Failure Mechanisms and Texture Evolution of Wrought AZ31B ...
Transcript of Failure Mechanisms and Texture Evolution of Wrought AZ31B ...
Brigham Young University Brigham Young University
BYU ScholarsArchive BYU ScholarsArchive
Theses and Dissertations
2012-03-14
Failure Mechanisms and Texture Evolution of Wrought AZ31B Failure Mechanisms and Texture Evolution of Wrought AZ31B
Magnesium at Temperatures Ranging from 25 C to 125 C Magnesium at Temperatures Ranging from 25 C to 125 C
Jonathan Michael Scott Brigham Young University - Provo
Follow this and additional works at: https://scholarsarchive.byu.edu/etd
Part of the Mechanical Engineering Commons
BYU ScholarsArchive Citation BYU ScholarsArchive Citation Scott, Jonathan Michael, "Failure Mechanisms and Texture Evolution of Wrought AZ31B Magnesium at Temperatures Ranging from 25 C to 125 C" (2012). Theses and Dissertations. 3135. https://scholarsarchive.byu.edu/etd/3135
This Thesis is brought to you for free and open access by BYU ScholarsArchive. It has been accepted for inclusion in Theses and Dissertations by an authorized administrator of BYU ScholarsArchive. For more information, please contact [email protected], [email protected].
Failure Mechanisms and Texture Evolution of Wrought AZ31B
Magnesium at Temperatures Ranging from 25 C to 125 C
Jonathan Scott
A thesis submitted to the faculty of Brigham Young University
in partial fulfillment of the requirements for the degree of
Master of Science
David T. Fullwood, Chair Mike P. Miles Eric R. Homer
Department of Mechanical Engineering
Brigham Young University
April 2012
Copyright © 2012 Jonathan Scott
All Rights Reserved
ABSTRACT
Failure Mechanisms and Texture Evolution of Wrought AZ31B
Magnesium at Temperatures Ranging from 25 C to 125 C
Jonathan Scott Department of Mechanical Engineering, BYU
Master of Science
Failure mechanisms were studied in wrought AZ31B magnesium alloy under different strain paths and various temperatures. Optical micrographs were used to observe the formation of shear bands and regions of high twin density in samples strained under uniaxial, biaxial and plane strain conditions. Interrupted testing at 4% effective strain increments until failure was used to observe the evolution of the microstructure. Results showed that shear bands with a high percentage of twinned grains appeared early in the samples strained under biaxial or plane strain tension. These bands are similar to the failure region in uniaxial tension specimens. A forming limit diagram for AZ31B was developed from the strain data, showing that plane strain and biaxial tension had very similar limit strains, in contrast to materials like steel or aluminum alloys which typically have greater ductility in biaxial tension compared to plane strain tension.
When the experiments are repeated at elevated forming temperatures of 75 C and 125 C there is no observable shear band formation. The forming limit diagrams for each temperature were created and showed a marked increase in biaxial tension formability, compared to plane strain tension. Optical microscopy showed no sign of any compression twins and very few tensile twins in samples strained in biaxial or plane strain tension, up to 12% effective strain. The lack of compression twins at these strain levels shows that the effect of temperature on critical resolved shear stress for <c+a> slip is greater than previously expected.
Keywords: Jonathan Scott, optical microscopy, CRSS, shear band, compression twin
ACKNOWLEDGEMENTS
I would like to thank Dr. Raj Mishra for all the support throughout this research. His
input was vital in understanding the microstructure evolution. Thank you to General Motors as
well, who supplied all the materials as well as provided me with an internship opportunity.
Thanks to Ali Khosravani who helped me with most of my research. I would have been lost on
the SEM without him. Thanks, last of all, to my wife for all her support while her husband was
spending nights with microscopes and polishing tables.
This material is based upon work supported by the National Science Foundation under
Grant No. 0928923. Any opinions, findings, and conclusions or recommendations expressed in
this material are those of the author and do not necessarily reflect the views of the National
Science Foundation.
v
TABLE OF CONTENTS
LIST OF FIGURES ................................................................................................................................ vi
1 Development of Shear Bands with High Twin Density in Wrought Magnesium AZ31B Sheet at Room Temperature.......................................................................................................... 1
1.1 Introduction ........................................................................................................................... 1
1.2 Experimental Procedure ........................................................................................................ 3
1.3 Results .................................................................................................................................. 6
1.4 Discussion of Results .......................................................................................................... 15
1.5 Conclusions ......................................................................................................................... 17
2 Formability and Failure Mechanisms of Wrought AZ31B Magnesium at Slightly Elevated Temperature .................................................................................................................. 19
2.1 Introduction ........................................................................................................................ 19
2.2 Experimental Procedure ...................................................................................................... 22
2.3 Results ................................................................................................................................. 24
2.4 Discussion of Results .......................................................................................................... 32
2.5 Conclusion ........................................................................................................................... 35
REFERENCES ....................................................................................................................................... 37
vi
LIST OF FIGURES
Figure 1-1: AZ31B specimen pulled to 19% uniaxial tensile strain along the RD. A high density of compression twins is seen in the area around the crack tip. ........................6
Figure 1-2: SEM image showing compression twins in the necked failure region of a uniaxial strained specimen. ...........................................................................................7
Figure 1-3: SEM image showing compression twins 2mm from the necked failure region of a uniaxial strained specimen. ...........................................................................................8
Figure 1-4: IPF + IQ image of a compression twin with accompanying grain boundary misorientation graph. ....................................................................................................9
Figure 1-5: Shear band in a 7% strain uniaxial specimen composed of a single line of tension twins. .............................................................................................................................9
Figure 1-6: Tension twin boundaries, shown in red, from a shear band in a uniaxial tension specimen strained to 7%. ............................................................................................10
Figure 1-7: Shear band forming on TD plane, in specimen stretched to 4% effective strain along the RD in plane strain. Dotted white line added for easy identification. .........11
Figure 1-8: Multiple shear bands forming on TD plane in specimen stretched to 8% effective strain along the RD in plane strain. Dotted white lines are added parallel to shear bands for ease of identification. ..................................................................................12
Figure 1-9: Image of the ND plane on a sample pulled in biaxial tension to 8% strain, showing three shear bands, with a high density of compression twins in each band. ............................................................................................................................13
Figure 1-10: Image taken from the same area as Figure 6 at greater magnification, showing compression twins within a shear band. .....................................................................14
Figure 1-11: Forming limit diagram for 2 mm thick alloy AZ31B at room temperature. .............15
vii
Figure 2-1: Strain history for a point undergoing plane strain .......................................................25
Figure 2-2: Strain history for a point undergoing biaxial strain. ...................................................26
Figure 2-3: Microstructure near the center of a biaxial sample strained to 12% at 75 C. .............27
Figure 2-4: Microstructure near the edge of a biaxial sample strained to 12% at 75 C. ................28
Figure 2-5: Microstructure near the edge of a plane strain sample strained to 5%at 75 C. The twins shown appear to be the stunted beginnings of a shear band that never fully developed. ...................................................................................................................29
Figure 2-6: Forming Limit Diagram for AZ31B at 75 C ...............................................................30
Figure 2-7: Microstructure at the edge of a plane strain sample strained to 5% at 125 C. ............31
Figure 2-8: Forming Limit Diagram for AZ31B at 125 C. ............................................................32
Figure 2-9: Forming limit diagram comparison of AZ31B at 25, 75 and 125 C. ..........................34
1
1 DEVELOPMENT OF SHEAR BANDS WITH HIGH TWIN DENSITY IN WROUGHT MAGNESIUM AZ31B SHEET AT ROOM TEMPERATURE
1.1 Introduction
Over the last 10 years magnesium consumption has increased 20% [1-2]. This is due to
the fact that magnesium has many mechanical properties that are desirable in structural parts
such as high specific strength, high fracture toughness, good electromagnetic shielding, good
heat transfer characteristics, recyclability and vibration and noise damping, among others [3].
However, magnesium has poor formability at room temperature and this has limited its use for
applications where complex shapes need to be formed.
The poor formability of magnesium at room temperature has been attributed to the lack of
slip systems in the HCP crystal structure [4–7]. Additionally, in a rolled AZ31B sheet there is a
strong basal texture present with no easy slip planes to accommodate strain of the c-axis at room
temperature [6, 8, 9]. The typical approach to overcoming the low formability of AZ31B is to
form the material at elevated temperatures, usually 250 °C or higher [10, 11]. Forming at these
high temperatures is expensive and difficult to implement in a high-volume production
environment. It is therefore important to study the formability of wrought magnesium at room
temperature in order to understand the failure initiation mechanisms that precede fracture. If an
understanding of failure mechanisms can be gained, then it is possible that the starting
microstructure or the forming path can be manipulated in order to increase formability.
2
Previous studies on formability and failure mechanisms have been performed on samples
that undergo either compression or uniaxial tensile strains, where measurements were taken in a
highly deformed or necked region. Although compression is vital to the sheet rolling process
and uniaxial tension is a good t ool in understanding material behavior, neither strain path is
crucial in the part forming process. Complex strain paths, on the other hand, are vital in the
sheet forming and stamping processes and ultimately control the geometry of any parts made
from AZ31B magnesium. It has been shown that the majority of part failures during forming or
stamping occur in plane strain regions [12]. Since prior work on c rack nucleation and failure
initiation has depended primarily on uniaxial tension specimens with strains well above 10%, the
present study focused on failure mechanisms for biaxial and plane strain tension, with
comparisons to uniaxial tension.
The primary mechanism for both through-thickness deformation and failure in AZ31B in
samples strained by either compression or uniaxial tension, is thought to be {101�1} compression
twinning [4, 13]. The strong basal texture in AZ31B means that any in-plane tensile deformation
of wrought sheet causes compression of the c-axis, which is accommodated by compression
twinning. At high strains, compression twins also act as a major contributing factor in fracture
initiation [14–16]. T here have been observations linking crack nucleation with {101�1}
compression twins. Cracks are seen near and parallel to compression twins in extruded AZ31B
and ZK60 bar as well as rolled AZ31B when pulled in tension [15, 17–19].
One study measured a steady decrease in the number of {101�2} tension twins as strain, or
cold work reduction, increased [20]. This reduction of twins was attributed to the fact that the
measurements excluded shear-banded regions of the microstructure. The same study showed a
steady increase in shear banded area with increasing strain. B ecause of limitations of EBSD
3
microscopy, the study could not identify microstructures inside the shear bands and therefore the
relationship between twin fraction and shear band area had to be implied rather explicitly
measured.
A study by Kang et al. [21] showed that for tensile strains above 10% in rolled AZ31B
formations of compression twins that quickly clustered into shear bands were observed. There
was a strong correlation between the formation of compression twin shear bands and the onset of
diffuse necking.
To investigate the relationship between the strain path and the failure initiation
mechanisms, twin formations during uniaxial, biaxial and plane strain conditions in annealed
AZ31B were studied. Samples were strained at increments of 4% until failure. Microstructure
evolution was evaluated using optical microscopy. A forming limit diagram (FLD) for AZ31B
at room temperature was constructed to give a complete picture of formability under the three
strain paths.
1.2 Experimental Procedure
The material investigated was composed of 2.05 mm thick sheets of cold rolled AZ31B
magnesium with 3% Al and 1% Zn by weight alloying elements. The average linear intercept
grain size in 10µm. All samples were annealed at 350 °C for a minimum of 15 minutes and then
air-cooled outside the oven. Specimens were rough cut using a water-jet before annealing.
Samples for microscopy were cut from the strained specimens using wire electrical discharge
machining (EDM).
Tension specimens were prepared from the annealed sheets using the ASTM E8 standard
[22], with one exception, and pulled at a constant crosshead speed of 16 mm/min. The other
4
tensile specimen was prepared with a variable width gauge section similar to samples prepared
by Kank et al. [21]. The sample was 30mm long and tapered so that the center of the sample was
3mm wide, increasing to a width of 5mm at the ends of the gauge length.
Biaxial and plane strain specimens were strained using the limiting dome height (LDH)
tooling, with a standard 100 mm diameter punch and punch speed of 38 mm/min. Plane strain
was achieved by using the LDH tooling with a 100 mm wide specimen, which was determined
by experimenting with specimens of various widths until plane strain tension was achieved in the
area between the top of the dome and the clamped region of the sheet. Clamping dies were
modified for magnesium sheet by using three small lock beads instead of one large lock bead, as
it typical for steel or aluminum. The die changes were made in order to avoid fractures in the
clamped area and are described in greater detail in [19].
Strain was measured by coating the surface of each sample with a white paint. B lack
paint was then applied in a speckle pattern for use with a digital image correlation (DIC) system
for strain measurement. A Trilion DIC [23] system was used to map the strain during the
stretching experiments. For tensile specimens a facet size of 15 pixels was used. For biaxial and
plane strain specimens the smallest facet size that would not produce gaps in the data was used
and ranged from 29 to 21 pixels. Variation was caused by the uneven size distribution within the
black speckled pattern.
Biaxial and plane strain samples were lubricated by placing a thin layer of boron nitride
paste on the surface of the sample in contact with the punch. Biaxial strain measurements were
taken from the surface of the dome, in the pole area. Plane strain measurements were taken from
the area of the sheet that was determined, experimentally, to be in plane strain, using surface
5
strain measurements, and was located approximately midway between the dome peak and the
clamp.
Optical micrographs were taken in order to observe a much larger area than can be seen
using EBSD. Samples were chemically etched with a solution of 100 ml ethanol, 10 ml water, 5
ml acetic acid and 6 g picric acid.
For convenience RD, TD and ND refer to rolling direction, transverse direction and
normal direction respectively. A plane identified as the RD plane is one whose normal is along
the rolling direction.
Effective strain (𝜖𝑒𝑓𝑓) was calculated using the following equation:
𝜖𝑒𝑓𝑓 = �23
(𝜀12 + 𝜀22 + 𝜀32) (1-1)
where 𝜀1 is the measured strain in the major axis, 𝜀2 is the measured or calculated strain in the
minor axis and 𝜀3 is the measured or calculated strain through the thickness of the sheet. The
following results were used for each strain case:
𝜀𝑢𝑛𝑖𝑎𝑥𝑖𝑎𝑙 ≈ 𝜀1 (1-2)
𝜀𝑏𝑖𝑎𝑥𝑖𝑎𝑙 = 2 ∗ 𝜀1 (1-3)
𝜀𝑝𝑙𝑎𝑛𝑒 = 2√3∗ 𝜀1 (1-4)
6
1.3 Results
Uniaxial tension specimens behaved as previously reported [4] and demonstrated limited
{101�1} compression twinning until the onset of necking, which happens suddenly before failure.
After failure the region in the necked area of the specimen was examined. There was a high
density of compression twins in the area near the crack tip, as shown in figure 1-1. Individual
shear bands were difficult to see because of the large number of twins present.
Figure 1-1: AZ31B specimen pulled to 19% uniaxial tensile strain along the RD. A high density of compression twins is seen in the area around the crack tip.
Using a Scanning Electron Microscope (SEM), a higher magnification image was taken
from a uniaxial sample at the failure site as shown in figure 1-2. The areas with compression
twins are circled in red. When a similar image is taken 2mm farther away from the failure site a
sharp decrease in compression twin activity can be seen as shown in figure 1-3. Again,
RD
TD
7
compression twins are highlighted with red circles. Images taken at 4mm or farther from the
fracture site show little to no compression twins.
Figure 1-2: SEM image showing compression twins in the necked failure region of a uniaxial strained specimen.
These features were inspected with Transmission Electron Microscopy (TEM). A sample
was cut from a region of the material that contained features that are consistent with compression
twins, that is thin straight twins with multiple twins in a single grain. TEBSD analysis with a
step size of 10nm verified that the features were indeed compression twins with a misorientation
angle of 56.8° as shown in figure 1-4.
8
At lower strains of less than 10% shear bands were observed with optical microscopy.
Figure 1-5 shows a shear band that formed in a variable width uniaxial specimen, similar to the
ones used by Kang [21]. The area shown is in the 7% strain region. EBSD analysis showed that
the twins in the shear band were tension twins, as seen in figure 1-6.
Figure 1-3: SEM image showing compression twins 2mm from the necked failure region of a uniaxial strained specimen.
9
Figure 1-4: IPF + IQ image of a compression twin with accompanying grain boundary misorientation graph.
Figure 1-5: Shear band in a 7% strain uniaxial specimen composed of a single line of tension twins.
10
When specimens were stretched in plane strain tension, shear bands formed at much
lower strain levels than in the uniaxial tension case. Figure 1-7 shows a shear band containing
compression twins in a specimen stretched to 4% effective strain. At 8% effective strain, which
is essentially the limit strain for this material in plane strain tension, numerous shear bands have
formed, as can be seen in figure 1-8.
For specimens strained in biaxial tension similar results were seen. At 4% effective
strain shear bands began to form and then became increasingly easy to identify at higher strains.
Figure 1-9 shows three parallel shear bands on the ND surface from a sample stretched to 8%
effective strain. F igure 1-10 is a magnified image of one of the shear bands showing that the
band is indeed populated by a large number of compression twins.
Based on uniaxial tension, plane strain, and biaxial tension testing experiments a forming
limit diagram (FLD) was constructed for AZ31B, as shown in figure 1-11.
Figure 1-6: Tension twin boundaries, shown in red, from a shear band in a uniaxial tension specimen strained to 7%.
11
Figure 1-7: Shear band forming on TD plane, in specimen stretched to 4% effective strain along the RD in plane strain. Dotted white line added for easy identification.
12
Figure 1-8: Multiple shear bands forming on TD plane in specimen stretched to 8% effective strain along the RD in plane strain. Dotted white lines are added parallel to shear bands for ease of identification.
13
Figure 1-9: Image of the ND plane on a sample pulled in biaxial tension to 8% strain, showing three shear bands, with a high density of compression twins in each band.
14
Figure 1-10: Image taken from the same area as Figure 6 at greater magnification, showing compression twins within a shear band.
15
Figure 1-11: Forming limit diagram for 2 mm thick alloy AZ31B at room temperature.
1.4 Discussion of Results
A study conducted by Barnett [20] established an inverse relationship between the
number of twins and the area of shear bands in compressed AZ31B; however, EBSD methods
did not allow for observing the microstructure of the shear bands. T he optical micrographs
shown in figures 1-9 and 1-10 show that shear bands in biaxial and plane strain tension
specimens are populated by compression twins. The study conducted by Kang [21] showed that
the onset of shear bands with a high density of compression twins in uniaxial samples
0
5
10
15
20
25
30
-20 -15 -10 -5 0 5 10
% m
ajor
Str
ain
% minor Strain
Forming Limit Diagram
Parrallel to RD
Perpendicular to RD
Bi-axial
16
immediately precedes diffuse necking and failure. The work by Barnett and Kang and Figures
1-1 through 1-3 show that compression twins appear in large quantities only in the necked region
of uniaxially strained samples where the local strain is very high. The shear bands observed in
biaxial and plane strain specimens are similar to those seen in uniaxial specimens near a failure,
but occur at much lower effective strains.
Kang [21] also observed that specimen thinning occurred very little before failure,
because of the difficulty of c-axis compression in AZ31B. The formation of compression twin
shear bands directly correlates to the reduction of thickness and onset of necking. D uring
uniform straining of a uniaxial tension specimen the strain in the minor direction allows for
strain to occur without an immediate need for c-axis compression. However, in biaxial and plane
strain specimens thinning cannot be delayed because of the imposed strains in the plane of the
sheet. Therefore, the reduction in sheet thickness is accommodated by e arly formation of
compression twin shear bands, compared to uniaxial tension. It is reasonable to conclude that the
relatively low limit strains observed for biaxial and plane strain tension experiments were caused
by early development of shear bands composed of compression twins, compared to the uniaxial
tension case.
The room temperature FLD for AZ31B is not typical for ductile metals because the limit
strain for biaxial tension is about the same as it is for plane strain tension. Typically for
aluminum or steel alloys the biaxial limit strain is significantly greater than it is for plane strain
[24-25]. Similar results to those reported here for AZ31B were reported in another study by
Chino [26], except that the limit strains in both biaxial and plane strain tension were 2% lower
than those measured in the present study. This discrepancy is possibly caused by a combination
of small punch radius and large grid pattern used by Chino.
17
1.5 Conclusions
The microstructure of magnesium alloy AZ31B was examined by optical microscopy for
specimens strained in uniaxial, biaxial and plane strain tension, and the following conclusions
were reached:
(1) Shear bands with a high concentration of compression twins begin to appear at about
4% effective strain in biaxial and plane strain tension specimens.
(2) These shear bands are similar to the shear bands that occur in tensile specimens near a
fracture. Their early appearance contributes to lower limit strains in the biaxial and
plane strain tension specimens.
(3) The measured FLD for alloy AZ31B shows an inverse relationship between shear
band formation and formability. Shear band formation is similar in the plane strain
and biaxial tension cases, and both of these exhibit similar levels of formability. This
behavior is in contrast with other ductile metals, which typically show greater levels
of formability in biaxial tension compared to plane strain tension.
19
2 FORMABILITY AND FAILURE MECHANISMS OF WROUGHT AZ31B MAGNESIUM AT SLIGHTLY ELEVATED TEMPERATURE
2.1 Introduction
Magnesium alloys are gaining popularity in the automotive manufacturing world as a
structural material and as an alternative to steel and aluminum. Over the last 10 years this
increased interest has translated into 20% increase in magnesium consumption [1-2]. Magnesium
is favored due to the fact that it has many mechanical properties that are desirable in structural
parts such as high specific strength, high fracture toughness, good e lectromagnetic shielding,
heat shedding characteristics, recyclability, and vibration and noise damping, among others [3].
However, magnesium’s greatest drawback is its poor formability at room temperature and this
has limited its use for applications where complex shapes need to be formed. The present study
focuses on expanding the usefulness of AZ31B by investigating the relationship between failure
mechanisms and forming temperature for biaxial and plane strain tension, with comparisons to
uniaxial tension.
The poor formability of magnesium at room temperature has been attributed to the HCP
crystal structure and lack of active slip systems [4–7]. A ccording to Taylor, a crystalline
material needs at least 5 independent active slip systems in order to deform into an arbitrary
shape. Metals with a BCC or FCC crystal structure have 6 or 8 independent slip planes due to
their symmetry. Steel, aluminum and copper are examples of these. AZ31B has 5 independent
20
slip systems. If it didn’t it would be as brittle as ceramics. However, the critical resolve shear
stress (CRSS) of these slip systems differs greatly and not all slip systems are active all the time.
AZ31B has a strong basal texture aligning the c-axis of the HCP crystals with the sheet
normal. This texture can be as high as 20x random [26]. Slip in the basal plane is the easiest slip
system to activate with a CRSS of 1 MPa at room temperature in a single crystal [27]. Models
predict that the CRSS for polycrystalline AZ31B is about 10 MPa. The basal plane contains two
independent slip systems, both of which accommodate slip only in the <a> direction. Prismatic
slip allows for a third independent slip system with a component of slip on the basal plane but it
also has a CRSS that is 55Mpa or higher at room temperature. Any thinning of the sheet requires
slip with <c+a> components. The prismatic and pyramidal slip systems each have a slip plane
with <c+a> components, but the CRSS is reported to be as high as 100 Mpa [28].
Measurements of the CRSS of compression twinning have been inconclusive [15, 29-31].
It has been shown that <c+a> slip and compression twinning are competitive modes of c-axis
compression, but the estimates for compression twin CRSS vary greatly, ranging from 25 MPa to
well over 100 MPa [15, 29–31]. According to Ono [32], the primary mechanism for through-
thickness deformation is <c+a> slip which can account for 40% of the total dislocations at room
temperature. H owever, many researchers have observed significant compression twinning
whenever c-axis compression is present [4, 13].
At high strains, compression twins also act as a major contributing factor in fracture
initiation[14–16]. T here have been observations linking crack nucleation with compression
twins. Cracks are seen near and parallel to compression twins in extruded AZ31B and ZK60 bar
as well as rolled AZ31B when pulled in tension [15, 17, 18, 33].
21
Effectively identifying compression twins is rather difficult due to the nature of the twin.
Compression twins are much thinner than tension twins, often reaching only 1µm in width.
Compression twins also do not grow in width as tension twins do when strain is increased.
Multiple parallel twins can also appear in a single grain. Because the surface of the twin is so
small, EBSD data from these areas is often interpreted as noise. With a fine enough scan many
researchers have been able to identify compression twins and verify that the optical features
coincide with compression twin boundaries [15, 20]. O ther researchers have simply
characterized compression twins based on their optical properties [34].
The usual approach to overcoming the low formability of AZ31B is to form the material
at an elevated temperature, typically 250 C or higher [10–12]. When the temperature increases
there is a g reater availability of energy and slip planes activate at lower strains, effectively
increasing the formability of the material. Forming at these high temperatures is expensive and
difficult to implement in a high-volume production environment. It i s therefore important to
study the formability of wrought magnesium near room temperature in order to understand the
failure initiation mechanisms that precede fracture. More importantly, since the majority of part
failures during forming or stamping occur in plane-strain tension [35], focus will be given to
complex strain paths like biaxial and plane strain tension. If an understanding of failure
mechanisms can be gained, then it is possible that the starting microstructure or the forming path
can be manipulated in order to increase formability with limited increase in forming temperature.
To understand the relationship between forming temperature and failure initiation
mechanisms, samples will be deformed at 2% surface strain increments until failure for both 75C
and 125 C . O ptical microscopy will be used to observe large areas of the microstructure.
Forming Limit Diagrams (FLD’s) will also be constructed from the data.
22
2.2 Experimental Procedure
2.05 mm thick sheets of cold rolled magnesium with 3% Al and 1% Zn by w eight
alloying elements (AZ31B) were annealed at 350 C for a minimum of 15 minutes and then air-
cooled outside the oven. Specimens were rough cut using a water-jet before annealing.
Tension specimens were then prepared from the annealed sheets using the ASTM E8 standard
[22, 24]. A gage length of 4 inches was chosen so the heater attached to the Instron would heat
the entire gage length evenly. Tension samples were pulled at a constant crosshead speed of 16
mm/min.
Biaxial and plane strain specimens were strained using the limiting dome height (LDH)
tooling, with a standard 100 mm diameter punch and punch speed of 38 mm/min. Plane strain
was achieved by using the LDH tooling with a 88 mm wide specimen, which was determined by
experimenting with specimens of various widths until plane strain tension was achieved in the
area between the top of the dome and the clamped region of the sheet. Clamping dies were
modified for magnesium sheet by using three small lock beads instead of one large lock bead, as
it typical for steel or aluminum. The die changes were made in order to avoid fractures in the
clamped area and are described in greater detail in the work by Miles [19].
Strain was measured by coating the surface of each sample with a white paint. B lack
paint was then applied in a speckle pattern for use with a digital image correlation (DIC) system
for strain measurement. A Trilion DIC [25] system was used to map the strain during the
stretching experiments. For tensile samples a facet size of 15 pixels was used. For biaxial and
plane strain samples the smallest facet size that would not produce gaps in the data was used and
23
ranged from 29 t o 21 p ixels. Variation was caused by the uneven size distribution within the
black speckled pattern.
Biaxial and plane strain samples were lubricated by placing a thin layer of boron nitride
paste on the surface of the sample in contact with the punch. Biaxial strain measurements were
taken from the surface of the dome, in the pole area. Plane strain measurements were taken from
the area of the sheet that was determined experimentally to be in plane strain, approximately
midway between the dome peak and the clamp.
Tension samples were heated through radiation by placing heating coils on both sides of
the tension sample while it is clamped in the Instron. The coils were maintained at a constant
percentage of their heating capacity until a steady state temperature was achieved. At 75 C there
was a variation of 1 C along the gage length. At 125 C there was a variation of 4 C along the
gage length.
Bi-axial and plane strain samples were heated by conduction with belt heaters attached to
the forming die. Two thermocouples were welded to a sample to measure the temperature, one
at the center of the sample and the other approximately two inches away near the clamp. Again
the heaters were maintained at a specific power level until a steady state temperature was
achieved. At 75 C there was a variation of 1 C within the forming area and at 125 C there was a
variation of 5 C.
Samples for microscopy were cut from the strained specimens using wire electrical
discharge machining (EDM). Cut samples were mounted in a two part epoxy. EBSD samples
had nickel nano-strands added to the epoxy for additional conductivity.
Samples were polished using a series of fine grit sandpapers. Once cleaned, the samples
were exposed only to distilled water for all subsequent polishing steps. O il based
24
monocrystaline diamond suspensions of decreasing size ending with 1 µm diamonds were used
next followed by OP-S colloidal silica.
Optical micrographs were taken in order to observe a much larger area than can be seen
using EBSD. After polishing, samples were chemically etched with a solution of 100 ml ethanol,
10 ml water, 5 ml acetic acid and 6 g picric acid.
For convenience RD, TD and ND refer to rolling direction, transverse direction and
normal direction respectively. A plane identified as the RD plane is one whose normal is along
the rolling direction.
2.3 Results
DIC was used to verify that the stain conditions stated were indeed achieved. Figure 2-1
shows the strain history at a point assumed to be in plane strain. The specimen was deformed at
75 C to failure. As you can see there was very little strain in the minor direction throughout the
deformation. Likewise, when the strain history is seen for a biaxially strained specimen at 75 C,
as shown in figure 2-2, you c an see that the point of interest demonstrated biaxial strain
throughout the experiment.
When a biaxial sample strained to 12% effective strain at 75 C is viewed under an optical
microscope we see a distinct absence of shear band formations as shown in figures 2-3 and 2-4.
In room temperature samples strained to 4% shear bands formed near the edge and propagated
inward with twins occupying almost the entire edge. With 3 times more strain we would expect
25
Figure 2-1: Strain history for a point undergoing plane strain.
to see at least as much twinning and shear band formation. However, all twin formation appears
to be very limited. Tension twins formed throughout the material but at a much lower frequency
and never in groups or shear bands.
26
In plane strain samples strained to 5% at 75 C we see similar results as shown in figures
2-3 and 2-4. In figure 2-5 we see a much stronger concentration of twins and they almost appear
to be forming a shear band. However, there is no continuity between the twins except proximity.
Outside the image there are no more twins. This group of twins appears to be forming the
earliest stages of a shear band that never fully developed. Images elsewhere in the sample also
show no indication of shear band formation.
Figure 2-2: Strain history for a point undergoing biaxial strain
27
Figure 2-3: Microstructure near the center of a biaxial sample strained to 12% at 75 C
An FLD for AZ31B at 75 C is constructed from the forming data and is presented in
figure 2-6. T he FLD departs from the unusual FLD shape as t he biaxial strain increases
significantly. The limit strain in plane strain tension also increases, but to a much lower degree.
As can be expected, samples strained at 125 C showed no shear band formation. Figure
2-7 shows the greatest concentration of twins that could be found on any sample strained at 125
C. The FLD shown in figure 2-8 shows that formability has continued to increase in a manner
similar to the 75 C results.
29
Figure 2-5: Microstructure near the edge of a plane strain sample strained to 5%at 75 C. The twins shown appear to be the stunted beginnings of a shear band that never fully developed
30
Figure 2-6: Forming Limit Diagram for AZ31B at 75 C
0
5
10
15
20
25
30
35
40
-30.00 -25.00 -20.00 -15.00 -10.00 -5.00 0.00 5.00 10.00 15.00
Maj
or S
trai
n
Minor Strain
FLD for 75 C
32
Figure 2-8: Forming Limit Diagram for AZ31B at 125 C
2.4 Discussion of Results
The dominant failure mechanism in AZ31B for samples strained in tension or
compression at room temperature is thought to be void nucleation near compression twins [4,
13]. When the material starts to thin, which is late in the evolution of the microstructure, the
compression of the c-axis causes compression twin nucleation and eventually void nucleation.
When forming tests are conducted at 75 C we see st artling differences. T here is no
evidence of shear band formation or compression twin nucleation. Tension twins form
throughout the material but at a relatively low density and never in groups or bands. This
apparent lack of compression twins prior to failure may be justified by Agnew [36] when he
proposed that the ratio between the CRSS for <a> and <c+a> slip in AZ31B could be as small as
0
5
10
15
20
25
30
35
40
45
-40 -30 -20 -10 0 10 20
Maj
or S
trai
n
Minor Strain
FLD for 125 C
33
2-2.5. This countered his own earlier results which asserted that the value of the CRSS for non-
basal slip was significantly higher.
In a study conducted by Chapuis and Driver [31], a single crystal of AZ31B was
preferentially aligned for <c+a> slip and then strained. T he crystal twinned instead, forming
compression twins until the sample was heated to over 200 C. T he activation energy of the
compression twin was over 100 MPa. Further study by the same researchers estimated the CRSS
for <c+a> slip at 20 MPa. In addition, they showed a strong inverse correlation to the forming
temperature and the CRSS of compression twinning and <c+a> slip [30], both of which
decreased at roughly the same rate.
There is no disputing the fact that there is some <c+a> slip at room temperature [32], but
compression twinning was seen to be the primary means of c-axis compression when thinning
was imposed during biaxial and plane strain forming experiments. However, at even modestly
elevated temperatures of 75C and 125C, compression twinning ceases to be the dominant
mechanism of c-axis compression and instead it is reasonable to assume (based on a l ack of
compression twinning and shear band development) that <c+a> slip becomes dominant, thus
increasing the forming limits of AZ31B significantly, compared to that seen at room
temperature. The CRSS for <c+a> slip reaches a t hreshold, below which the local stresses at
each grain is relieved by <c+a> slip and never builds high enough for compression twinning.
The change in biaxial formability when forming temperature is increased by 50 C is
surprising. Figure 2-9 shows the results of all three forming limit diagrams for comparison.
During biaxial forming, the strain of the sheet thickness must accommodate for stretching of the
plate in both directions. This increase in formability is likely due to the decrease in local stress
in biaxial samples
34
The basal texture of AZ31B causes the c-axis to be parallel to the sheet normal, but it has
no effect on the orientation of the grain around that axis. Grains have a random rotation around
the sheet normal. To an individual grain, this makes little difference. Slip directions can rotate
only 30 de grees from the global strain before they start becoming redundant. O n a
polycrystalline scale, however, this means that the second independent <a> slip component of
each grain is plus or minus 30 from parallel to the global strain, increasing the CRSS for that slip
system. With every grain stressed in the same direction, polycrystalline deformation become
more difficult. Local strains in a biaxial sample relieve stress in every direction. B ut local
strains in plane strain samples can only relieve stress transverse to the major strain axis. Because
the global strain in the minor direction for a plane strain sample has to be zero, any strain along
Figure 2-9: Forming limit diagram comparison of AZ31B at 25, 75 and 125 C
35
the major strain axis increases the local stress on the neighboring grains. This stress increases
until a grain deforms in the transverse direction, relieving the built up stress. This stress
concentration is a likely site for void nucleation, which would cause premature fracture.
Disrupting the texture of AZ31B has little effect. In fact, Delvalle [37] determined that
the optimal forming criterion for AZ31B is basal texture with 10 µm grain size, which was the
microstructure of every sample used in this study. If a high forming temperature is undesirable,
then the most effective way to increase the formability of wrought AZ31B magnesium is to
increase the forming temperature by 50 C and design a forming process that avoids plane strain
while emphasizing biaxial and uniaxial tensile strains.
2.5 Conclusion
Forming AZ31B at elevated temperatures has led to the following conclusions:
1) Compression twins were not visible in samples strained in biaxial and plane
strain.
2) The temperature dependence of the CRSS for <c+a> slip and compression
twinning is roughly the same. However, the decrease in CRSS for <c+a> slip
reaches a threshold where compression twinning is no l onger needed for c-
axis compression, even near the limit strains of the material.
3) With compression twinning no longer present the failure initiation mechanism
is no longer tied to void nucleation along compression twins. A t higher
strains the local stress states in plane strain tension specimens cannot be
relieved as easily as in biaxial tension specimens, resulting in relatively
greater forming limits for biaxial tension.
37
REFERENCES
[1] S. Das, “Primary magnesium production costs for automotive applications,” Jom, vol. 60, no. 11, pp. 63-69, Nov. 2008.
[2] A. a. Luo, “Magnesium: Current and potential automotive applications,” Jom, vol. 54, no. 2, pp. 42-48, Feb. 2002.
[3] G. Ambrogio, C. Bruni, S. Bruschi, L. Filice, a. Ghiotti, and M. Simoncini, “Characterisation of AZ31B magnesium alloy formability in warm forming conditions,” International Journal of Material Forming, vol. 1, no. 1, pp. 205-208, Apr. 2008.
[4] M. Barnett, “Twinning and the ductility of magnesium alloysPart I: ‘Tension’ twins,” Materials Science and Engineering: A, vol. 464, no. 1-2, pp. 1-7, Aug. 2007.
[5] S. R. Agnew, J. a. Horton, and M. H. Yoo, “Transmission electron microscopy investigation of ⟨c+a⟩ dislocations in Mg and α-solid solution Mg-Li alloys,” Metallurgical and Materials Transactions A, vol. 33, no. 3, pp. 851-858, Mar. 2002.
[6] L. L. Chang, E. F. Shang, Y. N. Wang, X. Zhao, and M. Qi, “Texture and microstructure evolution in cold rolled AZ31 magnesium alloy,” Materials Characterization, vol. 60, no. 6, pp. 487-491, Jun. 2009.
[7] J. Jiang, A. Godfreyb, and Q. Liu, “Microstructure and texture evolution during warm compression of the magnesium alloy AZ31,” Science in China Series E: Technological Sciences, vol. 52, no. 1, pp. 186-189, Jan. 2009.
[8] H. Jeong and T. Ha, “Texture development in a warm rolled AZ31 magnesium alloy,” Journal of Materials Processing Technology, vol. 187-188, pp. 559-561, Jun. 2007.
[9] I. Kim and S. Akramov, “the Development of the Microstructure and Texture in Cold Rolled Az31 Mg Alloy Sheets,” International Journal of Modern Physics B, vol. 22, no. 31 & 32, p. 5925, 2008.
[10] F. Chen, “Formability of stamping magnesium-alloy AZ31 sheets,” Journal of Materials Processing Technology, vol. 142, no. 3, pp. 643-647, Dec. 2003.
[11] A. Jager, “Tensile properties of hot rolled AZ31 Mg alloy sheets at elevated temperatures,” Journal of Alloys and Compounds, vol. 378, no. 1-2, pp. 184-187, Sep. 2004.
38
[12] Y. Lee, M. Kim, S. Kim, Y. Kwon, S. Choi, and J. Lee, “Experimental and analytical studies for forming limit of AZ31 alloy on warm sheet metal forming,” Journal of Materials Processing Technology, vol. 187-188, pp. 103-107, Jun. 2007.
[13] H. Li, E. Hsu, J. Szpunar, R. Verma, and J. T. Carter, “Determination of Active Slip/Twinning Modes in AZ31 Mg Alloy Near Room Temperature,” Journal of Materials Engineering and Performance, vol. 16, no. 3, pp. 321-326, Apr. 2007.
[14] B. Kim, J. Do, S. Lee, and I. Park, “In situ fracture observation and fracture toughness analysis of squeeze cast AZ51–xSn magnesium alloys,” Materials Science and Engineering: A, vol. 527, no. 24-25, pp. 6745-6757, Sep. 2010.
[15] M. Barnett, “Twinning and the ductility of magnesium alloysPart II. ‘Contraction’ twins,” Materials Science and Engineering: A, vol. 464, no. 1-2, pp. 8-16, Aug. 2007.
[16] M. H. Yoo, “Slip, twinning and fracture in polycrystalline 3% silicon iron,” Acta Metallurgica, vol. 14, no. 1. pp. 35-41, Jan-1966.
[17] T. Alsamman and G. Gottstein, “Room temperature formability of a magnesium AZ31 alloy: Examining the role of texture on the deformation mechanisms,” Materials Science and Engineering: A, vol. 488, no. 1-2, pp. 406-414, Aug. 2008.
[18] X. Li, P. Yang, L.-N. Wang, L. Meng, and F. Cui, “Orientational analysis of static recrystallization at compression twins in a magnesium alloy AZ31,” Materials Science and Engineering: A, vol. 517, no. 1-2, pp. 160-169, Aug. 2009.
[19] M. P. Miles, D. Fullwood, B. L. Adams, a. Khosravani, J. Scott, and R. K. Mishra, “Room Temperature Ductility and Microstructure of Magnesium AZ31B Sheet,” Journal of Materials Engineering and Performance, Mar. 2011.
[20] M. Barnett, M. Nave, and C. Bettles, “Deformation microstructures and textures of some cold rolled Mg alloys,” Materials Science and Engineering A, vol. 386, no. 1-2, pp. 205-211, Nov. 2004.
[21] J. Kang, D. Wilkinson, and R. K. Mishra, “Inhomogeneous Deformation of AZ31 MAgnesium Sheet in Uniaxial Tension,” in TMS. 2011, pp. p. 307-311.
[22] “ASM Online Handbook,” 2011. [Online]. Available: http://products.asminternational.org/hbk/index.jsp.
[23] “ARAMIS 3D Digital Image Correlation Strain & 3D Deformation,” Trillion Optical Test System, 2011. .
[24] A. Rezaee-Bazzaz, H. Noori, and R. Mahmudi, “Calculation of forming limit diagrams using Hill’s 1993 yield criterion,” International Journal of Mechanical Sciences, vol. 53, no. 4, pp. 262-270, Apr. 2011.
39
[25] M. Butuc, “A theoretical study on forming limit diagrams prediction,” Journal of Materials Processing Technology, vol. 142, no. 3, pp. 714-724, Dec. 2003.
[26] Y. Chino, K. Kimura, and M. Mabuchi, “Deformation characteristics at room temperature under biaxial tensile stress in textured AZ31 Mg alloy sheets,” Acta Materialia, vol. 57, no. 5, pp. 1476-1485, Mar. 2009.
[27] W. B. Hutchinson and M. R. Barnett, “Effective values of critical resolved shear stress for slip in polycrystalline magnesium and other hcp metals,” Scripta Materialia, vol. 63, no. 7, pp. 737-740, Oct. 2010.
[28] S. Agnew, “Study of slip mechanisms in a magnesium alloy by neutron diffraction and modeling,” Scripta Materialia, vol. 48, no. 8, pp. 1003-1008, Apr. 2003.
[29] D. Brown, S. Agnew, M. Bourke, T. Holden, S. Vogel, and C. Tome, “Internal strain and texture evolution during deformation twinning in magnesium,” Materials Science and Engineering: A, vol. 399, no. 1-2, pp. 1-12, Jun. 2005.
[30] A. Chapuis and J. H. Driver, “Temperature dependency of slip and twinning in plane strain compressed magnesium single crystals,” Acta Materialia, vol. 59, no. 5, pp. 1986-1994, Mar. 2011.
[31] A. Chapuis and J. Driver, “A fundamental study of the high temperature deformation mechanisms of magnesium,” in Journal of Physics, 2012, p. 420.
[32] N. Ono, R. Nowak, and S. Miura, “Effect of deformation temperature on Hall–Petch relationship registered for polycrystalline magnesium,” Materials Letters, vol. 58, no. 1-2, pp. 39-43, Jan. 2004.
[33] M. Barnett, S. Jacob, B. Gerard, and J. Mullins, “Necking and failure at low strains in a coarse-grained wrought Mg alloy,” Scripta Materialia, vol. 59, no. 10, pp. 1035-1038, Nov. 2008.
[34] M. Knezevic, A. Levinson, R. Harris, R. K. Mishra, R. D. Doherty, and S. R. Kalidindi, “Deformation twinning in AZ31: Influence on strain hardening and texture evolution,” Acta Materialia, vol. 58, no. 19, pp. 6230-6242, Nov. 2010.
[35] M. Miles, J. Siles, W. Wagoner, and K. Narasimhan, “A Better Sheet Formability Test,” Metallurgical and Materials Transactions A, vol. 24, no. May, pp. 1143 - 1151, 1993.
[36] S. R. Agnew and Ö. Duygulu, “Plastic anisotropy and the role of non-basal slip in magnesium alloy AZ31B,” International Journal of Plasticity, vol. 21, no. 6, pp. 1161-1193, Jun. 2005.