Effect Of Heat Treatments In The Silicon Eutectic Crystal Evolution In Al-Si Alloys

A.Forn1, Mª T. Baile2, E. Martin3 and E. Rupérez4

Light Alloys and Surface Treatments Design Centre (CDAL)

Universitat Politècnica de Catalunya, 08800 Vilanova i la Geltrú. SPAIN.

1 antonio.forn@upc.es,

2 maite.baile@upc.es,

3 enric.martin@upc.es,

4 elisa.ruperez@upc.es

Keywords: Aluminium, SSM, Thixocasting, Rheocasting.

Abstract. This paper describes the heat treatment effect on the eutectic silicon evolution in the

A357 alloy, obtained by semisolid forming process (SSM). The coarsening rate of the silicon was

determined by Image Analysis Technique in specimens from rheocasting ingots and thixocasting

components. The study was realized in the temperature range from 450 to 550ºC by applying

heating times between 1 and 24 hours. The results show that during the heat treatment the

coarsening and sphereodization of the silicon particles is produced and the fragmentation stages,

which are observed in conventional alloys, do not appear. Kinetic silicon growth has been adjusted

to the Oswald’s ripening equation.

Introduction

Semisolid forming is an alternative to the traditional casting processes of aluminium alloys [1].

These processes consist of the material pseudoplasticity control when it keeps in semisolid state.

Under these conditions viscosity decreases with shear rate to which the material is agitated, and the

thixotropic behaviour relies on the shear stress and time. The resultant structure consists of globular

alpha particles dispersed in the eutectic liquid. The ingot obtained by this process and showing these

characteristics, is called rheocasting ingot. For conformation by means of thixocasting, the ingots

are cut at adequate size and are heated to the semisolid range, about 7 minutes at 585ºC, through an

induction furnace. Immediately after, the ingot is automatically handled like a solid to the injection

machine. The resulting microstructure contains a globular alpha phase surrounded by a very fine

eutectic microconstituent. Through heat treatments the eutectic silicon can grow and become

globular as it happens in conventional casting [2]. The formation of silicon spherulites and growth

of larger particles at the expense of smaller ones, and the subsequent redistribution of the crystals in

the Al-matrix, brings about an improvement in plastic behaviour and workability.

Grain growth occurs in polycrystalline materials to decrease the system energy by decreasing the

total grain boundary energy. The total surface energy of the silicon particles is reduced by

coarsening. Under ideals conditions competitive grain growth kinetics should obey a power law

proposed by Hu and Rath [3-4] that describe s the growth of the average particle size as a function

of time, according to the Ostwald ripening model [5]:

ktDDm

O

m =− (1)

where D is the length scale, e.g., mean grain diameter, after annealing for time t, k is a temperature-

dependent rate parameter and DO is the length scale or the grain size at t = 0

ktDDn

O

n =− /1/1 (2)

where n is an empirical constant typically ≤ 0.5 .The grain growth exponent is also often defined as

m = 1/n and is then typically ≥ 2. The value of m depends on the grain growth mechanism. The

classical value of grain growth exponent for pure metals or ceramics is 2. The coarsening exponent

Materials Science Forum Vols. 480-481 (2005) pp. 367-372online at http://www.scientific.net© 2005 Trans Tech Publications, Switzerland

Licensed to Martín (enric.martin@upc.es) - SpainAll rights reserved. No part of the contents of this paper may be reproduced or transmitted in any form or by any means without thewritten permission of the publisher: Trans Tech Publications Ltd, Switzerland, www.ttp.net. (ID: 147.83.156.214-29/03/05,08:44:39)

for classical theoretical papers on Oswald ripening [6], of grains in a stable polycrystalline matrix, is

m= 3 in case of volume diffusion controlled growth, m = 4 if the grain growth is controlled by grain

boundary diffusion, and m = 5 for diffusion on dislocations; but in practice much higher values for

m (5-11) are often found [7]. These higher values of m are usually ascribed to particle growth due to

coalescence. In spite of the many studies devoted to this subject there is, however, still no

satisfactory understanding of this ageing behaviour.

Assuming that Do<<D

n

ktD = (3)

The growth rate k varies with temperature according to an Arrhenius-type equation of the form [8]

−=RT

Qkk o exp (4)

from the rate parameter k one gets access to the activation energy Q, which can be used to deduce

the grain growth mechanism.

Experimental procedure

The material used in the present study was Al alloy A357 manufactured by rheocasting and

thixocasting processes. This alloy contains 7,75 wt-%Si, 0,65 wt-%Mg and 0,024 wt-%Sr. Two

types of materials were tested. A rheocast A357 aluminium alloy ingot was the starting material.

Two areas were studied in this material: periphery and core of ingot. After, the A357 ingot was

formed in semisolid state using a thixocasting method. The semisolid state was obtained in an

induction industrial furnace heating the ingot for 7 minutes before water quenching. The specimens

obtained using this procedure are named component. This component has a homogeneous

microstructure and that is why only a zone is studied.

The heat treatments were carried out on samples with a volume of 2 cm3 in a tubular electric ∅

50mm furnace, with a heat treating region about 200mm. The samples were being heated to 450,

500 and 550oC in the range from 15 minutes to 24 h and immediately after they were cooled in the

air.

By means of image analysis the equivalent diameter and the aspect ratio of eutectic silicon were

determined. A statistical analysis was made starting from a study of 20 regions belonging to every

sample.

Results and Discussion

In Fig. 1 the ingot microstructure is shown. The first micrography shows a globular α-phase and a

very thin eutectic; this structure corresponds to the ingot centre. The second one reveals the ingot

periphery with a greater size dendritic structure and the eutectic ratio becomes smaller. The

component microstructure is observed in Fig. 2. The silicon size is larger and besides, the ingot

heating provokes the appearance of fine eutectic spheroids inside the α grains. That is because

when the temperature increases over the eutectic isotherm, the silicon solubility in the α phase

decreases, and liquid drops are produced. These occluded drops in the α grains suffer an

undercooling process and solidify fast [9]. Obviously, this phenomenon isn’t carried out in the

liquid surrounding the α-phase.

Cross-Disciplinary Applied Research in Materials Science and Technology368

a) b)

Fig. 3 Microstructure of Rheocasting Ingot at 450ºC, 6h: a) core and b) periphery

The micrographs obtained after several heat treatments, at 450oC, are shown in Fig. 3

These treatments produce a change in matrix composition, the silicon particles size increase and

become globular. It is possible to see (Fig. 3) how silicon becomes spherical and grows quickly.

This fact does not only take place at the core but at the periphery, with different speed in everyone.

Besides, a shape and size variation of the α phase spheres is also to be observed. The α-phase grows

increasing temperature and time of annealing. The original grain boundaries are defined by silicon

eutectic crystals and they are blurred as the silicon crystals grow and coalesce the α-phase grains. As

a) b)

Fig. 1 Rheocasting ingot: a) core and b) periphery

Fig. 2 Microstructure of Thixocasting

component

Materials Science Forum Vols. 480-481 369

Component

0,0

0,5

1,0

1,5

2,0

2,5

3,0

0 5 10 15 20 25 30

Time /h

Dia

me

ter

/ µµ µµm

450 ºC

500 ºC

550 ºC

the treatment time goes on, the big silicon crystals become an homogeneous distribution inside the

continuous alpha matrix .

The silicon equivalent diameter evolution at different temperatures, in the component, has been

described in figure 4 and the corresponding power law is detailed in table 1. The silicon grain size

grows quickly at the first heating phase and reaches values between 1,5 µm in the core and 1,7 µm

in the periphery.

By applying the equation 3 the m value from table 1 is obtained. For a mean value of n=0,20 and

applying other time the equation 3, the values of kT are obtained (Fig. 5). These values are related

in Table 1.

a) b)

c)

Table 1

Temperature /ºC Function (INGOT CORE) m kT

450 23,0409,0 tD = 4.3 0.409

500 15,0605,0 tD = 6.6 0.605

550 20,0774,0 tD = 5.0 0.774

Function (INGOT PERIPHERY)

450 20,0565,0 tD = 5.0 0,565

500 23,0603,0 tD = 4.3 0,603

550 19,0973,0 tD = 5.2 0,973

Function (COMPONENT)

450 21,0720,0 tD = 4.7 0.721

500 22,0924,0 tD = 4.5 0.935

550 20,0394.1 tD = 5.0 1.347

Fig. 4 The silicon equivalent diameter evolution at different T and times: a) Ingot core,

b) Ingot periphery and c) component

Ingot (periphery).

0,0

0,2

0,4

0,6

0,8

1,0

1,2

1,4

1,6

1,8

0 5 10 15 20 25 30

Time /h

D /

µµ µµm

550 ºC

500 ºC

450 ºC

Ingot (core)

0,0

0,2

0,4

0,6

0,8

1,0

1,2

1,4

1,6

0 5 10 15 20 25 30

Time /h

Dia

me

ter

/ µµ µµm

550 ºC

500 ºC

450 ºC

Cross-Disciplinary Applied Research in Materials Science and Technology370

k and Q/R values in the equation 4 are determined by the graphic function of k in front of 1/T

(Fig. 6).

The obtained activation energy values appear in table 2

Ingot (core) y = 1,0171x - 0,2558

R2 = 0,901

y = 0,7812x - 0,5026

R2 = 0,9984y = 1,1656x - 0,8932

R2 = 0,9846

-1

-0,8

-0,6

-0,4

-0,2

0

0,2

0,4

0,6

0 0,2 0,4 0,6 0,8

n lnt

lnD

450 ºC 500 ºC 550 ºC

Ingot core

0,0

0,2

0,4

0,6

0,8

1,0

1,2

1,4

0,0012 0,0013 0,0014

1/T

k

−=T

k1

3808exp7.80

Ingot (periphery) y = 0,9607x - 0,0272

R2 = 0,9115

y = 1,1737x - 0,5053

R2 = 0,8223

y = 1,0249x - 0,5702

R2 = 0,9166

-0,8

-0,6

-0,4

-0,2

0

0,2

0,4

0,6

0,8

0 0,2 0,4 0,6 0,8

n lnt

lnD

450 ºC 500 ºC 550 ºC

Ingot periphery

0,0

0,2

0,4

0,6

0,8

1,0

1,2

1,4

0,0012 0,0013 0,0014

1/T

k

−=

Tk

13177exp6.42

Component y = 1,0867x + 0,2985

R2 = 0,9177

y = 1,0793x - 0,0668

R2 = 0,9488

y = 1,0606x - 0,3272

R2 = 0,9624

-0,4

-0,2

0

0,2

0,4

0,6

0,8

1

1,2

0 0,2 0,4 0,6 0,8

n lnt

lnD

450 ºC 500 ºC 550 ºC

Component

0,0

0,2

0,4

0,6

0,8

1,0

1,2

1,4

0,0012 0,0013 0,0014

1/T

k

−=T

k1

3700exp118

c)

Fig. 5 Graphic function of diameter in

front of 1/T: a) ingot core, b) ingot

periphery and c) component

c)

Fig. 6 Graphic function of k in

front of 1/T: a) ingot core,

b) ingot periphery and

c) component

Materials Science Forum Vols. 480-481 371

Table 2

Q/R Q kJ/mol

ingot core 3808 32

ingot

periphery 3177 26

component 3700 31

Conclusion

The resulting average activation energy to the silicon coarsening is 30 kJ.mole-1. This value is in

concordance with the one described in the bibliography [10], which is Q = 32.75 kJ/mol in the

range of temperatures 480 ÷ 620 ºC. Therefore the silicon growth exponent is 5 and, consequently,

the mechanism can be considered as volume diffusion in matrix.

References

[1] M. C. Flemings: Metallurgical Transactions Vol. 22A (1991), p. 957

[2] A. J. Criado, J. A. Martinez, J. M. Gómez de Salazar, F. Molleda: Praktische-Metallographie,

24(4) (1987), p. 175

[3] H. Hu, B. B. Rath: Met. Trans. 1, (1970), p. 3181

[4] J. D. Verhoeven: Fundamentos de Metalurgia Física (Limusa, Mexico 1987).

[5] R. E. Smallman: Modern Physical Metallurgy 4th

ed (Butterworth, London 1985).

[6] V. S. Solomatov: Lunar and Planetary Science XXXIII, (2002).

[7] J. J. Martin: Thesis: An Atomic Scale View on a Model Catalyst Pd Nanoparticles on TiO2

(2002).

[8] L. M. Lifschitz, V. V. Slyozov: J. Phys. Chem. Solids,19, (1961), p. 35

[9] J. M. Bastidas, A. Forn, M. T. Baile, J. L. Polo and C. L. Torres:Materials and Corrosion 52,

(2001), p. 691

[10] E. A. Brandes: Smithells Metalls Reference Book, 6th

Ed, (Ed. Butterworths, 1983)

Cross-Disciplinary Applied Research in Materials Science and Technology372