Fabrication of SiC p /Cu–Al electronic packaging material by pressureless infiltration method

Abstract

The SiC_p/Cu–Al composites with 72·7 vol.-%SiC particles were prepared by pressureless infiltration, and Cu coated SiC particles (SiC/Cu composite powder) were used as reinforcements for aluminium matrix. The effects of moulding pressure, infiltration temperature and infiltration time on the infiltration depth were studied by orthogonal test. The morphology and phase structure of the composites were analysed by scanning electron microscopy and X-ray diffraction. The results show that under moulding pressure of 10 MPa, infiltration temperature of 850°C and infiltration time of 3 h conditions, the SiC_p/Cu–Al composite structure is uniform and dense without obvious porosity defects, and the thermal expansion coefficient is close to thermal expansion coefficient of the Turner model.

Keywords

Eectronic packaging SiC/Cu composite powder Pressureless infiltration Thermal expansion coefficient

Introduction

With excellent high quality of thermal conductivity, low coefficient of thermal expansion (CTE), high strength, high chemical stability, low density and wear resistance, and high volume fraction (40–70 vol.-%), SiC_p/Al composite enjoys great potential for application in the realm of electronic packaging.1^–5 Currently, the infiltration of liquid aluminium alloy into SiC preforms with open pores is one of near net shape process accepted by most material researchers. Pressure infiltration routine has been successfully applied in industry by several companies. Pressureless infiltration attracts considerable attention in recently years, as it has the advantage of low cost, simple practice and great availability of industrialisation. However, the realisation of pressureless infiltration is very difficult in this system owing to the poor wettability between SiC and Al. In addition, Al₄C₃ was formed in uncoated SiC particle reinforced aluminium matrix composites prepared by pressureless infiltration. It is undesirable because Al₄C₃ is a hydrophilic and brittle phase and prone to act as thermal diffusion barrier at the interface. So, the long term stability of uncoated SiC_p/Al composites during thermal cycling and exposure to moist environments remains to be evaluated.

In the present paper, the SiC/Cu composite powders were fabricated by chemical plating technology and pressed into performs, and then, the SiC_p/Cu–Al composites with near net shape and high volume fraction were manufactured by pressureless infiltration. The microstructure and thermal properties of SiC_p/Cu–Al composites were studied.

Experimental

The 6066 aluminium alloy is used as a matrix material. The main chemical composition of the 6066 aluminium alloy is Al–0·13Si–4·13Cu–0·08Zn–0·21Fe–1·58Mg–0·54Mn–0·02Ni–0·04Ti (wt-%). The average size of the β-SiC particulates is ∼10 μm.

The Cu coating on the SiC particles was fabricated via chemical plating technology. The formulation of plating liquid is shown in Table 1. The plating liquid was prepared according to Table 1 and then placed in a water bath at 50°C and was added with a certain amount of SiC particles. Using a constant temperature magnetic stirrer for mechanical stirring, the stirring speed is 20 rev min⁻¹. After 20 min, the final deposition of particles was filtered, rinsed, and dried at 80°C.

Table 1

Formulation of plating liquid

ρ (CuSO₄.5H₂O)/g L⁻¹	ρ (KNaC₄H₄O₆.4H₂O)/g L⁻¹	ρ (EDTA)/g L⁻¹	ρ [K₄Fe(CN)₆.3H₂O]/g L⁻¹	ρ (NaOH)/g L⁻¹	σ (HCHO)/mL L⁻¹
40	42	0·8	2	32	6–10

The SiC/Cu composite powders were processed into performs by mould pressing method. The binder is polyvinyl alcohol. Samples were press formed at 10, 15 and 20 MPa for 30 s respectively. In order to prevent perform deformation and cracking, the method of phased warm drying was used to dry performs in the process. The infiltrated experiment was carried out using a ZM-25-16 vacuum molybdenum wire furnace. The infiltration temperatures are 700, 800 and 850°C, and infiltration times are 1, 2 and 3 h. To establish the optimum preparation condition, the orthogonal experiment was used. The factors and levels of orthogonal design are shown in Table 2.

Table 2

Factors and levels of orthogonal design

Levels	Factors
Levels	Moulding pressure/MPa	Infiltration temperature/°C	Infiltration time/h
1	10	700	1
2	15	800	2
3	20	850	3

The specimen dimension of SiC_p/Cu–Al composite is Φ15×15 mm. The samples were cut longitudinally along the middle line by electrical discharge machining method, and infiltration depth was measured by vernier calliper.

The surface structure of the SiC/Cu composite powders was investigated by D/max-2400 X-ray diffractometer (XRD) and 6460LV scanning electron microscopy. The XRD was also applied to investigate the phase composition of SiC_p/Cu–Al composites. The microstructure of composites was observed by GX-71 optical microscopy. The coefficient of thermal expansion (CTE) of composites was measured by PCY-III coefficient of linear expansion tester.

Results and discussion

Orthogonal experiment

Orthogonal experiment method is a scientific method of arranging and analysing multifactor experiment. The reliable regularity can be discovered by fewer experiments. For the orthogonal array of L₉ (3⁴), it means that the orthogonal experiment has three levels, four factors and nine orthogonal trials. The content of L₉ (3⁴) and the analysis of the experimental results are given in Table 3. Where K_ij is the sum of infiltration depth for factor i at level j in all orthogonal trials, k_ij is the average of K_ij, and R_i is defined as the distance between the extreme values of factor i, which reflects the effect of different levels of factor i on the infiltration depth. If the R of one factor is big, it means different levels of this factor have the great effect on the infiltration depth. Usually, this factor is the main factor. On the contrary, if the R of one factor is small, then it means different levels of this factor have less effect on the infiltration depth. Usually, this factor is the minor factor.

Table 3

Design matrix L₉ (3⁴) of orthogonal experiment and analysis of experimental results

Trials	Factors						Infiltration depth/mm
Trials	Level number	Moulding pressure/MPa	Level number	Infiltration temperature/°C	Level number	Infiltration time/h	Infiltration depth/mm
1	1	10	1	700	1	1	4·21
2	1	10	2	800	2	2	4·46
3	1	10	3	850	3	3	4·74
4	2	15	1	700	2	2	3·78
5	2	15	2	800	3	3	4·35
6	2	15	3	850	1	1	3·32
7	3	20	1	700	3	3	3·56
8	3	20	2	800	1	1	3·10
9	3	20	3	850	2	2	3·23
K1	13·41		11·55		10·63
K2	11·45		11·91		11·47
K3	9·89		11·02		12·65
k1	4·47		3·85		3·54
k2	3·82		3·97		3·82
k3	3·30		3·67		4·22
R	1·17		0·30		0·68

The visual analysis diagrams for each factor are shown in Fig. 1. It can be seen that with the increase in moulding pressure, the infiltration depth decreases gradually. This is because the preforms were prepared by cold briquetting process, and the greater pressure is, the closer pattern materials get. The more contact point and contact surface of accumulation are, the greater the perform density is, and the smaller the porosity is. The molten aluminium is difficult to enter the preforms.

Along with the increase in temperature, the infiltration depth increases firstly, and then decreases. From 700 to 800°C, with the increase in temperature, the infiltration depth increases.

Because of the rise in temperature, on the one hand, the viscosity of molten aluminium alloys decreases, and the mobility is better, On the other hand, the incubation period shortens in the process of infiltration, and the effective infiltration time increases within the same time. Above 800°C, the infiltration depth decreases with increase in temperature, which may be atomic diffusion for copper coating layer of SiC/Cu preforms in incubation period leading to densification. It can reduce the size of the pores, even make them closed, and further drop infiltration rate of liquid aluminium.

The size of SiC/Cu composite powders is varied. There are two kinds of pores in the process of formation of performs. One is the large pore among the neighbouring larger particles; the other is the small pore among smaller particles. In the early infiltration of the preform, molten aluminium alloy mainly fills in the larger pores by fluxion. However, in the process of densification, molten aluminium alloy fills in the smaller pores by capillary action. The dynamics theory of capillary action shows that spontaneous infiltration depth h and time t have the following relationship6 (1) where ΔP is the capillary pressure, λ is geometric factor and K is the penetration rate of the preform. It shows that the infiltration depth is proportional to t^1/2. The molten aluminium alloy infiltrate fully into performs by extending infiltration time appropriately. At the same time, the reaction releases large amounts of heat, which increases the infiltration temperature, reduces the fluidity of molten aluminium and accelerates spreading of aluminium on the surface of particles. Therefore, the infiltration depth increases at the macro level with increasing time.

In the present paper, using SAS software to analyse variance for each parameter, the command ‘PROC ANOVA, CLASS A B C; MODEL y = A B C.’ was entered in SAS programming window; the analysis results are shown in Table 4, where A is the moulding pressure, B is the infiltration temperature, C is the infiltration time, and DF, SS, MS and Sig. respectively represent degree of freedom, standard deviation square, mean square and level of significance test.

Table 4

ANOVA

Source	DF	SS	MS	Sig.
Model	6	2·85948889	0·47658148	<0·0001
Error	2	0·00000000	0·00000000
Corrected total	8	2·85948889
A	2	2·07395556	1·03697778	<0·0001
B	2	0·14569722	0·07284861	<0·0001
C	2	0·68648889	0·34324444	<0·0001

As can be seen from Table 4, the Sig. values are <0·0001 for A, B and C. The results show that three factors have significant influence on infiltration depth. To study the impact of moulding pressure, infiltration temperature and infiltration time on infiltration depth have an extremely important significance.

From Fig. 1 and Table 4, it also can be seen that the optimum conditions are as follows: the moulding pressure is 10 MPa, infiltration temperature is 850°C and infiltration time is 3 h.

As can be seen from the range results, moulding pressure is the most important factor in all process parameters, then infiltration time, and then infiltration temperature. The infiltration depth is the highest for no. 3. Therefore, the no. 3 is optimal condition in the nine orthogonal experiments.

SiC/Cu composite powders

Scanning electron microscopy analysis

To improve the interfacial bonding strength between SiC particles and aluminium matrix and to inhibit the formation of Al₄C₃ resulting from direct contact between the SiC particles and the Al melt, copper was deposited on the surfaces of the SiC particles to obtain the Cu coated SiC particles. Figure 2 illustrates the surface morphology of original SiC particles and SiC/Cu composite powders. In Fig. 2, it can be seen that the Cu coated SiC particles feature anomalous shapes with sharp edges and obvious cleavage steps resulting from mechanical milling. Cu coating is continuous and compact, spreading on the surface of the SiC particles uniformly. Moreover, coating is denser in the edge of SiC particles. This is because the specific surface area of particle edges is large, the surface energy is high and the energy of Cu non-uniform nucleation is smaller, so as to promote the Cu nucleation.

X-ray diffractometer analysis

Figure 3 displays the phase composition of SiC/Cu composite powders by XRD analysis. As shown in Fig. 3, there are two main diffraction peaks: SiC and Cu. The latter diffraction peak is high and sharp. This indicates that the content of Cu is higher in SiC/Cu composite powders, which shows good crystallinity.

SiC_p/Cu–Al composites

Microstructure of SiC_p/Cu–Al composites

The optical microstructure of SiC_p/Cu–Al composites is shown in Fig. 4. As illustrated, the SiC particles are distributed throughout the Al matrix uniformly. Complete infiltration of the packed SiC particles is achieved, and no pores or obvious defects are observed in the composites. The intermediate layer (Cu coating) can be combined with the SiC metallurgically and wetted by Al matrix very well. At the same time, the Cu coating and molten aluminium interact with each other, resulting in interdiffusion and dissolution action and formation of CuAl₂ compounds at the interface, which lead to a decrease in interfacial energy.7 Molten aluminium is more easily impregnated into the preform. The uniform and dense structure are advantageous to the composite material for electronic packaging, which not only can improve the thermal conductivity of the composite materials but also can increase the strength, elastic modulus and dimensional stability and extend the lifecycle of composite materials. Generally speaking, the infiltration quality of the Cu coated SiC/Al composites is excellent, and the SiC particles appear to be well bonded to the Al matrix.

The volume fraction of SiC_p/Cu–Al composites is 72·7 vol.-% under the optimum conditions by Image Pro Plus software. The high SiC particles restrain the thermal expansion of matrix. With the increase in SiC particles, this inhibition is more obvious. The lower the CTE of composite, the better the performance of electronic packaging. The well known Young–Dupre equation is described below (2) where γ is the specific surface energies of subscripted interfaces between solid (s), liquid (l) and gas (g) phases respectively. It shows that when γ_ls and γ_lg reduce, the wetting angle θ decreases. The Cu coating on SiC particles can reduce γ_ls and γ_lg. So, Cu coating results in beneficial improvements in wettability, compared to uncoated SiC.

Phase analysis of SiC_p/Cu–Al composites

Figure 5 shows XRD spectra of SiC_p/Cu–Al composites. As shown in Fig. 5, there are four main diffraction peaks: SiC, CuAl₂, AlCu and Cu₉Al₄. The contents of the phase are 32·95, 44·22, 8·57 and 14·26% respectively. The brittle phase Al₄C₃ is not found in SiC/Cu–Al composites. The Cu coating on SiC particles is shown to be effective to prevent the formation of Al₄C₃. Meanwhile, the Cu coating can enhance the interfacial bonding strength between SiC particles and Al matrix and improve the thermal stability of SiC particles as well.

Thermal expansion coefficient of SiC_p/Cu–Al composites

According to the measure of elongation of the specimen, by equation (3), the relation of CTEs of SiC_p/Cu–Al composites and temperature is shown in Fig. 6 (3) where α is CTE (K⁻¹), L is the original length of specimen (mm), ΔT is the temperature interval (°C) and ΔL is the increment of length in range of ΔT (mm).

It can be seen that the CTEs of the composites increase with increasing temperature (Fig. 6), due to the thermal expansion nature of matrix alloy and the restriction at the interface between the Cu coated SiC and the Al matrix. Compared with CTE of the matrix alloy (23×10⁻⁶ K⁻¹), that of SiC_p/Cu–Al composites is significantly reduced.

On the assumption of only stresses existing in the interface, thermal expansion model for composites advanced by Turner8 could be expressed as follows (4) where α, V and K are CTE (K⁻¹), volume fraction (vol.-%) and bulk elastic modulus (GPa) respectively. The c, m and p subscripts represent composites, metal matrix and SiC reinforcement respectively.

Considering that stress and shear stress both exist in the interface, a mended model represented by Kerner9 is shown as follows (5) (6) (7) (8) where G is the shear modulus (GPa), E is the elastic modulus (GPa) and ν is the Poisson's ratio.

The physical parameters of matrix alloy and silicon carbide particle are shown in Table 5. The experimental CTE of SiC_p/Cu–Al composite (at 80°C) and theoretical CTEs are given in Table 6.

Table 5

Physical parameters of matrix alloy and silicon carbide particle

Material	Thermal expansion coefficient/×10⁻⁶ K⁻¹	Elastic modulus/GPa	Poisson's ratio	Vol.-%
6066 aluminium alloy	23·6	69	0·33	28
SiC	4·0	400	0·20	72

Table 6

Experimental CTE of SiC_p/Cu–Al composite (at 80°C) and theoretical CTEs (×10⁻⁶ K⁻¹)

Turner model	Kerner model	Experimental value
6·07	12·82	6·39

From Table 6, it can be seen that the Kerner model theoretical calculated value is higher than the experimental value. This is mainly because the reinforcing phase is spherical in the model, while the SiC particles are irregular in practice. However, the experimental CTE of SiC_p/Al composite is close to Turner model. In addition, the CTE is lower than that of the matrix alloy and higher than that of SiC, significantly.

Conclusions

Al matrix composites reinforced with Cu coated SiC particles are fabricated by pressureless infiltration successfully. The SiC particles are dispersed in the Al matrix uniformly, and the interfacial bonding between the SiC particles and the Al matrix is tight.

The Cu coating on SiC particles is effective to prevent the formation of Al₄C₃ and to improve the interfacial bonding strength between SiC particles and Al matrix. Furthermore, it solves the problem of wettability between SiC and molten aluminium.

The moulding pressure is the most important factor in all process parameters, then infiltration time, and then infiltration temperature.

The SiC_p/Cu–Al composites are manufactured by pressureless infiltration. The optimum conditions are as follows: moulding pressure of 10 MPa, infiltration temperature of 850°C and infiltration time of 3 h. The volume fraction of SiC particles is 72·7 vol.-% in SiC_p/Cu–Al composites.

The CTE of SiC_p/Cu–Al composites is lower than that of the matrix alloy and higher than that of SiC. It is close to CTE of the Turner model.

Footnotes

Acknowledgements

The author of the present article would like to express profound gratitude to Professor S. M. Du for the valuable comments and helpful cooperation during the course of this research. The author also appreciates valuable services made by all personnel of Xi'an University of Science and Technology, China.

References

Carl

: ‘Advanced materials for optoelectronic packaging’, Optoelectronics, 2002, 42, (9), 37–40.

Kwangjun

, Suk

: ‘Effect of process parameters on the fabrication of Al–SiC composite sheets by atmospheric plasma spraying method’, J. Korean Inst. Met. Mater., 2003, 41, (6), 342–349.

Aghajanian

, Rocazella

, Burke

: ‘The fabrication of metal matrix composites by a pressureless infiltration technique’, J. Mater. Sci., 1991, 26, (2), 447–454.

Aghajanian

, Burke

, White DR , Nagelberg

: ‘A new infiltration process for the fabrication of metal matrix composites’, SAMPE Q, 1989, 20, (4), 43–46.

Johnston

, Young

: ‘Advanced thermal management material’, Int. Newslett. Microsyst. MEMs, 2000, 2, (1), 14–15.

Gui

: ‘High volume fraction SiCP/Al composites pressureless melt infiltration: processing, properties and applications’, Key Eng. Mater., 2003, 249, 45–48.

Zweben

: ‘Metal-matrix composites for electronic packaging’, Int. J. Miner. Metall. Mater., 1992, 44, (7), 25–28.

Sharma

: ‘Effect of albite particles on the coefficient of thermal expansion behavior of the Al 6061 alloy composites’, Metall. Mater. Trans. A, 2000, 31A, (3), 773–780.

Vaidy Rajendra

, Chawla

: ‘Thermal expansion of metalmatrix composites’, Compos. Sci. Technol., 1994, 50, 13–22.