Trans Indian Inst Met

DOI 10.1007/s12666-016-0842-7

TECHNICAL PAPER

Development of Cube Recrystallization Texture in Strip Cast

AA3004 Aluminium Alloy
Rohit Singh1 • Rajesh K. Khatirkar1 • R. N. Chouhan2 • Sanjay G. Sapate1

Received: 30 July 2015 / Accepted: 31 January 2016

 The Indian Institute of Metals - IIM 2016

Abstract AA3004 aluminium alloy is most widely used contrast to the DC casting route, where strong cube
in the beverage industry due to its excellent deep drawing recrystallization texture is obtained.
properties and corrosion resistance. During deep drawing
of the AA3004 sheets, the ‘earing phenomenon’ has to be Keywords AA3004  Twin roll casting  Cube texture 
minimized for economic advantage. In AA3004 and Annealing  Cold rolling  Strip casting
modified AA3004 alloys (like AA3104), this can be
achieved by tailoring the crystallographic texture of the
fully recrystallized sheet. Conventionally, AA3004 sheets 1 Introduction
are produced through direct chill (DC) casting route. In the
present study, a twin roll cast AA3004 sheet is used for AA3004 aluminium alloy constitutes a significant propor-
further processing (cold rolling and annealing). The texture tion of aluminium sheet products world-wide due to its
comprised mainly of Cu ({112} h111i), Brass ({110} large use in the beverage industry (as can stock) [1]. It has
h112i) and S ({123} h634i) orientations in the as-cast as very good deep drawability and corrosion resistance [2, 3].
well as cold rolled condition (Cu, Brass and S components The presence of magnesium (Mg) provides strength, while
being stronger in the cold rolled condition as compared to the distribution of second phase particles (Al6Mn and
as-cast material). After complete recrystallization of 90 % others) stabilizes the microstructure during the thermo-
cold rolled sheet, cube texture are deviated from ideal cube mechanical processing operations [2, 3]. Conventionally,
({100} h001i) (5–10) along with formation of Goss AA3004 alloy is processed through direct chill (DC)
({110} h001i) and P{011} h122i components. This was in casting route. The billets are hot rolled, cold rolled and
finally annealed to obtain the desired combination of
strength, microstructure and second phase particle distri-
& Rajesh K. Khatirkar bution for excellent deep drawing properties [2, 3]. In twin
[email protected]; [email protected] roll casting (TRC), there is very high solute supersaturation
Rohit Singh in processed alloys due to rapid solidification during the
[email protected] casting process [4]. Aluminium alloys produced through
R. N. Chouhan TRC route can be subsequently cold rolled and annealed,
[email protected] eliminating the need for hot rolling (as in DC casting route)
Sanjay G. Sapate and thus can result in cost savings if properly optimized.
[email protected] For the industrial production of beverage cans, deep
1
drawing is one of the most important property require-
Department of Metallurgical and Materials Engineering,
ments. During deep drawing of cold rolled and annealed
Visvesvaraya National Institute of Technology (VNIT),
South Ambazari Road, Nagpur 440010, India aluminium sheet, the tendency of the material to form ears
2 (i.e. the undulations in the circumference of the can body)
Jawaharlal Nehru Aluminium Research Development and
Design Centre (JNARDDC), Amravati Road, Wadi, Nagpur is major cause of problems. There is very stringent
440023, India requirement of the earing tendency in the can stock (it

123
 Trans Indian Inst Met

should not exceed *3 %) [5]. This earing tendency 250 to 450 C. In (b), temperature was kept constant at
depends on the crystallographic texture, size of the second 350 C and time was varied in the range of 20 min to
phase particles and their distribution in the can stock [2, 3]. 35 min. Microstructural characterization of all the samples
In cold rolled sheet, the position of ears is usually ±45 to was carried out by using scanning electron microscopy
the rolling direction (four positions) around the drawn cup. (SEM) and electron backscattered diffraction (EBSD). For
The fcc cold rolled texture components attributed with this SEM (secondary electron imaging, Jeol 6380A, Japan),
position are S {123} h634i, Brass {110} h112i and after metallographic polishing, the samples were etched
Cu{112} h111i (also known as b-fibre) [2, 3, 6]. In con- with Keller’s reagent [8]. Energy dispersive spectrometer
trast, fully recrystallized sheet shows ears at 0 and 90, (EDS, Bruker XFlash, Germany) attached to SEM was
which is the position of cube texture component [2, 3, 6]. used to obtain chemical composition of the second phase
The superimposition of ±45 and 0/90 ears can result in particles. Bulk texture was measured using X-ray diffrac-
lower anisotropy and balanced earing profile. The most tion in a PANalytical X’Pert PRO MRD system. Three
important parameter which needs to be controlled is the incomplete pole figures ({200}, {111} and {113}) were
cube, i.e. {100} h001i orientation, which develops during measured with Cu-Ka radiation in the back reflection
recrystallization of the cold rolled sheet [2]. It has been mode. The diffracted radiation was measured using a Xe
reported that with the increase in cold rolling reduction, the filled point detector. The Orientation distribution function
intensity of recrystallization texture decreases, whereas the (ODF) was obtained from the three incomplete pole fig-
intensity of the rolling texture component increases [2, 3, ures using the standard series expansion method and the
7]. This suggests that, cold rolling reduction as well as software MTM-FHM [9]. The standard u2 = 0, 45 and
recrystallization, both needs to be optimized for tailoring 65 constant ODF sections were used for the representation
the texture and consequently the earing behaviour. The aim of texture [2, 3]. For volume fraction measurements, X-ray
of the present investigation is to study microstructural ODFs were convoluted with suitable model function (with
developments during cold rolling and annealing of a TRC integrated ODF value of 1 and using 16.5 Gaussian
AA3004 alloy. spread). Micro-texture was measured by EDAX, USA
EBSD system fitted on a FEI Quanta 200HV SEM. The
EBSD data was acquired using a step-size of 0.1 and
2 Experimental 0.5 lm. The EBSD data for deformed sample was acquired
using a step size of 0.1 lm, while a step size of 0.5 lm was
TRC AA3004 aluminium alloy sheet with initial thickness used for partially and fully recrystallized samples. An
of 6.5 mm and chemical composition as shown in Table 1 accelerating voltage of 25 kV, sample tilt of 70 and a
was used for the present study. The AA3004 alloy was working distance of 15 mm were used. For EBSD mea-
received in the form of plate of size approximately surements, samples were electropolished at -20 C and
210 mm 9 190 mm. Strips of approximate size 11 V DC for 1 min using an electrolyte of methanol: per-
52 mm 9 190 mm were extracted from the plate by elec- chloric acid (80:20) after standard metallographic polish-
trical discharge machining (EDM) for cold rolling. Cold ing. The SEM and EBSD was carried out on long
rolling was done in a laboratory rolling mill in multiple transverse plane (plane containing RD and normal direc-
passes (with reduction being approximately constant for tion (ND)), while bulk texture measurements were done on
each rolling pass) in single direction [unidirectional, along the rolling plane [plane containing RD and transverse
RD (rolling direction)] up to 90 % thickness reduction. RD direction (TD)].
represents the initial direction along which the sheet was
pulled during TRC. After cold rolling, samples were
annealed in a salt bath furnace. Recrystallization was car- 3 Results
ried out in two ways—(a) by keeping time constant and
changing the temperature and (b) by keeping temperature Figure 1 shows the secondary electron scanning electron
constant and changing the time. In (a), time was kept micrograph of the as-received strip cast AA3004 alloy
constant at 20 min and the temperature was varied from (long transverse direction). The figure shows segregation at
the centre line in the microstructure. The morphology and
type (channel, deformation and banded segregates) are
Table 1 Nominal weight percent composition of AA3004 aluminum usually a result of combined solidification and rolling
alloy
processes. When the sheet is thick and is under very high
Alloy Mg Mn Fe Si Cu Al load, defect free structures are formed. If the casting
velocity increases, the resultant deformation of the semi-
AA 3004 1.1 1.3 0.4 0.15 0.1 Balance
solid material causes the liquid metal to be squeezed back

123
Trans Indian Inst Met

Fig. 1 Secondary electron

scanning electron microscopy
image of strip cast AA3004
aluminium alloy showing
channel segregation. Energy
dispersive spectroscopy (EDS)
of the second phase particles at
channel segregation region is
also shown

towards the sump. Channels are formed in situations where lamellar structure, while a thin lamellar structure forms
liquid metal flows between the dendrites [10, 11]. The after 90 % cold rolling. The color coded IPF map indicates
partial hot deformation during the TRC process is expected that a strong crystallographic texture develops after 90 %
to be the main reason for its occurrence. Other factors are cold rolling. This has been confirmed by X-ray pole fig-
cooling rate (which depends on water flow rate through the ures (PFs) and orientation distribution functions (ODFs).
rolls) and the rapid solidification of the trapped solute rich Figure 3 shows the {200}, {111} and {113} X-ray pole
liquid [10, 11]. These segregates are reported to have figures of the strip cast and 90 % cold rolled sample, while
eutectic like structure consisting mainly of Al, Mn and Fe Fig. 4 shows the /2 = 0, 45 and 65 constant ODF
cells formed due to the incoming of the solidification front sections for the strip cast and 90 % cold rolled samples.
from the edges towards the centre [10, 11]. In the present Both the pole figures and ODF shows the presence of
study, EDS spectrum (left side EDS spectrum in Fig. 1) at copper (Cu, {112} h111i), S ({123} h634i) and brass
the central portion reveals the presence of Al, Cu, Si, Mn, ({110} h112i) texture components in strip cast as well as
Mg and Fe (in very small amount), while, that away from 90 % cold rolled sample, the texture components being
the central region (right side in Fig. 1) shows predomi- stronger in cold rolled sample as compared to strip cast
nantly Cu and Si. The EBSD inverse pole figure (IPF) maps sample. The volume fractions of different orientations are
of strip cast and 90 % cold rolled AA3004 aluminium alloy shown in Fig. 5. After cold rolling, cube and Goss com-
samples are shown in Fig. 2. The grains are elongated ponents decreases, while Cu, S and brass components
along RD not only in the cold rolled sample, but also in the increases. The brass component increased *1.4 times, Cu
strip cast sample. The strip cast material has a coarse *1.6 times and S *1.6 times. The volume fraction of S

123
 Trans Indian Inst Met

Fig. 2 Electron backscattered

diffraction (EBSD) inverse pole
figure (IPF) map for a strip cast
and b 90 % cold rolled sample

Fig. 3 {200}, {111} and {113} X-ray pole figures for a strip cast and b 90 % cold rolled sample

component has been highest (*47 %), while that of Cu fibers become stronger in the 90 % cold rolled sample.
and brass has been *24 % each in the 90 % cold rolled Texture development has already been initiated in the strip
sample. The notable feature of the strip sample is the casting process, which become substantially stronger after
presence of moderately intense a and b fibers. Both the cold rolling. The a-fiber extends from Goss {110} h100i at

123
Trans Indian Inst Met

Fig. 4 u2 = 0, 45 and 65

constant section of orientation
distribution function (ODF) for
a strip cast and b 90 % cold
rolled sample. The standard
components are also marked on
the ODF

keeping temperature constant at 350 C are shown in

Fig. 7. At higher temperatures, 350–450 C, the grains
nucleate and grow to cause appreciable change in the
microstructure. Complete recrystallization has been
observed after annealing at 400 C and 450 C for 20 min.
This indicates that the temperature has dominant role as
compared to time. The deformation and recrystallization of
metals also lead to changes in local substructure which
manifests itself in the form of misorientation in EBSD
scans. Grain orientation spread (GOS, the average value of
the spread of orientation (i.e. misorientation) from the
average orientation for a particular grain), grain average
misorientation (GAM, average misorientation between
each neighboring pair of pixels for a particular grain) and
kernel average misorientation (KAM, average misorienta-
tion between all the neighboring points/pixels (not
exceeding a user defined value) for a particular grain) are
Fig. 5 Volume fractions of different texture components for strip the most widely used measures of deformation and
cast and 90 % cold rolled samples recrystallization in metals and alloys [13, 14]. The varia-
tion of average GOS, average GAM and average KAM for
/ = 45, u2 = 90, u1 = 0 to brass {110} h112i at / the strip cast, 90 % cold rolled, partially recrystallized and
= 45, u2 = 90, u1 = 35, while the b- fiber starts from fully recrystallized samples are shown in Fig. 8. The
brass {110} h112i at / = 45, u2 = 90, u1 = 35 and average values, that has been plotted in Fig. 8, are from the
goes through S{123} h634i at / = 37, u2 = 63, distribution of GOS, GAM and KAM for the entire EBSD
u1 = 59 to copper (Cu) {112} h111i at / = 35, map. The strip cast sample has already been in the strained
u2 = 45, u1 = 90 [2, 3, 12]. Hot rolling recrystallization condition due to the casting process (higher GOS, GAM
texture components are cube and Goss [2, 3]. The EBSD and KAM values). After 90 % cold rolling, the values of
IPF map of 90 % deformed samples annealed at different GAM and KAM are approximately constant, but the GOS
temperatures at constant time (20 min) are shown in Fig. 6, value increases considerably. During deformation, grain
while the samples annealed at different time intervals fragmentation, generation of dislocations (statistically

123
 Trans Indian Inst Met

Fig. 6 EBSD IPF maps of 90 % cold rolled AA3004 aluminum alloy samples annealed at a 250 C, b 300 C c 350 C, d 400 C and e 450 C
for 20 min each

stored and geometrically necessary) and grain rotation cast material does not show the presence of 0 and 90
takes place giving rise to misorientations. After complete ears after annealing. The 0 and 90 ears are formed due to
recrystallization, GOS/GAM/KAM, all decreases. Their the presence of cube recrystallization texture [2, 3]. In the
values have been intermediate for partially recrystallized DC cast AA3004 alloy, the hot rolled sheet has correct
samples. Figure 9 shows the u2 = 0, 45 and 65 constant amount of cube, which in cold rolling reduces and then
sections of the ODF for fully recrystallized samples. The after recrystallization of the cold rolled sheet again
annealing temperature and time are also shown in Fig. 9. increases. This gives a desired mix of cube (formed pri-
These ODFs show that the cube component increases after marily due to recrystallization) and rolling texture (b-
recrystallization and the rolling texture components (Cu, S fiber) necessary for reduced ears in the deep drawn cup.
and brass) are still present within it, but in lesser fractions The formation of cube texture in Al alloys has been a
as compared to 90 % cold rolled sample. The volume subject of both scientific and industrial importance. It is
fraction of different texture components is given in now more or less accepted that the cube oriented grains
Table 2. The cube component is stronger for annealing at nucleate during recrystallization after hot rolling at the
350 C for 35 min as compared to other two cases (350 C persistent cube bands [2, 3, 15]. The cube grains are sta-
for 30 min and 450 C for 20 min). For all the fully bilized by non-octahedral slip system ({110} h011i) at
recrystallized samples (Fig. 9), exact cube texture is not high temperatures [16]. At low temperatures, the cube
formed. grains are unstable and tend to break up and rotate away
from ideal cube. The cube bands usually occur adjacent to
the S bands [17]. In the present study, the ODF shows
4 Discussion deviated cube components (around 15–20 away from the
exact cube). Another reason behind the suppression of
The principal difference in aluminium alloys produced by cube component can be generation of other orientations
strip casting and DC casting route is, very high level of [{110} h100i (Goss), {011} h122i (P) and {013} h231i
solute super-saturation and small size of the second phase (Q)] [3] and particle stimulated nucleation (PSN). Large
particle in the former. It has been reported [4] that the strip particles (having size i2.5–3 lm) will promote PSN and

123
Trans Indian Inst Met

Fig. 7 EBSD IPF maps of

90 % cold rolled AA3004
aluminum alloy samples
annealed at 350 C for
a 21 min, b 22 min, c 24 min,
d 25 min and e 35 min

by decreasing the recrystallized grain size will generate a bar) values for the strip cast, 90 % deformed and fully
weaker cube. The presence of P-component and ND-ro- recrystallized AA3004 samples. It can be observed that the
tated cube oriented grains has been reported in 3XXX value of Dr is lowest for the fully recrystallized sample
aluminium alloy. The occurrence of these texture com- annealed at 350 C for 30 min and its r-bar value is also
ponents has been related to the non-deformable constitu- considerable. The cube component has been highest for
tive particles and particle stimulated nucleation (PSN) of this condition and it seems to be the optimum condition.
recrystallization [2, 3, 17] resulting in the reduction of the The present investigation demonstrates that PSN may play
cube in recrystallization texture. The occurrence of large a very important role in TRC AA3004 and needs to be
sized intermetallic particles is often seen in TRC alloys investigated further. The particle size of precipitates can
than in DC cast aluminium alloys. The presence of second only be controlled during the initial casting process which
phase particles are major sites for the PSN in TRC alloys is one of the serious drawbacks for TRC aluminium
giving rise to randomly oriented grains. Huang et al. [17] products. In contrast, in DC casting route, particles size of
have reported the formation of {011} h566i and normal intermetallic precipitates can be very accurately controlled
direction (ND)- rotated cube texture component {001} at various stages and hence the formability of the final
h310i during recrystallization. Table 3 shows the texture sheet. The difference in recrystallization behavior in both
estimated planar anisotropy (Dr) and normal anisotropy (r- the methods may be bridged by increasing the annealing

123
 Trans Indian Inst Met

temperatures for strip cast product. The presence of Mn in

the 3004 alloy matrix decreases the tendency of alloy to
anneal properly. Along with this, Fe present in the solid
solution reduces the growth of cube grains due to its
segregation in the cube oriented grain boundaries [18].

5 Conclusions

• The strip cast material consisted of weak b-fiber (brass,

copper and S), which increased significantly after 90 %
cold rolling. After complete recrystallization, the cube
texture was deviated from the ideal cube orientation
and an ND-rotated cube texture was observed. Goss and
P components were also present in the fully recrystal-
lized material.
Fig. 8 Variation of grain average misorientation (GAM), grain
orientation spread (GOS) and kernel average misorientation (KAM) • Control of intermetallic particles during strip casting
for as-received, cold rolled, partially recrystallized and fully recrys- process and annealing process was necessary to obtain
tallized samples. It needs to be noted that the values reported are area cube texture after recrystallization. This study demon-
average values

Fig. 9 u2 = 0, 45 and 65

constant section of ODF of fully
recrystallized samples

123
Trans Indian Inst Met

Table 2 Volume fraction (in %) of different texture components

Sr. no. Condition Cube Brass S Cu Goss Rotated cube

1. As-received 3.0 16.1 30.7 15.6 5.7 2.8

2. 90 % cold rolled 1.0 23.1 48.3 24.7 2.3 3.1
3. Annealed at 350 C for 30 min 5.8 5.5 12.5 6.1 1.3 7.3
4. Annealed at 350 C for 35 min 3.4 6.4 13.0 6.1 1.7 6.6
5. Annealed at 450 C for 20 min 4.0 7.4 15.0 6.8 1.4 6.4

Table 3 Texture estimated r-bar and Dr values for different Series- series, (ed) Cahn R W, Elsevier, Amsterdam (2007),
conditions ISBN 978-0-08-044497-0.
3. Humphreys F J, Hatherly M, Recrystallization and Related
Condition r-bar Dr Annealing Phenomena, 2nd edition, Pergamon-Elsevier (2004).
4. Es-Said O S, Zeihen A, Ruprich M, Ouattrocchi J, Thomas M,
As-received 1.15 -0.98 Shin K H, O’Brlen M, Johansen D, Tijoe W H, Ruhl D, J Mater
90 % cold rolled 2.46 -4.06 Eng Perform 3 (1994) 123.
Annealed at 350 C for 30 min 0.68 -0.57 5. Naess S E, Earing and Texture in Strip-Cast 3004 Type Alloys,
(ed) Morris J G, Merchant H D, Westerman E J, and Morris P L,
Annealed at 350 C for 35 min 0.88 -0.96
TMS (1993) p 274.
Annealed at 450 C for 20 min 0.85 -0.90 6. Engler O, Mater Sci Eng A 538 (2012) 69.
7. Cheng X-M, JMEPEG (2001) 399.
8. ASM Handbook, Metallogr Microstruct, 9 (2004).
9. Van Houtte P, in Manual the ‘‘MTM-FHM’’ Software System-
strated that strip cast material also had a potential to be Version 2, Department of Metallurgy and Materials Engineering,
used for applications requiring texture control. KU Leuven, Belgium (2009).
10. Lockyer S A, Yun M, Hunt J D, Edmonds D V, Mater Charact 37
(1996) 301.
Acknowledgments The authors are thankful to Director, VNIT 11. Yun M, Lockyer S, Hunt J D, Mater Sci Eng A280 (2000) 116.
Nagpur for providing the necessary facilities and constant encour- 12. Randle V, Engler O, Introduction to Texture Analysis: Macro-
agement to publish this paper. The authors also wish to acknowledge texture, Microtexture and Orientation Mapping, Taylor and
the use of ‘National Facility of Texture and OIM (a DST-IRPHA Francis, USA (2010).
facility)’ for EBSD and bulk texture measurements. 13. Wright Stuart I, Nowell Matthew M, Field David P, Microsc
Microanal, 17 (2011) 316.
14. TSL OIM Manual, version 5.3 USA, (1997–2007).
References 15. Daaland O, Nes E, Acta Mater 44 (1996) 1389.
16. Maurice C, Driver J, Acta Metall Mater, 41 (1993) 1653.
1. Aluminum alloy- Selection and applications The Aluminum 17. Huang K, Li Y J, Marthinsen K, Mater Charact 102 (2015) 92.
Association, Inc. (1998). 18. Merchant H D, Morris J G, Metall Trans A, 21A (1990) 2643.
2. Verlinden B, Driver J, Samajdar I, Doherty R D, Thermo-Me-
chanical Processing of Metallic Materials, Pergamon Materials

123