You are on page 1of 9

Materials Science and Engineering A 435436 (2006) 521529

Mechanical and wear properties of rheocast and conventional


gravity die cast A356 alloy
A.K. Dey a , P. Poddar a , K.K. Singh b , K.L. Sahoo a,
a

National Metallurgical Laboratory, Jamshedpur 831007, India


National Institute of Foundry and Forge Technology, Ranchi 834003, India

Received 8 May 2006; received in revised form 10 July 2006; accepted 28 July 2006

Abstract
A356 alloy produced by means of conventional gravity die casting and rheocasting has been investigated and their microstructure, mechanical
and tribological properties were compared. The microstructure of conventional cast sample is fully dendritic in contrast to spheroidal morphology
in rheocast sample. The mechanical properties of the rheocast samples are considerably higher than the conventional cast samples. The volumetric
wear loss and coefficient of friction in rheocast samples are always less than those in conventional cast samples at all loads. The wear occurs mainly
by ploughing mechanism.
2006 Elsevier B.V. All rights reserved.
Keywords: Rheocasting; Conventional die casting; Microstructure; Mechanical property; Wear property

1. Introduction
It is reported [1,2] that semi-solid casting processes posses
many advantages over the conventional casting process. The
morphology of growing solidliquid interface in conventional
casting processes is typically dendritic. The conventional casting often contains internal structural defects such as oxide and
gas entrapment, shrinkage porosity that leads to poor mechanical properties. To achieve better properties, there is an increasing
trend to produce common Al-alloys (e.g., A356 and A357) automotive components by semi-solid processing route [3,4]. Two
casting technologies, namely, thixocasting and rheocasting, have
been developed for the production of metal components by
semi-solid processing route. Rheocasting process offers several
advantages than the thixocasting process including reduced process complexity, increases shot size flexibility and effective solid
fraction tailoring [5,6].
In the rheocasting process, molten alloy is cooled from the
liquid sate to the mushy state followed by stirring the alloy
during the solidification to produce semi-solid slurry, then pouring/injecting the slurry directly into the die. The formation
and the evolution of non-dendritic rheocast microstructure are

Corresponding author. Tel.: +91 657 2271709; fax: +91 657 2270527.
E-mail address: klsah@nmlindia.org (K.L. Sahoo).

0921-5093/$ see front matter 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.msea.2006.07.148

associated with the breakup of dendrite arms into small pieces


followed by agglomeration and sintering of these pieces to form
clusters [1,7]. However, the rheocasting process has been largely
explored to produce complex-shaped Al-alloy components [8].
For wide application of this process in various engineering fields,
the mechanical and wear behaviour of rheocast products need
to be well understood.
The objective of this study was to examine the mechanical
and wear behaviour of rheocast A356 alloy, which is a popular Al-alloy for automotive components. These properties were
compared with those of conventional gravity die cast counterpart.
2. Experimental
In this study, commercial A356 aluminium alloy was used.
The composition of the A356 alloy is given in Table 1. The
liquidus and the solidus temperature of the alloy were found to
be 615 C and 538.5 C, respectively.
The melting of the alloy was carried out in an electric resistance furnace in a clay bonded graphite crucible coated with
alumina paints. After melting sufficient time was given for complete homogenisation of the melt. The melt was then degassed
with dry argon.
In the rheocasting process, the melt was continuously cooled
and stirred under a constant rotational speed of 320 rpm using a

522

A.K. Dey et al. / Materials Science and Engineering A 435436 (2006) 521529

Table 1
Composition of A356 alloy used for the present investigation
Element

wt.%

Si

Mg

Mn

Fe

Ni

Ti

Zn

Sr

Al

7.22

0.45

0.01

0.15

0.016

0.13

0.04

0.01

Balance

mechanical stirrer type rheocaster with bottom pouring arrangement. The average cooling rate during stirring was about
4 min1 . When the temperature was reached about 580 C, the
rheocast slurry was poured in a cast iron rectangular type mould
of dimension 20 mm 100 mm 300 mm.
Conventional gravity die casting, poured from 30 C above
the liquidus temperature, was also made in the same mould, for
comparison purposes. The samples were cut from the middle
of the rectangular plate for structural and mechanical properties
investigations. The polished samples were etched with modified
Kellers reagent (2 cm3 HF, 3 cm3 HCl, in 175 cm3 H2 O) for
microstructure examination.
The plate type castings of both conventional and rheocast
samples were cut into thickness of approximately 10 mm for hot
(300 C) rolling. The rolling was carried out in a four high mill
with reverse rolling facilities. Tensile specimen from as cast and
rolled sheet were prepared as per sketch given elsewhere [9].
The specimens were tested at a strain rate of 8.3 104 s1 .
The fractured surfaces of the specimen were examined under
scanning electron microscopy (SEM). The Vickers hardness of
conventional and rheocast samples were taken using a load of
5.0 kg.
The sliding wear tests were carried out using a pin-on-disc
machine. The pins of 8 mm diameter and 40 mm length were
fabricated from 20 mm plate and made to slide against a low
alloy steel disc (material: 103 Cri-Eng-31HRS60W61, equivalent to AISI 4340) of diameter 215 mm and hardness 62 Rc. The
track radius and the disc speed were maintained at 55 mm and
350 rpm, respectively, to maintain at constant sliding velocity of
2.0 m/s. Three loads, namely, 19.6 N, 29.4 N and 49.0 N were
applied for each test materials. Tangential force and hence, the
coefficient of friction were measured continuously with an electronic sensor attached to the machine and recorded. Frictional
force in kg and cumulative wear loss in m were measured from
the sensor output as a function of time. The wear test was carried
out for a total sliding distance of about 2.4 km. The specific wear
of the pins, defined as the cumulative volumetric wear loss suffered by the pin per unit sliding distance per unit load [10,11],
were calculated from the cumulative wear data. Wear tests were
carried out at room temperature without any lubrication. After
the completion of wear test, the Vickers hardness of the worn
surfaces of each pin was measured under a load of 5 kg. The
worn surfaces of pin samples were examined under SEM to
understand the wear mechanism.

fied in the cast iron plate mould (average cooling rate 10 C/min).
As cast microstructure of both castings show that phases are
uniformly distributed. The figures clearly show a morphological change in the microstructures. In conventional cast sample,
the microstructure is fully dendritic whereas in rheocast sample, the primary dendrites are fragmented due to mechanical
stirring. Fig. 1b clearly shows that some grains are nearly spherical and some are agglomerated together to form a bigger grain.
The primary Al dendrites are plastically deformed during rheocasting processing. However, with the continued stirring, the
plastic strains within the fragmented grains would be considerably less and process of coarsening or ripening will start. Since
the coarsening/ripening is driven by interfacial energy [12,13],
the process will lead to a reduction in the surface area and eventually spheroidal morphology is obtained. It was observed that

3. Results and discussion


Fig. 1a and b shows the optical microstructure of as cast
ingots (both conventional gravity die cast and rheocast) solidi-

Fig. 1. Optical microscopy of as cast: (a) conventional cast and (b) rheocast
A356 samples.

A.K. Dey et al. / Materials Science and Engineering A 435436 (2006) 521529

523

Fig. 2. Optical microstructure of 70% reduced sample: (a) conventional cast and
(b) rheocast.

not only the primary Al grains (white portions in Fig. 1) but


also eutectic Si particles (black portions) are globular in rheocast samples in comparison to conventional cast samples. As the
structure contains good amount of eutectic phases it should give
a range of mechanical properties when thermo-mechanically
processed. The plate castings are subsequently processed by
rolling in order to examine its improvement in microstructure
and thereby, any escalation in mechanical properties. Fig. 2
shows optical microstructures of rolled products (70% reduction) for both the samples. It was observed that in both cases
primary -Al grains are elongated in the direction of rolling.
The Al grains are softer than the eutectic Si phases.
Fig. 3a and b shows the ultimate tensile strength (UTS) of
samples in as cast and in rolled condition (70% reduction).
The rheocast sample shows better strength and elongation than
the conventional cast samples. Table 2 shows the 0.2% proof

Fig. 3. Comparative plots of stressstrain: (a) as cast and (b) rolled samples
(70% reduction).

stress (YS), UTS and percent elongation (f ) of both the samples. Rheocasting shows better strength and elongation than
the conventional cast samples because less shrinkage and associated porosity are expected in the rheocastings as pouring
occurs at a temperature below the liquidus. The globular primary
solid structure in the mushy rheocast material would be more
favourable to liquid penetration for feeding, in comparison to the
complex feeding through dendritic structure in conventional process [14]. The globular shaped Si particles in the rheocast sample
may effectively reduce the stress concentration at the boundary

Table 2
Mechanical properties of conventional cast and rheocast samples
Sample and condition

YS (MPa)

UTS (MPa)

Elongation (%)

Hardness (VPN)

Conventional, as cast
Rheocast, as cast
Conventional, 70% reduction
Rheocast, 70% reduction

79.0
117.0
179.0
217.0

143.0
203.5
196.5
232.2

4.12
6.30
2.20
2.60

83.3
88.6
87.2
94.0

524

A.K. Dey et al. / Materials Science and Engineering A 435436 (2006) 521529

Fig. 4. SEM microstructure showing fractured surface of tensile specimen: (a)


conventional cast and (b) rheocast.

between Si particles and matrix under an applied stress. Therefore, the globular shape would improve the tensile strength as
well as ductility in the rheocast products. The Vickers hardness
of the rheocast samples at 5.0 kg load is also higher than the
conventional cast sample (Table 2). The samples in the form of
plates of thickness 10 mm are homogenised at 400 C for 2 h
for rolling. The object of hot rolling was to fragment the larger
particles and to disperse them in the matrix as well as to seal
the micro-porosity/voids formed due to interdendritic shrinkage.
The rolling operation was carried out in several steps in order to
bring down the reduction level of 70%. A considerable improvement in strength was observed for both the samples after rolling.
Fig. 4a and b shows the SEM microscopy of the fractured surfaces for both the samples. In both samples, the tensile fracture
paths tended to follow the primary -phase boundaries. Also
the shearing of primary -phase is observed in the conventional
cast samples. From the fractographic view, it is also evident that
the tensile fracture occurred by dimpled rupture with the void
initiation at the eutectic Si particles for both samples.

Fig. 5. The variation of cumulative wear loss vs. sliding distance of conventional
and rheocast A356 alloy under the load of: (a) 19.6 N, (b) 29.4 N and (c) 49 N.

The plate castings of 20 mm thickness were machined down


to 8 mm diameter for wear behaviour study by pin-on-disc tests.
The microstructure of both samples show that phases are uniformly distributed in Al matrix. No porosity was observed by
optical microscope. Fig. 5ac represents the cumulative wear
loss as a function of sliding distance at different loads, namely,

A.K. Dey et al. / Materials Science and Engineering A 435436 (2006) 521529

525

19.6 N, 29.4 N and 49.0 N, respectively. After a transient period,


the wear loss was found to increase linearly with increase in
sliding distance. The volumetric wear loss also increases rapidly
with increasing load. The wear loss in case of conventional cast
samples is distinctly higher than that of rheocast alloys at all
loads. The wear test at the highest load (49.0 N) in conventional
cast alloy could not be carried out for longer duration due to
high rate of wear. It suffered a wear loss of above 2 mm in travelling around 1 km distance. The sudden jump in wear rate in the

Fig. 7. The variation of cumulative wear loss vs. sliding distance of conventional
and rheocast A356 alloy under the load of 19.6 N.

above sample indicates a change in wear mechanism from mild


to severe mode. The specific wear (understood as the wear loss
divided by the normal force and the sliding distance) at room
temperature is plotted in Fig. 6 as a function of sliding distance
at different loads. The specific wear rate of rheocast sample is
significantly lower as compared to that of conventional cast samples due to combined influence of lower porosity and globular
nature of grains. The wear rates at 19.6 N and 29.4 N loads are
observed to be nearly constant or slightly increase after a sliding distance of 1300 m for both samples whereas at the applied
load of 49.0 N, the wear rate of rheocasting sample decreases
and that of conventional cast sample rapidly increases. The high
wear rate at the initial stage (Fig. 6) is due to the adhesive nature
of wear which is associated with high material loss and high
value of coefficient of friction (). The representative graph for
is presented in Fig. 7 for both the alloys at a load of 19.6 N.
The average value of at different loads is shown in Fig. 8. The

Fig. 6. Specific wear of conventional cast and rheocast alloys as a function of


sliding distance at a load of: (a) 19.6 N, (b) 29.4 N and (c) 49.0 N.

Fig. 8. Histogram showing the average coefficient of friction at different loads.

526

A.K. Dey et al. / Materials Science and Engineering A 435436 (2006) 521529

Fig. 9. Hardness of as cast and worn surfaces of the samples at a load of 5 kg.

coefficient of friction of rheocast samples is always less than


that of conventional cast samples. The variation of coefficient
of friction from its mean values in case of conventional cast
samples is comparatively larger than that of rheocast samples
because during wear run the large particles piled up and forged
together (shown later). This gives some obstruction to wear run.
The hardness (VPN) of the worn surfaces was determined for
both the alloys at three different wear loads (Fig. 9). The hardness of the worn surfaces of the pin materials increased with
increasing load. This is due to work hardening of the materials.
Thus, extent of work hardening is more with increase in load.
However, the hardness of the worn surfaces of the rheocast alloy
is always higher than that of conventional cast alloy. From the
above results and discussion, it can be said that casting method
has a great influence on the properties of the alloys. The rheocasting offers better mechanical properties and wear resistance
than the conventional one.
The SEM micrograph of the worn surfaces of the samples has
depicted some feature to understand the wear mechanism. Low
magnification micrographs of the worn surfaces of the pin materials at 19.6 N and 49.0 N load are presented in Figs. 10 and 11,
respectively. The higher magnification micrographs at 49.0 N
load are shown in Fig. 12.
The worn surfaces of the pin materials show diverse topographical features. Most of the worn surfaces consist of smooth
strips, the surfaces of which are characterized by fine scoring
marks. These smooth strips were extend uninterrupted from one
end of the pin specimen to the other. The worn surfaces of
the conventional cast samples were more heavily scored than
the rheocast counterparts. Scoring may be due to abrasion by
entrapped debris, work-hardened deposits on the counterface or
hard asperities on the hardened steel counterface [1518]. The
wear by scoring mechanism, however, may not lead to a large
volume of wear since the amount of material removed from a
fine groove is small. The worn surfaces of the conventional cast
alloy at 19.6 N load (Fig. 10a) also shows extensive and long continuous grooves on the surfaces parallel to the sliding direction

Fig. 10. Low magnification SEM micrograph of worn surface at 19.6 N load:
(a) conventional cast and (b) rheocast.

whereas the rheocast alloy at the same condition shows much


lesser grooves. The width of the continuous grooves in conventional cast samples is larger than that of the grooves on rheocast
samples. As the load increases, the width of the grooves enlarges
significantly. Irregular plastic flow lines can be seen (Fig. 12)
indicating the occurrence of extensive plastic deformation during wear run. Cracks and strips of roughened surfaces are also
seen in the same micrograph. The cracks propagate in the sliding direction (Fig. 11a). A number of cracks inclined to the
sliding direction are shown in Fig. 12a. Crack nucleation, generally, occurs at some depth below the surface rather than very
near to the surface, owing to high hydrostatic compressive pressure acting near the asperity contact [19]. Cracks may initiate

A.K. Dey et al. / Materials Science and Engineering A 435436 (2006) 521529

Fig. 11. Low magnification SEM micrograph of worn surface at 49.0 N load:
(a) conventional cast and (b) rheocast.

in the highly work-hardened layer, particularly in the subsurface region. When cracks grow and get interconnected, a layer
of metal is removed leaving a crater. This is delamination wear.
Fig. 12a shows such evidence. At the highest load, the mixed
mode of wear, i.e., ploughing with local delamination wear were
observed.
During wear run the worn particle debris agglomerated and
stick to the grooved surfaces. This is very prominent in case
of a conventional cast alloy. Fig. 13 shows such areas. During
the wear run, with increasing load the frictional heat generated
is quite appreciable. The localized heating actually facilitates
sticking of the pin to the disc surface. The interface temperature
increases as the wearing couples slide against each other that
facilitate the sticking of wear debris at the grooved surfaces

527

Fig. 12. Higher magnification SEM micrograph of worn surfaces at 49.0 N load:
(a) conventional cast and (b) rheocast.

[20]. The increase of friction coefficient at elevated temperature


has also been reported particle-reinforced cast alloy composites
[2123]. The rise in friction coefficient with increase in load
may be attributed to enhanced accumulation of wear debris at
the pin and disc surface.
Archard [24] defined adhesive wear volume loss as a function of sliding speed, normal load and materials hardness. This
was based on a mechanism of adhesion at the asperities and the
materials removal process was related to the cohesive failure of
asperities. The initial microstructure, the process of crack nucleation and the subsequent growth were ignored in this theory.
With the assumption of hemispherical wear particles of same
radius as contact area, Archard [24] developed an equation of

528

A.K. Dey et al. / Materials Science and Engineering A 435436 (2006) 521529

Fig. 14. Correlation between experimental and theoretical values of wear loss:
(a) conventional cast and (b) rheocast alloy.

between hardness and the wear rate were not always observed
in sliding wear system.
The experimental wear data obtained in the present study
is used to develop a generalized equation (a modified version
of Archards equation) to estimate wear loss of materials. The
relation proposed is
W = b0 Lb1 P b2 H b3
Fig. 13. SEM micrograph showing the particle agglomeration on the wear track:
(a) conventional cast and (b) rheocast.

wear rate as:


W=

kLP
3H

(1)

where W is volume of materials worn out, k the wear coefficient,


L the sliding distance, P the applied load and H is bulk hardness of the materials. The proposed proportionality between the
applied load and the wear rate, and the inverse proportionality

(2)

where W is the volume wear loss (mm3 ), b0 the wear coefficient,


L the sliding distance (m), P the applied load (N), H the bulk
hardness (BHN) and b1 , b2 and b3 are the fitting coefficients.
These values for both the alloys are tabulated in Table 3. The
correlation (R) for the fitting is very good (0.9937). The values
of wear loss predicted by using Eq. (2) are compared with experimental data for both conventional cast and rheocast alloys as
shown in Fig. 14a and b, respectively. The graphs indicate a good
general agreement between predicted and experimental data
with an error range of 8%. The experimental data obtained can
be fitted to an exponential equation W = 0.49L0.6 P1.2 H1.3 for

Table 3
Results of fitting data for the conventional cast and rheocast alloys
Materials

Conventional cast
Rheocast

Coefficients
b0

b1

b2

b3

0.49
0.49

0.6
0.9

1.29
0.8

1.35
1.4

R2

Final loss (observed predicted)2

0.98
0.98

146
96

A.K. Dey et al. / Materials Science and Engineering A 435436 (2006) 521529

conventional cast alloy and W = 0.49L0.9 P0.8 H1.4 for rheocast


alloy with good correlation and high coefficient of determination. As expected, the exponent of hardness is negative. All the
exponents deviate from unity. The reason for this deviation is the
ignorance of the initial microstructure. The variation of load and
hardness exponents clearly indicates that the initial microstructure plays an important role to the wear behaviour. The load
exponent deviates much from unity in conventional cast samples because wide variation of rate wear with increasing load.
4. Conclusions
The microstructure of conventional cast samples is fully dendritic whereas in rheocast samples the primary phases are of
nearly spheroidal morphology.
The rheocast alloy showed significant improvements in
mechanical properties. After hot rolling, it shows further
increase in strength mainly due to plastic deformation and work
hardening.
Rheocasting offers superior wear resistance compared to
conventional castings. The wear loss increases with increasing
load in all samples. Wear occurs by ploughing mechanism. At
the highest load, in conventional cast samples the wear mode
changes to mixed mode of ploughing and delamination wear.
Coefficient of friction of the rheocast samples is lesser than the
conventional cast samples at all loads. The hardness of the worn
surfaces increases with load indicating work hardening of the
materials.
The experimental data obtained can be fitted to an exponential
equation W = 0.49L0.6 P1.2 H1.3 for conventional cast alloy and
W = 0.49L0.9 P0.8 H1.4 for rheocast alloy with good correlation.

529

References
[1] M.C. Flemings, Metall. Trans. A 22 (1991) 957.
[2] K.P. Young, R. Fitze, Proceedings of the Third International Conference
on Semi-Solid Processing of Alloy and Compounds, Tokyo, Japan, 1994,
p. 155.
[3] K.P. Young, Light Metals, The Minerals, Metals and Materials Society,
Warrendale, PA, 1996, p. 775.
[4] H.V. Atkinson, Prog. Mater. Sci. 50 (2005) 341.
[5] Z. Fan, Int. Mater. Rev. 47 (2002) 1.
[6] T. Haga, P. Kapranos, J. Mater. Proc. Technol. 130131 (2002) 594.
[7] R.D. Doherty, H.I. Lee, E.A. Feest, Mater. Sci. Eng. A 65 (1984)
181.
[8] C. Park, S. Kim, Y. Kwon, Y. Lee, J. Lee, Mater. Sci. Eng. A 391 (2005)
86.
[9] K.L. Sahoo, C.S.S. Krishnan, A.K. Chakrabarti, J. Mater. Proc. Technol.
112 (2001) 6.
[10] K.L. Sahoo, C.S.S. Krishnan, A.K. Chakrabarti, Wear 239 (2000) 211.
[11] M.A. Martinez, A. Martin, J. Llorca, Scr. Metall. Mater. 28 (1993)
207.
[12] R.D. Doherty, H.I. Lee, E.A. Feest, Mater. Sci. Eng. A 65 (1984) 181.
[13] N. Apaydin, K.V. Prabhakar, R.D. Doherty, Mater. Sci. Eng. A 46 (1980)
145.
[14] D. Brabazon, D.J. Browne, A.J. Carr, Mater. Sci. Eng. A 326 (2002)
370.
[15] B.N. Pramila Bai, S.K. Biswas, Scr. Metall. 39 (1991) 833.
[16] B.N. Pramila Bai, S.K. Biswas, Wear 87 (1983) 237.
[17] A.S. Reddy, B.N. Pramila Bai, K.S.S. Murthy, S.K. Biswas, Wear 181183
(1995) 658.
[18] S. Wilson, A.T. Alpas, Wear 225229 (1999) 440.
[19] S. Jahanmir, N.P. Suh, Wear 44 (1977) 17.
[20] R.T. Spurr, Wear 55 (1979) 289.
[21] C.-C. Yang, W.-M. Hsu, E. Chang, Mater. Sci. Technol. 13 (1997) 687.
[22] M. Roy, B. Venkataraman, V.V. Bhanuprasad, Y.R. Mahajan, G. Sundararajan, Metall. Mater. Trans. A 23 (1992) 2833.
[23] D. Mandal, B.K. Dutta, S.C. Panigrahi, Wear 257 (2004) 654.
[24] J.F. Archard, J. Appl. Phys. 24 (1953) 981.