227

The Effect of Texture on the Serrated Flow

in Peak-Aged 2090 Al-Li Alloy

Y.Z. Shena, K.H. Ohb and D.N. Leec

Research Institute of Advanced Materials and School of Materials Science and Engineering,

Seoul National University, Seoul 151-744, Korea

ashenliu8@snu.ac.kr, bkyuhwan@snu.ac.kr, cdnlee@snu.ac.kr

Keywords: 2090 Al-Li alloy, solution-treatment, peak-aging, serrated flow, dynamic strain aging, Portevin-Le Chatelier effect, texture.

Abstract. Tensile specimens cut from the surface layer to the center layer of a 12.7 mm thick 2090 Al-Li alloy plate were solution treated at 550°C for 30 min and subsequently peak-aged at 190°C for 18 h. They were tensile tested along the rolling direction at 25°C at various strain rates. The

solution-treated specimens gave rise to serrated flows at a strain rate of 2×10-4 s-1. On the other hand, for the peak-aged alloy, the surface-layer and subsurface-layer specimens underwent complex, serrated flows (fine and coarse types superimposed each other), whereas the center-layer and near-center-layer specimens were devoid of serrated flows. The textures of the surface-layer and subsurface-layer specimens were approximated by the {001}<110> orientation, while those of the center-layer and near-center-layer specimens were approximated by the {011}<211> orientation. The different flow behaviors were discussed based on the crystallographic textures, microstructures and the strain rates.

Introduction

Explanation of the phenomenon of serrated flow, or the Portevin-Le Chatelier (PLC) effect in Al-Li alloys are based on two lines of thought: One is dynamic strain aging (DSA) involving dynamic interaction between dissolved lithium solute atoms and mobile dislocations [1-4] and another is

shearing of δ′ (Al3Li) precipitates by mobile dislocations [5-8]. In general, aluminum alloys are known to exhibit a considerable anisotropy of mechanical

properties. This anisotropy is usually related to the crystallographic texture and in turn active slip systems. Since the PLC effect involves the interactions between gliding dislocations and other crystal defects, one should expect that its intensity may also exhibit anisotropy [9]. Such anisotropy has been reported on a commercial Al-Mg alloy by Cheng and Morris [10] and on an Al-Li-Cu-Zr model alloy by Mizera and Kurzydlowski [9]. They accounted for the observed difference in the PLC effect intensity in terms of anisotropy of the microstructure or the morphology of the grains. On the other hand, little work has been made of the textural effect on the PLC effect.

The previous study [11] indicated that serrated flow behavior in a peak-aged 2090 Al-Li alloy was strongly affected by the textures of specimens. The purpose of this study is to investigate the flow behavior of peak-aged specimens compared with that of solution-treated specimens.

Experimental procedure

The material used in this study was a near peak-aged 2090-T81 Al-Li alloy (2.05% Li, 2.86% Cu, 0.12% Zr, balance Al) plate of 12.7 mm in thickness, produced by Alcoa, UK. Two types of specimens were prepared from the plate. One is RT-specimen, in which the tensile axis is aligned with the rolling direction (RD) while the width direction is parallel to the transverse direction (TD), and another one is RN-specimen, in which the tensile axis is aligned with RD while the width direction is parallel to the normal direction (ND) as shown in Fig. 1. In order to investigate properties along the ND of the plate, it was sliced into four sheets of about 1.2 mm in thickness from the surface

Solid State Phenomena Vol. 105 (2005) pp. 227-232online at http://www.scientific.net© 2005 Trans Tech Publications, Switzerland

Licensed to D.N. (dnlee@snu.ac.kr) - KoreaAll rights reserved. No part of the contents of this paper may be reproduced or transmitted in any form or by any means without thewritten permission of the publisher: Trans Tech Publications Ltd, Switzerland, www.ttp.net. (ID: 147.46.69.28-04/05/05,02:50:54)

228

Fig. 1. Designation of tensile specimens.

layer to the center layer as shown in Fig.1. RT type tensile specimens of 1 mm×6 mm×25 mm (ASTM B557M-94) in gauge dimension were cut from the sheets. It is noted that RT0-specimen and RT3-specimen are from the center and surface layers, respectively, and RN-specimen is equivalent to a specimen consisting of one RT0-specimen and about two RT1-specimens.

The specimens were solution treated at 550˚C for 30 min. Some of the solution treated specimens were peak-aged at 190˚C for 18 h. All the treatments were performed in a salt bath.

Tensile tests were carried out at initial strain rates of 2×10-4 s-1 to 10-2 s-1 at 25˚C. Thin foil specimens were examined using a Philips CM20 transmission microscope operated at 200 kV. The textures of the specimens were measured with an x-ray texture goniometer in the back reflection

mode with Ni filtered Cu-Kα radiation. The (111), (200), and (220) partial pole figures were measured and used to calculate the orientation distribution functions (ODFs) by the WIMV method [12]. Other experimental details are described in Ref. [11].

Results and Discussion

Microstructures. Fig. 2 shows an optical microstructure and dark-field TEM micrographs of the solution-treated and peak-aged specimens. Elongated grains can be seen in the optical microstructure. The small white particles in the TEM micrograph of solution-treated specimen are

thought to be δ′ (Al3Li) precipitates, because δ′ precipitation takes place even during quenching from the solution temperature or rapidly at room temperature [13]. The white particles in the TEM

micrograph of the peak-aged specimen are δ′ (Al3Li) precipitates. The δ′ phase is ordered fcc and

has an L12 crystal structure. It is also known that the δ′ phase is fully coherent and has a very small

lattice mismatch with the fcc α matrix [14]. Textures of surface and center layers. The textures of the solution treated specimens were the

same as those of the peak-aged specimens. Fig. 3 shows the orientation distribution functions (ODFs) of various specimens from the peak-aged Al-Li alloy plate. The textures of the RT3 and RT2 specimens are almost same and can be approximated by the {001}<110> orientation. The textures of the RT0 and RT1 specimens are almost the same and can be approximated by the {011}<211> orientation. The texture of the RN specimen can be approximated by the {111}<211> orientation. With reference to Fig. 1, the RT-specimen plane is normal to ND of the plate and ND is parallel to the width direction of the RN-specimen. Therefore, the {111}<211> orientation of the RN-specimen is equivalent to the {011}<211> orientation of the RT-specimen. Since the textures of the RT0 and RT1 specimens are almost same, the RN-specimen, which comprises one RT0 specimen and about two RT1 specimens, has almost the same texture as that of the RT0-specimen.

From the fact that the {001}<110> orientation is the major component in shear deformation texture of aluminum alloys, and the surface layer of aluminum undergo shear deformation (e.g., [15]), the surface and subsurface layers (RT3 and RT2) of the Al-Li alloy plate must have been shear deformed during rolling. The shear texture remains little changed even after annealing [16]. On the other hand, the center and near-center layers (RT0, RT1, and RN), which undergo plane strain compression, show a texture approximated by the {011}<211> orientation, which is obtained in many copper alloys with low stacking fault energies [17] and Cu-Mn alloys with high strain-hardening rates [18]. These results indicate that the deformation texture of the starting plate

RT

RN

RD

TD

ND

RD

TD

ND

center RT0 RT1 RT2RT3

Texture and Anisotropy of Polycrystals II228

229

Fig.2. (a) Optical microstructure and dark-field TEM micrographs of (b) solution-treated, (c) peak-aged 2090 Al-Li alloy specimens. L, LT, and ST stand for longitudinal (rolling),

long transverse (transverse), and short transverse (normal) directions of plate.

remained almost unchanged even after the solution heat treatment and subsequent aging treatments.

Flow stresses of surface and center layers. Fig. 4 shows the flow curves of RT0, RN, and RT3-specimens taken from the solution-treated Al-Li alloy specimen. The RT0-specimen has the highest flow stress, and the RT3-specimen has the lowest flow stress. The flow curve of the RN-specimen is located between the RT0 and RT3-specimens. This behavior is understandable from the fact that the RN-specimen is equivalent to a specimen consisting of one RT0 and about two RT1-specimens. Fig. 5 shows the flow curves of various specimens taken from the peak-aged Al-Li alloy plate. The differences in flow curves of specimens from different depth layers of the same plate could be caused by different crystallographic textures. The solution-treated specimens and the peak-aged specimens showed the same texture. It indicates that the texture of matrix did not change by precipitation. The flow stress of each specimen decreases with increasing strain rate, implying a negative strain rate sensitivity.

Regardless of heat-treatments (solution treatment, peak-aging), the center-layer specimens show about 1.3-1.4 times higher flow stresses than the surface-layer ones at a given strain rate. This can be qualitatively explained using the approximate orientations of the surface and center layers,

which are the (001)[110] orientation and the )110( [211] orientation, respectively. The tensile

directions of the (001)[110] and )110( [211] oriented specimens are [110] and [211], respectively.

For an fcc single crystal, the yield stresses under a tensile stress along the [110] and [112] axes

will be same, because the largest Schmid factors for the two axes are all 6/6 (Table 1).

However, for an fcc polycrystalline material with a well developed texture, its yield behavior will not be controlled only by slip systems with the largest Schmid factors. Lee and Oh [19] calculated the plastic anisotropy of polycrystalline metals by assuming that all the slip systems contributed to the deformation but that their contributions were proportional to their Schmid factors. Similarly, the yield stress or flow stress of a polycrystalline may be calculated by

avgc

i

i

c Mmn

ττ

σ == ∑ (1)

where σ , cτ , n, and im are the flow stress, the critical resolved shear stress, the total number of

active slip systems, and the reciprocal of the Schmid factor on the ith slip system, respectively. The flow stress for the [110] tensile axis is given by

=]110[σ 6/6 cτ (2)

If all the slip systems with non-zero Schmid factors are assumed to be active, the flow stress for the [112] tensile axis is given by

100nm

cb

nm 100

a

μm 500

Solid State Phenomena Vol. 105 229

230

6/5.10]112[ cτσ = (3)

RT0 RT1 RT2 RT3 RN

Fig. 3. ODFs of RT0, RT1, RT2, RT3, RN specimens from peak-aged Al-Li alloy plate. Textures of RT0 and RT1 are approximated by {011}<112>. Textures of RT2 and RT3 are approximated by {001}<110>. Texture of RN is approximated by {111}<112>. Here {hkl} indicates lattice planes parallel to specimen plane and <uvw> indicates lattice directions parallel to tensile direction.

Plane 111 111 111 111 Direct. 101 110 101 110 101 110 110 110 011 101 101 011

[110] 0 6 /6 6 /6 0 0 0 0 0 0 0 6 /6 6 /6

[112] 0 6 /9 6 /9 6 /9 6 /6 6 /18 6 /9 6 /18 6 /6 0 0 0

Table 1. Calculated Schmid factors for tensile axes of [110] and [112] of fcc crystal having {111}<110> slip systems.

If the slip systems with the Schmid factor of 18/6 are assumed to be non-active, because the

factor is small compared with 9/6 and 6/6 ,

6/8]112[ cτσ = (4)

Thus, 75.1/ ]110[]112[ =σσ for ]112[σ in Eq. (3) and 33.1/ ]110[]112[ =σσ for ]112[σ in Eq. (4). It can be

seen that 33.1/ ]110[]112[ =σσ is in qualitative agreement with the experimental results.

Types of serrated flows. The experimental results in Figs. 4 and 5 show two different types of serrated flows, fine (F) type and coarse (C) type (Fig. 6). The flow curves of the solution-treated specimens show F-type serration (Fig. 4). On the other hand, F-type serration is drastically reduced and C-type serration appears in the flow curves of the peak-aged specimens (Fig. 5). The peak-

aging brings about the precipitation of δ′ (Al3Li) and reduces the concentrations of dissolved solutes. It follows from this result that F-type serration is associated with dynamic strain aging (DSA) (or Portevin-Le Chatelier effect) due to repeated locking of moving dislocations by solute

atoms, especially Li, and C-type serration seems to be associated with shearing of δ′ precipitates. When a specimen containing precipitates is elongated, dislocations pile up on the slip planes at

the precipitates. If the stress on the dislocation at the head of the dislocation pile-up reaches the shear yield stress of precipitates, the precipitates are sheared and a stress drop can take place.

Such processes are repeated to form C-type serration. If the stress on the dislocation at the head of dislocation pile-up reaches the fracture stress of precipitates, cracking of the precipitates takes

Max: 9.5Max: 9.5Max: 9.5Max: 9.5

1 3 5 7 9 1 3 5 7 9 1 3 5 7 9 1 3 5 7 9

Max: 9.5Max: 9.5Max: 9.5Max: 9.5

1 3 5 7 9 1 3 5 7 9 1 3 5 7 9 1 3 5 7 9

Max:10.2Max:10.2Max:10.2Max:10.2

1 3 5 7 9 1 3 5 7 9 1 3 5 7 9 1 3 5 7 9

Max:10.2Max:10.2Max:10.2Max:10.2

1 3 5 7 9 1 3 5 7 9 1 3 5 7 9 1 3 5 7 9

Max:10.8Max:10.8Max:10.8Max:10.8

1 3 5 7 9 1 3 5 7 9 1 3 5 7 9 1 3 5 7 9

Max:10.8Max:10.8Max:10.8Max:10.8

1 3 5 7 9 1 3 5 7 9 1 3 5 7 9 1 3 5 7 9

Max:11.8Max:11.8Max:11.8Max:11.8

1 3 5 7 9 1 3 5 7 9 1 3 5 7 9 1 3 5 7 9 11 11 11 11 Max:11.8Max:11.8Max:11.8Max:11.8

1 3 5 7 9 1 3 5 7 9 1 3 5 7 9 1 3 5 7 9 11 11 11 11 9,7,5,3,1

3.10:max

Texture and Anisotropy of Polycrystals II230

231

place. The precipitate cracking will likely to bring about fracture of specimen, resulting in a reduction of its elongation.

0 5 10 15 20 25 30 350

100

200

300

400

RN

RT3

RT0

2090 Al-Li Alloy

Solution-treated : 550oC x 30 min

Tension : 2 x 10-4s

-1, 25

oC

Str

ess(M

Pa)

Strain (%)0 2 4 6 8 10 12

0

100

200

300

400

500

600

(8)(7)

(1) RT3, 2 x 10-4 s

-1

(2) RT3, 10-3 s

-1;

(3) RT3, 2 x 10-3 s

-1; (6) RT1, 2 x 10

-4 s

-1

(4) RT2, 2 x 10-4 s

-1; (7) RT0, 2 x 10

-4 s

-1

(5) RT2, 10-2 s

-1; (8) RN, 2 x 10

-4 s

-1

S

tress (

MP

a)

Strain (%)

(6)

(4)

(5)(1)

(2)(3)

However, these two types of serrations are not independent each other, because solute atoms can pin stagnant dislocations surrounding the precipitates, if any. This makes it impossible to simply add the effects of solute atoms and precipitates on serration. Nevertheless, such a rough classification is convenient for understanding the difference in serration before and after peak-aging.

Texture effect on serration. For the peak-aged alloy plate, the surface-layer and subsurface-layer

specimens show serration at strain rates from 2×10-4 s-1 to 10-2 s-1 (Curves 1-5 in Fig. 5), whereas the center-layer and near-center-layer specimens are almost devoid of serration (Curves 6-8 in Fig. 5). When the (001)[110] oriented surface-layer and subsurface-layer specimens are elongated, it

can be supposed that the ],110)[111( ],101)[111( ],101)[111( and ]011)[111( slip systems are

equally activated, because the Schmid factors on the systems are same (Table 1), and dislocations

pile up on the slip planes at δ′ spherical precipitates (Fig. 2c). When the stress on the dislocation at the head of the dislocation pile-up reaches the shear yield stress of the precipitates, the precipitates are sheared, resulting in the stress drop (C-type).

However, tensile straining of the )110( [211] oriented center-layer and near-center-layer

specimens activates the (111) ]110[ , (111) ]101[ , )111( [110], )111( [101], )111( [110], and )111(

[011] slip systems (Table1). The slip systems having √6/18 are neglected. Thus, the Schmid factors for the active slip systems are different unlike the (001)[110] oriented surface-layer and

subsurface-layer specimens. The slip systems on which the Schmid factor is √6/6 will first be

activated. If the slip systems of which the Schmid factor is √6/9 are activated by cross slip before the stress on the dislocation pile-up on the first slip systems reaches the shear yield stress of the precipitates, particles are not likely to be sheared. In this case, no C-type serration will take place.

This may be the case of the flow curve of the center layer specimen, which is devoid of serration. This does not negate the possibility of shearing of the precipitates. If a sufficient number of precipitates are not sheared, the serration cannot be detected.

The smaller elongations of the peak-aged center-layer specimens imply that dislocations on the

slip systems with the smaller Schmid factors are difficult to shear δ′ precipitates without cracking,

Fig. 4. Flow curves of RT0, RN, TR3 specimens from solution-treated 2090 Al-Li

alloy plate at tensile strain rate of 2×10-4 s-1.

Fig. 5. Flow curves of RT3, RT2. RT1, and RN specimens from peak-aged 2090

Al-Li alloy at various strain rates at 25°C.

F

C

Fig. 6. Schematic of F and C types of flow curves.

Solid State Phenomena Vol. 105 231

232

because the precipitates associated with the smaller Schmid factors are subjected to higher tensile stresses than those with the smaller Schmid factors.

Strain rate effect on serration. According to flow curves at different strain rates of the peak-aged specimens, the serration characteristics of the surface-layer specimen (RT3), appear to be

composed of C-type and F-type. At a strain rate of 2×10-4 s-1, the flow curve contains both C-type and F-type serrations, but as the strain rate increases to the order of 10-3 s-1, F-type serration disappears. For the subsurface-layer specimens (RT2), C-type and F-type serrations are observed at

a strain rate of 2×10-4 s-1, whereas F-type serrations disappear as the strain rate increased to 10-2 s-1. That is, for the peak-aged surface-layer (RT3) and subsurface-layer (RT2) specimens, F-type serration decreases with increasing strain rate while C-type serration is insensitive to strain rate, as shown in Curves 1-5 of Fig. 5.

The evolution of DSA is a function of strain rate. If the dislocation velocity is too fast compared with the diffusion speed of solute atoms, the solute atoms cannot catch and lock dislocations, resulting in no DSA. If the dislocation velocity is too slow compared with the diffusion speed of solute atoms, the solute atoms can catch but move together with dislocations, resulting in no DSA.

Acknowledgement

This study has been supported by Texture Control Laboratory (NRL), Seoul National University.

References

[1] L.P. Kubin, A. Styczynski, Y. Estrin: Scripta Metall. Mater. Vol. 26 (1992), p. 1423.

[2] J.C. Huang, G.T. Gray: Scripta Metall. Mater. Vol. 24 (1990), p. 85.

[3] S. Kumar, J. Król, E. Pink: Scripta Mater. Vol. 35 (1996), p.775.

[4] N. Ilić, D.J. Drobnjak, V. Radmilović, M.T. Jovanović, D. Marković: Scripta Mater. Vol. 34

(1996), p.1123.

[5] S.J. Zambo, J.A. Wert: Scripta Metall. Mater. Vol. 29 (1993), p.1523.

[6] S. Kumar, E. Pink: Scripta Metall. Mater. Vol. 32 (1995), p. 749.

[7] F. Chmelík, E. Pink, J. Król, J. Balík, J. Pešička, P. Lukáč: Acta mater. Vol. 46 (1998), p. 4435.

[8] S. Kumar, E. Pink: Acta mater. Vol. 45 (1997), p. 5295.

[9] J. Mizera, K.J. Kurzydlowski: Scripta Mater. Vol. 45 (2001), p. 801.

[10] X.-M. Cheng, J.G. Morris: Scripta Mater. Vol. 43 (2000), p. 651.

[11] Y.Z. Shen, K.H. Oh and D.N. Lee: Scripta Mater. Vol. 51 (2004), p. 285.

[12] S. Matthies: Phys stat sol. Vol. 101 (1980), p 111.

[13] M.H. Tosten, A.K. Vasudevan, P.R. Howell: Metall. Trans. Vol. 19A (1988), 51.

[14] D.B. Williams, J.W. Edington: Metal Sci. Vol. 9 (1975), p.529.

[15] K.-H. Kim and D.N. Lee: Acta mater. Vol. 49 (2001), p. 2583.

[16] D.N. Lee and K.-H. Kim, Innovations in Processing and Manufacturing of Sheet Materials

(edited by M.Y. Demeri, TMS, 2001), p. 219.

[17] P.A. Beck and H. Hu: Trans. AIME Vol. 194 (1952), p. 83.

[18] O. Engler: Acta mater. Vol. 48 (2000), p. 8427.

[19] D.N. Lee and K.H. Oh: J Mater. Sci. Vol. 20 (1985), p. 3111.

Texture and Anisotropy of Polycrystals II232