You are on page 1of 9

j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 3 ( 2 0 0 8 ) 333341

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

The effect of average cooling rates on the microstructure of the Al20% Si high pressure die casting alloy used for monolithic cylinder blocks
H. Yamagata a , W. Kasprzak b,c , M. Aniolek b , H. Kurita a , J.H. Sokolowski b,
R&D Operations, Yamaha Motor Co., Ltd., 2500 Shingai, Iwata, Shizuoka 438-8501, Japan Light Metals Casting Technology Group, Department of Mechanical, Automotive and Materials Engineering, University of Windsor, 401 Sunset Avenue, Windsor, Ont., Canada N9B 3P4 c CANMET Materials Technology Laboratory, 568 Booth Street, Ottawa, Ontario K1A 0G1, Canada
b a

a r t i c l e
Article history:

i n f o

a b s t r a c t
The effect of average cooling rates on the microstructure of the hypereutectic Al20% Si alloy was investigated using the novel Universal Metallurgical Simulator and Analyzer Platform. The quantitative measurements of the primary Si size and the Secondary Dendrite Arm Spacing of the non-equilibrium -aluminum as a function of the cooling rates was performed for the laboratory test samples. This research was carried out in order to analyze the microstructure of the high pressure die cast cylinder block and to understand its complex solidication process. The Equivalent Diameter of the primary Si decreased from 89.7 17.3

Received 22 June 2007 Received in revised form 12 October 2007 Accepted 15 October 2007

Keywords: Al20% Si alloy High pressure die casting Cylinder block Rapid solidication Microstructure Primary silicon

to 16.5 3.8 m and the Secondary Dendrite Arm Spacing from 22.1 5.9 to 5.1 0.8 m with an increase in the cooling rate from 4.9 to 82.9 C/s. Observations of the cylinder block microstructures revealed that the primary Si size was nearly identical at the subsurface and the centre locations of the bore wall. The Secondary Dendrite Arm Spacing of the nonequilibrium -aluminum phase as well as the eutectic Si size was signicantly smaller at the subsurface of the bore wall. Based on the UMSA laboratory measurements it was determined that the primary Si in the engine bore wall (both at the subsurface and the centre) nucleated as a rst phase from the liquid melt at a cooling rate of approximately 7274 C/s. It was found that the non-equilibrium -aluminum dendrites at the engine bore wall subsurface nucleated from the semi-solid melt at a cooling rate of approximately 85 C/s, while at the centre of the bore wall at approximately 49 C/s. Research revealed that some primary Si particles nucleated from the beginning of the melt pouring into the shot sleeve prior to the injection process while the -aluminum dendrites and eutectic Si nucleated in the die cavity. Therefore, it was proven that the injected hypereutectic Al20% Si liquid melt had solid primary Si particles. 2007 Elsevier B.V. All rights reserved.

1.

Introduction

Current automotive engines require low exhaust gas emissions and adequate output power. To generate a higher output

power, the engine cylinder block requires high cooling performance. Compared to the conventional cast iron cylinder block and/or the aluminum block with a cast iron liner, the monolithic cylinder block (Yamagata, 2005) made of a hyper-

Corresponding author. E-mail address: jerry@uwindsor.ca (J.H. Sokolowski). 0924-0136/$ see front matter 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.jmatprotec.2007.10.023

334

j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 3 ( 2 0 0 8 ) 333341

eutectic AlSi alloy is an attractive alternative. This block can provide superior cooling and tribological performance as well as a reduction in weight. The co-authors (Kurita and Yamagata, 2004; Yamagata and Kurita, 2005) recently enabled a mass production of Al20% Si cylinder blocks using a vacuum assisted High Pressure Die Casting (HPDC) process. The monolithic cylinder block had a liner-less bore surface with uniformly dispersed primary Si crystals inside the aluminum matrix. The Al20% Si alloy was developed for severe operational conditions and has superior wear resistance characteristics. Tests performed on the actual engine revealed that the bore wall surface temperature decreased by 30 C and oil consumption decreased by half in comparison with the aluminum cylinder block having a cast iron liner. Exceptional tribological properties of the bore surface were obtained by selectively exposed primary Si crystals during the honing operation (Kurita and Yamagata, 2004; Yamagata and Kurita, 2005). Since additional operations like heat treatment or machining could not change the primary Si morphology, the control of the ascast microstructure was of primary importance. Laboratory and casting plant experiments demonstrated that the casting parameters during the HPDC process affected the morphology of the Si particles (size and distribution). The control of the microstructure was a technologically challenging task in order to keep these parameters within the required range. To date, the detailed relationship between the cooling and/or solidication rate and the corresponding microstructure modication and renement of the hypereutectic AlSi alloys was not reported. Particularly the solidication rates signicantly exceeding equilibrium conditions were not fully investigated or scientically understood (Backerud et al., 1990; Djurdjevic et al., 2001). Recently the authors of this paper performed a series of experiments for the Al20% Si alloy using the novel Universal Metallurgical Simulator and Analyzer (UMSA) platform and reported the thermal characteristics at a relatively slow solidication rate, i.e., 1 C/s. It was demonstrated that the UMSA platform can generate a superior metallurgical database by linking thermal and structural characteristics of the alloys and cast components that directly contribute to rapid technological advancements (Chen et al., 2006; Kasprzak et al., 2001, 2002; Yamagata et al., 2008). The microstructure formation in the HPDC process is very complex and is inuenced by many factors associated with liquid and semi-solid melt processing, ow and heat transfer phenomena. Since microstructure control is a key factor to generate a tribologically excellent cylinder bore, therefore determination of its structural and thermal characteristics is of primary importance for development of novel HPDC processes. The goal of this paper is to present the new methodology to determine the thermal and structural characteristics of the HPDC cylinder block based on customized UMSA computer controlled rapid solidication experiments. In addition, to estimate selected engine bore section solidication rates based on the comparative study of the Secondary Dendrite Arm Spacing (SDAS) and the Equivalent Diameter (ED) of the primary Si obtained from the rapidly solidied laboratory test samples.

Table 1 Average chemical composition of the hypereutectic Al20% Si alloy (wt%) Si


20.0

Cu
3.0

Mg
0.5

Zn
0.1

Fe
0.5

Mn
0.1

Ni
0.1

Ti
0.001

P
0.01

2.
2.1.

Experiments
Material

The material used for the HPDC cylinder blocks and for the UMSA test samples was a phosphor modied Al20% Si alloy with a chemical composition as listed in Table 1. Ingot metallurgical analyses revealed that the microstructure consisted of coarse primary Si crystals and unmodied eutectic AlSi as well as Cu, Mg based inter-metallic phases dispersed in the aluminum matrix (Yamagata et al., 2008). The average ED of the primary Si particles of the ingot was approximately 45 m for a 100 ppm addition of phosphor.

2.2.

UMSA test sample conguration

Ring-shaped test samples with an outer diameter OD = 44.7, an inner diameter ID = 35.4 and a height of 12 mm were machined out from the ingot. The test sample mass was kept in a very narrow range of 18 0.2 g allowing for the highest level of repeatability and reproducibility of the structural and thermal characteristics. Each sample had a predrilled hole to accommodate a temperature sensor. In order to assure rapid temperature response, very low thermal mass K type thermocouples were utilized. The ring-shaped test sample was placed inside a crucible made of ultra thin steel sheet offering a negligible thermal mass and consequently a very rapid sample temperature response to the imposed experimental thermal conditions. The top and bottom parts of the test sample were insulated. This design allowed for radial heat extraction conditions. Therefore, it could be assumed that the test sample represented a hollow cylinder with an innite length (Djurdjevic et al., 2001; Kasprzak et al., 2001).

2.3.

Rapid solidication experiments

The melting and rapid solidication experiments were performed using the UMSA technology platform under an ambient atmosphere (Kasprzak et al., 2002). The experimental procedure was as follows: First, the test sample taken from the pre-modied ingot was heated to 785 0.2 C and isothermally kept at this temperature for a period of 2 min in order to stabilize and homogenize the melt conditions. The thermal analysis experiments showed an absence of any metallurgical reactions above 710.9 C, therefore, an approximate superheat temperature of 75 C was satisfactory to achieve some degree of thermal modication of the primary Si crystals (Yamagata et al., 2008). The melting rate was identical for all experiments and was calculated between the non-equilibrium Solidus (beginning of the melting process) and the non-equilibrium Liquidus (end

j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 3 ( 2 0 0 8 ) 333341

335

Fig. 1 Overall view of the machined cylinder block cast using the vacuum HPDC process. The as-cast bore wall thickness measured approximately 7 mm. The microstructural observations were carried out at the contour enclosed by the white dashed box.

HPDC machine was used to cast the experimental cylinder block samples. The melt was kept at 785 10 C in the holding furnace and was ladled into the shot sleeve and injected into the die cavity. Since the Al20% Si alloy Liquidus temperature was 691 C (Yamagata et al., 2008), the melt holding temperature of 785 10 C ensured a superheat of approximately 94 C. The shot sleeve temperature before melt dosing was set at approximately 200 C to prevent excessive melt cooling prior to injection. In the production cycle, the time from the onset of injection to the completion of die lling was approximately 1.5 s. The other best machine parameters including die temperature, injection speed of the melt and lubricant were selected to enable the mass production of the cylinder blocks. The measured gas content of the cylinder block was 0.3 cm3 /100 g and was signicantly lower in comparison with other HPDC technologies due to melt degassing and the vacuum assisted die casting process. The detailed manufacturing procedures were described in the previous paper (Kurita and Yamagata, 2004).

2.5. Metallographic assessment of the UMSA and HPDC cylinder block samples
After completion of the UMSA slow and rapid solidication experiments each test sample was removed from the crucible, sectioned in a perpendicular direction and prepared for metallographic observations. Quantitative assessment of the primary Si particles was performed using an automated Leica Light Optical Microscope (LOM) equipped with a QW-550 Image Analysis (IA) system. Thirty analytical elds were analyzed and the Length, Width and ED of the primary Si particles were measured. It was found that in some locations primary Si particles were agglomerated, i.e., the individual crystals touch each other. During the IA measurements the primary Si crystals that were in contact with adjacent neighbours were interactively separated. As a consequence the measured ED values of the individual primary Si particles might appear smaller than observed in the micrographs containing Si agglomerates. The ED was selected as the most reliable stereological descriptor of the primary Si particles. The ED value was expressed as a diameter of the circle having the same area as the analyzed feature. Next, the average values and corresponding standard deviation of the ED were calculated. The SDAS was measured by plotting the line across the dendrite arms. Next, the line length was divided by the number of intercepted dendrite arms. Thirty analytical elds were analyzed and the average values of the SDAS with corresponding standard deviation were calculated. As above, identical measurements were performed for the test samples taken from the cylinder block (Fig. 1). Two analytical locations were selected for analysis i.e., 0.5 and 3.5 mm away from the bore wall surface termed subsurface and centre locations respectively. The measurements were performed to determine the effect of the CR on the microstructural characteristics in these locations. The representative LOM micrographs under 100, 200 and 1000 magnications were taken.

of the melting process) temperatures, i.e., 502.8 and 710.9 C, and was equal to approximately 0.75 C/s (Yamagata et al., 2008). Next, the samples were cooled at rates of 4.9, 14.7, 23.9, 52.7 and 82.9 C/s. Computer controlled rapid melting and short isothermal holding of the pre-modied alloy prior to solidication secured high process repeatability. This procedure avoided phosphor fading and consequently its effect on chemical modication potency of the primary Si. High Cooling Rates (CR) were obtained using a controlled ow of compressed gas and/or atomized water sprayed from the multi-functional UMSA coil on to the test crucibles radial surface. The average CR was calculated between 730 and 380 C to closely match the industrial process in which the 730 and 380 C temperatures corresponded to the melt injection into the die and casting removal respectively. Please note that the term Cooling Rate is intentionally used to quantify both the alloy cooling (above liquid and below semi-solid states) and the solidication process taking place between the non-equilibrium Liquidus and Solidus temperatures. The conventional term Solidication Rate might be technologically incoherent since it only relates to the process between the Liquidus and the Solidus temperatures. The thermal analysis signal represented by heating and cooling curves was recorded during melting and solidication cycles at an acquisition rate of up to 20 Hz. The Temperature vs. Time curves were automatically calculated and plotted and used for detailed thermal characterization of these cycles.

2.4.

HPDC cylinder block samples

A machined monolithic cylinder block for the 250 cm3 single cylinder engine is presented in Fig. 1. The bore wall portion (dashed rectangular box) was extracted from the as-cast cylinder block for metallurgical analysis. An eight MN cold chamber

336

j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 3 ( 2 0 0 8 ) 333341

Fig. 2 Temperature vs. Time curves recorded during UMSA rapid solidication experiments. The average CR calculated between 730 and 380 C were as follows: (a) 4.9, (b) 14.7, (c) 23.9, (d) 52.7 and (e) 82.9 C/s.

It was found that increasing the CR resulted in the modication of the primary Si particles. The ED of primary Si was reduced from approximately 89.7 17.3 to 16.5 3.8 m when the average CR was increased from 4.9 to 82.9 C/s (Fig. 4). In addition, the distribution of primary Si was more homogenous for a higher CR and size variation was signicantly less as expressed by the standard deviation. The size of the SDAS of the -aluminum dendrites is shown in Fig. 5 as a function of the average CR. The SDAS was reduced from approximately 22.1 5.9 to 5.1 0.8 m when the average CR was increased from 4.9 to 82.9 C/s. It was observed that standard deviation was reduced as well. This coincided with the metallographic observations that the primary dendrites were more uniform and evenly distributed. Test samples that solidied at a maximum CR, i.e., 82.9 C/s had primary Si particles homogeneously distributed and had signicantly lower numbers of agglomerates as well as a dendrite less metal matrix (Fig. 3e). In addition, it was observed that an increased CR thermally modies the AlSi eutectic (Fig. 3a and e).

3.
3.1.

Results
UMSA solidication experiments

3.2.

Microstructure of the HPDC cylinder block samples

The solidication experiments performed using ring test samples showed that the slowest average CR of 4.9 C/s was achieved for the sample cooled down using compressed air. The maximum average CR of 82.9 C/s was achieved for the test sample solidied in an atomized water environment. Fig. 2 summarizes the UMSA experimental cooling curves. Detailed analysis of the Temperature vs. Time curves showed a visible temperature arrest at approximately 560570 C. This plateau corresponded to the nucleation of the AlSi eutectic and was caused by latent heat release. The latent heat of eutectic Si was approximately 1810 kJ/kg as compared with 395 kJ/kg for aluminum (Jorstad and Apelian, 2004) and was sufcient to increase the temperature of the test specimen/cast component during the solidication process. In the literature this phenomenon was related to the so-called recalescence effect (Backerud et al., 1990; Djurdjevic et al., 2001; Kasprzak et al., 2001; Yamagata et al., 2008). The thermal arrest was clearly observed on the Temperature vs. Time curves recorded for the test samples solidied at cooling rates of 4.9, 14.7 and 23.9 C/s (Fig. 2). The Temperature vs. Time curves spatial resolution obtained for the test samples solidied at 52.7 and 82.9 C/s was not fully satisfactory (for the given test sample/crucible congurations) for the advanced curve analysis using the rst derivative technique (Djurdjevic et al., 2001; Kasprzak et al., 2001; Yamagata et al., 2008) allowing for quantication of the nucleation of the AlSi eutectic characteristics. However, UMSA test samples were used to perform comprehensive metallographic studies. Metallographic analysis revealed that structural modication of the Al20% Si alloy depended on the CR. Representative LOM micrographs taken from the test sample solidied at an average CR varying from 4.9 to 82.9 C/s and are presented in Fig. 3ae. Due to the fact that the size and distribution of the primary Si particles inuences the engine cylinder bore tribological properties the quantitative Si measurements were of primary interest in this study.

It was reported (Jorstad and Apelian, 2004; Tsumagari and Mobley, 1995) that the hypereutectic A390 alloy showed a primary Si depleted zone near the cast surface facing the die. For the analyzed HPDC component the depleted area was present but was microscopically small. The reason for the surface depletion is not well understood (Jorstad and Apelian, 2004) but might be related to the complexity of the HPDC solidication process and the resultant intricate microstructure. It is expected that the casting bore subsurface adjacent to the die must have a higher CR during the solidication process than the bore wall centre section. The cylinder bore wall (Fig. 1) representative microstructures of the subsurface and centre locations are presented in Figs. 68. It can be observed that the size of the primary Si particles in both locations is very similar. The average ED values obtained from the IA measurements were 20.6 5.9 and 20.0 6.2 m for the subsurface and centre respectively (Fig. 6 a and b). This difference was not statistically signicant. The SDAS of non-equilibrium -aluminum dendrites was found to be 4.6 0.7 m at the subsurface and 7.9 1.2 m in the centre of the engine bore section (Fig. 7 a and b). Correspondingly, the AlSi eutectic had a ne coral-like shape at the bore subsurface with a coarse needle-like shape morphology in the centre (Fig. 8 a and b). In addition, it was observed that some locations with distinctively developed non-equilibrium -aluminum dendrites were free from primary Si particles. Based on the average CR vs. SDAS/ED relationships obtained from the UMSA rapid solidication experiments an attempt was made to estimate the CR for the HPDC engine bore wall sections. The SDAS and ED values were considered as key parameters linking the microstructural/thermal characteristics of the UMSA test samples with the actual cast component. The CR rate was calculated based on the regression equations presented in Figs. 4 and 5. The results are summarized in Table 2. Based on the ED measurements it was determined that the average CR for the bore wall subsurface and centre (0.5 and 3.5 mm distance) locations was approximately 72.6 and

j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 3 ( 2 0 0 8 ) 333341

337

Fig. 3 (a) LOM micrograph of the Al20% Si alloy test sample solidied at a CR of 4.9 C/s. (b) LOM micrograph of the Al20% Si alloy test sample solidied at a CR of 14.7 C/s. (c) LOM micrograph of the Al20% Si alloy test sample solidied at a CR of 23.9 C/s. (d) LOM micrograph of the Al20% Si alloy test sample solidied at a CR of 52.7 C/s. (e) LOM micrograph of the Al20% Si alloy test sample solidied at a CR of 82.9 C/s.

74.5 C/s respectively. The average CR estimated based on the SDAS values was 84.7 and 48.9 C/s for subsurface and centre locations of the engine bore wall.

4.
4.1.

Discussion
Microstructures of the UMSA test samples

The UMSA experiments showed a strong statistical relationship between the CR and the ED and SDAS values as indicated

by the regression coefcient R2 that is 0.98 and 0.99 respectively. In general, the reduction of the primary Si particle size was caused by the combination of both thermal and chemical modications. However, in the UMSA experiments the level of phosphor addition was constant therefore, it can be assumed that its effect on Si modication was identical for all analyzed test samples. In addition, melt processing conditions (melting rate, superheat temperature and isothermal holding time) prior to the rapid solidication experiments were kept constant. Consequently it can be concluded that the size of the primary Si crystals was a function of the CR only. Forced

338

j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 3 ( 2 0 0 8 ) 333341

Fig. 4 The ED of the primary Si crystals as a function of the average CR obtained from the UMSA experiments for the Al20% Si alloy test samples. Vertical bars represent the standard deviation values.

Table 2 The ED of the primary Si and the SDAS of the HPDC cylinder block made from the Al20% Si alloy. The corresponding cooling rates were estimated based on the UMSA laboratory experiments Units ED of primary Si Subsurface
m CR ( C/s) 20.6 5.9 72.6

SDAS Subsurface
4.6 0.7 84.7

Centre
20.0 6.2 74.5

Centre
7.9 1.2 48.9

melt quenching by atomized water was the most effective in extracting the heat from the solidifying test sample and resulted in the highest CR of 82.9 C/s. Rapid cooling of the liquid melt most likely resulted in a higher number of nucleation sites for primary Si crystals and restricted their growth by reducing the time necessary for the diffusion process. Consequently, the size of the primary Si crystals was reduced as a function of the CR. The SDAS strongly depends on the increased CR that reduces the time necessary for coarsening of the -aluminum dendrites and results in multiplication of the secondary dendrite arms that distribute the solute content in front of the

Fig. 6 (a) LOM micrograph of the Al20% Si alloy HPDC cylinder block microstructure taken at the bore wall subsurface. Note the primary Si crystals with an average ED of 20.6 5.9 m and the non-equilibrium -aluminum dendrites. (b) LOM micrograph of the Al20% Si alloy HPDC cylinder block microstructure taken from the centre of the bore wall. Note the primary Si crystals with an average ED of 20.0 6.2 m and the non-equilibrium -aluminum dendrites.

Fig. 5 The SDAS as a function of the average CR obtained from the UMSA experiments for the Al20% Si alloy test samples. Vertical bars represent the standard deviation values.

solidifying interface. It was observed that the pockets of -aluminum dendrites were less distinguishable from the non-dendritic metal matrix for a low CR as compared with the test samples that solidied under a high CR. It is believed that the nucleation of the non-equilibrium dendrites was due to rapid solidication (Wang et al., 2004). Similar observations were conrmed for the laboratory test samples cast into the steel mould. The common metallurgical practice is to estimate the microstructural characteristics like SDAS based on the solidication/cooling rates. According to Nishi (2006), the empirical relationship was as follows: SDAS = BCRc where CR was the cooling rate and B and C were constant values. Based on the literature (Nishi, 2006; Salas et al., 2000) it was found that the c value for aluminum based alloys was in the range from 0.3 to 0.4. For example the c value was equal to 0.36 for the commercial die cast alloys like JIS-ADC10 (Al8.5% Si3Cu) and JIS-ADC12 (Al11% Si2.5Cu). For the permanent casting alloy like JIS-AC4C (Al7.12% Si0.31Mg) the c value was approxi-

j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 3 ( 2 0 0 8 ) 333341

339

Fig. 7 (a) LOM micrograph of the Al20% Si alloy HPDC cylinder block microstructure taken at the bore wall subsurface. Note the non-equilibrium -aluminum dendrites with an average SDAS of 4.6 0.7 m. (b) LOM micrograph of the Al20% Si alloy HPDC cylinder block microstructure taken from the centre of the bore wall. Note the non-equilibrium -aluminum dendrites with an average SDAS of 7.9 1.2 m.

Fig. 8 (a) LOM micrograph of the Al20% Si alloy HPDC cylinder block microstructure with the ne AlSi eutectic taken at the bore wall subsurface. (b) LOM micrograph of the Al20% Si alloy HPDC cylinder block microstructure with the coarse AlSi eutectic taken from the centre of the bore wall.

mately 0.49 when cast into the steel die and approximately 0.6 when cast into the copper die (Oonishi et al., 1996). The c value is calculated based on the current results and is approximately 0.6 and appears to be to some extent larger than reported. It is worth mentioning that the c value might change based on the denition of the CR. The data presented in Table 2 was obtained for the average CR calculated between 730 and 380 C. This might explain the difference between the calculated values based on the current data vs. the data cited in the literature.

4.2.

HPDC cylinder block microstructures

Recently, the authors carried out a metallurgical and thermal analysis assessment of the Al20% Si alloy under an approximate 1 C/s solidication rate. It was found that the Liquidus temperature was 691 C, the nucleation temperature of the AlSi eutectic was 567.1 C and the nucleation temperature of Cu, Mg enriched eutectic was 513.6 C (Yamagata et al., 2008). The temperature interval for the development of the primary

Si crystals in the investigated alloy was very wide and was equal to approximately 124 C. The Fraction Solid (FS) vs. Temperature curve analysis revealed that the primary Si increased the volume fraction reaching a maximum of 21.6% at the beginning of AlSi eutectic nucleation temperature (567.1 C). During AlSi eutectic growth the overall FS rose to 70% within a very narrow temperature range of 14 C. It is worth mentioning that these thermal characteristics were obtained for close to equilibrium cooling conditions and that they might be inuenced by the signicantly higher solidication rates in the present UMSA experiments. To date there is no thermal analysis data available that pertains to the highly non-equilibrium solidication conditions. Metallurgical comparison of the investigated test samples indicated more complex microstructural characteristics for the engine bore wall sections as compared with the UMSA designed experimental test samples. This is due to the fact that in the present stage the UMSA simulations did not address the melt thermal conditions in the HPDC shot sleeve and the runner system prior to injection into the die cavity. This aspect is presently being tackled.

340

j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 3 ( 2 0 0 8 ) 333341

Fig. 9 Schematic drawing of the HPDC machine with a description of the major components. The plunger injects the melt from the shot sleeve into the die cavity through the gate and runner.

Metallographic analysis revealed that the engine bore wall sections had statistically insignicant ED differences in the primary Si particles at the subsurface and in the centre. The SDAS and eutectic Si size were signicantly ner at the subsurface and coarser at the centre of the bore section. One factor affecting the sequential nucleation and growth of these structural features was the instantaneous temperature gradient across the bore section associated with the thermal management of the die. However, a higher level of complexity for this phenomenon occurred from the specic alloy solidication path during the HPDC operations that until now was not addressed in the open literature. Fig. 9 schematically illustrates the HPDC system. The holding time of the dosed melt in the shot sleeve was as short as a few seconds prior to injection. However, during this time the melt temperature exhibited a step gradient associated with rapid heat extraction in the shot sleeve which temperature was controlled. Temperature measurements of the melt in the shot sleeve prior to injection revealed approximately 700 C in the centre and 662 C close to the wall. For the investigated Al20% Si alloy the primary Si crystals started to nucleate at 691 C (Liquidus temperature). A further temperature drop increased the FS of the primary Si and reached a maximum of 21.6% at the nucleation temperature of the AlSi eutectic temperature (567.1 C). The melt temperature in the shot sleeve prior to the injection process near the shot sleeve wall was considerably lower than the nucleation temperature of the primary Si. Therefore, the solidication process for a relatively thick layer resulted in the formation of primary Si particles. The melt with some pre-existing primary Si crystals owed through the shot sleeve, gate and runner and most likely further nucleation of the primary Si crystals took place prior to the lling of the die cavity (Fig. 9). Most likely during this period, nucleation of new primary Si crystals and the growth of the already existing primary Si crystals took place progressively. Therefore, the size of these pre-existing primary Si crystals could not be effectively controlled by the accelerated heat transfer after the semi-solid melt entered the die cavity, thus the primary Si particle(s) size was almost identical at the subsurface and in the centre of the bore wall.

The estimated average CR based on the ED measurements of the primary Si for the centre and the subsurface locations were 74.5 and 72.6 C/s respectively. This suggests that the primary Si nucleated prior to the nal solidication stage in the die cavity. However, the estimated CR was based on the SDAS and was signicantly different, i.e., 48.9 and 84.7 C/s for the centre and subsurface locations respectively. The removal of heat from the die surface explained the difference in the SDAS and eutectic Si size and morphology. Rapid cooling generated ne aluminum dendrites and coral-like eutectic Si at the bore wall subsurface, while moderate cooling generated coarse dendrite and needle-like silicon on the centre. It is worth pointing out that most likely the melt temperature before entering the steel die cavity was above the nucleation temperature of the AlSi eutectic, i.e., 567.1 C. For this reason the remaining liquid melt responded well to the high heat extraction capabilities of the steel die. It can be concluded that the high pressure die casting of the Al20% Si alloy included melt injection with the presence of a solid primary Si phase. The whole casting route from the melt holding furnace to the die cavity through the shot sleeve, gate and runner system was responsible for the nal as-cast microstructure. The UMSA thermal data from the computer controlled melting and solidication cycles allowed for the rapid establishment of the statistically condent relationship between the processing parameters and the microstructural analysis results for both the test samples and the actual cast components.

5.

Summary

1. The effect of the average CR on the as-cast microstructures of the Al20% Si alloy was investigated using the novel UMSA platform. Compressed air cooling and atomized water quenching of the thin walled test specimen achieved average cooling rates ranging from 4.9 to 82.9 C/s (calculated between 730 and 380 C). Microstructure measurements revealed that the ED of primary Si decreased from 89.7 17.3 to 16.5 3.8 m and the SDAS from 22.1 5.9 to 5.1 0.8 m with the increase in the CR. These relationships were used to analyze the microstructures of the HPDC cylinder block and to assess the average CR for the given bore wall sections. 2. The HPDC cylinder block metallographic analysis revealed that the primary Si size was nearly identical at the subsurface and at the centre locations of the bore wall. The SDAS of the non-equilibrium -aluminum and the eutectic Si size was signicantly smaller at the bore wall subsurface. This indicated that the temperature gradient across the cylinder bore section had a negligible effect on the primary Si particle size, while the morphology of both the aluminum phase and the eutectic Si strongly depended on the temperature gradient conditions. 3. Based on the UMSA experiments it was found that the primary Si crystals in the engine bore wall section nucleated as a rst phase from the liquid melt at an average CR of approximately 74 C/s. The non-equilibrium -aluminum dendrites at the subsurface of the bore wall nucleated from

j o u r n a l o f m a t e r i a l s p r o c e s s i n g t e c h n o l o g y 2 0 3 ( 2 0 0 8 ) 333341

341

the semi-solid melt at an average CR of approximately 85 C/s and at the centre location at approximately 49 C/s. 4. This research revealed that the primary Si crystals nucleated prior to melt injection into the die cavity, while the non-equilibrium -aluminum phase and the eutectic Si nucleated and grew inside the die cavity. These observations were supported by the presence of a ner microstructure at the bore wall subsurface where the cooling rate was faster. Therefore, it could be concluded that the high pressure die casting of the Al20% Si alloy included the injection of the semi-solid melt.

Acknowledgements
The authors would like to thank Dr. M. Kasprzak from the Electro and Metallurgical Engineering Company (EME, S.C.) in Poland for his valuable contributions to the UMSA technology platforms conguration and experimental design. The authors would also like to acknowledge the long term support for the University of Windsors Light Metals Casting Technology Group by AUTO21, a member of the Networks of Centres of Excellence of Canada for allowing us to undertake this spin-off collaborative research with the Yamaha Motor Co. Ltd., Japan.

references

Backerud, L., Chai, G., Tamminen, J., 1990. Solidication Characteristics of Aluminum Alloys, vol. 2. AFS/Skanaluminium, Stockholm, pp. 338. Chen, X., Kasprzak, W., Sokolowski, J.H., 2006. Reduction of the heat treatment process for al-based alloys by utilization of heat from the solidication process. J. Mater. Process. Technol. 176 (13), 2431. Djurdjevic, M.B., Kasprzak, W., Kierkus, C.A., Kierkus, W.T., Sokolowski, J.H., 2001. Quantication of Cu enriched phases in

synthetic 3XX aluminum alloys using the thermal analysis technique. AFS Trans. 16, 112. Jorstad, J.L., Apelian, D., 2004. Hypereutectic AlSi alloys: practical processing techniques. Die Cast. Eng. 48-3, 5257. Kasprzak, M., Kasprzak, W., Kierkus, W.T., Sokolowski, J.H., 2002. Method and Apparatus for Universal Metallurgical Simulation and Analysis, Patent, PCT/CA02/01903, pp. 111. Kasprzak, M., Kasprzak, W., Kierkus, C.A., Kierkus, W.T., Sokolowski, J.H., Evans, W.J., 2001. Structure and matrix microhardness of the 319 aluminum alloy after isothermal holding during the solidication process. AFS Trans. 16, 111. Kurita, H., Yamagata, H., 2004. Hypereutectic Al20% Si alloy engine block using high-pressure die-casting. In: Proceedings of the SAE 2004 World Congress & Exhibition, Detroit, USA (2004-01-1028). Nishi, N., 2006. Daikasuto Imonono Netsushorini Kansuru Bunkenchousa Houkokusho. Die Cast. 123, 4561 (in Japanese). Oonishi, N., Takaai, T., Nakayama, Y., Ninomiya, K., 1996. Effect of casting conditions on the strength of the AC4CH aluminum casting alloy specimen. J. Jpn. Inst. Light Metals 46, 365370. Salas, G.F., Noguez, M.E., Ramirez, J.G., Robert, T., 2000. Application of secondary dendrite arm spacing cooling rate equation for cast alloys. AFS Trans. 108, 593597. Tsumagari, T., Mobley, C.E., 1995. Determination of skin thickness for 390 AlSi alloy die-casting. AFS Trans. 103, 431436. Wang, H., Ning, Z.Z., Davidson, C.J., StJohn, D.H., Xie, S.S., 2004. Thixotropic structure formation in A390 hypereutectic AlSi alloy. In: Proceedings of the 8th International Conference of Semi-Solid Processing, Limassol, Cyprus. Worcester Polytechnic Institute. Yamagata, H., Kurita, H., 2005. Effect of elastic deformation of the honing stone on the exposure of Si-crystals in a hyper-eutectic-Si aluminum cylinder block. In: Proceedings of the SETC 2005 World Congress & Exhibition, Bangkok, Thailand (SETC 20056577 and SAE 2005-32-0056). Yamagata, H., 2005. The Science and Technology of Materials in Automotive Engines. Woodhead Publishing, Cambridge, pp. 3738. Yamagata, H., Kurita, H., Aniolek, M., Kasprzak, W., Sokolowski, J.H., 2008. J. Mater. Process. Technol. 199, 8490.

You might also like