Journal of Materials Processing Technology 224 (2015) 156168

Contents lists available at ScienceDirect

Journal of Materials Processing Technology

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

Die design method for thin plates by indirect rheo-casting process and
effect of die cavity friction and punch speed on microstructures and
mechanical properties
Chul Kyu Jin a , Chang Hyun Jang a , Chung Gil Kang b,
a
Precision Manufacturing System Division, Graduate School, Pusan National University, San 30 Chang Jun-dong, Geum Jung-Gu,
Busan 609-735, South Korea
b
Engineering Research Center for Net Shape and Die Manufacturing, School of Mechanical Engineering, Pusan National University, San 30 Chang Jun-dong,
Geum Jung-Gu, Busan 609-735, South Korea

a r t i c l e

i n f o

Article history:
Received 12 November 2014
Received in revised form 28 April 2015
Accepted 1 May 2015
Available online 12 May 2015
Keywords:
Semi-solid slurry
Indirect rheo-casting
Electromagnetic stirring
Thin plate
Filling simulation
A356 alloy

a b s t r a c t
Thin plates with a thickness of 1.2 mm are fabricated from semi-solid A356 alloy through an indirect
rheo-casting process both with and without an electromagnetic stirrer (EMS). The thin die cavity for
forming is designed with the uid analysis software MAGMA. A semi-solid slurry with a solid fraction of
40% is prepared and then injected into the die of a 200 t hydraulic press. Forming tests are performed on
the thin plates at two punch speeds (30 and 300 mm/s) and two cavity friction conditions (mf = 0.4 and
mf = 0.9). The formability, mechanical properties, and microstructure are then evaluated. The semi-solid
slurry obtained with an EMS contains ne and globular solid particles; the semi-solid slurry produced
without an EMS reveals rosette particles and coarser globular solid particles. At high friction (mf = 0.9),
the cavity is mainly lled with the liquid phase. At a higher punch rate, the thin plates show better
formability and a microstructure with ne and even solid particles. The tensile strength and elongation
of the thin plate formed with a punch speed of 300 mm/s in the cavity with graphite lubrication (mf = 0.9)
are 216 MPa and 10%, respectively. These values are 57 MPa and 5.5% higher, respectively, than those of
the thin plate formed at a punch speed of 30 mm/s.
2015 Elsevier B.V. All rights reserved.

1. Introduction
The die casting process for aluminium involves the high-speed
injection of molten metal, which leads to internal defects because
of remaining gas or air in the molten metal, which in turn deteriorates the mechanical properties. Niu et al. (2000) found that the
volume of gas porosity and the pore sizes in the castings are significantly reduced through the use of a vacuum during die casting.
This markedly improves the density and mechanical properties,
particularly the tensile strength and ductility. The forging process has limited formability of near-net shapes and reduces the
after-treatment productivity and die life, which makes eco-friendly
production impossible. Squeeze casting is a metal forming process
where molten metal is solidied under a relatively high pressure to
reduce gas or shrinkage porosity. However, this process produces
a rosette and dendrite structure (Yue and Chadwick, 1996) and has

Corresponding author. Tel.: +82 51 510 1455; fax: +82 51 518 1456.
E-mail address: cgkang@pusan.ac.kr (C.G. Kang).
http://dx.doi.org/10.1016/j.jmatprotec.2015.05.002
0924-0136/ 2015 Elsevier B.V. All rights reserved.

the disadvantages of a shortened die life, limited shape complexity,

difculty with producing thin parts, and limited maximum size and
weight (Ghomashchi and Vikhrov, 2000).
Flemings et al. (1976) developed a rheological (semi-solid)
material and the rheocasting process as an alternative to metal
forming processes such as die casting and forging. Their process
produces a highly uid slurry of solid spheroids dispersed in liquid.
Joly and Mehrabian (1976) showed that the viscosity of the slurry
at a given volume fraction of solids decreases with a decreasing
cooling rate and increasing shear rate. The rheo-forming method,
which is performed on a material in the semi-solid state (i.e., the
temperature is above the solid line but below the liquid one), is
a solution to the problems of casting and forming processes. In
the rheo-forming process, molten aluminium is stirred as the temperature is decreased in order to create semi-solid slurry with a
controlled grain size, which is then injected into a die and formed
with a press. Kapranos et al. (2000) described the process of producing and assessing a high-quality thixoformed component using
an aluminium alloy and showed that thixoforming clearly has
near net-shape capability. Ji et al. (2001) developed a twin-screw

C.K. Jin et al. / Journal of Materials Processing Technology 224 (2015) 156168

rheo-moulding process, and Fan et al. (2005) presented a rheodiecasting (RDC) process which directly uses liquid Al alloys.
Their results indicated that the RDC samples had close to zero
porosity and a ne and uniform microstructure throughout the
entire sample under the as-cast condition. Atkinson (2005) summarised routes to spheroidal microstructures, types of semi-solid
processing, the advantages and disadvantages of these routes, the
background rheology, mathematical theories of thixotropy, the
transient behaviour of semi-solid alloy slurries, and computational
modelling.
The one major drawback of rheo-forming or thixoforming for
processing semi-solid metals is controlling the liquid or solid segregation (i.e., separation of the solid and liquid phases or uneven
distribution of the solid phase). When a semi-solid slurry lls the
die, the material comes in contact with the cavitys wall. This
causes uneven ows leading to segregation of the solid phase
(primary -Al particles) from the liquid phase. Chen and Tsao
(1997) proposed semi-solid deformation mechanisms and predicted the segregation phenomenon based on deformation of a
phenomenological model. Kang et al. (2007) investigated the effect
of changing the injection velocity on the globular microstructure
and mechanical properties of a product from semi-solid die casting. They found that the difference in the solid fraction between
samples with and without liquid segregation was approximately
1520%. For thin plates, segregation of the solid and liquid phases
can be more severe, which makes their mechanical properties
uneven at different locations. Because of the problems associated
with segregation and the low initial forming temperatures for
semi-solid metal processing, no research has been carried out so
far on the fabrication of thin plates with the rheo-forming process.
In this study, an indirect rheo-casting process was applied to
compensate for the disadvantages of indirect squeeze casting and
produce aluminium thin plates for electric and automobile parts
and fuel cell bipolar plates. The indirect rheo-casting process for
thin plates involves a low pouring temperature in a thin die cavity. This makes it highly likely that the material will fail to ll the
cavity entirely and solidify starting from the centre, which will
result in incomplete forming. Therefore, the aim of this study was
to nd ways to design a die for indirect rheo-casting which are
appropriate for the thin plates shape. A gate shape and overow
adequate for rheological behaviour were designed with the software MAGMA to allow the semi-solid slurry to ll the cavity. The
A356 alloy with a wide solid-liquid coexistent region was used as
the semi-solid slurry. A semi-solid slurry with ne and globular
solid particles was fabricated through the use of an electromagnetic stirrer (EMS) to control the grain size of the A356 alloy.
The semi-solid slurry was injected into a die installed in a 200t hydraulic press to form a thin plate. Experiments for forming
thin plates were performed at two punch speeds and two cavity
friction conditions, and the effect of the punch speed and friction on the formability, microstructure, and mechanical properties
was analysed. The microstructure and mechanical properties of
the formed thin plate samples were measured at different locations.

2. Experimental procedure

formula called a power law is used (Kim and Kang, 2000; Atkinson,
2005):
 = k n1

(1)

where  is the shear stress,  is the shear rate, k is the power law factor, and n is the power law index. When n = 1, the material becomes
a Newtonian uid whose viscosity  is the same as k. Kim and Kang
(2000) set n = 1 for the Newtonian uid model. For the Ostwaldde
Waele uid model, which is for the semi-solid state, the experimentally obtained n value was 0.48 to 0.45 (shear rate = 32500 s1 ).
This was applied in MAGMASOFT for comparison of the lling analysis results within the die. Their results demonstrated that the
Ostwaldde Waele uid model is consistent with the experimental results. The viscous behaviour model of MAGMASOFT uses the
Ostwaldde Waele model, which expresses the non-Newtonian
aspect of semi-solid materials through a power law:
n = m n1

(2)

where  is the apparent dynamic viscosity, m is the Ostwaldde

Waele coefcient, n is the Ostwaldde Waele exponent, and  is
the density.
The governing equations of MAGMASOFT are the control volume
nite difference method, continuity equation, NavierStokes equation, energy equation, and volume of uid (VOF) method. These are
the same used for liquids.
2.2. Simulation preparation for die design
When a semi-solid slurry is compressed, the liquid phase moves
towards the surface of the slurry, which leads to surface cracks.
The void content and segregation of the solid and liquid phases
become more severe at the side of the compressed specimen. Seo
et al. (2002) conducted compression experiments to investigate
the deformation behaviour of a semi-solid material with varying processing parameters, such as the test specimen size and
strain-rate. They suggested that the rheo-forging die for thin plates
must be designed as an indirect type of structure. To form thin
plates using a semi-solid slurry, a die for indirect rheo-casting was
designed where lling is carried out by compressing the semi-solid
slurry in closed upper and bottom dies with a punch. Because this
structure is similar to the die casting process, the design of the gate
system and overow is a critical variable. Analysis of the behaviour
of the semi-solid slurry as it goes through the gate and lls the cavity is a major factor for the design of thin plates. Seo et al. (2007)
analysed how the gate shape affects the liquid segregation of a
semi-solid slurry and concluded that a wider gate makes lling
more likely to be done in order and liquid segregation less likely
to form.
In order to examine the lling behaviour of a semisolid slurry according to the gate shape, a thin plate cavity
(150 mm 150 mm 1.2 mm) was simulated with different gate
shapes. The simulations were performed by using the A356 thixomodule (Ostwaldde Waele model) of MAGMA. Table 1 lists the
Table 1
Simulation parameters.
Parameters
Molten metal

2.1. Semi-solid ow model

The uid model of semi-solid materials shows that the viscosity depends on the shear rate. Semi-solid materials lose viscosity
drastically as the shear rate increases but have nearly constant viscosity at very low shear rates. In order to express the dependence of
viscosity on the shear rate in a high shear rate region, an empirical

157

Latent heat (Q)

Die
Heat transfer coefcient

Values
Material
Liquidus temperature (TL )
Solidus temperature (TS )
Initial temperature (TM )
430 kJ/kg
Material
Initial temperature (TD )
Material and die
Die and die

A356
617 C
547 C
596 C
SKD 61
300 C
7000 W/m2 K
1000 W/m2 K

158

C.K. Jin et al. / Journal of Materials Processing Technology 224 (2015) 156168

Table 2
Chemical composition of A356 alloy (wt%).
Si

Mg

Ti

Fe

Ni

Mn

Zn

Pb

Al

7.08

0.35

0.17

0.08

0.07

0.01

0.01

0.01

Bal.

conditions and heat transfer coefcient values used in the simulation.

2.3. Fabrication of semi-solid slurry
A356 alloy was used for the semi-solid slurry, and an EMS was
used to control the size of the solid particles. Because A356 alloy
has great uidity at the two-phase mushy zone and can enhance the
mechanical strength with heat treatment, it is used in automobile
parts such as knuckles, arms, and housing that require reliability.
In particular, the mechanical properties of A356 are closely related
to the size of the primary particles, secondary dendrite arm spacing
(SDAS), and Si particle shape and distribution within the eutectic
matrix. Table 2 lists the chemical compositions of A356 alloy. Fig. 1
shows the solid fraction versus temperature of the A356 alloy. The
solid phase fraction of A356 alloy at different temperatures was
determined by using the obtained differential scanning calorimetry (DSC) curve. The liquidus and solidus temperatures for A356
alloy were 617 and 547 C, respectively, and the solid fraction for a
temperature of 596 C was 40%.
Fig. 2(a) shows a photo of the EMS used in this study. The EMS
consisted of three phases (P, R, S) and three poles with a coil placed

Fig. 1. Solid fraction versus temperature of A356 alloy.

Fig. 2. Electromagnetic stirrer: (a) real picture and (b) schematic diagram.

Fig. 3. Variations in magnetic induction density at three positions as function of

stirring current.

vertical to the core. The core for xing the coil position was fabricated by piling up several 0.35 mm thick plates. The core consisted
of 240 unit laminations of SiZn alloy plates, and the coil was wound
around the core. Each phase was placed in a cylindrical direction,
as shown in Fig. 2(b), to let the current move alongside the coil, and
an electromagnetic force was generated in the cylindrical direction to stir the molten metal. The electromagnetic force of the EMS
was measured with a gaussmeter at three positions. Fig. 3 shows
the variations in the magnetic induction density measured at three
positions inside the EMS as a function of the current. The measured
magnetic induction density was proportional to the increase in the
current at each position. At a current of 60 A, the magnetic induction densities at the upper, middle, and lower positions were 640,
680 and 1120 G, respectively. The stirring force induced shear stress
in the molten aluminium, which controlled the growth of dendrite
arms which form during solidication. Thus, it controlled the grain
size of the solid phase and made the grains globular.
Fig. 4 shows the process to make a semi-solid slurry. First, the
cup is inserted into the EMS, and a ladle is used to scoop the molten
metal from the furnace. Then, stirring starts as the electric current
is applied while the molten metal is poured into the EMS cup. (The
molten metal is at a temperature of 680 C in the furnace, 635 C
in the ladle, and 620 C in the cup.) Stirring is performed until the
temperature of the molten metal in the cup cools to 596 C, i.e., the
solid fraction (fs) is 40%. It takes about 78 s of stirring to reach this
level. The variables for the stirring experiment were the molten
metal temperature at the start of stirring (TS ), the stirring current
(A), and the stirring time (t). Bae et al. (2007) suggested that ne
and globular solid particles can form when the molten metal temperature at the start of stirring is below 655 C, the stirring current
is 60 A, and the stirring time is 60 s based on an electromagnetic
stirring experiment using A356. Therefore, in the present experiment, the molten metal temperature was 620 C, and the stirring
current was 60 A, as given in Table 3.
Stainless steel 304 was used for the EMS cup. Stainless steel is
nonmagnetic because it is austenitic, so it is not affected by electromagnetic forces. In addition, stainless steel 304 does not deform
even at temperatures over 700 C. Seo et al. (2002) performed a
compression experiment on a semi-solid slurry and examined how
changes in the height and diameter of the billet affect the liquid segregation. They concluded that a larger billet diameter makes it less
likely for pores and liquid segregation to occur. In addition, a greater
billet length increases the void content gets. Thus, after the amount
of material required for thin plate forming was considered, the EMS
cup was designed to have a similar diameter to the inner diameter
of the die sleeve (60 mm) and a much lower height. Fig. 5(a) and
(b) illustrates the shape dimensions of the EMS cup and actual cup,

C.K. Jin et al. / Journal of Materials Processing Technology 224 (2015) 156168

159

Fig. 4. EMS process for fabricating semi-solid slurry with ne and globular solid particles.

Table 3
Experimental conditions for semi-solid slurry fabrication.
Parameters

Values

Stirring method
Stirring current (A)
Temperature at the start of
stirring (TS )
Temperature at the nish
of stirring (TF )/solid
fraction (fs)
Stirring time (t)

Electromagnetic stirring
60 A
620 C
596 C/40%

60 s

Fig. 6. 200-t hydraulic press with die installed for thin plate fabrication.

Table 4
Experimental conditions for indirect rheo-casting with semi-solid slurry.
Fig. 5. Geometries of stirring cup and slurry: (a) section of cup, (b) photo of cup,
and (c) photo of semi-solid slurry.

respectively. Fig. 5(c) shows the semi-solid slurry. The cup is 2 mm

thick and 100 mm long. Because the molten metal rotates during
stirring because of the stirring force, the molten metal will ow
over the cup if the cup is lled to the inlet. Therefore, the cup was
lled to a height of 90 mm with the molten metal. The volume of
the fabricated semi-solid slurry was 165,597 mm3 , and the volume
of the thin plate model calculated using the space nder function of
the software UG NX6 was 161,304 mm3 . The amount of semi-solid
slurry was adequate to ll the thin cavity.
2.4. Indirect rheo-casting process
The semi-solid slurry (fs = 40%) produced by the EMS experiment
was injected into the die in a 200 t hydraulic press for the thin
plate forming experiment. Fig. 6 shows the 200 t hydraulic press
with the installed die for thin plates. Table 4 and Fig. 7 present the
experimental conditions for the thin plate forming process and a
general diagram of the process, respectively. The dies temperature
was maintained at 280290 C with a cartridge heater. The pressure of the punch was set to 200 MPa, and two speed were used:
30 and 300 mm/s. The semi-solid slurry was injected into the die
and compressed with the punch; the pressure was maintained for
5 s. To ensure that the formed plate would not be deformed by the
ejector pin, the die was left open for 10 s for some solidication.
Then, the plate was taken out of the die and immediately cooled in
water.

Parameters

Value

Temperature and solid fraction

of semi-solid slurry
Temperature of die
Pressure of press punch (PP )
Velocity of press punch (VP )
Lubricant for die cavity
Condition 1
Condition 2
Condition 3

596 C and 40%

280290 C
200 MPa
30 and 300 mm/s
Graphite
VP = 300 mm/s, without lubricant (mf = 0.9)
VP = 300 mm/s, graphite lubricant (mf = 0.4)
VP = 30 mm/s, graphite lubricant (mf = 0.4)

Pinsky et al. (1984) discovered that segregation of the solid

and liquid phases increases with friction between the die and
material in compression and ring compression experiments on
semi-solid SnPb alloy. Kang et al. (1999) performed compression experiments on semi-solid materials at different strain rates
and concluded that the solid phase is evenly distributed at high
strain rates, which leads to less liquid segregation. Therefore, three
forming experiments were performed to analyse the formability,
microstructure, and mechanical properties for two friction conditions between the cavity and semi-solid slurry and two punch
speeds. A graphite lubricant was sprayed in the cavity for the forming experiment with two punch speeds, and the experiment was
performed with a non-lubricated cavity to analyse the segregation
of the solid and liquid phases as well as defects for different cavity
friction states. Experiment condition 1 was a non-lubricated cavity
(mf = 0.9) with a punch speed of 300 mm/s. Conditions 2 and 3 were
a graphite-lubricated cavity (mf = 0.4) with punch speeds of 300 and
30 mm/s, respectively. These conditions are listed in Table 4. Five

160

C.K. Jin et al. / Journal of Materials Processing Technology 224 (2015) 156168

Fig. 7. Schematic diagram of indirect rheo-casting process: (a) input semi-solid slurry, (b) forming, and (c) ejecting.

tests were performed under each experiment condition to produce

a total of 15 thin plate samples.
2.5. Measurement of microstructure and mechanical properties
The size and globularity of the solid particles in the microstructure were measured using an image analyser. The equivalent
diameter (average size) of the solid particles is dened by Eq. (3)
below, and the shape factor (globularity) represents the roundness
of the particles and is dened by Eq. (4). When R = 1, a particle has
a perfectly circular shape; for a less globular particle, R > 1 because

it has a longer circumference than a circular particle with the same

area.


D=

R=

4A


P2
4A

(3)

(4)

here D, R, A, and p are the diameter, radius, area, and perimeter,

respectively, of a globule.

Fig. 8. Mould-lling behaviour of semi-solid slurry according to gate shape: (a) straight gate and (b) fan gate.

C.K. Jin et al. / Journal of Materials Processing Technology 224 (2015) 156168

161

Fig. 9. (a) Detail design of fan gate system (unit: mm) and (b) material age.

In order to analyse the mechanical properties of the formed

thin plate, a tensile test was performed on specimens fabricated
with different cavity locations. The tensile specimens were at with
a 25 mm gage length and 6 mm width; there were fabricated in
accordance with the ASTM E 8 M specications (sub-size specimen). The specimens were set to be as thick as the sample thin
plate. For the tensile test, a 25-t MTS was used, and the strain rate

was set to 1 mm/min. For accurate measurement of the elongation,

an extensometer was used; this is a method to contact specimens. The tensile test specimens were used to measure the Vickers
hardness. Each specimen was tested three times and ve times
for tensile and hardness, respectively. The results were averaged,
and the maximum and minimum values are presented with an
error bar.

Fig. 10. Mould-lling behaviour of semi-solid slurry for model with two overows at side: (a) velocity and (b) temperature.

162

C.K. Jin et al. / Journal of Materials Processing Technology 224 (2015) 156168

3. Result and discussion

3.1. Simulation for die design
To design the gate shape for a thin plate, the biscuit was set
to 60 mm, which was equivalent to the inner diameter of sleeve.
Fig. 8(a) shows the lling behaviour of the semi-solid slurry in a
die cavity with a straight gate. As shown in the simulation results,
the slurry lled the cavity in a straight line along the gates shape
because of its viscosity. The sides of the cavity were lled by the
backow of the uid, which reached the end of the cavity and
returned. This may cause incomplete lling on the sides of the cavity. Thus, the gate was expanded to a fan type so that the material
could ll up the entire cavity after passing through the gate. The
length of the gate was shortened to avoid the semi-solid slurry
solidifying within the cavity during lling. Fig. 8(b) shows that the
entire width of the cavity was lled as the material passed through
the gate.
After a fan shape was selected, specic details were designed for
the gate. As shown by the gate cross-section in Fig. 9, the gate was
18 mm thick at the part connected to the biscuit and then narrowed
closer to the cavity. For easy removal of the formed thin plate, 5 and
6 gradients were set for the biscuits top and bottom parts, respectively, along the parting line of the die. Each corner was rounded.
When the molten metal cooled to a semi-solid slurry, the upper part
of the slurry had a poor surface state because of the air inow from
outside; the surface of the slurry was 3 C lower than the centre
because of heat transfer. Because the upper part of the semi-solid
slurry was injected into the die rst, the poor upper part of slurry
may ow into the cavity and cause incomplete lling. To account for

this possibility, the gate was located 4 mm higher than the bottom
part of the biscuit (i.e., around the semi-solid slurry upper part), as
shown in Fig. 9(a). The poor upper part (further solidied) of the
slurry does not ll the cavity and is placed at the area of the semisolid slurry upper part. Fig. 9(b) shows the material age results; the
initial material did not ow into the cavity. The material age refers
to the elapsed time for material inow; a longer time indicates an
earlier material inow.
The fan gate caused the material to ll the entire cavity. The ow
rate at the sides of the cavity increased as well, as shown in Fig. 8(b).
Because the ow at the sides of the cavity reaches the end of the
cavity faster than the ow at the centre of the cavity, the gas or air
in the centre cannot be released through the dies parting line and is
locked inside the cavity instead. The formed sample surface may be
rough at the end of the cavity because of defects such as pores, air
pockets, and ow marks. In addition, an uneven lling speed may
cause the solid and liquid phases to segregate. Therefore, to obtain
a uniform ow rate within the cavity, two overows at the sides
of the cavity were designed. Fig. 10 shows the simulation results
for a model with two overows placed at the sides of the cavity.
As the faster ow on the cavity sides was led to the overows,
the cavity was evenly lled overall. However, when the material
passed through the cavitys centre, the velocity and temperature
dramatically dropped. After the cavity was totally lled, the temperature at the end of the cavity was below 585 C. These may lead
to incomplete lling at the end of the cavity in an actual forming
experiment, which can cause casting defects such as surface cracks
and pores. Because the material portion that solidies at temperatures below 585 C at the end of the cavity can also be removed
with an overow, ve overows were placed at the cavity end, as

Fig. 11. Mould-lling behaviour (temperature) of semi-solid slurry for model with two overows at side and ve overows at end.

Fig. 12. Pressure distribution in cavity for models: (a) two overows at side and ve overows at end, (b) four overows at side and ve overows at end.

C.K. Jin et al. / Journal of Materials Processing Technology 224 (2015) 156168

163

Fig. 13. Indirect rheo-casting die for forming thin plates with semi-solid slurry: (a) photo of real die cavity and (b) die structure.

shown in Fig. 11. The material part that solidied at below 585 C
lled the ve overows at the cavity end, and the temperature in
the cavity was maintained at 590 C.
Fig. 12(a) shows the pressure distribution in the cavity. The
pressure distribution was signicantly uneven. The pressure at the
sides was about atmospheric pressure (1013 mbar). Applied pressure was lost, and some areas were subjected to pressures over
10,000 mbar. The ve overows at the end of cavity were subjected
to atmospheric pressure. The applied pressure was not transferred

to the cavity end because of the uneven pressure distribution. This

result shows that incomplete lling may occur at the entrance
of overows in an actual forming experiment. In order to solve
this problem, two overows were added at the sides, as shown
in Fig. 12(b). A pressure of over 4000 mbar was evenly distributed
over the cavity with the two added overows.
Based on the simulation results, the cavity shape and die structure were designed as shown in Fig. 13(a) and (b), respectively.
Because the dies temperature should stay at 280290 C during the

Fig. 14. Microstructures of semi-solid slurry (a) with and (b) without EMS.

164

C.K. Jin et al. / Journal of Materials Processing Technology 224 (2015) 156168

forming experiment, three and four 20 mm holes were machined

in the upper and bottom dies, respectively, as cartridge heater
holes. Then, 1.8 mm holes were machined at the centres of the
upper and bottom die, and a K-type thermocouple was inserted
for measurement and control of the die temperature. For easy
removal of the thin plate sample after forming, 14 mm holes were
machined for ejector pins: one at the sleeve, six at the cavity, and
one for each overow. The diameter of the punch for compressing the semi-solid slurry was 60 mm, which was the same as the
sleeves inner diameter. The upper part of the punch had a diameter of 100 mm to prevent the punch from buckling during repeated
experiments. To prevent the aluminium alloy from sticking to the
die surface and enhance the surface hardness, the punch, sleeve,
and cavity were heat-treated through nitriding.
Fig. 15. Volume fraction, equivalent diameter, and roundness of solid particles in
semi-solid slurry with and without EMS.

Fig. 16. Thin plate samples fabricated with different punch speeds (VP ) and friction states of cavity: (a) 30 mm/s, without lubricant; (b) 300 mm/s, graphite lubricant; and
(c) 30 mm/s, graphite lubricant.

C.K. Jin et al. / Journal of Materials Processing Technology 224 (2015) 156168

165

without an EMS were 119 m and 3.0, respectively. The semi-solid

slurry with an EMS had volume fractions of 45% at the centre and
41% at the sides. The semi-solid slurry without an EMS had volume
fractions of 43% at the centre and 40% at the sides.
The results conrmed that a semi-solid slurry with ne and
globular solid particles can be obtained with an EMS. If a semisolid slurry with uneven and coarse particles is injected into the
die and then compressed by a punch, incomplete lling or misrun
(short shot) can be caused by the poor uidity, and the mechanical properties of the formed sample will be low because of the
poor microstructure. Therefore, a semi-solid slurry produced with
an EMS was used in the experiment on thin plate forming.
3.3. Formability of thin plate

Fig. 17. Thickness of formed thin plates.

3.2. Microstructures of semi-solid slurry

Before the thin plate forming experiment, the semi-solid slurry
needed to be cooled down immediately in water for analysis of
the microstructure. Fig. 14 shows the microstructures of the semisolid slurry stirred with an EMS and not stirred but cooled down
to 596 C. The microstructure showed primary -Al for the solid
phase and a eutectic phase for what used to be the liquid phase
in the solidliquid state. In the microstructure with stirring, ne
and globular solid particles were distributed evenly at the centre
of the semi-solid slurry because the uneven growth of particles
was controlled (Fig. 14(a)). At the sides of the semi-solid slurry, the
cups surface was about 3 C lower at the centre because of the heat
transfer during stirring. Thus, the solid particles were coarser than
at the centre. For the microstructure without stirring, a large number of rosette particles, dendrites, and some ne globular particles
were distributed (Fig. 14(b)). The molten metal grew as a dendrite
structure until it cooled down to 596 C. The grain size was coarser
around the sides of the material.
Fig. 15 shows the volume fraction, equivalent diameter, and
roundness of the solid particles in the semi-solid slurry with and
without an EMS. At the centre, the equivalent diameter and roundness of the solid particles of the semi-solid slurry with an EMS were
75 m and 1.5, respectively. Those of the solid particles of the semisolid slurry without an EMS were 110 m and 2.3, respectively. On
the sides, the equivalent diameter and roundness of the solid particles of the semi-solid slurry with an EMS were 82 m and 1.65,
respectively. These of the solid particles of the semi-solid slurry

Fig. 16 illustrates a selection of the best thin plate samples fabricated under each condition. The thin plate samples were formed
at two punch speeds (30 and 300 mm/s) and cavity friction conditions (with and without the graphite lubricant). The black part on
the surface of the plate sample is the graphite lubricant. Fig. 16(a)
shows a sample formed at a punch speed of 300 mm/s without the
graphite lubricant in the cavity (mf = 0.9). Because semi-solid slurry
exhibits severe stickiness from its viscosity, the punch may get
stuck in the sleeve if the slurry is compressed without a lubricant.
Therefore, in this experiment, a small amount of graphite lubricant was sprayed within the die sleeve. The sample formed under
this condition had short shots at the sides of the end of the cavity. As shown by the simulation result for the pressure in Fig. 5(b),
this short shot can be attributed to the decreased pressure at the
entrance of the overow at the end of the cavity. Fig. 16(b) and (c)
shows the thin plate samples formed with a graphite lubricant in
the cavity (mf = 0.4) at punch speeds of 300 and 30 mm/s, respectively. The thin plate sample formed at a punch speed of 300 mm/s
with the graphite lubricant lled the cavity completely without
any short shots. Although only two overows were connected at
the end of the cavity, all ve overows were formed in the actual
experiment, and the entrance of the overow was broken and fell
off because of the ejection force from the sample being pulled out.
Forming at a punch speed of 30 mm/s led to a short shot at the end of
the cavity, and six overows resulted in short shots (Fig. 16(c)). As
shown by the simulation results for the lling temperature in Fig. 4,
this is the point where the materials temperature dropped below
585 C. Because the punch speed was 10 times less than the simulation condition, the temperature decreased even farther, which
may have caused incomplete forming.
Fig. 17 shows the thickness of the formed thin plate samples.
Groups A (A1, A2, and A3) and B (B1, B2, B3) indicate the centre
and side positions, respectively, of the sample. A1 and B1 (near the

Fig. 18. Simulation results of solidication behaviour.

166

C.K. Jin et al. / Journal of Materials Processing Technology 224 (2015) 156168

gate) were thicker than A3 and B3 (near the overow), and group B
was thicker than group A. This was presumed to be due to shrinkage upon solidication. Fig. 18 shows the simulation results of the
solidication behaviour after lling. Similar to the lling patterns of
the semi-solid slurry, the solidication patterns in a contour shape
formed toward the gate at the overow. Because positions B1 and
A1 underwent solidication slowly compared to the other positions
in the cavity, they can be expected to be thicker. In contrast, B3 and

A3 solidied faster than the other positions, which suggests that

they should be thinner than the other positions. The thickness of
the thin plate fabricated with graphite lubricant in the cavity at
a punch speed of 300 mm/s was the nearest to the thickness of
the die cavity; thus, it was the most precisely formed product. The
results for thin plate forming using a semi-solid slurry suggest that
the optimal conditions are a high punch speed of 300 mm/s and
graphite lubricant in the cavity.

Fig. 19. Microstructure of thin plate samples fabricated with different punch speeds (VP ) and friction states of cavity: (a) 30 mm/s, without lubricant; (b) 300 mm/s, graphite
lubricant; and (c) 30 mm/s, graphite lubricant.

C.K. Jin et al. / Journal of Materials Processing Technology 224 (2015) 156168

167

3.4. Microstructures of thin plate

Fig. 19(a) illustrates the microstructure of the sample formed
with a punch speed of 300 mm/s and with no graphite lubricant in
the cavity (condition 1). The microstructure shows that the solid
and liquid phases were severely segregated at positions F and E.
At position F (near gate), there were a few coarse dendrites from
coalescence; at position E (far from gate), there were few solid particles. When there was high friction, the ow of solid particles was
difcult. Thus, a small part of the solid phase moved into the cavity,
and the liquid phase, which had better uidity, lled the cavity. In
other words, high friction between the semi-solid material and die
cavity cause a high degree of segregation between the solid and
liquid phases.
Fig. 19(b) and (c) shows the microstructures of the samples
formed at punch speeds of 300 and 30 mm/s (conditions 1 and
2), respectively, with graphite lubricant in the cavity. The punch
speed of 300 mm/s produced a ner grain size and denser distribution of the solid phase than the punch speed of 30 mm/s. At
30 mm/s, the gap between solid particles was large because there
was a low number of solid particles with a diameter of over 100 m.
The porosity was high at position E. Fig. 20 shows the volume fraction, equivalent diameter, and roundness of solid particles in the
samples formed at punch speeds of 300 and 30 mm/s with graphite
lubricant. For the sample formed at 300 mm/s, the equivalent diameter, roundness, and volume fraction of solid particles at position
F were 65 m, 1.25, and 45.5%, respectively. At position E, they
were 62 m, 1.27, and 43.5%, respectively. For the sample formed
at 30 mm/s, the equivalent diameter, roundness, and volume fraction of the solid particles at position F were 72 m, 1.45 and 36%,
respectively. At position E, they were 70 m, 1.4, and 35%, respectively. At position E, the solid phase was slightly smaller and more
globular, but the volume fraction was lower than at position F. The
results show that the microstructure of thin plate samples formed
at 300 mm/s and 30 mm/s were ner and more globular compared
with the microstructure of the semi-solid slurry with EMS.
When a semi-solid slurry was compressed with a punch, coarse
solid particles from coalescence broke off, and single solid particles were deformed. These particles and the liquid phase lled the
die cavity. Plastic deformation of the solid particles occurred by
forced convection during lling. A small and globular grain structure then formed. The moving velocity and distance of the solid
phase increase for smaller and more globular grains.
3.5. Mechanical properties of thin plate
Fig. 21 shows the tensile strength, elongation, and hardness values of the samples formed under the three conditions. The sample

Fig. 21. Mechanical properties of fabricated samples.

formed under condition 2 had a tensile strength and elongation

of 216 MPa and 10%, respectively, at position F and 209 MPa and
8.5%, respectively, at position E. This sample had Vickers hardness
values of 72 HV and 69 HV at positions F and E, respectively. The
sample formed under condition 3 had a tensile strength, elongation,
and Vickers hardness of 160 MPa, 4.5%, and 69.5 HV, respectively,
at position F and 148 MPa, 2.5%, and 68 HV, respectively, at position
E. These values were 56 MPa, 5.5%, and 2.5 HV lower at position F
and 61 MPa, 6%, and 1 HV lower at position E compared to the sample under condition 2. The sample formed under condition 1 (i.e.,
forming without graphite lubrication of cavity) had very low tensile
strength and elongation of 132 MPa and 2%, respectively, at position F and 126 MPa and 2%, respectively, at position E compared
to condition 2. However, the Vickers hardness was much higher
under condition 3 than condition 2 because of the large eutectic
phase which used to be the liquid phase. The Vickers hardness of
the solid phase was about 56 HV, and that of the liquid phase was
about 98 HV.
The tensile strength and elongation of the sample formed under
condition 2 at 300 mm/s were higher than those of the sample
formed under condition 3 at 30 mm/s. This was presumed to be
because the solid particles were slightly smaller and more densely
distributed. In particular, the elongation of the sample under conditions 1 and 3 were signicantly lower than that of the sample
under condition 2. This indicates that fewer solid particles in the
microstructure decrease the elongation. Under condition 3, even
though the solid particles in the microstructure were mostly globular, there was a large number of solid particles with a diameter
of over 100 m. This indicates that the number of solid particles
within the 1.2 mm thick thin plate was much smaller than that
under condition 2. For the microstructure under condition 1, solid
particles were rarely distributed and were concentrated as dendrite
structures in one place. With regard to the difference in mechanical properties at different positions, position F (close to the gate)
showed slightly improved properties compared to position E under
all three conditions, but the microstructures were similar at the two
locations.
Based on the analysis of the formability, mechanical properties,
and microstructure of the thin plates formed at different punch
speeds and with and without lubrication, condition 2 (with graphite
lubricant and a punch speed of 300 mm/s) was found to be the most
appropriate.
4. Conclusions

Fig. 20. Volume fraction, equivalent diameter, and roundness of solid particles in
samples formed at 300 and 30 mm/s punch speeds with graphite lubricant.

An indirect rheo-casting die for 1.2 mm thick thin plates was

designed through a lling simulation. Thin plates were formed

168

C.K. Jin et al. / Journal of Materials Processing Technology 224 (2015) 156168

under two friction conditions for the cavity and two punch speeds.
The effects of these parameters on the formability and mechanical
properties were analysed. The analytical and experimental results
are as follows:
(1) A fan-shaped gate allows the semi-solid slurry to evenly ll the
entire thin cavity.
(2) The increased ow rate along the cavitys sides can be controlled
by using overows located at the sides of the cavity.
(3) The placement of overows at the end of the cavity allows for
an even temperature and pressure (greater than atmospheric
pressure) distribution within the cavity.
(4) A semi-solid slurry having a solid fraction of 40% with ne
and globular solid particles can be obtained with an EMS. The
semi-solid without EMS had rosette particles and coarser globular -Al particles. The equivalent diameter and roundness
of the solid particles of the semi-solid slurry with EMS were
about 75 m and 1.5, respectively. Those of the solid particles of the semi-solid slurry with EMS were 110 m and 2.3,
respectively.
(5) The cavitys surface state experiences greater segregation
between the solid and liquid phases with increased friction.
When the friction of the cavity is high, solid particles nd it difcult to move into the cavity, so the liquid phase with relatively
good uidity ows into the cavity.
(6) A higher punch speed for compression of the semi-solid
slurry produces better formability for the thin plate and
a microstructure with densely and evenly distributed solid
particles.
(7) The microstructure of formed thin plate was ner and more
globular compared with the microstructure of the semi-solid
slurry with EMS. Plastic deformation of the solid particles
occurred by forced convection during lling. A small and globular grain structure then formed. The moving velocity and

distance of the solid phase increase for smaller and more globular grains.
(8) When the semi-solid slurry was formed at a punch speed of
300 mm/s with a graphite lubricant, a thin plate without segregation between the solid and liquid phases or other defects was
obtained with a tensile strength of 216 MPa and elongation of
10%.
Acknowledgements
This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government
(No. 2013R1A1A2062759). Ministry of knowledge Economy (No.
20104010100540), This study was also supported by the Korea
Research Foundation (KRF) grant funded by the Korea government
(MEST) (grant No. 2012-0001204).
References
Atkinson, H.V., 2005. Prog. Mater. Sci. 50, 341412.
Bae, J.W., Kim, T.W., Kang, C.G., 2007. J. Mater. Process. Technol. 191, 165169.
Chen, C.P., Tsao, C.-Y.A., 1997. Acta Mater. 45, 19551968.
Fan, Z., Fang, X., Ji, S., 2005. Mater. Sci. Eng., A 412, 298306.
Flemings, M.C., Riek, R.G., Young, K.P., 1976. Mater. Sci. Eng. 25, 103117.
Ghomashchi, M.R., Vikhrov, A., 2000. J. Mater. Process. Technol. 101, 19.
Ji, S., Fan, Z., Bevis, M.J., 2001. Mater. Sci. Eng., A 299, 210217.
Joly, P.A., Mehrabian, R., 1976. J. Mater. Sci. 11, 13931418.
Kang, C.G., Bae, J.W., Kim, B.M., 2007. J. Mater. Process. Technol. 187188, 344348.
Kang, C.G., Choi, J.S., Kim, K.H., 1999. J. Mater. Process. Technol. 88, 159168.
Kapranos, P., Ward, P.J., Atkinson, H.V., Kirkwood, D.H., 2000. Mater. Des. 21,
387394.
Kim, N.S., Kang, C.G., 2000. J. Mater. Process. Technol. 103, 237246.
Niu, X.P., Hu, B.H., Pinwill, I., Li, H., 2000. J. Mater. Process. Technol. 105, 119127.
Pinsky, D.A., Charreyronv, P.O., Flemings, M.C., 1984. Metall. Mater. Trans., B 15B,
173181.
Seo, P.K., Youn, S.W., Kang, C.G., 2002. J. Mater. Process. Technol. 130131, 551557.
Seo, P.K., Kim, D.U., Kang, C.G., 2007. J. Mater. Sci. Eng., A 445446, 2030.
Yue, T.M., Chadwick, G.A., 1996. J. Mater. Process Technol. 58, 302307.