Texture and Grain Boundary Evolutions of Continuous Cast and Direct Chill Cast AA 5052 Aluminum Alloy During Cold Rolling

ABSTRACT

INTRODUCTION

Compared with aluminum alloy sheet made by conventional direct chill (DC) cast ingots, aluminum alloy sheet made by using continuous cast (CC) technology possess advantages of both energy and economic savings while reducing environmental emissions that have become an urgent issue in today’s world. However, there exist substantial differences in the microstructures of CC and DC materials as a result of the difference in the casting processes. The cast structures are almost completely broken up in DC material by extensive thermomechanical processes such as homogenization and hot rolling in reversing mills. After hot rolling, the DC hot band has a uniform distribution of the constituent particles and solid state produced precipitates. On the contrary, the CC material reaches the final stage directly from the cast state without any homogenization or significant degrees of hot working. As a result, CC hot band has a banded intermetallic structure in which the spatial distribution of the particles is not uniform. This situation leads to different recrystallization and texture behaviors and therefore different mechanical property behaviors in the various temper conditions of DC and CC materials. Usually, the formability of CC material is inferior to DC material.

AA 5052 is one of the AA 5XXX aluminum alloys that finds a variety of applications in the automotive industry because it lowers the vehicle weight and hence is beneficial for fuel economy. Previous research work on AA 5052 alloy has addressed the texture evolution during either hot rolling^{1, 2} or cold rolling³, formability⁴ and earing behavior^{5, 6}. However, little attention has been paid to the differences in cold rolling texture evolution between CC and DC AA 5052 alloy. The differences could be strongly related to different formability characteristics between CC and DC aluminum alloys. The aim of this study is two fold. First, the cold rolling texture evolution of both CC and DC AA 5052 alloy will be traced and the differences will be highlighted. Second, the grain boundary evolution of both materials during the early stages of cold rolling (≤40% thickness reduction) will be observed and compared.

EXPERIMENTAL

A. Materials and Procedures

The materials used in this work were industrially produced hot bands of CC and DC AA 5052 alloy. The chemical compositions are given in Table I. Plates of 3.7×101.6×127.0 mm³ were cut from the DC hot band and plates of 4.1×101.6×127.0 mm3 from the CC hot band. The cut plates were then annealed at 550°C for 2 hours followed by cold rolling in order to generate a completely recrystallized microstructure before cold rolling. The annealed plates were homogeneously cold rolled to different reductions from 10% through 90% on a two-high laboratory rolling mill with rolls 100 mm in diameter.

B. Microstructure Observation

Samples for microstructure observation were cut from plates in the normal direction (ND) and rolling direction (RD) cross-section, cold mounted and mechanically polished per standard metallographic processes. Prior to anodizing using Barker’s reagent (5 vol. pct HBF₄ in Methanol) the samples were electropolished using 1.5 vol. pct HNO₃-5.0 vol. pct HClO₄ acids in Methanol to remove the deformation layer. The microstructures were observed under polarized light using an Olympus inverted metallurgical microscope.

C. Texture Measurement

1. Macrotexture

Samples for macrotexture measurements were sectioned in the rolling plane (RD and the transverse direction (TD) cross-section) at the mid-thickness position of the plate. The surface for measurement was carefully polished to minimize surface stresses.

Texture measurements were carried out on a Rigaku D/MAX X-ray goniometer using Cu K_a radiation by means of the Schulz reflection method⁷. Three incomplete pole figures {111}, {200}, {220} up to a tilting angle of 75° (a_max = 75°) were measured. All incomplete pole figure data were corrected for defocusing error and background intensity. Three-dimensional orientation distribution functions (ODFs) f(g) were calculated by using the arbitrarily defined cell (ADC) method⁸. ODFs are expressed by using Bunge’s notation system⁹. Orientations, g, are described by three Euler angles j₁, F, j₂ which transform the crystallographic orientation into the sample coordinate system specified by RD, TD, and ND. The ODFs are represented in three-dimensional Euler space in the range of 0° ≤ j₁, F, j₂≤ 90° by way of iso-intensity contour lines in different sections with j₂ constant. Each texture component is fitted by using a number of Gauss-type scattering functions for quantitative analysis¹⁰. Therefore, the volume fraction M_i of each texture component i is calculated by determining the central orientations g_i, the orientation intensity f_i, and the scattering width y_i.

2. Microtexture

The electron backscatter diffraction (EBSD) technique is based on the discovery¹¹ and application¹² of Kikuchi patterns. Remarkable progress^13-15 in the use of this technique has been made in the past 20 years. The stationary beam is focused to a fine point on a highly tilted sample and the diffracted back scattering electrons form Kikuchi patterns on a phosphor screen. The orientation of the local lattice can be obtained by digitizing two appropriate zones in the Kikuchi pattern. In F.C.C. material, two zones on the band passing through <111>, <112>, <114> and <332> are usually selected for digitizing. Furthermore, the orientation information of an interested area can be collected by an electron beam movement controlled by a computer with selected step size.

Samples for microtexture measurements were also sectioned in the rolling plane at the mid-thickness position of the plate. The surface for measurement was mechanically polished followed by electropolishing using the same solution used for microstructure observation to remove the deformation layer.

Figure 1. Geometry of the sample setup for EBSD measurement.

In this study, EBSD work was carried out using a Hitachi 3200N SEM interfaced to a Unix workstation with orientation image microscopy (OIM) software from TexSEM Laboratories, Inc. (Draper, UT) installed. The sample was mounted on a pre-tilted sample holder with tilt angle of 72° for better pattern quality. The geometry of the sample setup is shown in Figure 1. Note that TD points out of the paper. The acceleration voltage and working distance were maintained at 20 kV and 20 mm, respectively. Image scan was adopted and each image contains at least 10,000 pixels (orientations). In order to ensure the reliability of the data, each data set was subject to a cleanup step using an algorithm that checks to determine if the orientation is different from its immediate neighbors. All nearest neighbor pixels must differ in orientation more than a tolerance angle that is set to 5°. Less than 5% of the data was changed in this cleanup operation.

D. Characterization of (Sub)Grain Boundaries

(Sub)Grain boundaries, especially coincidence site lattice (CSL) types, can be characterized and quantified from misorientation information obtained by running a EBSD scan. The misorientation is correlated because the calculation of misorientation is based on the neighbor grains. The (sub)grain boundary can be classified by different neighbor misorientations: low-angle boundaries (LABs, q < 5°), moderately misorientedboundaries (MMBs, 5° ≤ q < 15°) and high-angle boundaries (HABs, 15° ≤ q). Owing to the spatial resolution of the EBSD system, misorientations of less than 1.5° were not identified. CSL boundaries are defined following Brandon’s criterion¹⁶. All misorientation and CSL distributions are presented in manner of histogram plot form.

RESULTS

A. Microstructures of Hot Bands

The as-received hot bands of both CC (Figure 2a) and DC (Figure 2b) materials show a typical pancake structure that occurs at conditions of high hot rolling reduction. There exist some fine recrystallized grains in the DC hot band. Compared with CC hot band, DC hot band underwent significantly higher rolling reduction and hence it displays more flat grain boundaries along the RD. After annealing at 550 °C for 2 hours, CC and DC hot bands are completely recrystallized with equiaxed grain structures (Figures 2c and 2d). It appears that there are more grain clusters (neighbor grains with close orientations) in the DC hot band than in the CC hot band.

a		b		c		d
Figure 2. Microstructures of as-received AA 5052 (a) CC hot band, (b) DC hot band, (c) CC hot band annealed at 550 °C for 2 hours and (d) DC hot band annealed at 550 °C for 2 hours.

B. Textures of Hot Bands

1. Macrotextures

Figures 3a and 3b show the ODFs of the as-received hot bands of CC and DC materials, respectively. The intensities of typical orientations have been indexed. The texture of the CC hot band shows a typical rolling texture with a well developed b fiber starting from the Copper orientation {112}<111> through the S orientation {123}<634> and ending at the Brass orientation {011}<211>. The maximum orientation intensity along the b fiber is found at the Brass orientation with a value of 9.2. The texture of the DC hot band displays a combination of a rolling texture and a recrystallization texture (Figure 3b) with a Cube orientation {001}<100> which is also confirmed by microstructure observation (cf. Figure 2b). As can be seen from Figure 3b, the maximum orientation intensity on the b fiber is located close to the S orientation while the intensities of the Copper and Brass orientations are very close. The recrystallization texture in the CC hot band (Figure 4a) is characterized by a weaker Cube orientation accompanied by a Cube_ND fiber. However, the recrystallization texture of the DC hot band (Figure 4b) contains a strong Cube orientation with an intensity up to 48.0.

a		b
Figure 3. Complete ODFs of as-received AA 5052 (a) CC and (b) DC hot bands.
a		b
Figure 4. Complete ODFs of AA 5052 (a) CC and (b) DC hot bands after annealing at 550 °C for 2 hrs.

2. Microtextures

Figure 5. Inverse pole figure (IPF) maps of completely recrystallized AA 5052 (a-top) CC and (b-bottom) DC hot bands. White segments indicate the twin (S3) and high order twin boundaries (S9, S27a&b).

Figure 5 presents the inverse pole figure (IPF) maps of completely recrystallized hot bands with respect to ND. Each grain is painted with color based on its crystal orientation. The grain size of recrystallized CC hot band (Figure 5a) is larger than that of recrystallized DC hot band (Figure 5b). Cube clusters, neighbor grains with their <001> crystal directions parallel to the ND, are dominant in DC hot band. This indicates the existence of a large amount of Cube clusters in DC hot band. Cube clusters, however, are found to a less degree in CC hot band than in DC hot band. White segments in Figure 5 indicate the annealing twin boundaries (S3, S9, S27a & 27b) in both hot bands. Quantitative analyses show that the fractions of annealing twin boundaries are 9.8% and 7.1% for recrystallized CC and DC hot bands, respectively.

C. Macrotexture Evolution During Cold Rolling

ODFs of CC and DC materials with increasing cold rolling reductions are shown in Figures 6a, 6b, 6c, 7a, 7b, and 7c, respectively. Three sections of j₂ = 0°, j₂ = 45° and j₂ = 65° are selected to show the formation of the typical deformation orientations Brass, Copper and S along the b fiber. By watching the section of j₂ = 0°, the evolution of the a fiber can be tracked. Volume fractions of Gauss-type texture components are calculated and plotted as a function of the cold rolling reduction in Figures 8a and 8b. The orientation intensities vs. cold rolling reductions along the a fiber are displayed in Figurse 9a and 9b. The orientation intensities along the b fiber (skeleton line) and their positions in Euler space j₁, F as function of j₂are given in Figures 10a, 10b, 11a, and 11b for CC and DC materials, respectively.

1. CC Material

As shown in Figure 6a, the intensity of the Cube orientation decreases with increase of cold rolling reduction. With an increase in cold rolling reduction up to 50%, orientations continuously evolve along the Cube_RD to the Goss orientation {110}<001> and then flow along the a fiber to the Brass orientation. After 60% cold rolling, the Brass component is well developed and becomes sharper during further cold rolling. Figure 6b displays the formation process of the Copper component. The intensity of the Copper orientation increases as the intensity of the Cube orientation decreases. Orientations spread out from the Cube position to the Copper position during cold rolling. Furthermore, a fiber is formed with Cube and Copper orientations as two ends between cold rolling reductions of 30% and 70%. Finally, the intensity of the Copper orientation increases to 7.0 while the intensity of the Cube orientation goes to zero. Similarly, the S orientation develops at the expense of the Cube orientation during cold rolling (Figure 6c).

Volume fractions of typical orientations are plotted versus cold rolling reduction in Figure 8a. It can be seen that the volume fraction of the Cube orientation decreases from about 11% before cold rolling to zero after 90% cold rolling. The volume fraction of the Goss orientation increases from zero to about 5% after 10% cold rolling and then is stable with increasing cold rolling reductions. After 60% cold rolling, there is a jump in volume fraction of the deformation orientations Copper, Copper/S, S, S/Brass and Brass, especially Copper which increases by 10%. Correspondingly, the volume fraction of the random orientations decreases rapidly after 60% cold rolling.

a		a
b		b
c		c
Figure 6. Complete ODFs of completely recrystallized (RX) AA 5052 CC hot band under various cold rolling reductions at sections of (a) j2 = 0°, (b) j2 = 45° and (c) j2 = 65°.		Figure 7. Complete ODFs of completely recrystallized (RX) AA 5052 DC hot band under various cold rolling reductions at sections of (a) j2 = 0°, (b) j2 = 45° and (c) j2 = 65°.

The development of the a fiber is illustrated in Figure 9a. The a fiber becomes well formed beyond 60% cold rolling. The intensity along the a fiber is uniformly distributed before the cold rolling reduction reaches 60% beyond which the Brass orientation becomes stronger and sharper.

Figures 10a and 10b give the intensity distributions and positions of the b fiber (skeleton line) during cold rolling. At low cold rolling reductions (<50%), orientation intensities are homogeneously distributed along the b fiber. An increase of orientation intensity is observed along the b fiber after 60% cold rolling. However, the rate of increase of orientation intensity is not uniform along the b fiber. The intensities of the Copper/S and Brass orientations are higher than all other orientations along the b fiber. This trend becomes more pronounced beyond 80% cold rolling. The position of the b fiber (skeleton line) is revealed in Figure 10b. The Copper orientation is sharp at F = 30° at all cold rolling reductions. There is a shift of the Brass orientation from j₁ = 35° to j₁ = 30°. The j₁ value of the S orientation shifts from j₁ = 65°at low cold rolling reductions (<50%) to j₁ = 60° at high ones. The position of the b fiber (skeleton line) becomes sharp at high cold rolling reductions (>50%).

a		a		a
b		b		b
Figure 8. Dependence of the volume fraction of various texture components in AA 5052 (a) CC and (b) DC materials on the cold rolling reduction.		Figure 9. Orientation intensity of orientations of AA 5052 (a) CC and (b) DC materials along the a fiber under different cold rolling reductions.		Figure 10. Orientation intensity of orientations of AA 5052 CC material along the b fiber (skeleton line) (a) and their exact positions in Euler space j1, F as function of j2 (b).

2. DC Material

Figures 7a, 7b, and 7c show the ODFs of DC material during cold rolling. Starting from a strong Cube orientation, the orientation intensity around the Cube position becomes weak when the cold rolling reduction is increased (Figure 7a). The intensity of the Brass orientation reaches 2.4 after 60% cold rolling and continues to increase after further cold rolling. It is worth noting that the Cube orientation, accompanied by a Cube_RD fiber, is weak but still remains even after 90% cold rolling reduction. The evolutions of the Copper (Figure 7b) and S (Figure 7c) orientations in DC material during cold rolling are similar to those previously described for CC material. The intensities of the Copper and S orientations in DC material are however lower than that in the CC material.

Volume fraction of the Cube orientation reaches about 35% after annealing of the hot band (Figure 8b). At low cold rolling reductions (<50%), the volume fraction of the Cube drops quickly by about 22%. After 90% cold rolling, the volume fraction of the Cube drops to about 3%. Orientations start to flow to the b fiber at Copper, Copper/S, S, Brass/S and Brass positions after 20% cold rolling and then increase during further cold rolling. It is interesting to note that the volume fractions of the Copper/S, S, Brass/S and Brass orientations converge to about 12% while the volume fraction of Copper orientation, increases by about 10% after 70% cold rolling, and reaches about 18% after 90% cold rolling. There exists an increase of random orientations before 20% cold rolling. This may be explained by the randomizing of the strong Cube orientation during the early stages of cold rolling.

As shown in Figure 9b, the development of the a fiber in DC material follows the same pattern as in the CC material. However, the a fiber in the DC material is less intense.

a		b
Figure 11. Orientation intensity of orientations of AA 5052 DC material along the b fiber (skeleton line) (a) and their exact positions in Euler space j1, F as function of j2 (b).

Figures 11a and 11b present the evolution of the intensity and position (skeleton line) of the b fiber. Copper/S orientation becomes more intense than other typical deformation orientations along the b fiber (Figure 11a). Compared with the position of the b fiber in CC material, the position of the b fiber in DC material is sharper during the whole cold rolling process (Figure 11b).

D. (Sub)Grain Boundary Evolution During Cold Rolling

Misorientation distributions of recrystallized CC and DC hot bands are shown in Figures 12a and 12b, respectively. Two peaks are noteworthy for both CC and DC materials: the first is located at 40° to 45° and the second at 55° to 62.8°. The first peak can be ascribed to randomly orientated Cube grains¹⁷, which were also reported in AA 2004¹⁸ and AA 5083 aluminum alloys^19-21. Smaller grain boundary populations with misorientations between 40° and 45° in DC material can be explained by more Cube grain clusters (<001>//ND). The second peak corresponds to the S3 (annealing twin) boundaries. Compared with grain boundary distributions in CC material, the grain boundary distributions in DC material appear random. It can also be seen that the fractions of low-angle boundaries (LABs, q < 5°) and moderately misoriented boundaries (MMBs, 5° ≤ q < 15°) are smaller in CC material than that in DC material. This can be verified in Figures 13a and 13b that show the fraction of S1 boundaries (LABs/MMBs) is about 5% in CC material (Figure 13a) while it reaches about 12.5% in DC material (Figure 13b).

a		a
b		b
Figure 12. Grain boundary misorientation distributions of completely recrystallized AA 5052 (a) CC and (b) DC hot bands.		Figure 13. CSL Grain boundary distributions of completely recrystallized AA 5052 (a) CC and (b) DC hot bands.

Figures 14a, 14b, 14c, 14d, 14e, and 14f show coincident site lattice (CSL) grain boundary distributions in CC and DC materials after 10% (Figures 14a & 14b), 20% (Figures 14c & 14d) and 30% (Figures 14e & 14f) cold rolling reductions. In CC material, the fraction of S1 boundaries keeps increasing with increase in cold rolling reduction (Figures 14a, 14c & 14e) and even reaches about 45% after 30% cold rolling. However, there is no evidence of the development of either twin boundaries (S3) or high order twin boundaries (S9, S27a & 27b). In DC material, the fraction of S1 boundaries increases to about 37% after 20% cold rolling (Figure 14d) and then decreases dramatically to about 26% after 30% cold rolling (Figure 14f). Further quantitative analyses show that the fraction of S1 boundaries drops to 12% while the fraction of twin boundaries (S3, S9, S27a & 27b) is kept constant in DC material after 40% cold rolling.

a		b		c
d		e		f
Figure 14. CSL Grain boundary distributions after 10% cold rolling of AA 5052 (a) CC, (b) DC; 20% cold rolling of AA 5052 (c) CC, (d) DC; and 30% cold rolling of AA 5052 (e) CC, (f) DC materials.

Figures 15a & 15b display grain boundary maps in CC material after 30% and 40% cold rolling reductions, respectively. From 30% to 40% cold rolling, the LABs decrease by about 10% while the high-angle boundaries (HABs, 15° ≤ q) increase about 10%. Therefore, it is reasonable to conclude that the HABs develope at the expense of the LABs. The same changes of grain boundaries can also be observed in DC material in which the LABs drops from 29.6% after 20% cold rolling (Figure 16a) to 21.4% after 30% cold rolling (Figure 16b) while the HABs increase from 63.5% to 74%.

Figure 15. Grain boundary maps of AA 5052 CC material after (a-top) 30% cold rolling and (b-bottom) 40% cold rolling.		Figure 16. Grain boundary maps of AA 5052 DC material after (a-top) 20% cold rolling and (b-bottom) 30% cold rolling.

DISCUSSION

A. Texture Evolution During Cold Rolling

The evolution of rolling textures of various F.C.C. metals have been intensively investigated and discussed in the literature^22-26 by using ODFs. It is well accepted^22-26 that at low degrees of rolling, the orientations along the a and b fibers are homogeneously developed. This homogeneity, however, is destroyed with increase in the degree of rolling. With increase in the degree of rolling orientations flow along the a fiber to the Brass position and therefore promote a more intense Brass orientation. The a fiber disappears and this leaves a peak around the Brass orientation at very high degrees of rolling. Simultaneously, orientations flow into the b fiber as the rolling reduction increases. The development of orientations along the b fiber, is not uniform, the orientations mainly concentrate at the Copper, Brass and S positions. The evolution of textures of AA 5052 CC and DC materials during cold rolling follow the above rules. At low cold rolling reductions (<50%), the a and b fibers, though weak, are uniformly developed in both CC and DC materials (Figures 9a, 9b, 10a and 11a). At high cold rolling reductions (>50%), the intensities around Brass, Copper and S orientations increase rapidly and form peaks. Finally, the S orientation becomes the strongest orientation along the b fiber, which is also observed in Al²³, Copper²⁴, and Al-Cu alloys²⁷ at high rolling reductions.

Figure 17. Schematic map for transformation from the Cube orientation to the Copper orientation during cold rolling of AA 5052 aluminum alloy. {111} pole figures show the transformation steps (small red crosses indicate the positions of corresponding Euler angles in pole figures).

Zhou et al.^28,29 investigated the formation of rolling textures for F.C.C. polycrystals by using a rate-dependent crystal plasticity model. They show that during deformation, orientations move either directly into the b fiber or first into the a fiber, then along the a fiber to the b fiber and finally towards the corresponding stable orientations. For the latter path, orientations near the Cube orientation rotate towards the a fiber. This explains the formation of the Cube_RD and a fibers for CC and DC materials (Figure 6a and Figure 7a, respectively) at most cold rolling reductions. It is known that the Cube orientation transforms to the S orientation during rolling deformation³⁰, and is observed in both CC (Figure 6c) and DC (Figure 7c) materials during cold rolling. An interesting path that runs from the Cube orientation (001) [] to the Copper orientation (112) [] is observed at j₂ = 45° section in both CC (Figure 6b) and DC (Figure 7b) materials. This suggests that the Cube orientation rotates towards the Copper orientation during cold rolling. A schematic map of the rotation is drawn in Figure 17. It can be seen that starting from point a of the Cube orientation, the orientation passes through points b, c, d, e, and finally reaches point f of the Copper orientation. This transformation is finished after 90% cold rolling for CC material but not for the DC material. In general, the development of cold rolling textures in AA 5052 CC and DC materials follows the same path.

B. (Sub)Grain Boundary Evolution During Cold Rolling

The evolution of grain boundaries has been traced in both CC and DC materials during the early stages of cold rolling (≤ 40% thickness reduction). In CC material, S1 boundaries are well developed before 30% cold rolling and thereafter HABs develop at the expense of S1 boundaries. Grain boundary evolution in DC material follows the same path as in CC material except that HABs start to increase after 20% cold rolling. There is not a remarkable change in twin boundaries (S3, S9, S27a & 27b) during the early stages of cold rolling. Therefore, mechanical twinning is not an acting mechanism in these early stages. Instead, grains are subdivided by the process of forming cell walls (S1 boundaries). These cell walls transform to HABs during further cold rolling by misorientation increases. It can be seen from Figure 15 and Figure 16 that the misorientation gradient along the TD becomes larger than that along the RD with increase in cold rolling reduction, which indicates an elongated grain structure.

CONCLUSIONS

In this work, texture and grain boundary evolutions in industrially produced hot bands of CC and DC AA 5052 aluminum alloy during cold rolling have been studied. The following conclusions can be drawn.

1. After complete recrystallization, a stronger Cube orientation is observed in DC hot band than in CC hot band. a and b fibers become well developed beyond a 50% cold rolling reduction in both CC and DC materials. The highest intensity along the b fiber (skeleton line) is located near the S orientation in both materials. The Cube orientation is retained in DC material even after 90% cold rolling. Cold rolling texture evolutions for CC and DC materials follow the same path.
   
2. There exists a path by which the Copper orientation (112) [] develops at the expense of the Cube orientation (001) [] with increasing cold rolling reductions.
   
3. In both CC and DC materials, a cell structure develops with the indication of increasing LABs/MMBs during the early stages of cold rolling (≤ 40% thickness reduction). However, LABs/MMBs decrease while HABs increase after 40% and 30% cold rolling for CC and DC materials, respectively.
   
4. There is no evidence of the development of twin boundaries (S3, S9, S27a & 27b) in both CC and DC materials when cold rolling reductions are less than 40%.

ACKNOWLEDGEMENTS

Financial support from the U.S. Department of Energy (DOE) (under Contract No. DE-FC07-01ID14193) is gratefully acknowledged. The author would like to acknowledge Dr. James G. Morris for stimulating discussions and for his review of this manuscript.

References