See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/297678196

Influence of Process Parameters and Sr Addition on the Microstructure and

Casting Defects of LPDC A356 Alloy for Engine Blocks

Article in Journal of Materials Science and Technology -Shenyang- · March 2016

DOI: 10.1016/j.jmst.2016.03.010

CITATIONS READS

42 1,426

3 authors:

Giulio Timelli Daniele Caliari

University of Padova University of Padova
163 PUBLICATIONS 1,546 CITATIONS 23 PUBLICATIONS 78 CITATIONS

SEE PROFILE SEE PROFILE

Jovid Rakhmonov
Northwestern University
24 PUBLICATIONS 201 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

High Tech Die Casting 2020 Conference View project

Fluxing and de-coating treatments in aluminium recycling View project

All content following this page was uploaded by Giulio Timelli on 31 January 2019.

The user has requested enhancement of the downloaded file.

 Journal of Materials Science & Technology 32 (2016) 515–523

Contents lists available at ScienceDirect

Journal of Materials Science & Technology

j o u r n a l h o m e p a g e : w w w. j m s t . o r g

Influence of Process Parameters and Sr Addition on the

Microstructure and Casting Defects of LPDC A356 Alloy for Engine
Blocks
Giulio Timelli *, Daniele Caliari, Jovid Rakhmonov
Department of Management and Engineering, University of Padova, Stradella S. Nicola, 3 I-36100 Vicenza, Italy

A R T I C L E I N F O
The effects of Sr addition and pressure increase on the microstructure and casting defects of a low-pressure
Article history: die cast (LPDC) AlSi7Mg0.3 alloy have been studied. Metallographic and image analysis techniques have
Received 25 September 2015 been used to quantitatively examine the microstructural changes and the amount of porosity occurring at
Received in revised form different Sr levels and pressure parameters. The results indicate that an increase in the filling pressure induces
1 December 2015 lower heat dissipation of the liquid close to the die/core surfaces, with the formation of slightly greater
Accepted 4 December 2015 dendrite arms and coarser eutectic Si particles. On the other hand, the increase in the Sr level leads to finer
Available online 8 March 2016 microstructural scale and eutectic Si. The analysed variables, within the experimental conditions, do not
affect the morphology of eutectic Si particles. Higher applied pressure and Sr content generate castings
Key words: with lower amount of porosity. However, as the filling pressure increases the flow of metal inside the die
Aluminium alloys
cavity is more turbulent, leading to the formation of oxide films and cold shots. In the analysed range of
Engine block
experimental conditions, the design of experiment methodology and the analysis of variance have been
Microstructure
Casting defects
used to develop statistical models that accurately predict the average size of secondary dendrite arm spacing
Sr addition and the amount of porosity in the low-pressure die cast AlSi7Mg0.3 alloy.
Process parameters Copyright © 2016, The editorial office of Journal of Materials Science & Technology. Published by
Elsevier Limited.

1. Introduction of lower strength of Al alloys compared to cast iron. Today, Al alloys

are commonly used for gasoline engine blocks and their applica-
The engine block is the largest and most intricate single com- tion is strongly growing for the production of diesel engine blocks[1,2].
ponent used in an internal combustion engine. It affects 3%–4% of Base materials commonly used for engine blocks are hypoeu-
the total weight of the average vehicle. Thus it plays a key role in tectic Al–Si foundry alloys, which include AlSi8Cu3 and AlSi6Cu4
all weight-reduction considerations. As a consequence of the current alloys. These alloys are mostly applied for engine blocks produced
engine evolution trend in overall new vehicles[1], cylinder blocks must with gravity casting processes. In contrast, almost all high-pressure
permit increased power and reduced displacement, which is almost die cast engine blocks are produced with the wide diffused
entirely due to the shift towards engines with fewer cylinders; future AlSi9Cu3(Fe) alloy.
expected powers are up to 65 kW/l for direct injection diesel engines Engine blocks cast from the AlSi7Mg0.3 alloy achieve very high
and 75 kW/l for boosted gasoline direct injection engines[2]. strength and ductility values at room temperature after a T6 heat
In order to resist to high combustion pressure, the develop- treatment. Higher resistance to cracking in the plastic regime offered
ment of high strength materials and innovative processes for the by this alloy enables engine blocks to operate under severe thermal
production of cylinder blocks is demanded[3]. fatigue loading conditions[4]. In order to satisfy the increasing product
Aluminium alloys are increasingly used to produce cylinder blocks requirements, a proper control of the as-cast microstructure by the
due to the favourable ratio between mechanical strength and low application of preliminary processes, such as eutectic addition, is
density, and good thermal conductivity. Aluminium casting alloys generally necessary[5]. It is well known how the addition of certain
as a substitute for the traditional cast iron can mean a reduction elements, such as sodium and strontium, completely changes the
in engine block weight in the range between 40% and 55%, in spite morphology of eutectic Si crystals from faceted acicular flakes to a
fibrous rod-like form, thereby improving mechanical properties, es-
pecially fracture toughness and elongation[6].
* Corresponding author. Ph.D.; Tel.: +39 444 998769; Fax: +39 444 998889. Many foundry processes can be potentially used for the pro-
E-mail address: [email protected] (G. Timelli). duction of engine blocks. However, low-pressure die casting (LPDC)

http://dx.doi.org/10.1016/j.jmst.2016.03.010
1005-0302/Copyright © 2016, The editorial office of Journal of Materials Science & Technology. Published by Elsevier Limited.
516 G. Timelli et al. / Journal of Materials Science & Technology 32 (2016) 515–523

Table 1
Chemical composition of the experimental alloys (wt%); Alloy 0 corresponds to the supplied AlSi7Mg0.3 alloy

Alloy Si Fe Cu Mn Mg Zn Cr Ni Ti B Sr Al

0 6.96 0.111 0.016 0.016 0.301 0.01 0.001 0.003 0.112 0.0016 0.0003 Bal.
1 6.95 0.105 0.015 0.015 0.430 0.01 0.001 0.003 0.187 0.0154 0.0084 Bal.
2 6.89 0.110 0.017 0.016 0.432 0.01 0.001 0.003 0.179 0.0156 0.0166 Bal.

is one of the most promising processes. This foundry technology values in the range of 2.63–2.65 g/cm3 were measured during the
is characterized by controlled filling of die cavity and solidifica- experiments.
tion under pressure, with the solidification front moving from the The analysed component is an in-line 4-cylinder diesel engine
most distant point of the casting to the ingate. Therefore, LPDC is block (Fig. 1), which was produced by LPDC technology using ex-
an enhanced process that is generally used for parts with premium perimental alloys 1 and 2, respectively. The casting was multiple-
requirements[7]. The principle of this process is based on Pascal’s gated from the side of the crankshaft support to ensure good filling
pressure theory. The permanent die and the filling system are placed and proper feeding during solidification. A steel gauze filter was
over a sealed holding furnace containing the molten metal. To fill placed before the gating system after each casting. The weight of
the die, gas pressure is applied over the metal bath and the metal the machined Al alloy casting was about 21.1 kg.
is then forced to rise through one or more riser tubes, runners and The casting process was cyclic and began with the pressuriza-
gating system, and consequently feeds the die cavity. The pres- tion of the furnace. The excess pressure in the holding furnace forced
sure is then maintained to pressurize the casting and improve feeding the molten metal to fill the die cavity. The casting cycle was divided
during late-stage solidification. Once the casting is completely so- into the following steps: (1) Filling the riser tube in 10 s up to a pres-
lidified, the overpressure in the furnace is removed, and the residual sure of 0.02 MPa; (2) Filling the die cavity up to two different final
molten metal in the tube flows back down to the furnace by the pressures (also called holding pressure), 0.035 and 0.05 MPa, re-
action of gravity. spectively, at a constant pressure ramping time (35 s); thus, the
Many process parameters may affect the casting quality in LPDC, pressurizing speed (P/t) was set up at 429 and 857 Pa/s, respec-
such as exerting pressure velocity, holding pressure, pressure holding tively. A rapid increase of the pressure up to the final value at
time, melt temperature, die temperature[8–11]. A key process pa- constant pressure ramping time allows obtaining higher metal ve-
rameter during LPDC is the setting up of the exerting pressure in locity at the gate vg according to the following equation[12]:
the crucible that allows controlling the filling time of the die cavity
and to ensure a laminar flow of molten metal through the feed tube dP (t ) ⎛ A ⎞A
= ρg ⎜ 1 + c ⎟ t v g (1)
into the die. If the filling speed is not adequately determined and dt ⎝ Af ⎠ Ac
adjusted, the quality of the die cast part will be rather poor.
In this study, the effects of Sr addition and pressure increase on where P(t) is the pressure as function of time t, ρ is the density of
the microstructure and casting defects of a LPDC A356 alloy have molten metal, g is the gravitational constant, A c is the cross-
section area of the cavity, Af is the cross-section area of the crucible
been investigated. A statistical approach based on the analysis of
and At is the cross-section area of the riser tube. Higher metal flow
variance has been implemented to develop functional equations
velocity allows decreasing the filling time, which is defined as V/(Ag
useful to estimate the scale of secondary dendrite arm spacing and
vg), where V and Ag are the volume of die cavity and the gate area,
the amount of porosity as function of the initial Sr content in the
respectively; (3) Holding the final pressure for 160 s to assist the
alloy and the pressure parameters used during LPDC.
feeding of solidification shrinkage in the die cavity. Therefore,
the material solidifies under the holding pressure; (4) Discharging
2. Experimental

2.1. Alloys and casting parameters

In the present work, an AlSi7Mg0.3 cast alloy (EN AB-42100,

equivalent to the US designation A356) was supplied as commer-
cial alloy ingots and used as a baseline. The ingots were pre-
heated at 150 ± 10 °C in a tower furnace with a capacity of 50 t, and
subsequently brought to the melting temperature at 780 ± 10 °C. The
chemical composition of the supplied alloy (Alloy 0), measured on
separately poured samples by optical emission spectrometry, is
shown in Table 1.
The molten metal was tapped from the furnace and N-degassed
for 8 min with a rotary degasser (Foseco MTS Rotor Ø190, at
600 rpm). During degassing, weighed AlMg20 waffle ingots and
AlTi5B1 rods were added into the base alloy to ensure that the Mg
level reached the desired content about 0.45 wt%, and the Ti and
B levels were nominally 0.2 wt% Ti and 0.018 wt% B. Furthermore,
Sr addition was obtained by introducing AlSr10 rods in the melt en-
suring the Sr content reached two different levels, nominally 90 and
170 ppm (Alloy 1 and 2, respectively). The chemical compositions
of the experimental alloys are shown in Table 1. The liquid metal
was then transferred to an electrical holding furnace inside a
LPDC unit and held at 715 ± 5 °C. A reduced-pressure test was per-
formed to evaluate the molten quality in the holding furnace. Density Fig. 1. CAD section of the in-line 4-cylinder diesel engine block analysed.
 G. Timelli et al. / Journal of Materials Science & Technology 32 (2016) 515–523 517

Table 2
Independent variables selected in the DOE to study the effect on SDAS

Parameter Variable Lower Higher

designation level level

A Sr (wt%) 0.009 0.017

B Pressurizing speed (Pa/s) 429 857

has a significant impact on the process output. An analysis of vari-

ance (ANOVA) should be performed by considering the independent
factors (inputs) and their interaction that influence the response
(output). The ANOVA is a statistical methodology that enables to
investigate and to model the relationship between the output and
one or more input variables[13].
In the present study, a factorial ANOVA approach was imple-
mented to investigate the effects of the Sr level and the pressurizing
Fig. 2. Locations where the specimens for metallographic investigations were drawn. speed on the microstructural scale (SDAS), additionally the influ-
ence of the Sr level and the holding pressure on the amount of
microporosity.
the pressure of the furnace in 5 s; (5) After the complete solidifi- Each independent variable was analysed at two levels. The in-
cation, the side dies were opened, the top die was raised vertically dependent variables, along with their values at selected levels used
and the casting was ejected (40 s); (6) The die was blown (30 s), in the statistical approach, are given in Tables 2 and 3. With the aim
the steel gauze filter and the sand cores were placed appropriate- of taking into account the highest degree of interaction, a full bal-
ly (95 s); (7) The die was closed and the cycle was restarted. anced factorial plan was implemented. Besides the Sr content and
Typical cycle time was about 460 s. During solidification, circu- the pressure parameters as independent variables, their interac-
lating water at room temperature through internal channels in the tion was also considered and studied.
dies was used to control the cooling rates of the casting and to sta-
bilize the temperature of the dies in the range of (300–450) ± 5 °C. 4. Results and Discussion
The dies were made of AISI H11 tool steel.
4.1. Microstructural investigations
2.2. Metallographic characterization
The following discusses the microstructural investigations in-
In order to evaluate the combined effect of pressure increase and cluded in the present study.
Sr addition on the microstructure and casting defects, the cylin-
der blocks were sectioned transversally along the in-line cylinders. 4.1.1. General microstructure
In general, the cylinder bridges show relatively thin thickness and Typical microstructures of the analysed region drawn from the
are thermally and mechanically stressed regions. Therefore, the eval- cylinder block are shown in Fig. 3, which refer to AlSi7Mg0.3 alloy
uation of microstructure is mandatory, as required by most of the after 170 ppm Sr addition and obtained with different pressure
customers’ specifications[2]. Samples from the region indicated in ramps. In general, the microstructure consists of a primary phase,
Fig. 2 were drawn for metallographic investigations from three dif- α-Al solid solution, and an eutectic mixture of aluminium and silicon.
ferent castings, mechanically prepared to a 3 μm finish with diamond The α-Al precipitates from the liquid as the primary phase in the
paste and, finally, polished with a commercial fine silica slurry form of dendrites. Intermetallic compounds, such as Fe- and Mg-
(Struers OP-S). rich intermetallics, are also observed. The Fe-bearing particles are
To quantify the microstructural features, the image analysis was present in the microstructure mainly in the form of needle-
focused on the secondary dendrite arm spacing (SDAS) measured shaped β-Al5FeSi, even if some Chinese script α-Al15(Fe,Mn)3Si2
by the line intercept method, and on the size and aspect ratio of particles were detected. A high fraction of Mg-rich structures was
the eutectic Si particles. Size is defined as the equivalent circle di- detected in the form of Chinese script Mg2Si and π-Al8Si6Mg3Fe as
ameter (d), and the aspect ratio (α) is the ratio of the maximum to products of binary eutectic reactions. In addition to these binary
the minimum Ferets. To obtain a statistical average of the distri- eutectic structures, some complex eutectic clusters were also ob-
bution, a series of at least 15 micrographs of each specimen was served on the polished sections, as formed from the ternary L →
taken, and each measurement included more than 1000 particles. α-Al + Mg 2 Si + π-Al 8 Si 6 Mg 3 Fe and quaternary L → α-Al + π-
The secondary phases, such as Mg2Si particles, were excluded from Al8Si6Mg3Fe + Mg2Si + Si eutectic reactions, respectively.
the measurements.
Porosity was also studied in the same region, which corre- 4.1.2. Effect of pressure increase on SDAS
sponds to a hot spot in the casting. To obtain a statistical average, The scale of the microstructure was evaluated by means of SDAS
a series of 20 micrographs of each specimen were taken covering measurements (see Table 4). Considering the relatively low heat ex-
an area of about 6.6 mm × 2 mm. Quantitative metallographic tech- traction in the analysed region which is surrounded by sand cores,
niques were also used to determine average pore area and roundness.
The latter is defined as p2/(4πA), where p and A are perimeter and
area of the pore, respectively. Table 3
Independent variables selected in the DOE to study the effect on microporosity

3. Design of Experimental Matrix and ANOVA Parameter Variable Lower Higher

designation level level

The design of experiments (DOEs) is a statistical approach to the A Sr (wt%) 0.009 0.017
B Holding pressure (MPa) 0.035 0.050
experimental investigation that allows analysing which process input
518 G. Timelli et al. / Journal of Materials Science & Technology 32 (2016) 515–523

point of metals and alloys. This relationship has been described with
Clausius–Clapeyron equation[17]. For Al–Si binary alloys, the calcu-
lated increase in melting temperature has been estimated to be about
0.06 °C/MPa[15]. Therefore, when pressure is applied to a liquid alloy
at temperature just above its melting temperature, some degree of
undercooling results, thus providing a greater amount of nuclei.
Higher cooling rate, especially if coupled with a prompt undercool-
ing, can cause significant refinement in the microstructure.
In the present work, the increase in the melting temperature has
been calculated to be 0.0021 and 0.003 °C for the holding pres-
sures of 0.035 and 0.05 MPa, respectively, and therefore it is
negligible. On the other hand, the highest pressure used here is defi-
nitely low, if compared to that applied in squeeze casting, for
increasing the cooling rate by improving the contact between the
casting and the sand cores placed in the die cavity.
The rapid increase of the pressure up to the final value at con-
stant pressure ramping time (35 s) allows obtaining rapid filling
speed of the metal inside the die cavity and thus lower filling time.
This results in lower heat dissipation of the liquid close to the die/
core surfaces. Therefore, the local cooling rate is reduced, which
means that the local solidification time increases. The slower cooling
rates obtained with steeper pressure ramp lead to significant coars-
ening in the microstructure if compared with that obtained by
smoother pressure ramp.
Decreasing the filling time allows reducing the dissipation of melt
superheat. It is reported how the grain size and the SDAS coarsen
with increasing melt temperature, and the main cause of these phe-
Fig. 3. Typical microstructures of 170 ppm Sr modified AlSi7Mg0.3 alloy (Alloy 2)
obtained at (a) 429 and (b) 857 Pa/s pressurizing speed, respectively. nomena is the decreasing cooling rate, which affects the nucleation
and growth of grains[18]. In addition, the increased melt tempera-
ture deactivates available solidification sites, decreasing the frequency
significant coarsening of the dendrite arms occurred during solid- of grain nucleation.
ification as revealed from the coarse microstructural scale. While To include the effect of melt superheat on the solidification time
a lower filling pressure leads to a finer microstructure, the appli- ts, Tiryaltioglu et al.[19] revised the Chvorinov Rule’s and proposed
cation of a higher pressure, which is further held during casting the following equation:
solidification, induces the formation of slightly greater dendrite arms,
independent of the Sr level (see Fig. 3). t s = B ′k 1.31V 0.67 expξU (B ′ = 9.6 s cm −2, ξ = 0.00276°C −1
(3)
The range of cooling rates was estimated by secondary den- for A 356 alloy )
drite arm spacing, λ2, which strictly depends on the cooling rate
according to the empirical equation given below[14] where B′ is a mould constant (s cm‒2), k is the shape factor which
is defined as the ratio between the surface area of the sphere having
λ2 = 39.4R −0.317 (2) the same volume as the casting and the surface area through which
heat is lost by the casting, V is casting volume (cm3), ξ is an em-
where R represents the mean cooling rate of the primary α-Al den- pirical constant (°C‒1), U is superheat (°C).
drites during solidification. The average SDAS values and the Therefore, the pouring temperature directly affects the solidifi-
estimated cooling rates are listed in Table 4. cation time of the alloy and thus the growth of the primary α-Al
These results seem to be in contradiction with that reported in crystals.
literature where it is stated how a significant refinement in the struc- On industrial casting production it is difficult to obtain high so-
ture can be obtained in Al–Si foundry alloys by increasing the applied lidification rate throughout the casting, and therefore fine SDAS, for
pressure[15,16]. Even if these conclusions were drawn by applying several reasons: (1) the complexity of the shape of cylinder blocks,
squeeze casting technologies where the applied pressures are greater having many cavities in close proximity; (2) the need for direction-
than those used in LPDC, the structure refinement is generally at- al solidification to avoid casting macrodefects, and thus satisfactory
tributed to several factors. First, a significant increase in the cooling quality in the final component; (3) efficient heat removal is possi-
rate is obtained through improved contact between the casting and ble only if close to the die walls; (4) the gating, and eventually riser,
the die surface. The second factor is related to the thermody- designs are influenced by the part geometry.
namic nature of the applied pressure, which changes the melting Several investigations have been published about the factors that
influence SDAS and the relationship between SDAS and mechani-
Table 4 cal properties of Al alloy castings[20–22].
Average SDAS values measured under different experimental conditions (standard
deviation in parentheses); cooling rate, calculated according to Eq. (2) is reported 4.1.3. Effect of pressure increase on eutectic Si
Sr (wt%) Pressurizing SDAS Cooling rate, Due to Sr addition, the eutectic Si exhibits a fibrous morpholo-
speed (Pa/s) (μm) R (°C/s) gy for the analysed pressure ramps with coarser particles by
0.009 429 58 (4) 0.30 decreasing the filling time, i.e. increasing the pressurizing speed,
0.017 429 48 (5) 0.54 at the same Sr level (Fig. 4). The size distributions of the eutectic
0.009 857 67 (7) 0.19 Si particles were investigated and found to follow similar lognor-
0.017 857 55 (5) 0.35
mal distributions, independent of the filling time. The hypothesis
 G. Timelli et al. / Journal of Materials Science & Technology 32 (2016) 515–523 519

Fig. 4. Silicon crystals in the eutectic region of 170 ppm Sr modified AlSi7Mg0.3 alloy
(Alloy 2) obtained at (a) 429 and (b) 857 Pa/s pressure, respectively.

that the equivalent diameter of eutectic Si follows the lognormal

distribution cannot be rejected with the Anderson–Darling test sta-
tistic for all four data sets. The fitted distributions are shown in
Fig. 5(a). The coefficient of determination R2 was used to evaluate
the quality of the fitting, and it is 0.99, suggesting a good agree-
ment of the particle distribution with the lognormal distribution
adopted.
Fig. 5. Distributions of (a) equivalent diameter and (b) aspect ratio for eutectic Si
By decreasing the cooling rate, that is, increasing the pressur- particles estimated as function of the pressurizing speed and Sr levels. Lognormal
izing speed, the distribution of the eutectic Si size becomes more fitting curves are drawn.
spread for both the Sr levels analysed. The equivalent diameter of
eutectic Si particles with the maximum frequency shifts to higher
d values and the absolute value of the maximum frequency de- Recently, Sui et al.[23] reported a general tendency of SDAS to grad-
creases. The average size, at 90 ppm Sr level (Alloy 1), moves from ually decrease and then increase by Sr addition in a gravity die cast
about 1.0 to 2.5 μm for 429 and 857 Pa/s pressure increasing, re- AlSi12Cu4Ni2Mg0.8 alloy and the minimum value was measured
spectively, and from about 0.7 to 1.7 μm for the 170 ppm Sr modified at 0.02 wt% Sr. Fabrizi et al.[24] observed similar results for SDAS and
alloy (Alloy 2). It can be established that increasing the filling speed, grain size in a high-pressure die cast AlSi9Cu3(Fe) after 130 ppm
the microstructure is characterized by slightly coarser eutectic Si, Sr addition.
while reducing the flow rate, the formation of a high number of fine The size of the dendrites is, besides the cooling rate of solidifi-
Si particles is predominant. cation, dependent on the level of alloying elements present in the
The distribution of aspect ratio of eutectic Si particles also follows melt, even if this effect is not easily recognized due to the leading
a lognormal distribution with a threshold, since the minimum value effect of the cooling rate. Rearranging Eq. (2) into a general form
of α is 1. The threshold was here taken as 0.9, slightly below unity, by taking into account the solute concentration, the secondary den-
so that the fit can allow the presence of particles with an aspect drite arm spacing can be expressed as[25]
ratio of 1. The impact of the filling time, within the analysed range,
λ2 = G × M × R −n (4)
on the morphology of eutectic Si is negligible. Therefore, an in-
crease in the filling speed is related more to the size of the particles where G is a constant (also called geometric factor) and M is ex-
than to the aspect ratio, which is steady in the range of 1.9 to 2.1. pressed for a multicomponent alloy through:
n
4.1.4. Effect of Sr addition on SDAS
M = B × ∑ c 0bi,i (5)
The average SDAS values measured at different Sr levels are re- j =1
ported in Table 4 with the corresponding cooling rates calculated
according to Eq. (2). In general, the microstructural scale de- where B is a constant developing on the alloy and bi is the expo-
creases by increasing the Sr content in the alloy, independent of the nent of the concentration c0 of the i-th component.
applied pressure ramp. The lowest SDAS value is ~48 μm when Sr During the solidification of primary phase, the alloying ele-
level is 0.017 wt% and the pressure increase is set up at 429 Pa/s. ments are not uniformly distributed between the solid and liquid
520 G. Timelli et al. / Journal of Materials Science & Technology 32 (2016) 515–523

phases. Excess of solute moves away from the solid–liquid inter-

face into the melt increasing the solute content among the dendrite
arms already formed. This supersaturation affects the constitution-
al undercooling at the solid–liquid interface, which is the driving
force for the spacing adjustment mechanism in the region between
the dendrite arms. The dendrite is able to branch sufficiently to
reduce this undercooling to a very low value. It is also expected that
the elements with a higher solubility in the Al melt are less effec-
tive in reducing the size of SDAS[26]. Higher concentration of alloying
elements, e.g. Cu and Si as analysed by Sivarupan et al.[27], and Al–
Si eutectic modifiers, e.g. Sb as reported by Boontein et al.[28], will
cause the formation of finer dendrites with lower SDAS.
Therefore, the reduction in SDAS after the addition of a rela-
tively low Sr level (~0.017 wt%) is believed to be the result of the
constitutional undercooling promoted by Sr which segregates at the
solidification interface due to very low partition coefficient in
the Al–Sr system (<0.0046 as calculated from Reference 29).

4.1.5. Effect of Sr addition on eutectic Si

The influence of Sr level on the eutectic Si size can be deduced
from the lognormal distributions shown in Fig. 5(a). The distribu-
tions of the eutectic Si size become more spread by decreasing the
Sr content, independent of the pressure parameters. However, the
change of the eutectic Si size through variation of the Sr level is more
evident at lower filling time (i.e. greater pressure increase), where
slower cooling rates occur. The refining effect after Sr addition is
almost comparable in the experimental conditions where the cooling
Fig. 6. Typical pore morphology in the analysed region of the 90 ppm Sr modified
rate is higher. Upon increasing the Sr content from 90 to 170 ppm, AlSi7Mg0.3 alloy (Alloy 1): (a) shrinkage porosity and (b) gas porosity. The micro-
the average size of eutectic Si particles decreases about 29% at lower graphs refer to a casting produced with 0.035 MPa holding pressure.
pressurizing speed, and about 32% at the highest pressure increase.
As it is possible to observe in Fig. 5(b), the variation in the aspect
ratio of eutectic Si is negligible within the analysed Sr range. There- content. In castings solidified under 0.035 MPa, the porosity frac-
fore, in addition to the eutectic Si refinement, Sr is responsible for tion is in the range between 0.3% and 0.5%, and the components
the morphological addition. However, in contrast to the refining obtained with the highest holding pressure (0.05 MPa) show a po-
mechanism, the applied pressure ramp seems not to compromise rosity level between 0.1% and 0.4%. In the present work, the
the Sr addition of eutectic Si morphology. experimental results indicate that the applied pressure reduces the
driving force for pore nucleation and growth.
4.2. Casting defects The average area and roundness of pores in the analysed region
of three different castings for each experimental condition are
Fig. 6 shows the typical morphologies of various porosities in given in Table 5. The pores in the castings produced with higher
the polished surfaces of the samples. The pores have typical “shrink-
age” (Fig. 6(a)) or “gas” (Fig. 6(b)) porosity aspects and were
homogeneously distributed throughout the analysed section. While
the former are coupled with a lack of interdendritic feeding during
mushy zone solidification, gas porosity is the evolution of hydro-
gen due to a sudden decrease in hydrogen solubility during
solidification[5]. Microporosity is a leading cause in the reduction
of mechanical properties, particularly fatigue resistance[30], as well
as a loss of pressure tightness, which is a key requirement in the
cylinder blocks, especially in the space between the cylinders.
To assess the efficiency of feeding for the different experimental
conditions, the area fraction of porosity was measured in the analysed
region of the cylinder blocks. The method based on the shape of pores
is theoretically able to discriminate between the two forms of po-
rosity. However, such a distinction was not considered to be relevant
because most porosity is in any case assumed to be closely related to
the presence of oxide films[31]. The difficulty of differentiating accu-
rately between these defects was therefore avoided[29].

4.2.1. Effect of holding pressure on microporosity

Fig. 7 summarizes the influence of holding pressure and Sr
content on porosity in the castings. In general, the amount of po-
rosity is low, therefore confirming LPDC as a useful foundry process
for the production of Al blocks for high performance engines. Further, Fig. 7. Influence of holding pressure and Sr levels on the porosity in the analysed
higher applied pressure generates castings with lower porosity region of cylinder blocks.
 G. Timelli et al. / Journal of Materials Science & Technology 32 (2016) 515–523 521

Table 5
Influence of holding pressure and Sr levels on the average size and roundness of pores

Sr (wt%) Holding Average pore Roundness

pressure (MPa) size (μm2)

0.009 0.035 6843 2.5

0.017 0.035 3054 2.2
0.009 0.050 3498 1.9
0.017 0.050 1649 1.7

pressure are smaller and more rounded than the pores in the cast-
ings obtained at lower holding pressure.
The final applied pressure during casting solidification is aimed
to feed and transport further material into the die cavity to com-
pensate for the subsequent solidification shrinkage and thermal
contraction, thus producing sound castings. In LPDC, the risers or
feeders are limited with respect to sand casting or gravity die casting.
Contrary to high-pressure die casting or squeeze casting where the
intensification pressures are greater (up to ~120 MPa), the pres-
sure applied in LPDC is low enough to prevent the evolution of
dissolved hydrogen or to effectively compress any entrapped air.
Pore growth is governed by Eq. (6), which represents a balance
between the pressures that promote and oppose pore formation[32]:

Pg + Ps ≥ Papp + PH + Ps-t (6)

where Pg is the equilibrium pressure of dissolved gases, Ps is the pres-

sure drop due to solidification shrinkage, Papp is the applied pressure, Fig. 8. Typical filling-related defects, (a) oxide films and (b) cold shot, observed in
PH is the pressure due to metallostatic head, and Ps-t is the pres- the castings produced with higher pressurizing speed.
sure due to pore–liquid surface tension. Thus, the pore formation
can occur when the sum of the pressures referred to solidification
shrinkage and hydrogen rejection is larger than the combined effects ty by restricting pore growth[33]. It is clear from Eq. (7) that the size
of applied pressure, metallostatic head and surface tension. For a of the pores, and hence the amount of porosity, is not linearly pro-
wetting liquid, Ps-t can be expressed as 2γ/r where γ is the surface portional to the applied pressure.
tension and r is the pore radius[32]. Therefore, the pore size can be
calculated as follows: 4.2.2. Effect of Sr addition on microporosity
The effect of Sr on the amount of porosity can be observed in
2γ Fig. 7. Contrary to expectations derived from the literature, the
r= (7)
Pg − (Ps + Papp + PH ) average porosity decreases by increasing Sr content from 90 up to
170 ppm in the alloy, and this behaviour seems to be independent
The size and amount of porosity is expected to be inversely pro- of the applied holding pressure.
portional to the applied holding pressure. Strontium addition is generally associated with an increase in
On the other hand, it can be reasoned that impurities may affect porosity, especially above that of the unmodified alloy, and also a
porosity formation. They may alter the surface tension, so the term redistribution of porosity from macroscopic shrinkage porosity to
Ps-t, and the stability of the protective oxide layer, affecting the term well-dispersed microporosity[5]. Further, higher amount of poros-
Pg; they may act as active nucleate site for pores allowing pore for- ity was generally detected by increasing the Sr content in the liquid
mation at a lower gas pressure, Pg, or hydrostatic shrinkage pressure, bath[34]. Recently, Dinnis et al.[35] demonstrated how a different dis-
Ps, than would otherwise be the case; they may affect the evolu- tribution of eutectic solid throughout the dendritic network due to
tion of the microstructure during solidification, altering in this changes in nucleation and growth modes can affect feeding and po-
manner the term Ps. rosity formation. Gradually increasing the Sr content, the eutectic
In the present work, density measurements carried out by reduced- grains begin to nucleate and grow away from existing primary α-Al
pressure testing revealed a strict range of density values, 2.63–2.65 g/ dendrites, independent of them, with a relatively smooth solidifi-
cm3. Therefore, it can be assumed that the number of heterogeneous cation front[36,37]. This mechanism may lead to an increase in porosity
pore nucleation sites in the liquid bath inside the holding furnace because the growth of the eutectic grains can obstruct feed paths
remained constant over the entire experimental campaign. causing a decrease in permeability and feeding efficiency.
As the pressurizing speed increases, the flow of metal inside the In the present work, the porosity differences in the Sr-containing
die cavity can be more turbulent, eventually leading to the entrain- AlSi7Mg0.3 engine blocks can be related to the eutectic solidifica-
ment of inclusions and gases, and an increase in porosity. Fig. 8 shows tion mode by considering the ability of the riser tube in the LPDC
the typical filling-related defects, such as oxide films and cold shots, machine through the runners and gating system to supply feed metal
as revealed in the castings produced with greater pressure in- to the analysed region between the cylinders, which represents a
crease. As evidenced by Liu et al.[12], the gate velocity increases with hot spot in the casting. The multiple runners connected to the riser
the effective pressure under higher pressurizing speed, leading to tube and the ingates solidify early and cannot supply feed metal
gate velocities greater than critical value (0.5 m/s)[31]. This leads to to the region between the cylinders. After the runners’ solidifica-
greater turbulence which is the main cause of oxide film entrain- tion, the hot spot of the engine block acts as a riser for the casting.
ment and formation of cold shots. However, the application of higher In high Sr-modified alloy (Alloy 2), the low permeability of the
pressure may eventually prevent the potential increase in porosi- solidifying microstructure, due to independent nucleation of
522 G. Timelli et al. / Journal of Materials Science & Technology 32 (2016) 515–523

Fig. 9. (a) Main effects plot displaying the influence of Sr content and pressurizing
Fig. 10. (a) Main effects plot displaying the influence of Sr content and holding pres-
speed on SDAS; (b) Pareto chart evaluates the significant variables and their pos-
sure on the amount of porosity; (b) Pareto chart evaluates the significant variables
sible interactions (p-value of 0.05 is drawn as a reference line).
and their possible interactions (p-value of 0.05 is drawn as a reference line).

eutectic grains, reduces the flow of metal out of the hot spot to feed
the rest of the casting. In contrast, in the low Sr modified alloy (Alloy The reduced-pressure test, which has been shown to be a reason-
1), the relatively permeable microstructure enables liquid to flow able and useful method for hydrogen measurement[38], revealed
from the hot spot to feed the remainder of the casting. Therefore, almost a steady-state value during the experiments. Upon increas-
the porosity in the analysed region is lower by increasing the Sr level ing the Sr level, the pores become more rounded due to the
from 90 up to 170 ppm in the AlSi7Mg0.3 alloy. progressive change in the morphology of the solidification front.
The average area and roundness of pores measured at different Dahle et al. [37] explained how the eutectic grains, nucleating
Sr contents are listed in Table 5. Pores are smaller and more rounded independently of the α-Al dendrites, grow with a relatively smooth
by Sr increasing, independent of the applied pressure. This behaviour solidification front because the fibrous morphology of the eutec-
seems to not be associated with the dissolved hydrogen in the bath. tic silicon does not grow ahead of the eutectic aluminium.

Table 6
Analysis of variance for SDAS (coded units)

Source DF Seq SS Adj SS Adj MS F-value p-value

Main effects 2 1806.25 1806.25 903.13 32.74 0.000

Sr 1 1221.03 1221.03 1221.03 44.26 0.000
Pressurizing speed 1 585.22 585.22 585.22 21.21 0.000
2-Way interactions 1 9.03 9.03 9.03 0.33 0.571
Sr· pressurizing speed 1 9.03 9.03 9.03 0.33 0.571
Residual error 36 993.10 993.10 27.59
Pure error 36 993.10 993.10 27.59
Total 39 2808.38

Table 7
Analysis of variance for porosity (coded units)

Source DF Seq SS Adj SS Adj MS F-value p-value

Main effects 2 1.11289 1.11289 0.55644 32.96 0.000

Sr 1 0.92480 0.92480 0.92480 54.79 0.000
Holding pressure 1 0.18809 0.18809 0.18809 11.14 0.001
2-Way interactions 1 0.04601 0.04601 0.04601 2.73 0.103
Sr·holding pressure 1 0.04601 0.04601 0.04601 2.73 0.103
Residual error 68 1.14783 1.14783 0.01688
Pure error 68 1.14783 1.14783 0.01688
Total 71 2.30673
 G. Timelli et al. / Journal of Materials Science & Technology 32 (2016) 515–523 523

4.3. Analysis of variance (7) As the pressurizing speed increases, the flow of metal inside
the die cavity can be more turbulent, leading to the forma-
The analysis of the data by means of ANOVA revealed a com- tion of oxide films and cold shots.
plete and exhaustive evaluation of the relationship between the (8) In the considered experimental conditions, the average size
independent variables, such as pressure parameters and Sr content, of SDAS and the amount of porosity can be significantly de-
and the dependent variables, i.e. SDAS and amount of porosity, within scribed by a regression model that considers the concentration
the experimental conditions. Figs. 9 and 10 show the results of SDAS of Sr and the pressure parameters applied during LPDC.
values and amount of porosity, in terms of main effect diagrams and
Pareto charts. Further, Tables 6 and 7 summarize the results of Acknowledgments
ANOVA for SDAS and porosity. As generally reported in the ANOVA
tables[13], the number of degrees of freedom from each source (DF), The authors acknowledge Dr. R. Molina and Dr. M. Badiali from
the sum of squares (SS), the mean squares (MS), the results of Fisher Teksid Aluminum Spa (Carmagnola, Italy) for the support to this re-
test (F-value) and the p-values are shown. Furthermore, both the search, and thank Dr. A. Guglielmi for the experimental contribution
sequential (Seq SS) and the adjusted sums (Adj SS) and mean (Adj to this work.
MS) of squares are also indicated. A detailed description of the sta-
tistical method used was given elsewhere[13]. References
Figs. 9 and 10 show the relative power of each factor, or inter-
action, to affect the variation in the microstructural scale or porosity [1] J. Alson, A. Hula, A. Bunker, Light-duty automotive technology, carbon dioxide
level. At the studied levels, the variable most significantly affecting emissions, and fuel economy trends: 1975 through 2014, <http://www.epa
.gov/otaq/fetrends-complete.htm#summary>, Trends Report, October, 2014.
the analysed properties is the Sr content in the AlSi7Mg0.3 alloy. At [2] European Aluminium Association, The Aluminium Automotive Manual, 2nd ed.,
higher level, while Sr tends to decrease both SDAS and porosity, the 2011, pp. 1–23. <http://european-aluminium.eu/wp-content/uploads/2011/11/
effects of the pressure parameters depend on the specific charac- AAM-Applications-Power-train-2-Engine-blocks.pdf>.
[3] W. Kasprzak, D. Emadi, M. Sahoo, M. Aniolek, Mater. Sci. Forum 618–619 (2009)
teristic. Upon increasing the pressurizing speed during die cavity 595–600.
filling, the SDAS value increases; on the other hand, the porosity [4] G. Nicoletto, E. Riva, Metall. Ital. 103 (2011) 41–48.
decreases by increasing the holding pressure. The effects of the re- [5] J.E. Gruzleski, B.M. Closset, The treatment of liquid aluminum-silicon alloys,
Des Plaines: American Foundrymen’s Society, Illinois, 1990.
lationships among the factors, i.e. Sr·Pressurizing speed and Sr·Holding [6] S. Hegde, K.N. Prabhu, J. Mater. Sci. 43 (2008) 3009–3027.
pressure, appear statistically insignificant, with a p-value higher than [7] M. Merlin, G. Timelli, F. Bonollo, G.L. Garagnani, J. Mater. Process. Technol. 209
0.05 for both the features analysed (see Figs. 9(b) and 10(b)). (2009) 1060–1073.
[8] P.H. Fu, A.A. Luo, H.Y. Jiang, L.M. Peng, J. Mater. Process. Technol. 205 (2008)
On the basis of the significance level provided by the analysis
224–234.
of variance, a linear regression model can be developed describ- [9] J.H. Kuo, F.L. Hsu, W.S. Hwang, Sci. Technol. Adv. Mater. 2 (2001) 131–145.
ing the relationship between the response and the predictor [10] A.M. Du, N. Wei, Z.P. Zhu, K. Zhou, Adv. Mater. Res. 1004–1005 (2014) 1172–
variables. The reliability of the model is based on the coefficient of 1177.
[11] G. De Looze, Shape casting—The John Campbell Symposium—TMS 2005 Annual
determination, R2, which measures the quality of the least-squares Meeting, San Francisco, California, U.S.A., February 13–17, 2005.
fitting to the experimental data. In the considered range of exper- [12] S.G. Liu, F.Y. Cao, X.Y. Zhao, Y.D. Jia, Z.L. Ning, J.F. Sun, Mater. Sci. Eng. A 626
imental conditions, the size of SDAS and the amount of porosity can (2015) 159–164.
[13] D.C. Montgomery, Design and Analysis of Experiments, 8th ed., Hoboken, NJ,
be described by the following semi-empirical equations: 2012.
λ2 (μm ) = 59 − 1024 × (Sr ) + 0.025 × (Pressurizing speed) (8) [14] Q.G. Wang, D. Apelian, D.A. Lados, J. Light Met. 1 (2001) 85–97.
[15] M.R. Ghomashchi, A. Vikhrov, J. Mater. Process. Technol. 101 (2000) 1–9.
[16] T. Reguła, J.J. Sobczak, A. Fajkiel, P. Dudek, Arch. Foundry Eng. 11 (2011) 55–60.
Porosity ( % ) = 0.96 − 28.33 × (Sr ) − 6.81 × (Holding pressure) (9) [17] R.T. DeHoff, Thermodynamics in Materials Science, McGraw-Hill, New York,
1993, pp. 142–150.
where Sr is in wt%, Pressurizing speed is in Pa/s and Holding pres- [18] D.G. Eskin, V.I. Savran, L. Katgerman, Metall. Mat. Trans. A 36 (2005) 1965–1976.
sure is in MPa. In order to verify the reliability of the regression [19] M. Tiryaltioglu, E. Tiryakioglu, D.R. Askeland, AFS Trans. 97–95 (1997) 907–913.
models of Eqs. (8) and (9), the experimental Sr contents [20] F. Grosselle, G. Timelli, F. Bonollo, Mater. Sci. Eng. A 527 (2010) 3536–3545.
[21] L.A. Dobrzanski, R. Maniara, J. Sokołowski, W. Kasprzak, J. Mater. Process.
(see Table 1) and the experimental pressure parameters during LPDC Technol. 191 (2007) 317–320.
were used to predict the SDAS and the level of porosity. These results [22] L.A. Dobrzański, R. Maniara, J.H. Sokolowski, Arch. Mater. Sci. Eng. 28 (2007)
were compared with the measured ones to describe the quality of 105–112.
[23] Y. Sui, Q. Wang, G. Wang, T. Liu, J. Alloy. Compd. 622 (2015) 572–579.
the developed models by means of R2 values, which are 0.99 and [24] A. Fabrizi, S. Ferraro, G. Timelli, Mater. Charact. 85 (2013) 13–25.
0.95 for the average SDAS value and porosity, respectively. [25] V. Rontó, A. Roósz, Int. J. Cast Met. Res. 13 (2001) 337–342.
[26] Y. Wen, Dendrite and silicon particle refinement using Al-insoluble elements
5. Conclusions in hypoeutectic aluminum silicon alloys, M.Sc. Thesis, University of Windsor,
1996.
[27] T. Sivarupan, C.H. Cáceres, J.A. Taylor, Metall. Mater. Trans. A 44 (2013)
(1) The increase in the pressurizing speed reduces the filling time 4071–4080.
and induces the formation of slightly greater dendrite arms, [28] S. Boontein, N. Srisukhumbovornchai, J. Kajornchaiyakul, C. Limmaneevichitr,
independent of the Sr level. Int. J. Cast Met. Res. 24 (2011) 108–112.
[29] S.M. Liang, M. Engstler, V. Groten, J. Barrirero, F. Mücklich, A. Bührig-Polaczek,
(2) Coarsening of the eutectic Si particles is observed by increas- R. Schmid-Fetzer, J. Alloy. Compd. 610 (2014) 443–450.
ing the pressurizing speed. [30] J.Y. Buffière, S. Savelli, P.H. Jouneau, E. Maire, R. Fougères, Mater. Sci. Eng. A 316
(3) The microstructural scale decreases by increasing the Sr (2001) 115–126.
[31] J. Campbell, Complete Casting Handbook—Metal Casting Processes, Metallurgy,
content in the alloy, independent of the applied pressure ramp. Techniques and Design, 1st ed., Butterworth-Heinemann, Oxford, 2011.
(4) Upon increasing the Sr level in the alloy, the average size of [32] D.M. Stefanescu, A.V. Catalina, Int. J. Cast Met. Res. 24 (2011) 144–150.
eutectic Si decreases more at lower filling time, which is ob- [33] W. Jiang, Z. Fan, D. Liao, D. Liu, Z. Zhao, X. Dong, Mater. Des. 32 (2011) 926–934.
[34] S. Shivkumar, L. Wang, D. Apelian, JOM 43 (1991) 26–32.
tained by increasing the pressurizing speed.
[35] C.M. Dinnis, M.O. Otte, A.K. Dahle, J.A. Taylor, Metall. Mater. Trans. A 35 (2004)
(5) The impact of both applied pressure and Sr content, within 3531–3541.
the analysed conditions, on the morphology of eutectic Si is [36] K. Nogita, A.K. Dahle, Mater. Charact. 46 (2001) 305–310.
negligible. [37] A.K. Dahle, K. Nogita, J.W. Zindel, S.D. McDonald, L.M. Hogan, Metall. Mat. Trans.
A 32 (2001) 949–960.
(6) Higher holding pressure and Sr content generate cylinder [38] X.G. Chen, F.J. Klinkenberg, S. Engler, L. Heusler, W. Schneider, JOM 46 (1994)
bridges with lower porosity content. 34–38.

View publication stats