Materials Science & Engineering A 638 (2015) 2029

Contents lists available at ScienceDirect

Materials Science & Engineering A

journal homepage: www.elsevier.com/locate/msea

The role of microstructure and texture in controlling mechanical

properties of AZ31B magnesium alloy processed by I-ECAP
Michal Gzyl a,n, Andrzej Rosochowski b, Sonia Boczkal c, Lech Olejnik d
a

Advanced Forming Research Centre, University of Strathclyde, 85 Inchinnan Drive, Renfrew PA4 9LJ, United Kingdom
Design, Manufacture and Engineering Management, University of Strathclyde, James Weir Building, 75 Montrose Street, Glasgow G1 1XJ, United Kingdom
c
Institute of Non-Ferrous Metals in Gliwice, Light Metals Division, ul. Pilsudskiego 19, 32-050 Skawina, Poland
d
Institute of Manufacturing Processes, Warsaw University of Technology, ul. Narbutta 85, 02-524 Warsaw, Poland
b

art ic l e i nf o

a b s t r a c t

Article history:
Received 23 March 2015
Received in revised form
14 April 2015
Accepted 16 April 2015
Available online 27 April 2015

Mechanical properties of AZ31B magnesium alloy were modied in this work by various processing
routes of incremental equal channel angular pressing (I-ECAP) followed by heat treatment. Possible
strategies for improving ductility and strength of the alloy were investigated. Processing by routes A and
BC showed that texture plays predominant role in controlling mechanical properties at room temperature. Four passes of I-ECAP by route C followed by annealing enhanced ductility up to 0.35 of true strain.
It was found that tensile twinning was important in accommodating strain during tensile testing, which
resulted in a very good hardening behaviour. The yield strength was improved to 300 MPa by rening
grain size to 0.8 mm in I-ECAP at 150 C. The obtained structure and properties were shown to be stable
up to 150 C. True strain at fracture was increased to 0.2 after annealing at 150 C without lowering
strength.
& 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/).

Keywords:
Magnesium alloys
Equal channel angular pressing (ECAP)
Severe plastic deformation (SPD)
Ultra-ne grained (UFG) material
Electron backscatter diffraction (EBSD)

1. Introduction
Modern lightweight materials are dynamically entering new
markets. Especially the aerospace and automotive sectors are taking
advantage of fuel savings due to weight reduction [1,2]. Magnesium
alloys are of special interest because of their low density (1.7 g/cm3)
and a high specic strength, dened as a ratio of strength to density.
However, the yield stress of magnesium alloys is still lower than
commonly used aluminium alloys, which indicates that more research
is needed in this eld. Moreover, magnesium alloys suffer from limited
formability, especially at room temperature, and low corrosion resistance. However, the latter can be also an advantage as they have been
shown to be fully dissolvable within human body, which makes them
promising materials for biodegradable implants with a higher
strength than currently used polymers [3].
One of the possible approaches to improve strength of metals and
alloys is to reduce their grain size by severe plastic deformation
(SPD). Equal channel angular pressing (ECAP) [4,5] is among the
most developed SPD processes used worldwide. The concept of ECAP
is to press a billet through an angled channel without changing
billet's dimensions. Strain introduced by simple shear at the channel's intersection usually results in signicant grain renement.
n

Corresponding author.
E-mail address: [email protected] (M. Gzyl).

Therefore, it can be used to produce materials with a submicron

grain size, commonly known as ultrane-grained (UFG) metals.
However, conventional ECAP is not able to produce continuous billets as friction in the channel rises dramatically with the length of a
billet. Rosochowski and Olejnik proposed a solution to this problem
by separating stages of feeding and plastic deformation [6]. In their
cyclic process, called incremental ECAP (I-ECAP), the billet (bar [7],
plate [8] or sheet [9]) is delivered in consecutive steps to the
deformation zone and a punch, moving in a reciprocating manner,
deforms a small volume of material by simple shear in each cycle.
Conventional ECAP was shown to be effective in improving room
temperature ductility [10], strength [11] and corrosion resistance
[12,13] of magnesium alloys. However, the effect of grain renement
on mechanical properties is not clear as some researchers reported
decrease in strength [14,15] while it was possible to improve yield
stress in some other cases [16]. It was also emphasized that texture
produced during ECAP cannot be neglected due to hexagonal closepacked (hcp) crystallographic structure of the studied alloys [15,17].
Grain size produced by ECAP can be controlled by changing processing temperature while different textures are generated by different processing routes, known as A, BC and C. Route A indicates
processing without rotation of a billet between subsequent passes of
ECAP, route BC means that the billet is rotated by 90 always in the
same sense before putting back into the inlet channel and, nally, the
billet is rotated by 180 C in the case of route C. The mean grain sizes
reported in different studies for various ECAP temperatures was

http://dx.doi.org/10.1016/j.msea.2015.04.055
0921-5093/& 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

M. Gzyl et al. / Materials Science & Engineering A 638 (2015) 2029

0.37 mm at 115 C [11], 1 mm at 150 C [18], 2 mm at 200 C [11,19],

and 6 mm at 250 C [15]. Recent results obtained for I-ECAP of AZ31B
conrmed that grain renement at 250 C is similar to that obtained
by ECAP (56 mm) and texture generated by using different processing routes inuences strength and tensioncompression asymmetry of the yield stress [20]. Despite a lot of research conducted in
this eld, some questions still remain unanswered. The following
issues are addressed in this work:
1. What inuences mechanical properties more, grain size or texture?
2. How can room temperature ductility be improved by a heat
treatment after I-ECAP and what are deformation mechanisms
supporting enhancement of elongation?
3. How can strength of magnesium alloy be improved by I-ECAP
without lowering its ductility?

2. Material
The most common wrought magnesium alloy AZ31B (Mg3%Al1%
Zn0.5%Mn) was used in the current study. Billets for I-ECAP experiments were machined from a hot-extruded rod and a hot-rolled plate
along extrusion and rolling directions, respectively. The bar was supplied in as-fabricated conditions (AZ31B-F) while the plate was
delivered in annealed state (AZ31B-O) by Magnesium Elektron
(Manchester, UK), in accordance with ASTM-B107 and ASTM-B90,
respectively. EBSD analysis revealed that both materials exhibited
signicantly different grain sizes, bimodal structure 10 mm/50 mm for

21

the rod and 8 mm for the plate. It is apparent from Fig. 1 that microstructure of the rod is much more heterogeneous than microstructure
of the plate. The extruded rod is dominated by coarse grains as large as
70 mm surrounded by colonies of small grains with an average size of
10 mm (Fig. 1a). Microstructure of the plate is homogenous with only
little fraction of coarse grains  30 mm (Fig. 1b). Textures shown in
Fig. 1(c and d) are usual bre textures observed after extrusion and
rolling of magnesium alloys. It is apparent that c-axes of hexagonal
cells are aligned perpendicularly to the directions of extrusion as well
as rolling.

3. Experimental procedure
3.1. Details of I-ECAP
Double billet variant of I-ECAP [21] with a die angle of 90 was
carried out in this work on a 1 MN hydraulic servo press. A pair of
billets, with cross-sectional dimensions 10  10 mm2 each, was fed
using a motor driven screw jack whose action was synchronized
with the reciprocating movement of the punch. Process parameters
were the same as in the previous work on I-ECAP of AZ31B [20]:
feeding stroke 0.2 mm, punch movement frequency 0.5 Hz, peakto-peak amplitude of the punch movement 2 mm. The heating of
billets was realised by holding them for 15 min prior to processing in
the preheated die. The die temperature during processing was kept
constant within 72 C, based on the readings obtained from a
thermocouple located 15 mm from the deformation zone.

Fig. 1. Microstructures and textures of as-supplied extruded rod (a,c) and rolled plate (b,d). Symbol explanation: ED extrusion direction; ND normal direction of the
plate; RD in image (c) radial direction of the rod; RD in image (d) rolling direction of the plate.

22

M. Gzyl et al. / Materials Science & Engineering A 638 (2015) 2029

Table 1
Experimental plan of work conducted in this study.
Initial state Initial grain size, mm No. passes Route Temperaturea,C Comments
Part 1 Extruded
Part 2 Extruded
Part 3 Rolled
a

10/50 (bimodal)
10/50(bimodal)
8

4, 6
4
4

A, BC
C
A

250, 200
250
150

Initial four passes at 250 C

Additional annealing at 300 C
Initial two passes at 200 and 175 C, additional annealing at 150, 200, 250, and 300 C

Temperature in the nal pass.

3.2. Experimental plan

The experimental campaign was divided into three parts in order
to address the research questions stated in Section 1; details of
experiments are shown in Table 1. In the rst part, billets were
subjected to I-ECAP at 250 C and 200 C using two different routes,
A and BC. It was assumed that different temperatures will result in
different grain renement while the used routes will produce different textures. Four pairs of billets machined from the extruded rod
were rstly subjected to four passes at 250 C; then, two of them
were processed by additional two passes at 200 C. Conducting rst
pass at 200 C was not possible due to the occurrence of fracture.
The second part of the experimental plan was focused on
improving room temperature ductility. Our previous research has
conrmed that route C generates texture which results in
improved elongation compared to the initial extruded rod [20].
Additionally, route C was shown to produce more symmetrical
billets than route BC, which exhibited very similar mechanical
properties. Therefore, the billets were subjected to four passes of
I-ECAP by route C at 250 C. Thereafter, they were annealed for 6 h
at 300 C. Microstructures after I-ECAP, subsequent heat treatment
and tensile deformation were characterized using optical microscope and EBSD.
Strength improvement of AZ31B was attempted in the third
part of the experimental campaign. In order to achieve this goal,
gradual temperature reduction in consecutive I-ECAP passes was
used. Following our previous study [22], which revealed the effect
of initial grain size on formability during I-ECAP, a ne-grained
plate of AZ31B was selected as the initial material. Route A was
used in this part as it was shown to be less prone to fracture than
routes BC and C. The I-ECAP procedure was as follows: 1 pass at
200 C, 1 pass at 175 C, 2 passes at 150 C. Additionally, the
I-ECAPed billets were annealed for 1 h at 150, 200, 250, and 300 C
to investigate the effect of annealing temperature on mechanical
properties and grain size. In particular, ductility enhancement
without strength loss was sought. Annealing time was shortened
to 1 h as rst results showed that there was no difference in terms
of grain size and mechanical properties between 1 and 6 h.
3.3. Microstructural characterization and mechanical testing
Microstructural characterization was performed on Olympus
GX51 optical microscope and HRSEM Inspect F50. The as-received
specimens were characterized along extrusion and rolling directions. Preparation for optical microscopy included grinding using
SiC paper P1200, mechanical polishing using polycrystalline suspensions with particle sizes: 9, 3, and 1 mm, and nal polishing
with colloidal silica. Specimens were etched using acetic picral.
Samples for EBSD were prepared by ion milling on Leica RES 100.
Scans were performed with steps 0.3 m and 0.7 m and area
sizes 200  200 m2 and 400  400 m2, respectively. The larger
area was examined only in the case of coarse-grained rod to obtain
more reliable data for statistical analysis. The EBSD analysis was
performed on the EDND plane, as indicated in Fig. 2. Tensile ow
stress curves were obtained on Instron 5969 machine with the
maximum load capacity of 50 kN. Tests were carried out at room

Fig. 2. Schematic illustration of I-ECAP with the main direction indicated: ED

extrusion direction, ND normal direction, TD transverse direction.

temperature, with the initial strain rate 1  10 3 s 1 using 2 mm

thick at specimens with strain gauge dimensions 3 mm  12 mm.

4. Results
4.1. Microstructures and textures
Referring to the rst part of the experimental programme, the
average grain size of initially coarse-grained sample was rened to
4 mm by four I-ECAP passes at 250 C by route A (Fig. 3a). The
structure was relatively homogenous as standard deviation of the
measurement was 1.4 mm. Temperature reduction to 200 C in subsequent two passes led to further grain renement to 2.5 mm
(standard deviation 1 mm), as shown in Fig. 3b. As expected, smaller
grain size was obtained at lower temperature. It is apparent from
Fig. 3c that texture after four passes at 250 C is very similar to that
obtained after rolling of magnesium alloys (Fig. 1d) with c-axes tilted
at 90715 to ED in the I-ECAPed billet. Processing at 250 C/200 C
have not changed the texture signicantly, as shown in Fig. 3d. Two
overlapping peaks, indicating hcp cells tilted at 75 and 62 with
respect to ED, can be distinguished on (0001) pole gure with
maximum intensity of 11.4.
Processing by route BC led to slightly less effective grain
renement at 250 C than by route A as the grain size obtained
from EBSD was 5 mm (Fig. 4a). Moreover, structure was also more
heterogeneous since standard deviation was 3.1 mm. Texture
measurement of the sample revealed that hcp cells were tilted at
52 with respect to ED (Fig. 4c). Temperature reduction rened
grain size to 2.6 mm (standard deviation 1.2 mm) but texture has
not been changed signicantly, hcp cells are tilted at 57 and
maximum intensity increased from 16.1 to 19.5 (Fig. 4b and d).
Results obtained in the second part led to the conclusion that
microstructure after four passes of I-ECAP at 250 C via route C
was similar to that obtained using route BC as the mean grain size
was 4.5 mm. Standard deviation of 2.3 mm was also relatively high

M. Gzyl et al. / Materials Science & Engineering A 638 (2015) 2029

23

Fig. 3. EBSD maps and corresponding micro-textures of extruded samples after I-ECAP at 250 C (a) and (c) and at 250 C/200 C (b) and (d) using route A.

with a necklace-like structure displayed in Fig. 5a. Heat treatment

for 6 h at 300 C did not change the grain size signicantly as it
was increased to 6 mm; however, it is apparent from Fig. 6c that
structure became more homogenous as colonies of small grains
were no longer present. Microstructure characterization of tensile
samples pulled to true strain of 0.15 revealed many shear bands
parallel to shear plane of I-ECAP (Fig. 5b). It is not a new nding as
we have already reported similar bands in I-ECAPed samples
processed by routes A, BC and C [20]; it has been also shown in the
same article that fracture in tensile samples took place along those
shear planes. Deformation mechanism in the annealed sample was
different from that in the as-processed sample since severely
twinned structure was observed after pulling to true strain 0.15
(Fig. 5d).
In the rst part of the experimental campaign, the average
grain size obtained after I-ECAP at 250 C via route A was  4 mm
and further renement to 2.5 mm was achieved by lowering temperature to 200 C. In the third part of the programme, using a
rolled plate as the initial material, gradual reduction of temperature from 200 C in the rst pass to 175 C in the second pass and
to 150 C in the nal two passes produced an alloy with a submicron structure of 0.8 mm (Fig. 6a). Texture produced using this
procedure was similar to the sample processed for six passes via
route A at 250 C/200 C. However, two separate peaks in the
(0001) pole gure are clearly visible in this case (Fig. 6b), with one
group of grains perpendicular to ED and another one tilted at 57
to ED. It should be also noted that texture intensity was 6.2, which
was much lower than in the previous I-ECAPed samples.
Optical microscope images, presented in Fig. 7, showed that
grain size after nal I-ECAP at 150 C followed by heat treatment

was dependent on the annealing temperature. First samples were

kept in the furnace for 1 and 6 h; however no effect regarding
mechanical properties and microstructure was observed for different times. Therefore, the results shown here refer only to the
annealing time of 1 h. It is apparent from Fig. 7a that heat treatment at temperature equal to the last processing temperature
(150 C) did not introduce signicant changes to the structure, as
the measured grain size was  0.9 mm. Since a linear intercept
method was used here to measure the grain size, it could be a
systematic difference. When annealing temperature was increased
to 200 C and above, grain growth was observed (Fig. 7bd). The
average grain size measured was 2.7, 3.5 and 4.7 mm after
annealing at 200, 250, and 300 C, respectively.
4.2. Mechanical properties
The extruded rod, which was supplied in as-fabricated state
exhibited relatively high yield strength (220 MPa) and ultimate
tensile strength (295 MPa) accompanied by limited ductility at
room temperature as illustrated by tensile strain at fracture of 0.09
(Fig. 8a); true strain was used in this work as a measure of
deformation level. Four passes of I-ECAP at 250 C by routes A and
BC (part 1) reduced yield strength to 150 and 100 MPa, respectively, with simultaneous ductility enhancement to 0.14 and 0.22.
Lowering temperature of I-ECAP to 200 C in two subsequent
passes increased yield strength and fracture strain to 205 MPa and
0.2, when route A was used (Fig. 8a). Moreover, a good hardening
rate was obtained with ultimate tensile strength reaching 355 MPa
just before fracture. Completely different results were obtained for
route BC. The ow stress curve after additional two passes at

24

M. Gzyl et al. / Materials Science & Engineering A 638 (2015) 2029

Fig. 4. EBSD maps and corresponding micro-textures of extruded samples after I-ECAP at 250 C (a) and (c) and at 250 C/200 C (b) and (d) using route BC.

200 C looked almost exactly the same as after four passes at

250 C (Fig. 8b). Microstructural and textural changes taking places
during I-ECAP processing will be taken into account in Section 5 of
this article to explain the observed discrepancy between
mechanical responses of the billets subjected to different routes.
Mechanical properties after four passes at 250 C using route C
(part 2) were similar to route BC, with the yield strength of 60 MPa
and fracture strain of 0.22. The difference in ow behaviour of
billets subjected to various routes of I-ECAP has been already
discussed in our previous work [20]. In this study, an effect of heat
treatment of I-ECAPed billet is investigated. It is apparent from
Fig. 9 that signicant improvement of ductility was obtained after
annealing at 300 C. True strain at break rose from 0.22 to 0.35,
which is equivalent to 42% of engineering elongation, while yield
stress was increased to 80 MPa. It is also worth noting that the
sample was deformed uniformly almost to the fracture strain,
reaching 340 MPa of the ultimate strength.
Yield strength of the supplied rolled plate (165 MPa) was lower
than the extruded rod but fracture strain was 0.18, as shown in
Fig. 10a. Four passes of I-ECAP at temperatures lowering from 200 C
to 150 C (the third part of the experimental programme) signicantly increased yield strength to 295 MPa without a major loss
of ductility (true strain of 0.16). Although the hardening rate was
lowered due to I-ECAP, the yield stress was continuously increasing
during tensile testing and reached 360 MPa just before fracture. As
shown in Fig.10b, post I-ECAP annealing for one hour inuenced
signicantly strength and ductility at room temperature. The yield
strength obtained after I-ECAP at 150 C was slightly raised to
305 MPa along with fracture strain increased to 0.2 when annealing
temperature was the same as during processing. Keeping the

samples in the furnace heated to 200 C and above resulted in

decrease of yield strength to 265, 205, and 180 MPa after annealing
at 200, 250, and 300 C, respectively. However, reasonably good
elongations were obtained with true strains at fracture exceeding
0.2 and even reaching 0.3 after heat treatment at 250 C. Extraordinary mechanical properties were achieved after annealing at 150
and 200 C as both strength and ductility were simultaneously
improved compared to the commercially available rod and plate,
used in this work as reference materials.

5. Discussion
5.1. Effects of grain size and texture on mechanical properties
Grain renement in billets subjected to I-ECAP at 250 C is related
to temperature rather than processing route. The difference between
mean grain sizes obtained by routes A and BC was within 1 mm,
which does not explain signicantly different mechanical responses
of both samples. Therefore, higher yield strength of the billet processed by route A is attributed to the strong texture with c-axes of
hcp cells almost perpendicular to tensile direction while low yield
strength after route BC is a result of hcp cells alignment favourable
for slip on basal plane; this behaviour was already analysed for
routes A, BC and C in our previous article [20].
Further processing at 200 C resulted in a very similar grain
renement to 2.5 mm for both routes and the previous difference
in grain size was levelled. Texture after processing by route BC at
200 C was not changed signicantly as basal planes were still in
positions favourable for basal slip which was conrmed by almost

M. Gzyl et al. / Materials Science & Engineering A 638 (2015) 2029

25

Fig. 5. Optical microscope images of extruded samples after four passes of I-ECAP at 250 C by route C (a) and subsequent annealing for 6 h at 300 C (c), corresponding
microstructures after tensile testing to true train 0.15 are shown in (b) and (d), respectively.

the same ow stress curve as after four passes at 250 C. It clearly

shows that strong basal texture produced by route BC does not
allow taking advantage of grain reduction in terms of strengthening of the alloy. It has also revealed that HallPatch [23,24]
relation, expressing yield strength increase along with grain
renement, cannot be fullled in magnesium alloys when the
Schmid factor for slip on basal plane is signicantly larger than for
any other mode of deformation. Theoretical analysis carried out by
Kleiner and Uggowitzer [25] showed that Schmid factor for slip on
basal plane reaches its maximum when c-axes of hexagonal cells
are tilted at 57, which corresponds very well to results obtained
in this work by route BC.
In contrast to route BC, route A exhibited signicant yield
strength increase from 150 MPa after I-ECAP at 250 C to 205 MPa
after further processing at 200 C. Although grain size reduction
from 4 mm to 2.5 mm can be a reason for strength increase, it does
not explain ductility enhancement to 0.2 of fracture strain. It is
claimed here that distinctive texture produced after I-ECAP at
200 C by route A (Fig. 3d) resulted in better ductility. Two overlapping peaks can be distinguished in the corresponding (0001)
pole gure, with c-axes tilted at 75 and 62 to ED. Combination of
75 orientation of grains and their small size (2.5 mm) gave rise to
yield strength improvement to 205 MPa while another set of
grains, with basal planes tilted at 62 to ED, was deformed mainly
by basal slip which enabled ductility enhancement.
The conclusion drawn from this part of experimental work is
that I-ECAP by route A is more effective in improving strength
(without losing ductility) than commonly used route BC. This can
be explained by textural effects as it is shown that texture plays a
dominant role over grain size in controlling mechanical properties.
On the other hand, strongly textured AZ31B alloy, with c-axes of
hexagonal cells inclined at 57 to tensile direction, can be a

possible strategy for ductility improvement when strength is not

of a great importance.
5.2. Ductility enhancement by heat treatment
A large elongation, exceeding 40%, after ECAP via route BC followed by annealing was reported by Mukai et al. [10]. The
obtained ductility enhancement, accompanied by reduction of
yield strength, also observed in this work (increase of true strain at
fracture from 0.22 to 0.35), was attributed to grain growth to
15 mm during 24 h heat treatment at 300 C. This observation was
conrmed by Agnew et al. [26] who showed that post-deformation annealing did not change strong basal texture obtained after
ECAP, only grain size was increased to 19 mm. The authors of both
papers noticed that 45 tilting of c-axes with reference to ED
enabled slip on basal plane during tension which gave rise to
ductility enhancement. However, the papers did not explain how
grain growth during annealing improved formability compared to
as-ECAPed state.
Our previous research showed that route BC can lead to distortion of billet cross section [27], which has not been observed for
route C while mechanical properties after both processing paths
were very similar [20]. Therefore, in order to study the problem of
increased ductility, I-ECAP experiments were conducted at 250 C
using route C followed by annealing. Grain size after I-ECAP was
reduced to  4.5 mm and the following heat treatment for 6 h at
300 C increased it to only 6 mm. Based on these results, the effect
of grain growth during annealing can be considered negligible
from the view point of mechanical properties. One can also recall
that texture is not affected by annealing, according to [26].
Although microstructure looked very similar before and after heat
treatment, ductility was signicantly improved from 0.22 to 0.35

26

M. Gzyl et al. / Materials Science & Engineering A 638 (2015) 2029

Fig. 6. EBSD map and micro-texture of rolled sample after four I-ECAP passes by
route A with a nal temperature of 150 C.

Fig. 8. Tensile ow stress curves of extruded samples subjected to I-ECAP by routes

A (a) and BC (b) at 250 C and 250 C/200 C.

Fig. 7. Optical microscopy images of rolled samples after nal pass of I-ECAP at 150 C followed by annealing for 1 h at 150 C (a), 200 C (b), 250 C (c), and 300 C (d).

M. Gzyl et al. / Materials Science & Engineering A 638 (2015) 2029

Fig. 9. Tensile ow stress curves of extruded samples after four passes of I-ECAP at
250 C by route C and after subsequent annealing for 6 h at 300 C.

27

Post-I-ECAP annealing resulted in homogenization of grain size

distribution since necklace-like structure was no longer present
after six hours in the furnace (Fig. 5c). As a consequence, shear
banding, which led to fracture at true strain of 0.22, was avoided
and further deformation to true strain of 0.35 was enabled.
The dominant role of basal slip during tensile testing of the
annealed samples emphasized in the literature [26] should be discussed here as well. Despite I-ECAPed texture indicating importance
of this kind of deformation, it was shown that microstructure of the
annealed sample was severely twinned at true strain 0.15. It is especially interesting as twinning is usually expected to decrease ductility
[3133] while in this work it was revealed in the sample, which
exhibited extraordinary elongation along tensile direction. EBSD analysis of the deformed structure (Fig. 11) showed that almost all of the
twins are {1012} twins with 86 grain boundaries, which facilitate
extension along c-axis (so called tensile twins). This twinning variant
has been already reported in the literature as a deformation
mechanism giving rise to improvement of uniform elongation in
tensile samples tested along radial direction of the extruded rod of
AZ31 [34]. Thus, it seems that slip on the basal plane is the main
deformation mechanism at the initial stage of tension. However, further strain cannot be accommodated by the basal slip only and tensile
twining is activated, which leads to ductility enhancement. Additionally, good hardening rate of the annealed sample can be attributed
to the twinning occurrence since rotation of basal planes by 86
increases stress and twin boundaries can also act as barriers for dislocation movement. Finally, anisotropic mechanical properties of the
strongly textured I-ECAPed and annealed sample should be noted
here since Agnew et al. [26] showed that extraordinary elongation
along ED is not as spectacular during testing in other directions, which
should be taken into account during forming of the I-ECAPed billets.
5.3. Strength improvement

Fig. 10. Tensile ow stress curves of rolled samples after four passes of I-ECAP with
nal two conducted at 150 C (a) and after subsequent annealing for 1 h at 150,
200, 250 and 300 C (b).

of true strain, which is very similar to the values found in the

literature [10,26]. However, grain growth related explanation for
ductility enhancement proposed in those articles cannot be
applied here as the grain size was much smaller in our experiment.
An alternative explanation can be offered by analysing more
subtle features of the structure. A necklace-like structure (Fig. 5a)
is commonly observed when dynamic recrystallization (DRX)
takes place and this process is believed to control grain renement
of magnesium alloys in ECAP [11,16,28,29]. However, the heterogeneous microstructure can result in localization of strain along
colonies of ne grains (Fig. 5b), as it was reported in the past [30].

The results obtained in the rst part of experimental campaign

showed that route A at 250 C/200 C could be effective in strength
improvement of AZ31B magnesium alloy despite the yield
strength (205 MPa) being still lower than after direct extrusion
(220 MPa). Since our previous research revealed that smaller
initial grain size of a magnesium alloy allows processing at a lower
temperature [22], a procedure of gradual temperature decrease
was applied to the ne-grained billets obtained by hot rolling in
order to conduct I-ECAP below 200 C. I-ECAP at 150 C was successful in rening grain size to 0.8 mm as it was measured from
EBSD map. Generation of submicron structure resulted in signicant yield strength increase to 295 MPa, which was much
higher than in the cases of extruded rod and rolled plate.
The role of texture cannot be neglected as well since it was shown
in this work that texture produced by route A was more effective in
strength improvement than route BC. The pole gure obtained after
processing at 150 C was similar to that for 200 C but with more
visible peaks of maximum intensities. Although initial textures of the
extruded and rolled samples were different, large strain introduced
during I-ECAP via route A resulted in the very similar nal texture.
Interestingly, a very similar alignment of shear patterns on the EDND
plane after four passes of ECAP was predicted by a model developed
by Langdon [35]. He calculated that a maximum angle between shear
patterns introduced by subsequent ECAP passes reaches 37 after
fourth pass, which is in a very good agreement with the obtained pole
gure (Fig. 6b), where an angle between the two peaks of maximum
intensities is 3540. It supports the idea that the produced texture is
related directly to the processing route.
The same conclusion as for the billet subjected to I-ECAP at
250 C/200 C by route A can be drawn for 150 C, namely grains
oriented almost perpendicularly to ED attributed to strength increase
while 57 orientation helped with keeping reasonable ductility at

28

M. Gzyl et al. / Materials Science & Engineering A 638 (2015) 2029

strength at room temperature. One hour in the furnace at 300 C

was enough to reverse structural changes introduced by I-ECAP
and return to almost the same mechanical properties as in the assupplied state (after rolling). Interesting results were obtained
after annealing at 250 C where a signicant ductility enhancement was obtained (0.3 of true strain) with a reasonably high yield
stress of 205 MPa. However, we could not nd an explanation for
the lower ductility after heat treatment at 300 C compared to that
at 250 C; further research will be required to investigate this
behaviour. Finally, it was shown that only annealing at 150 C
allowed keeping moderate strength (even slightly increased) of
305 MPa with a relatively good true strain at fracture of 0.2. We
postulate that ductility enhancement from 0.16 after I-ECAP to
0.2 after subsequent annealing should be attributed to reduction
of internal stresses due to annihilation of dislocations. Additionally, low annealing temperature allowed keeping submicron
structure which resulted in sustaining of yield strength with a
decent hardening response. The results presented here are in good
agreement with the work by Strsk et al. [37] who showed that
structure of ultrane-grained AZ31 (  0.9 mm) was stable after
annealing at 170 C and grain growth was observed from 190 C.

6. Conclusions
Wrought magnesium alloy AZ31B was characterized after different versions of thermo-mechanical treatment, including incremental equal channel angular pressing (I-ECAP) and annealing.
Conclusions from the conducted study are listed below.

Fig. 11. EBSD (a) and twin grain boundaries (b) maps of the extruded and I-ECAPed
sample (4 passes at 250 C by route C) followed by annealing (6 h at 300 C) pulled
to true strain of 0.15. Tensile direction was horizontal.

room temperature due to activation of basal slip. It can be concluded

that combination of ultra-ne grain structure and proper texture is
required to increase strength to  300 MPa without losing ductility.
Moreover, much lower maximum intensity ( 6.2) of the (0001) pole
gure after processing at 150 C than for any other I-ECAPed sample
studied in this work is promising with respect to isotropy of
mechanical properties.
Heat treatment at 200 C and above of billets previously I-ECAPed
at 150 C resulted in grain growth and corresponding decrease of
strength at room temperature. The main mechanism for grain growth
during post-deformation annealing was static recrystallization. However, Barnett and Beer [36] claimed that microstructure restoration
can be also attributed to continued growth of nuclei formed by
dynamic recrystallization during plastic deformation. Despite the
annealing temperature in their study being higher (400 C), it could
have been also the case of the samples tested in this work.
Texture produced by I-ECAP was assumed to remain unaffected
by the following heat treatment, according to the experiments
conducted by Agnew et al. [26]. Thus, the grain size increase with
annealing temperature was shown to have a crucial effect on yield

1. Texture plays a dominant role in controlling mechanical properties after I-ECAP. It was shown that grain size reduction to
 2.5 mm has not changed mechanical properties of the billet
processed by route BC while strength improvement from 150 to
205 MPa was obtained after route A, despite the same nal
grain size as after route BC.
2. The room temperature ductility of AZ31B can be increased from
0.22 of true strain at fracture after I-ECAP at 250 C via route C to
0.35 (equivalent to 42% elongation) after subsequent annealing at
300 C. The achieved ductility enhancement was attributed to:
(1) the facilitation of basal slip due to texture developed during
I-ECAP and (2) the homogenization of grain size distribution after
heat treatment leading to suppression of shear banding. Tensile
twinning was shown to play an important role in accommodating
large strain during tensile testing of the annealed sample.
3. The yield strength of AZ31B alloy was increased to 305 MPa by
I-ECAP at 150 C using route A followed by annealing at the same
temperature. Additionally, decent hardening response was
obtained with ultimate tensile stress of 380 MPa and good ductility, expressed as 0.2 of true fracture strain.
4. I-ECAP was conrmed to be a promising tool for processing of
continuous magnesium alloy billets leading to a signicant
improvement of their mechanical properties.

Acknowledgments
Financial support from Carpenter Technology Corporation is
kindly acknowledged. Part of this research was supported by the
Engineering and Physical Sciences Research Council, United Kingdom [Grant no. EP/G03477X/1].

M. Gzyl et al. / Materials Science & Engineering A 638 (2015) 2029

References
[1] S. Schumann, H.E. Friedrich, in: H.E. Friedrich, B.L. Mordike (Eds.), Magnesium
Technology: Metallurgy, Design Data, Applications, Springer-Verlag, Heidelberg, 2006, pp. 499568.
[2] F.H. Froes, D. Eliezer, E. Aghion, in: H.E. Friedrich, B.L. Mordike (Eds.), Magnesium Technology: Metallurgy, Design Data, Applications, Springer-Verlag,
Heidelberg, 2006, pp. 603619.
[3] M.P. Staiger, A.M. Pietak, J. Huadmai, G. Dias, Biomaterials 27 (2006)
17281734.
[4] V.M. Segal, V.I. Reznikov, A.E. Drobyshevskiy, V.I. Kopylov, Russ. Metall. (Met.)
1 (1981) 99105.
[5] V.M. Segal, Mater. Sci. Eng. A 197 (1995) 157164.
[6] A. Rosochowski, L. Olejnik, in: E. Cueto, F. Chinesta (Eds.), Proceedings of the 10th
International Conference on Material Forming, Esaform 2007, 1820 April 2007,
Zaragoza, Spain, American Institute of Physics, 2007, vol. 907, pp. 653658.
[7] A. Rosochowski, L. Olejnik, Mater. Sci. Forum 674 (2011) 1928.
[8] A. Rosochowski, L. Olejnik, M. Richert, Mater. Sci. Forum 584586 (2008)
108113.
[9] A. Rosochowski, M. Rosochowska, L. Olejnik, B. Verlinden, Steel Res. Int. 81
(2010) 470473.
[10] T. Mukai, M. Yamanoi, H. Watanabe, K. Higashi, Scr. Mater. 45 (2001) 8994.
[11] S.X. Ding, C.P. Chang, P.W. Kao, Metall. Mater. Trans. A 40 (2009) 415425.
[12] J. Jiang, A. Ma, N. Saito, Z. Shen, D. Song, F. Lu, Y. Nishida, D. Yang, P. Lin, J. Rare
Earths 27 (2009) 848852.
[13] D. Orlov, K.D. Ralston, N. Birbilis, Y. Estrin, Acta Mater. 59 (2011) 61766186.
[14] W.J. Kim, C.W. An, Y.S. Kim, S.I. Hong, Scr. Mater. 47 (2002) 3944.
[15] S. Seipp, M.F.-X. Wagner, K. Hockauf, I. Schneider, L.W. Meyer, M. Hockauf, Int.
J. Plast. 35 (2012) 155166.
[16] M. Janecek, M. Popov, M.G. Krieger, R.J. Hellmig, Y. Estrin, Mater. Sci. Eng. A
462 (2007) 116120.

29

[17] S.R. Agnew, P. Mehrotra, T.M. Lillo, G.M. Stoica, P.K. Liaw, Mater. Sci. Eng. A 408
(2005) 7278.
[18] K. Bryla, J. Dutkiewicz, P. Malczewski, Arch. Mater. Sci. Eng. 40 (2009) 1722.
[19] L. Jin, D. Lin, D. Mao, X. Zeng, B. Chen, W. Ding, Mater. Sci. Eng. A 423 (2006)
247252.
[20] M. Gzyl, A. Rosochowski, R. Pesci, L. Olejnik, E. Yakushina, P. Wood, Metall.
Mater. Trans. A 45 (2014) 16091620.
[21] A. Rosochowski, L. Olejnik, M. Richert, Mater. Sci. Forum 584586 (2008)
139144.
[22] M. Gzyl, A. Rosochowski, L. Olejnik, A. Reshetov, Key Eng. Mater. 611612
(2014) 573580.
[23] E.O. Hall, Proc. Phys. Soc. Lond. Sect. B 64 (1951) 747753.
[24] N.J. Petch, J. Iron Steel Inst. 174 (1953) 2528.
[25] S. Kleiner, P.J. Uggowitzer, Mater. Sci. Eng. A 379 (2004) 258263.
[26] S.R. Agnew, J.A. Horton, T.M. Lillo, D.W. Brown, Scr. Mater. 50 (2004) 377381.
[27] M. Gzyl, A. Rosochowski, E. Yakushina, P. Wood, L. Olejnik, Key Eng. Mater.
554557 (2013) 876884.
[28] R.B. Figueiredo, T.G. Langdon, J. Mater. Sci. 44 (2009) 47584762.
[29] M. Gzyl, A. Rosochowski, A. Milenin, L. Olejnik, Comput. Methods Mater. Sci.
13 (2013) 357363.
[30] S.E. Ion, F.J. Humphreys, S.H. White, Acta Metall. 30 (1982) 19091919.
[31] B.C. Wonsiewicz, W.A. Backofen, Trans. Metall. Soc. AIME 239 (1967)
14221431.
[32] T. Al-Samman, G. Gottstein, Mater. Sci. Eng. A 488 (2008) 406414.
[33] M. Gzyl, R. Pesci, A. Rosochowski, S. Boczkal, L. Olejnik, J. Mater. Sci. 50 (2015)
25322543.
[34] M.R. Barnett, Mater. Sci. Eng. A 464 (2007) 17.
[35] T.G. Langdon, Mater. Sci. Eng. A 462 (2007) 311.
[36] A.G. Beer, M.R. Barnett, Scr. Mater. 61 (2009) 10971100.
[37] J. Strsk, M. Janeek, J. ek, J. Strsk, B. Hadzima, Mater. Charact. 94 (2014)
6979.