJ/mol for Ti-45Al-5Nb and 451.4 kJ/mol for Ti-43Al-9V-0.3Y[15]. That indicates the deformation mechanisms, such as dislocation sliding and climbing, can be hardly started for the present alloy when deformed in (α2+γ) field. That is also in coincidence with the existence of cavities and cracks in the deformed samples during microstructure examination. Therefore, the hot workability of TNM-1 alloy drops sharply when the deformation temperature decreases into (α2+γ) field. Fortunately, it can be improved by packing the samples with well selected cans, which will be discussed in the next part. 3.2 Effect of pack on hot workability

The parameter of importance in mechanical working of materials is termed as “workability”, which refers to relative ease with which a metal can be shaped through plastic deformation without any defects[16]. Workability of material is highly dependent on the processing parameters such as strain rate and temperature. Packed-forging has

many advantages, such as high efficiency, low cost, and low requirements for equipment. Therefore, in this work, the effect of the pack on hot workability of TNM-1 alloy was studied by packing the samples with well selected stainless steel cans according to Liu’s investigation results[17].

According to the hot workability map shown in Section 2.2 and microstructure examination in Section 2.3, the hot workability of TNM-1 alloy is improved by packing the samples with stainless steel cans. The packed samples are deformed more homogeneously and do not have any penetrating cracks or cavities even compressed in (α2+γ) phase filed. There are two explanations for this improvement of hot workability. Firstly, the outer pack can offset the secondary stress generated in the compression process and avoid cracks in billets[16]. Secondly, owing to the thermal insulation effect of the packs, there is a less temperature gradient in the packed samples, which causes inhomogeneity in microstructure and deformation, compared with the naked samples.

The thermal insulation effect of the packs has been mentioned in several studies and it can be confirmed in this study by microstructure examination. During the isothermal compression, the thermocouples are attached to the outside of the packs and the temperature of the packs is invariable. So, the temperature of the samples raised by plastic deformation can be used to check the thermal insulation effect of the packs. It can be concluded that the temperature of the sample after compression surpasses the α-transus temperature of TNM-1 alloy (about 1230 oC[18]) when compressed at 1150 oC with strain rate of 0.1 s-1 because there is no lamellar structure in the deformed microstructure (Fig.5d). However, the temperature of naked samples is much lower because of the existence of large part of lamellar structure in the deformed sample in the same deformation condition (Fig.4f), which indicates the temperature of the sample is below the α-transus temperature. Therefore, less temperature drop of the billets happens during packing forging processing due to the decrease of heat transfer between billets and the instrument, the thermal radiation and the heat convection between billets and surrounding environment.

4 Conclusions

1) When compressed below 1050 oC with strain rate larger than 0.1 s-1, the reason for the decrease of the flow stress is the occurrence of surface cracking. However, breaking-up of the lamellar structure and DRX are the main reasons for the decrease of the flow stress when the temperature increases and/or the strain rate decreases.

2) The deformation activation energy of TNM-1 alloy is 498 kJ/mol and little larger than that of conventional TiAl alloys because there is more α2 phase in TNM-1 alloy and α2 phase is harder to deform than γ phase. This also

Deng Zhihai et al. / Rare Metal Materials and Engineering, 2013, 42(7): 1356-1361

1361

indicates that TNM-1 alloy possesses relatively low hot workability when deformed in (α2+γ) phase filed.

3) The hot workability of TNM-1 alloy can be effectively improved by packing the samples with well designed cans. The packed TNM-1 alloy can be soundly and homogeneously deformed at 1050~1150 oC with strain rate of 0.1 s-1, which is more favorable to current industrial trials.

References

1 Lapin J. METAL 2009[C]. Czech Repubic: Hradec nad Moravicí, 2009: 19

2 Appel F, Brossmann U, Christoph U et al. Advanced Engineering Materials[J], 2000, 2(11): 699

3 Clemens H, Kestdr H. Advanced Engineering Materials[J], 2000, 2(9): 551

4 Clemens H, Chladil H F, Wallgram W et al. Intermetallics[J], 2008, 16(6): 827

5 Appel F, Oehring M, Paul J D H et al. Intermetallics[J], 2004, 12(7-9): 791

6 Wallgram W, Clemens H, Kremmer S et al. Advanced Intermetallic-Based Alloys for Extreme Environment and Energy Applications[C]. Warrende: Materials Research

Society, 2009: 109 7 Deng Z H, Li J S, Zhang T B et al. Proceedings of 12th World

Conference on Titanium[C]. Beijing: Science Press, 2011 8 Xu X J, Lin J P, Wang Y L et al. Materials Science and

Engineering: A[J], 2006, 416(1-2): 98 9 Sellars C M, McTegart W J. Acta Metallurgica[J], 1966, 14(9):

1136 10 Liu B, Liu Y, Li Y P et al. Intermetallics[J], 2011, 19(8): 1184 11 Laasraoui A, Jonas J. Metallurgical and Materials

Transactions A[J], 1991, 22(7): 1545 12 Kim J S, Lee Y H, Kim Y W et al. Materials Science

Forum[J], 2007, 539-543: 1531 13 Seetharaman V S S L. Metall Trans A[J], 1996, 27: 1987 14 Kim H Y, Sohn W H, Hong S H. Materials Science and

Engineering A[J], 1998, 251(1-2): 216 15 Kong F T, Chen Y Y, Li B H. Materials Science and

Engineering A[J], 2009, 499(1-2): 53 16 Semiatin S, Seetharaman V, Weiss I. Mat Sci Eng A[J], 1998,

243(1): 1 17 Liu Y. T Nonferr Metal Soc[J], 2002, 04: 596 18 Wallgram W, Schmolzer T, Cha L M et al. Int J Mater Res[J],

2009, 100(8): 1021