**3.1 As-cast wheel**

### **3.1.1 Thermography measurements**

A series of infrared (IR) thermographs was taken during wheel ejection from the top die and just prior the wheels were water cooled, to obtain 2D temperature maps of the casting surface. Fig. 4 shows an IR image of the wheel surface and the top die. The wheel stays on the transfer tray set under the top die. The top die shows a temperature between 435 and 495°C and the highest temperature is concentrated at the ingate, i.e. the hub region. These thermal values are comparable with the reading of the thermocouples which are located few mm under the die surface. The small differences can be related to the emissivity coefficient set up in the IR camera or some small variations in the experimental process parameters, influencing the thermal evolution and distribution of the die. High temperatures are localised in the die around the thickest regions of the casting, i.e. the hub and the spoke (to some extent), where the die receives a great quantity of solidification heat. Contrary, the surface temperature of the wheel shows a temperature range of 340-420°C, with the lowest temperature in the rim and the highest in the zone of the wheel between the spoke and the rim (Fig. 4). Under these conditions, the wheel was then automatically picked up by a robot and cooled.

Fig. 4. Infrared thermal mapping of the wheel during extraction from die

The amount of distortions of the wheels was carried out after post-cast cooling (ε) and after quenching (εt), by using a circular gauge, which allows to calculate the maximum variation of the diameter of the wheel along the rim. Generally, the maximum accepted distortion of a wheel is 1.5 mm, while wheels with higher distortions are normally rejected. This is a typical

The methodology to analyse and optimize the quality of A356-T6 18-inch wheels, in terms of casting distortion and hardness, and to optimise the whole process manufacturing is based

A series of infrared (IR) thermographs was taken during wheel ejection from the top die and just prior the wheels were water cooled, to obtain 2D temperature maps of the casting surface. Fig. 4 shows an IR image of the wheel surface and the top die. The wheel stays on the transfer tray set under the top die. The top die shows a temperature between 435 and 495°C and the highest temperature is concentrated at the ingate, i.e. the hub region. These thermal values are comparable with the reading of the thermocouples which are located few mm under the die surface. The small differences can be related to the emissivity coefficient set up in the IR camera or some small variations in the experimental process parameters, influencing the thermal evolution and distribution of the die. High temperatures are localised in the die around the thickest regions of the casting, i.e. the hub and the spoke (to some extent), where the die receives a great quantity of solidification heat. Contrary, the surface temperature of the wheel shows a temperature range of 340-420°C, with the lowest temperature in the rim and the highest in the zone of the wheel between the spoke and the rim (Fig. 4). Under these

conditions, the wheel was then automatically picked up by a robot and cooled.

Fig. 4. Infrared thermal mapping of the wheel during extraction from die

standard used for wheel manufacturing (Manente, 2008).

**3. Results and discussion** 



**3.1.1 Thermography measurements** 


on different steps:

**3.1 As-cast wheel** 

### **3.1.2 Microstructural observations of as-cast wheels**

The microstructure of the modified A356 alloy consists of a primary phase, α-Al solid solution, and an eutectic mixture of aluminium and silicon. The α-Al precipitates from the liquid as the primary phase in the form of dendrites. The scale of microstructure in different zones of the wheel was characterized by means of SDAS measurements. The coarseness of the microstructure varied inversely with the casting thickness, i.e. the solidification rate. Typical microstructure of the as-cast wheel is shown in Fig. 5, referred to the hub, spoke and rim zones, corresponding to 55, 36 and 22 μm in SDAS respectively. Local solidification times (tf) were estimated by means of SDAS measurements through the following relationship (Dantzig & Rappaz, 2009):

$$\text{SDAS} = 5.5 \left( -\frac{\Gamma\_{\text{sl}} \,\mathrm{D}\_{\text{l}} \ln \left( \frac{\mathbf{C}\_{\text{out}}}{\mathbf{C}\_{\text{o}}} \right)}{\mathrm{m}\_{\text{l}} (1 - \mathbf{k}\_{\text{o}}) (\mathbf{C}\_{\text{out}} - \mathbf{C}\_{\text{o}})} \,\mathrm{t}\_{\text{t}} \right)^{\mathrm{K}} \tag{1}$$

where Γsl is the Gibbs-Thomson coefficient, Dl the diffusion coefficient in liquid, ml the slope of the liquidus curve, k0 the partition coefficient, C0 and Ceut are the initial alloy concentration and the eutectic composition respectively. The solidification time was estimated to be 184 s in the hub, 52 s in the spoke and 12 s in the rim zone. The solidification sequence is approximately directional, starting at the outermost point of the wheel (rim) and continuing toward the centre of the wheel (hub), where the ingate is located.

Fig. 5. Microstructure of as-cast wheel with reference to the different positions analysed

Optimizing the Heat Treatment Process of Cast Aluminium Alloys 205

**20 40 60 80 100 Water temperature (°C)**

Fig. 7. Wheel distortion as a function of the cooling medium, i.e. air and water at different temperature, after ejection from the die; standard deviations are given as error bars

perfect. In the considered range of water temperature, the distortion ε can be described

 0.009 T 1.342 ε =− ⋅ + (3) where T is the water temperature in °C. Every wheel cooled in air and water shows a certain warp degree, which is more or less evident. The amount of distortions can reach critical level that compromises the functionality of the wheel too, as shown in Fig. 8. Residual

Fig. 8. Wheel distortion after post-cast cooling in water at a temperature of 30°C

**Linear fit (R2=0.94)**

**0**

**Air**

according to the following regression model (R2 = 0.94):

**0.2**

**0.4**

**0.6**

**0.8**

**Distortion,** ε **(mm)**

**1**

**1.2**

**1.4**

Coarse intermetallics compounds, such as Mg-rich particles and Fe-rich intermetallics, both in the form of coarse α-Al(FeMnSi) particles and needle-shaped β-Al5FeSi, were also observed, especially in the hub region where the solidification rate is lower (Fig. 6).

Fig. 6. Optical micrograph showing secondary phase particles in hub region; the eutectic silicon particles are present in the interdendritic channels, β-Al5FeSi phase appear with typical needle shape and β-Mg2Si particles as Chinese script

As it has been well established (Apelian et al., 2009; Kashyap et al.,1993), the eutectic Si phase in the microstructure of the Sr-modified alloy exhibits a fibrous morphology under ascast conditions. The mean equivalent diameter *d* of eutectic Si particles increases approximately from the rim (~0.8 μm) toward the spoke (~1.6 μm) and the hub region (~2 μm). It was established that by reducing the cooling rate, the microstructure is characterised by coarse eutectic Si particles, while by reducing the solidification time the formation of a high number of fine Si particles is predominant. Further, the size distribution of Si particles was investigated in several studies (Tiryakioglu, 2008; Shivkumar et al., 1989; Grosselle et al., 2009) and found to follow the three-parameters lognormal distribution as follows

$$f(d) = \frac{1}{(d-\tau)\sigma\sqrt{2\pi}} \exp\left[\frac{-\left(\ln(d-\tau) - \mu\right)^2}{2\sigma^2}\right] \tag{2}$$

where *d* is the diameter of Si particles, τ the threshold, σ the shape and μ is the scale parameter.

#### **3.1.3 Distortion behaviour in the as-cast temper**

The different cooling media produced different amount of distortions in the 18-inch wheels. Generally, the distortion was in the range between 0.6 and 1.1 mm. Fig. 7 compares the wheel distortion induced by air or water cooling. It is evidenced how water temperature higher than 70°C produces similar distortions as air cooling (ε ~0.6 mm).

By increasing the water temperature, the amount of distortions decreases. This relationship has been estimated by linear regression analysis, using the coefficient of the determination R2 to evaluate the quality of the least-squares fitting (Fig. 7). When R2 is equal to 1, the fit is

Coarse intermetallics compounds, such as Mg-rich particles and Fe-rich intermetallics, both in the form of coarse α-Al(FeMnSi) particles and needle-shaped β-Al5FeSi, were also

observed, especially in the hub region where the solidification rate is lower (Fig. 6).

Fig. 6. Optical micrograph showing secondary phase particles in hub region; the eutectic silicon particles are present in the interdendritic channels, β-Al5FeSi phase appear with

As it has been well established (Apelian et al., 2009; Kashyap et al.,1993), the eutectic Si phase in the microstructure of the Sr-modified alloy exhibits a fibrous morphology under ascast conditions. The mean equivalent diameter *d* of eutectic Si particles increases approximately from the rim (~0.8 μm) toward the spoke (~1.6 μm) and the hub region (~2 μm). It was established that by reducing the cooling rate, the microstructure is characterised by coarse eutectic Si particles, while by reducing the solidification time the formation of a high number of fine Si particles is predominant. Further, the size distribution of Si particles was investigated in several studies (Tiryakioglu, 2008; Shivkumar et al., 1989; Grosselle et al., 2009) and found to follow the three-parameters lognormal distribution as

> 1 (ln( )) ( ) exp ( )2 <sup>2</sup> ⎡− −τ −μ <sup>⎤</sup> <sup>=</sup> <sup>⎢</sup> <sup>⎥</sup> −τ σ π ⎣ σ ⎦

where *d* is the diameter of Si particles, τ the threshold, σ the shape and μ is the scale

The different cooling media produced different amount of distortions in the 18-inch wheels. Generally, the distortion was in the range between 0.6 and 1.1 mm. Fig. 7 compares the wheel distortion induced by air or water cooling. It is evidenced how water temperature

By increasing the water temperature, the amount of distortions decreases. This relationship has been estimated by linear regression analysis, using the coefficient of the determination R2 to evaluate the quality of the least-squares fitting (Fig. 7). When R2 is equal to 1, the fit is

higher than 70°C produces similar distortions as air cooling (ε ~0.6 mm).

2

2

*<sup>d</sup> f d d* (2)

typical needle shape and β-Mg2Si particles as Chinese script

**3.1.3 Distortion behaviour in the as-cast temper** 

follows

parameter.

Fig. 7. Wheel distortion as a function of the cooling medium, i.e. air and water at different temperature, after ejection from the die; standard deviations are given as error bars

perfect. In the considered range of water temperature, the distortion ε can be described according to the following regression model (R2 = 0.94):

$$
\varepsilon = -0.009 \cdot \text{T} + 1.342 \tag{3}
$$

where T is the water temperature in °C. Every wheel cooled in air and water shows a certain warp degree, which is more or less evident. The amount of distortions can reach critical level that compromises the functionality of the wheel too, as shown in Fig. 8. Residual

Fig. 8. Wheel distortion after post-cast cooling in water at a temperature of 30°C

Optimizing the Heat Treatment Process of Cast Aluminium Alloys 207

 Fig. 9. Eutectic Si particles in the hub region of A356-T6 18-inch wheels; the alloy has been solubilised at 540°C for various time: (a) as-cast, (b) 4, (c) 5, (d) 6, (e) 7 and (f) 8 hours. Silicon particles undergo necking and are broken down into fragments, then, spheroidization and

Fig. 10b shows the distribution of the eutectic Si particles as a function of the shape factor for samples heat treated at 540°C for various time. Pedersen observed how the particles undergo great changes in shape factor *α* distribution after short times (30 minutes) of solution heat treatment; the fraction of particles with a smaller *α* parameter is immediately reduced, while the number of particles with a greater *α* parameter is increased. Similar

coarsening mechanisms occur

stresses originate from differential thermal gradient and contraction during post-cast cooling. The wheel is extracted from the die at high temperature, as previously shown, and rapidly cooled. Therefore, the stress is so high that plastic deformation in the casting, free from the die, occurs. Generally, the casting distortion is more pronounced in casting ejected from the die at high temperature and in components showing drastic thickness changes (ASM Handbook, 1991). Further, higher the temperature difference between the casting and the cooling medium, greater will be the residual stresses and the casting distortion (Bates, 1987).

### **3.2 Solution heat treatment**

### **3.2.1 Evolution of eutectic silicon particles**

The influence of the solution heat treatment time on the microstructure of 18-inch wheels is shown in Fig. 9. The micrographs refer to the hub, which is the thickest zone of the wheel with a coarse microstructure, SDAS ~55 μm. In the range of solution temperature and times used, and due to the high diffusion rate of Mg in the α-Al matrix, the Mg bearing phases are completely dissolved and not more evident even in the coarse microstructure of the hub. These findings are in agreement with the results reported elsewhere (Rometsch et al., 1999; Zhang et al., 2002).After 4 h of solution heat treatment at 540°C, the Si particles become coarser and the interparticle distance increases (Fig. 9b). Rayleigh instability occurs; silicon particles undergo necking and are broken down into fragments. Due to the instability of the interfaces between the two different phases and a reduction in the total interface energy, spheroidization and coarsening processes occur. A prolonged solution treatment leads to extensive coarsening of the particles, with a small effect on the spheroidization level (Fig. 9cf). The interparticle spacing increases too. Because the coarsening and spheroidization are diffusion-controlled processes (Greenwood, 1956), they are directly proportional to the solution temperature and time. These findings are in agreement with Meyers (Meyers, 1985). Further, previous results (Zhang et al., 2002) showed there exists a decrease in average Si crystal size after short solution heat treatment, before the average size increases. From the literature, the most severe coarsening of eutectic Si particles takes place between 25 and 400 minutes of solution treatment of the unmodified alloy, while the average particle size increased more evenly in the modified alloy (Pedersen, 1999). It has been stated that the typical growth rates for gravity die castings are in the range of 0.02 to 0.07 μm/h (Pedersen, 1999).

The results of the Pedersen's work on the quantitative variation in the Si particle size and shape factor of an AlSi7Mg0.3 alloy with similar microstructural scale as the hub of the wheel (SDAS ~54 μm) as a function of solution time are reported in Fig. 10. The Si growth is estimated in terms of variation of the equivalent radius with respect to t1/3, as defined by the ordinary Lifshitz-Slyozow-Wagner model (Liftshitz & Sloyozov, 1961):

$$\mathbf{R}^3 - \mathbf{R}\_0^3 = \frac{8}{9} \frac{\mathbf{D} \cdot \mathbf{C}\_0 \mathbf{\mathcal{V}}^2}{\mathbf{R}\_{gas} \mathbf{T}} \mathbf{t} \tag{4}$$

where T and t are the temperature and time, respectively; R is the radius of the particle; R0 is the initial radius at t=0; Rgas is the gas constant; V is the molar volume; C0 is the equilibrium concentration of structures in matrix; γ is the surface energy of the particle; and D is the diffusion coefficient. The regression analysis leads to R2 equal to 0.97, indicating the reliability of the model.

stresses originate from differential thermal gradient and contraction during post-cast cooling. The wheel is extracted from the die at high temperature, as previously shown, and rapidly cooled. Therefore, the stress is so high that plastic deformation in the casting, free from the die, occurs. Generally, the casting distortion is more pronounced in casting ejected from the die at high temperature and in components showing drastic thickness changes (ASM Handbook, 1991). Further, higher the temperature difference between the casting and the cooling medium, greater will be the residual stresses and the casting distortion (Bates,

The influence of the solution heat treatment time on the microstructure of 18-inch wheels is shown in Fig. 9. The micrographs refer to the hub, which is the thickest zone of the wheel with a coarse microstructure, SDAS ~55 μm. In the range of solution temperature and times used, and due to the high diffusion rate of Mg in the α-Al matrix, the Mg bearing phases are completely dissolved and not more evident even in the coarse microstructure of the hub. These findings are in agreement with the results reported elsewhere (Rometsch et al., 1999; Zhang et al., 2002).After 4 h of solution heat treatment at 540°C, the Si particles become coarser and the interparticle distance increases (Fig. 9b). Rayleigh instability occurs; silicon particles undergo necking and are broken down into fragments. Due to the instability of the interfaces between the two different phases and a reduction in the total interface energy, spheroidization and coarsening processes occur. A prolonged solution treatment leads to extensive coarsening of the particles, with a small effect on the spheroidization level (Fig. 9cf). The interparticle spacing increases too. Because the coarsening and spheroidization are diffusion-controlled processes (Greenwood, 1956), they are directly proportional to the solution temperature and time. These findings are in agreement with Meyers (Meyers, 1985). Further, previous results (Zhang et al., 2002) showed there exists a decrease in average Si crystal size after short solution heat treatment, before the average size increases. From the literature, the most severe coarsening of eutectic Si particles takes place between 25 and 400 minutes of solution treatment of the unmodified alloy, while the average particle size increased more evenly in the modified alloy (Pedersen, 1999). It has been stated that the typical growth rates for gravity die castings are in the range of 0.02 to 0.07 μm/h (Pedersen,

The results of the Pedersen's work on the quantitative variation in the Si particle size and shape factor of an AlSi7Mg0.3 alloy with similar microstructural scale as the hub of the wheel (SDAS ~54 μm) as a function of solution time are reported in Fig. 10. The Si growth is estimated in terms of variation of the equivalent radius with respect to t1/3, as defined by the

> 3 3 0 0

8DC V R R <sup>t</sup> 9RT

where T and t are the temperature and time, respectively; R is the radius of the particle; R0 is the initial radius at t=0; Rgas is the gas constant; V is the molar volume; C0 is the equilibrium concentration of structures in matrix; γ is the surface energy of the particle; and D is the diffusion coefficient. The regression analysis leads to R2 equal to 0.97, indicating the

2

γ − = (4)

gas

ordinary Lifshitz-Slyozow-Wagner model (Liftshitz & Sloyozov, 1961):

1987).

1999).

reliability of the model.

**3.2 Solution heat treatment** 

**3.2.1 Evolution of eutectic silicon particles** 

Fig. 9. Eutectic Si particles in the hub region of A356-T6 18-inch wheels; the alloy has been solubilised at 540°C for various time: (a) as-cast, (b) 4, (c) 5, (d) 6, (e) 7 and (f) 8 hours. Silicon particles undergo necking and are broken down into fragments, then, spheroidization and coarsening mechanisms occur

Fig. 10b shows the distribution of the eutectic Si particles as a function of the shape factor for samples heat treated at 540°C for various time. Pedersen observed how the particles undergo great changes in shape factor *α* distribution after short times (30 minutes) of solution heat treatment; the fraction of particles with a smaller *α* parameter is immediately reduced, while the number of particles with a greater *α* parameter is increased. Similar

Optimizing the Heat Treatment Process of Cast Aluminium Alloys 209

right toward higher *α* values. Only after 8 hours solution time, the shape factor distribution moves to higher *α* values. The eutectic Si particles in AlSi7Mg gravity-cast alloys crack progressively with increasing applied plastic deformation, and the crack is favourable for the larger and longer particles, even if the progression of particle cracking is more gradual in a finer microstructure (Cáceres & Griffiths, 1996). In addition, it was observed that the population of cracked particles is distributed according to the *α·d* parameter and is

Since solidification rate has a dramatic effect on the size and morphology of eutectic Si particles, it is important to be aware of the influence of the solidification rate on the required minimum solution time for realizing the required coarsening and spheroidization. It was reported (Shivkumar et al., 1990c) that 3-6 h at 540°C is the optimal time for a Sr-modified sand-cast A356 alloy; while 30 min at 540°C is needed for a low-pressure die-cast Sr-modified A356 alloy with SDAS of 25 μm (Zhang et al., 2002). Fig. 11 shows the effect of a solution treatment at 540 °C for 6 h on the Si particle size in the different positions of the wheels ,where different microstructural scales were observed. The coarsening mechanism is faster in the rim and spoke region, where SDAS is about 22 and 36 μm respectively. While the coarse microstructure of the hub presents slower coarsening of Si particles, as indicated

by the values of equivalent diameter in the as-cast and solution heat treated temper.

The increase of solution temperatures for the heat treatment of the wheels would be desirable since it increases the diffusion rate of Si atoms in the Al matrix, leading to rapid fragmentation and coarsening mechanism of eutectic Si particles, and, therefore, to shorten the total time of the T6 heat treatment cycle. It was demonstrated that for a given short solution treatment time of 9.5 minutes, increasing the temperature from 540 to 550°C the number fraction of Si particles with a diameter of greater than 1 μm increases by more than 10%. Similar changes in the distribution of the shape factor for Si particles are observed by increasing the solution temperature, that is the number fraction of the particles with a shape factor of greater than 0.5 increases by approximately 10% (Zhang et al., 2002). Earlier works (Shivkumar et al., 1990b) showed that extremely high coarsening occurred at temperatures greater than 540°C for A356.2 alloys. However, the major problem associated with higher heat treatment temperatures remains the liquid phase formation, which increases with

In the present work, the possibility to heat the wheels at higher solution temperature was evaluated. A Fourier thermal analysis was carried out to determine the evolution of the solid fraction during solidification of the A356 alloy used for wheel production. A detailed description of the equipment, the casting procedure, and the process parameters is given elsewhere (Piasentini et al., 2005). The relationship between fraction of solid (fs) and temperature of solidifying A356 alloy is shown in Fig. 12 for a cooling rate of 1°C/s. With increasing solution temperature above 540°C (final solidification point), the amount of liquid phase (100 fs) increases slowly at first and then rapidly near the Al-Si eutectic reaction

At relatively lower solution temperatures, melting starts at grain boundaries and interdendritic regions. In alloys with a dendritic structure, local melting starts generally at interdendritic channels, since these often contain high concentrations of alloying elements/impurities. At higher solution temperatures, local melting may also start at grain

of ~560°C, at which point the fraction of liquid (100-fs) is about 15%.

characterized by its average *α·d* value.

**3.2.2 Partial melting** 

temperature.

Fig. 10. (a) Linear regression analysis of eutectic Si equivalent radius with t1/3; the point zero in the time axis represents the as-cast condition; (b) frequency distribution of the shape factor *α* after solution treatment at 540°C for different times (Pedersen, 1999)

Fig. 11. Average diameter *d* of the eutectic Si particles as a function of SDAS; data refer to the different positions of the as-cast and solution heat treated wheels

changes in particle distribution are not observed by increasing the solution times within 4 hours, even if the distribution curves flatten with solution time and their peaks move to the right toward higher *α* values. Only after 8 hours solution time, the shape factor distribution moves to higher *α* values. The eutectic Si particles in AlSi7Mg gravity-cast alloys crack progressively with increasing applied plastic deformation, and the crack is favourable for the larger and longer particles, even if the progression of particle cracking is more gradual in a finer microstructure (Cáceres & Griffiths, 1996). In addition, it was observed that the population of cracked particles is distributed according to the *α·d* parameter and is characterized by its average *α·d* value.

Since solidification rate has a dramatic effect on the size and morphology of eutectic Si particles, it is important to be aware of the influence of the solidification rate on the required minimum solution time for realizing the required coarsening and spheroidization. It was reported (Shivkumar et al., 1990c) that 3-6 h at 540°C is the optimal time for a Sr-modified sand-cast A356 alloy; while 30 min at 540°C is needed for a low-pressure die-cast Sr-modified A356 alloy with SDAS of 25 μm (Zhang et al., 2002). Fig. 11 shows the effect of a solution treatment at 540 °C for 6 h on the Si particle size in the different positions of the wheels ,where different microstructural scales were observed. The coarsening mechanism is faster in the rim and spoke region, where SDAS is about 22 and 36 μm respectively. While the coarse microstructure of the hub presents slower coarsening of Si particles, as indicated by the values of equivalent diameter in the as-cast and solution heat treated temper.
