E L S E V I E R Materials Science and Engineering A210 (1996) 114-122

MATERIALS SCIENCE &

ENGINEERING

A

Effect of type of processing on the microstructural features and mechanical properties of A1-Cu/SiC metal matrix composites

M. Gupta, M.O. Lai, C.Y. Soo Department of Mechanical and Production Engineering, National University of Singapore, 10 Kent Ridge Crescent, Singapore, Singapore

Received 1 May 1995; in revised form 24 October 1995

Abstract

In this study, an aluminum based metallic matrix (A1-2wt.% Cu) was reinforced with SiC particulates using a conventional casting technique and a new disintegrated melt deposition technique. Microstructural characterization studies conducted on the samples taken from disintegrated melt deposition technique revealed a more uniform distribution of SiC particulates and good interfacial integrity between SiC particulates and metallic matrix when compared to the conventionally cast composite samples. Results of ambient temperature mechanical tests demonstrate an increase in 0.2% YS and ultimate tensile strength of samples taken from disintegrated melt deposition technique when compared with the unreinforced and conventionally cast composite samples. The results of microstructural characterization and mechanical testing were finally rationalized in terms of the nature of processing technique employed to reinforce A1-2wt.% Cu metallic matrix with SiC particulates.

Keywords: Disintegrated melt deposition; Microstructure; Mechanical behaviour; Processing

1. Introduction

The flexibility associated with metal matrix com- posites (MMCs) in tailoring their physical and mechan- ical properties as required by the end application have made them suitable candidates for a spectrum of appli- cations related to energy, automobile and aeronautical sectors [1-2].

The suitability of MMCs as a viable replacement of the conventionally used monolithic materials, however, depends on the acquisition of scientific understanding in order to synthesize them with consistent reproducibility in microstructure and mechanical behaviour and their ability to exhibit enhanced performance based cost effectiveness in real time applications [1,3].

The enhanced performance from these rather unique materials depends on a careful selection of processing technique/parameters, matrix phase, reinforcing phase .and the heat treatment procedure. The full potential of the metal and ceramic combination, however, is strongly influenced by the processing associated microstructural evolution [4,5]. For applications where design criterion

is strongly influenced by the strength properties, it is imperative that the microstructure of the composite material should exhibit a uniform distribution of ceramic reinforcing phase and a strong interfacial integrity be- tween ceramic reinforcement and the metallic matrix. The non-uniform distribution of ceramic phase, for example, may result into inferior mechanical properties while a poor interfacial integrity may lead to inefficient load transfer from metallic matrix to ceramic reinforce- ment. One synthesis approach currently being investi- gated to circumvent these problems is to disintegrate the molten composite metal stream using linear inert gas jets and to deposit them on a metallic substrate or a shaped container. This process is designated as disintegrated melt deposition (DMD) technique and is a modification of spray atomization and deposition technique [3-6]. However, unlike spray deposition technique this process employs higher superheat temperatures, lower impinging gas jets velocity and the end product is only bulk composite material [4-6].

The objective of this study was to provide preliminary information regarding the microstructural evolution

0921-5093/96/$15.00 © 1996 -- Elsevier Science S.A. All rights reserved SSD1 0921-5093(95)10077-6

M. Gupta et al. / Materials Science and Engineering A210 (1996) 114-122 115

and tensile properties of the A1-2wt.% Cu/SiC metal matrix composites synthesized using disintegrated melt deposition technique and to compare the results with the conventionally cast composite samples of similar composition. Microstructural characterization studies were conducted using scanning electron microscopy while tensile testing was carried out using an auto- mated servohydraulic Instron testing machine. Particu- lar emphasis was placed to correlate the processing associated microstructural evolution with the mechani- cal properties of the composite materials synthesized in this study.

2. Experimental procedure

2.1. Materials

The nominal composition of the matrix alloy used in this study was (in wt.%): 2.0 Cu-Al (bal.). Silicon carbide (~-SiC) particulates with an average size of 23 /lm were selected as the reinforcement phase.

2.2. Processing

The synthesis of the metal matrix composites used in this study was carried out using conventional cast- ing and disintegrated melt deposition techniques. The synthesis of MMCs using conventional casting was carried out according to the following procedure. The metal ingots, prior to melting, were properly cleaned to eliminate surface impurities. The cleaned metal in- gots were melted to the desired superheating tempera- ture. SiC particulates, preheated to 900 °C, were then added into the molten metal stirred using an impeller. The composite melt thus obtained was poured into cylindrical steel molds (25 mm diameter and 178 mm height). In all the cases, stirring time of SiC particu- lates in the melt was maintained between 10 and 15 min. Regarding the disintegrated melt deposition of MMCs, the synthesis process involved: superheating the properly cleaned metal ingots to 940 °C in a graphite crucible, addition of SiC particulates pre- heated to 900 °C in the liquid metallic melt, argon gas assisted composite melt disintegration at 0.18 m from the melt pouring point and subsequent deposition on metallic substrate located at 0.25 m from the gas disin- tegration point. Both conventional casting and disinte- grated melt deposition processing of metal matrix composites were carried out in the ambient atmo- spheric conditions.

For the purpose of comparison, the base alloy was cast under similar processing conditions as described for conventional casting process.

2.3. Quantitative assessment of SiC particulates

Quantitative assessment of SiC particulates in the composite samples was carried out using a chemical dissolution method. This method involved: (i) measur- ing the mass of composite samples; (ii) dissolving the samples in hydrochloric acid, followed by (iii) filtering to separate the ceramic particulates. The particulates were then dried and the weight fraction determined [5].

2.4. Density measurement

Density measurements were carried out in order to ascertain the volume fraction of porosity in the unre- inforced and reinforced samples. Density measure- ments were carried out using Archimedes' principle following the procedure as discussed in Ref. [4].

2.5. Aging studies

Aging studies were carried out in order to obtain the peak hardness temperature and time conditions for the unreinforced, conventionally cast and DMD processed metal matrix composites. Unreinforced and convention- ally cast specimens (25 mm diameter x 7 mm height) were solutionized for 1 h at 490 °C, quenched in cold water and aged at 160 °C for various intervals of time. The DMD processed specimens, however, were solu- tionized for 1 h at 510 °C, quenched in cold water and aged at 155 °C for various intervals of time. It may be noted that 490 °C represents the optimal solutionizing temperature for unreinforced and conventionally cast composite samples while 510 °C represents the optimal solutionizing temperature for DMD processed com- posite samples used in the present study. Rockwell B hardness measurements were made using 1.58 mm di- ameter steel ball indentor with a 100 kg load.

2.6. Microstructural characterization

Microstructural characterization studies were con- ducted on the peak aged unreinforced and reinforced samples in order to investigate the distribution of SiC particulates and the presence of porosity. Particular emphasis was placed to examine the precipitation be- haviour and segregation of alloying elements in the interfacial region between the AI alloy matrix and ceramic particulates.

Microstructural characterization studies were pri- marily accomplished using a JEOL scanning electron microscope equipped with energy dispersive spec- troscopy (EDS). The composite samples were metallo- graphically polished prior to examination. Micro- structural characterization of the samples were con- ducted in both etched and unetched conditions. Etch- ing was accomplished using Keller's reagent (0.5 Hf, 1.5 HCI, 2.5 HNO3, 95.5 H20).

116 M. Gupta et aL / Materials Science and Engineering A210 (1996) I14-122

Table 1 Results of the acid dissolution tests and volume percent porosity determination

Matrix Condition Reinforcement size Wt. % SiC Vol.% porosity

A I - C u As-cast - - - - 0.6 A I - C u As-cast 23 gm 13.9 7.9 A1-Cu As-DMD processed 23 p-m 14.5 4.4

2. 7. Mechanical behaviour

The smooth bar tensile properties were determined on the peak aged monolithic and composite specimens following ASTM standard E8-81. Tensile tests were conducted using an automated servohydraulic Instron testing machine on 4 mm diameter specimens using a crosshead speed of 0.254 mm per min.

2.8. Fracture behaviour

Fracture surface characterization studies were carried out on the tensile fractured unreinforced and reinforced samples in order to provide insight into the various fracture mechanisms operative during tensile loading of the peak aged samples. Fracture surface characteriza- tion studies were primarily accomplished using a JEOL scanning electron microscope equipped with EDS.

3. Results

3.1. Macrostructure

Macrostructural characterization conducted on the as-processed, machined and polished conventionally cast specimens revealed the presence of macropores and the macrosegregation of SiC particulates. These fea- tures, however, could not be detected on the machined and polished surfaces of unreinforced and DMD com- posite specimens in the as-processed condition.

3.2. Quantitative assessment of SiC particulates

The results of acid dissolution experiments are sum- marized in Table 1. The weight percentages of SiC particulates was estimated to be approximately 13.9% for conventionally cast composite specimens and 14.5% for the DMD processed composite specimens.

3.3. Density measurement

The results of density measurements conducted on the as-processed unreinforced, conventionally cast and DMD processed composite specimens revealed density values of 2.71, 2.57 and 2.68 g cm-3, respectively. The

volume percent of the porosity computed using the experimentally determined density values and the re- sults of acid dissolution tests are shown in Table 1.

3.4. Aging studies

The results of aging studies conducted on the unrein- forced, conventionally cast and DMD processed com- posite samples are shown in Fig. 1. The results exhibit the presence of a well defined hardness peak at 9 h for the unreinforced and as-cast composite samples and at 6 h for the DMD processed composite samples. The results also reveal that the maximum peak hardness is achieved in the DMD processed composite samples followed by the conventionally cast composite samples and lastly, the unreinforced samples.

It may be noted that the peak aging conditions for unreinforced, conventionally cast and DMD processed composite samples were established in this study by measuring macrohardness (Rockwell B) of the bulk samples instead of microhardness of the metallic matrix at various aging times. This selection was made as a result of high scatter associated with microhardness measurements when made on the composite samples. This high scatter in microhardness measurements may primarily be attributed to the high sensitivity of the microhardness measurement techniques towards:

(a) the constitutional changes in the metallic matrix in the near vicinity of SiC particulates;

100

75

50

25

+ As-Cast (Aging Temp.: 160 C) + Unreinforced (Aging Temp.: 160 C)

As-DMD (Aging Temp,: 155 C)

Time (h)

Fig. I. Aging curves of conventionally cast A1-Cu, AI-Cu/SiC samples, and DMD processed AI-Cu/SiC samples.

M. Gupta et al. / Materials Science and Engineering A210 (1996) 114-122 117

weight percent of copper as a function of distance from the particulate matrix interface.

Fig. 2. Representative SEM image showing columnar-equiaxed grain structure, porosity and intercellular/interdendritically located Cu rich phases in the conventionally cast AI Cu samples.

(b) the microstructural features such as voids associ- ated with SiC particulates, and

(c) the uncertainty associated with the subsurface microstructural characteristics.

3.5. Microstructural characterization

Scanning electron microscopy conducted on unetched and etched unreinforced A1-Cu samples revealed the presence of partly dendritic and partly equiaxed matrix microstructure, minimal amount of micrometer sized porosity and interdendritically located Cu rich inter- metallic phase. A representative micrograph taken from the unreinforced sample is shown in Fig. 2.

The results of scanning electron microscopy con- ducted on conventionally cast AI-Cu/SiC samples re- vealed a partly dendritic and partly equiaxed matrix microstructure. The interdendritic/intercellular regions were found to be frequently associated with the pres- ence of Cu rich phases (see Fig. 3 (a)). The metallic matrix also revealed the presence of porosity predomi- nantly associated with the individual SiC particulates at the angular locations and with SiC clusters. The distri- bution of SiC particulates in the cast composite samples can be assessed from Fig. 3 (b). Predominantly, SiC particulates were present in the form of small clusters preferentially located at the grain boundaries. The in- terfacial integrity between SiC particulates and AI-Cu matrix was found to be poor and in some cases par- tially debonded interface was observed in the cast com- posite samples (see Fig. 3 (c)). In addition, the interface formed between the SiC particulates and A1-Cu matrix also revealed the presence of secondary phases. EDX analyses carried out at various locations in the AI--Cu/ SiC interfacial locations revealed the enrichment of Cu. The results of EDX semi-quantitative point analyses graphically represented in Fig. 4 shows the variation of

(a)

(b)

(c)

Fig. 3. Representative SEM images showing: (a) matrix microstructure (Cu rich phases are indicated by arrows); (b) distribution of SiC particulates, and (c) interfacial integrity between a SiC particulate and AI Cu metallic matrix, in conventionally cast composite samples.

118 M. Gupta et al. / Materials Science and Engineering A 210 (1996) 114-122

10. [] Cut/d-Ca/SiC 9 ~ D M D AI-Ca/SiC

g~

lb 1~ 2b 2~ 3b 35

Distance From SiC Interface (Bm)

Fig. 4. Segregation pattern of Cu observed at AI-Cu/SiC interfacial regions in case of conventionally cast and DMD processed AI-Cu/ SiC samples determined using semi-quantitative EDX analysis.

(a)

Finally, the results of microstructural characteriza- tion studies carried out on the DMD processed samples revealed the presence of columnar/equiaxed grains in the matrix microstructure, relatively lower amount of porosity particularly associated with short edges and angular locations of SiC particulates, Cu-rich phases at the interdendritic/intercellular regions, relatively more uniform distribution of SiC particulates and good inter- facial integrity between SiC particulates and the metal- lic matrix (see Fig. 5) when compared with the conventionally cast composite specimens. In addition, the results of EDX analyses revealed the copper content to decrease with increasing distance from the interface (see Fig. 4).

3.6. Mechanical behaviour

(b)

The results of ambient temperature testing on the unreinforced, conventionally cast and DMD processed composite samples, aged to peak hardness, are summa- rized in Table 2. The results in Table 2 reveal that 0.2% yield stress (0.2% YS) and ultimate tensile strength (UTS) of the DMD processed specimens are superior when compared with those of the unreinforced and conventionally cast composite samples. The ductility of the DMD processed samples was, however, found to be inferior than the unreinforced and conventionally cast composite samples. In addition, the results revealed the strength and ductility of the conventionally cast com- posites to be inferior even when compared to the unre- inforced material (see Table 2).

3. 7. Fracture behaviour

The tensile fracture surfaces of unreinforced and reinforced samples are shown in Fig. 6. The fracto- graphs taken from the unreinforced samples revealed

(c)

Fig. 5. Representative SEM images showing: (a) matrix microstruc- ture and distribution of ceramic particulates; (b) voids associated with SiC particulates, and (c) partially debonded AI-Cu/SiC interface and the presence of Cu rich phases in the immediate vicinity of AI-Cu/ SiC interface observed in DMD processed composite samples.

M. Gupta et al. / Materials Science and Engineering A210 (1996) I14-122 119

exhibiting interfacial debonding, SiC particulate crack- ing and evidence of matrix deformation between SiC particulates.

(a)

(b)

(c)

Fig. 6. Representative SEM images showing fracture surface charac- teristics of: (a) unreinforced samples; (b) conventionally cast com- posite samples, and (c) DMD processed composite samples.

the presence of dimples indicative of a predominantly ductile failure (see Fig. 6 (a)). Fracture studies con- ducted on the tensile fracture surface of the convention- ally cast (see Fig. 6 (b)) and DMD processed (see Fig. 6 (c)) specimens revealed a typical quasibrittle fracture

4. Discussion

4.1. Microstructure

The microstructure of the unreinforced samples, con- ventionally cast composite samples and DMD pro- cessed composite samples revealed three common salient features:

(a) presence of columnar / equiaxed matrix micro- structure;

(b) presence of porosity, and (c) presence of interdendritic Cu-rich phase. The columnar-equiaxed matrix microstructure, com-

monly referred as "ingot" type of structure [7] indicates that the remaining liquid temperature after the onset of solidification from the mold wall remained above the nucleation temperature. The underlying principles be- hind the development of "ingot" type of structure are well established and can be found elsewhere [7].

Another important microstructural feature observed in case of unreinforced and reinforced samples investi- gated in the present study was the presence of porosity. The formation of microporosity in the unreinforced and reinforced samples under the experimental condi- tions used in this study was inevitable primarily as a result of columnar-equiaxed type of solidification structure observed in this study. In analogous studies [7], it has been convincingly established that the mi- croporosity associated with the materials exhibiting columnar-equiaxed solidification structure can be at- tributed to the inability of the high viscosity liquid in the interdendritic regions to be "sucked" into the groove sufficiently fast to keep pace with the shrinkage that accompanies solidification. For further details, the reader is encouraged to refer to Ref. [7]. Regarding the amount of porosity, the results of this study revealed that the volume percent of porosity was 0.6% in case of unreinforced samples, 4.4% in case of DMD processed composite samples and about 7.9% determined for the conventionally cast composite samples. The higher vol- ume percent of porosity observed in composite samples can be attributed to the physical properties of the molten metallic matrix containing suspended ceramic particulates and the solidification associated distribu- tion of the particulates in the metallic matrix. The results of microstructural characterization conducted on the conventionally cast composite specimens clearly indicated the presence of metal free zones at the sharp corners of the SiC particulates and within the SiC clusters (see Fig. 3 (b)). The development of metal free zones at sharp corners of SiC particulates can primarily

120 M. Gupta et aL / Materials Science and Engineering A210 (1996) 114-122

Table 2 Results of room temperature mechanical properties

Material Processing Condition 0.2% YS UTS Ductility (MPa) (MPa) (%)

AI-Cu Cast Peak aged a 93.8 ___ 12.1 139.8 __+ 27.1 5.2 _+ 4.9 A1-Cu/SiC Cast Peak aged a 58.1 + 21.3 61.0 + 21.6 1.2 ___ 0.3 AI-Cu/SiC DMD Peak aged b 142.0 + 21.6 152.3 _+ 19.1 0.6 __+ 0.3

aThe samples were solutionized at 490 °C for 1 h, water quenched and aged at 160 °C for 9 h. bThe samples were solutionized at 510 °C for 1 h, water quenched and aged at 155 °C for 6 h.

be attributed to the inability of the high viscosity metallic alloy to negotiate sharp corners while the presence of metal free zones within SiC clusters can be attributed to the inability of liquid metallic alloy used in the present study to infiltrate the micrometer sized crevices in the inefficiently packed SiC clusters of the particulates formed ahead of moving solidification front during conventional casting. The qualitative mi- crostructural observations regarding the presence of porosity in case of DMD processed samples were simi- lar to that observed for conventionally cast specimens. However, a lower volume fraction of porosity obtained in DMD processed composite samples can be attributed to a more uniform distribution of SiC particulates and the reduced formation of clusters in the metallic matrix during DMD processing. The reduced cluster forming tendency may be attributed to the dynamic solidifica- tion conditions present during DMD processing. This observation is also consistent with the SiC particulates distribution results of other investigators obtained on spray atomized and deposited A1-Si/SiC metal matrix composites [4]. Further work is continuing in this area.

The presence of interdendritic/intercellular Cu-rich phase observed in case of unreinforced, conventionally cast and DMD processed composite samples (see Figs. 2, 3 and 5) can be attributed to the sluggish solidifica- tion front velocity achieved during primary processing of materials, rejection of Cu ahead of the moving liquid-solid interface and subsequent solidification when the temperature of the remaining liquid reached eutectic temperature [4,7,8]. The discontinuous nature of Cu-rich phase along the grain boundaries/interden- dritic regions may be attributed to the partial dissolu- tion during solutionizing step of the T6 heat treatment employed in this study.

4.2. Amount and distribution o f S iC particulates

In this study, 13.9 and 14.5 wt.% of SiC particulates were successfully incorporated in Al-2wt.% Cu metallic matrix using conventional casting and DMD processing techniques respectively. The successful incorporation of SiC particulates in the limits exceeding 10 wt.% using these techniques can be attributed to the enhanced wettability of SiC particulates as a result of the preheat-

ing of SiC particulates to 900 °C prior to the addition in the superheated liquid metallic melt. Preheating of SiC particulates has been shown to assist in: (i) remov- ing surface impurities; (ii) desorption of gases, and iii) altering the surface composition owing to the formation of thin oxide layer (SiO2) on the surface [9]. The ability of the oxide layer to improve the wettability of SiC particulates by alloy melt has previously been suggested by other investigators [10,I 1].

Regarding the distribution of SiC particulates, the following comments are in order. The preferential loca- tion of SiC particulates in case of cast composite sam- ples is consistent with the results of other investigators and was attributed to the sluggish solidification front velocity commonly associated with the casting route [4,7]. In the case of DMD processed composite samples, the improved distribution of SiC particulates can be attributed to the coupled effects of non-equilibrium conditions during melt disintegration and the dynamic events that follow during deposition as a result of impingement of disintegrated melt stream at the deposi- tion surface. The dynamic events during deposition bear a significant importance in achieving uniform distribu- tion since the solidification front velocity which is gov- erned by the rate of heat extraction is sluggish enough to engulf the SiC particulates during solidification even for the disintegration velocity of 296 m s-~ [12,13].

4.3. Interfacial characteristics

The results of microstructural characterization stud- ies conducted on the particulate/matrix interracial re- gion revealed a high concentration of Cu in both conventionally cast and DMD processed composite samples (see Fig. 4). This phenomenon can primarily be attributed to the presence of enhanced dislocation den- sity in the interracial region. The enhanced dislocation density results owing to the difference in coefficient of thermal expansion between SiC particulates and the aluminum matrix [14] and promotes the dislocation-as- sisted diffusion of the alloying elements from the adja- cent dislocation lean areas of the matrix. In related studies conducted on spray deposited aluminum based metal matrix composites, Gupta et al. [5] reported a similar enrichment of the main alloying element Cu in

M. Gupta et al. / Materials Science and Engineering A210 (1996) 114-122 12/

the SiC particulate/matrix interfacial region. These ex- perimental findings thus suggest that the interfacial segregation of the alloying elements is primarily gov- erned by the physical properties of the metallic and ceramic components of the metal matrix composites and is independent of the type of processing technique employed to synthesize the composites.

Another important microstructural characteristic as- sociated with the A1 Cu/SiC interfacial region in con- ventionally cast and DMD processed composite samples was the presence of secondary phases (see Figs. 3 and 5). These secondary phases were found to be located at and in near vicinity of SiC particulates demonstrating the heterogeneous nucleation capability of the interfacial region as a whole. These experimental results thus indicate the heterogeneous nucleation capa- bility of the SiC particulates and in addition the capa- bility of the dislocations defect structure in the interfacial region in providing preferential sites for het- erogeneous nucleation when compared to the bulk ma- trix. It can be noted that the heterogeneous nucleation capability of SiC particulates have been previously re- ported in case of SiC reinforced 7091 A1 alloy [15] while the presence of precipitates in the immediate vicinity of SiC particulates has been convincingly established in case of A1 based matrices by various investigators [5,14]. Further work is continuing in order to identify the composition and structural aspects of these sec- ondary phases.

Regarding the interfacial integrity, DMD processed composite samples exhibited minimal debonding or the presence of voids associated with SiC particulates in A1-Cu metallic matrix in the as-processed condition when compared with the conventionally cast specimens. The improved interfacial integrity in case of DMD processed composite samples indicates superior fluid flow characteristics of the composite slurry in case of DMD processing when compared with the conventional casting process.

4.4. Aging studies

The results of this study reveal that the aging kinetics remained the same for unreinforced and conventionally cast composite materials while it was accelerated in case of DMD processed composite samples. The similar aging kinetics exhibited by the unreinforced and con- ventionally cast composite samples are consistent with the findings and observations of other investigators [5,16,17]. Chawla et al. [16] and Salvo et al. [17], for example, showed a negligible difference in aging kinet- ics of unreinforced and reinforced materials aged at relatively low temperatures ( ~ 150 °C). These results are also supported by the findings of Gupta et al. [5] who reported the similar aging kinetics exhibited by spray processed A1-Cu, AI-Cu/SiC and AI-Cu/A1203

materials aged at 163 °C. These results are however in contradiction with the accelerated aging kinetics results obtained on reinforced aluminum based composite ma- terials by other investigators [18-19].

The accelerated aging kinetics observed in case of DMD processed composite samples aged at 155 °C and containing slightly higher (0.6%) weight percent of SiC particulates when compared with conventionally cast composite samples is strongly indicative of' the matrix microstructural variation brought about by the DMD processing technique. The dynamic events during melt disintegration and subsequently on the deposition sur- face [13] appear to be instrumental in significantly changing the solidification events affecting the mi- crostructural variation and the associated aging kinetics of the DMD processed samples. Further work is con- tinuing in this area.

4.X Mechanical behaviour

The results of this study revealed inferior strength (0.2% YS and UTS) and ductility of the conventionally cast composite samples when compared to the unrein- forced samples. This degradation in mechanical proper- ties exhibited by the conventionally cast composite samples can be attributed to the coupled influence of:

(a) presence of 7.9 vol.% porosity (see Table 1); (b) non uniform distribution of SiC particulates (see

Figure 3b), and (c) poor interfacial integrity between SiC particulates

and AI Cu matrix (see Fig. 3 (c)). The porosity associated reduction in strength has

been previously established by other investigators for steels, copper and aluminum based alloys [20,21]. Boc- chini [20] and Payne et al. [21], for example, asserted that the presence of pores lead to weakening of a material by reducing the amount of stress bearing area and therefore lower the amount of stress the material is able to withstand.

The degradation in mechanical properties as a result of non-uniform distribution of SiC particulates can be attributed to the tendency of early crack nucleation in the matrix at the clusters or agglomeration sites [5].

Another important feature that might have con- tributed towards the degradation in the strength of the conventionally cast composite material is the poor in- terracial integrity observed between A1 Cu matrix and SiC particulates. The poor interracial integrity prevents the effective load transfer across AI Cu/SiC interface thus reducing the role of SiC particulates as load carri- ers in the metallic matrix. In analogous studies [5], it has been indicated that a strong interracial bond con- tributes effectively towards the enhancement of strength of the composite materials.

Finally, the results of the mechanical properties char- acterization revealed a significant improvement in the

122 M. Gupta et al. / Materials Science and Engineering A210 (1996) 114-122

strength (0.2% YS and UTS) and reduction in ductility of the DMD processed composite samples when com- pared with the conventionally cast unreinforced and composite samples. For example, the results shown in Table 2 indicate an increase in UTS of the DMD processed samples by ~ 1.1 times when compared to the unreinforced samples and ~ 2.5 times when compared to the conventionally cast composite samples. The supe- rior strength properties thus exhibited by the DMD processed composite samples when compared to conven- tionally cast composite samples are attributed to the processing-associated improved microstructural unifor- mity, low volume fraction of porosity, relatively more uniform distribution of SiC particulates (minimum stress concentration sites) and improved interfacial integrity between the ceramic particulates and A1-Cu metallic matrix (see Fig. 5).

4.6. Fracture behaviour

The results of the fracture surface analysis indicate an increase in brittleness from the unreinforced to conven- tionally cast composite samples to DMD processed composite samples. These results are consistent with the mechanical properties results which show a reduction in elongation in same order (see Table 2). The presence of uniformly distributed dimples observed on the fractured surface of the unreinforced samples indicates a relatively ductile failure when compared with the composite sam- ples. Three salient features were observed on the fracture surface of composite samples. These are:

(a) interfacial debonding between SiC particulate and A1 matrix;

(b) SiC particulate cracking, and (c) presence of dimples (indicative of matrix plastic

deformation) in between the SiCparticulates. The results, in essence, did not reveal any single

dominant failure mechanism. In analogous studies con- ducted on cast A356 alloy reinforced with 20% SiC, investigators reported similar evidences of conjunct infl- uence of various failure mechanisms contributing to- wards the failure of composite material under tensile loading [4].

5. Conclusions

The primary conclusions that may be derived from this work are as follows:

(1) Aluminum based metal matrix composites contain- ing upto 14.5 wt.% of SiC particulates can be success- fully synthesized by disintegrated melt deposition route used in this study.

(2) A low volume percent (4.4%) of porosity observed in the disintegrated melt deposited samples when com- pared with the conventionally cast composite samples

(vol.% = 7.9) is indicative of the potential of the disinte- grated melt deposition technique to make near net shape products.

(3) The increase in strength of the disintegrated melt deposited composite samples can be attributed to the improved microstructural homogeneity and superior in- terfacial integrity between SiC particulates and A1-Cu matrix when compared with the conventionally cast material.

(4) The results of the fracture surface studies indicated similar failure mechanisms for composite samples pro- cessed using conventional casting and disintegrated melt deposition routes.

Acknowledgments

MG and MOL would like to thank NUS (grant # RP 3940619) for financial support during the course of this investigation. In addition, the authors would like to thank Mr. Thomas Tan, Mr. Boon Heng and Mr. Tung Siew Kong (National University of Singapore, Singa- pore) for their valuable assistance and for many useful discussions.

References

[1] P.S. Gilman, J. Met., 43(8) (1991) 7. [2] A.L. Geiger and J.A. Walker, J. Met., 43(8) (1991) 8. [3] I.A. Ibrahim, F.A. Mohamed and E.J. Lavernia, J. Mater. Sci.,

26(1991) 1137. [4] M. Gupta, C. Lane and E.J. Lavernia, Scr. Metall. Mater., 26

(1992) 825. [5] M. Gupta, T.S. Srivatsan, F.A. Mohamed and E.J. Lavernia, J.

Mater. Sci., 28 (1993) 2245. [6] M. Gupta, J. Juarez-Islas, W.E. Frazier, F.A. Mohamed and E.J.

Lavernia, Metall. Trans. B, 23 (1992) 719. [7] B. Chalmers, Principles of Solidification, Wiley, New York, USA,

1964, pp. 253-297. [8] Metallography and Microstructures, Metals Handbook, 9th edn.,

Vol. 9, ASM, Metals Park, OH, USA, 1986, p. 632. [9] C.M. Milliere and M. Suery, Mater. Sci. Technol., 4 (1988) 41.

[10] H. Ribes, R. Dasilva, M. Suery and T. Bretheau, Mater. Sci. Technol., 6 (1990) 621.

[11] L.N. Thanh and M. Suery, Scr. Metall., 25 (1991) 2781. [12] M. Gupta, F. Mohamed and E. Lavernia, Metall. Trans. A, 23,4

(1992) 831. [13] M. Gupta, F.A. Mohamed and E.J. Lavernia, Int. J. Rapid Solid.,

6 (1991) 247. [14] R.J. Arsenault and N. Shi, Mater. Sci. Eng., 81 (1986) 175. [15] D.J. Lloyd, Int. Mater, Rev., 39 (1994) 1. 116] K.K. Chawla, A.H. Esmaeili, A.K. Datye and A.K. Vasudevan,

Set. Metall. Mater., 25(6) (1991) 1315. [17] L: Salvo, M. Suery and F. Decomps, in J. Masounave and F.G.

Hamel (eds.), Fabrication of Particulates Reinforced Composites, ASM International, Materials Park, OH, 1990, p, 139.

[18] T.G. Nieh and R.F. Karlak, Scr. Metall., 18 (1984) 25. [19] J.M. Papazian, Metall. Trans., 19A (1988) 2945. [20] G.F. Bocchini, Int. J. Powder Metall., 22(3) (1986) 185. [21] R.D. Payne, A.L. Moran and R.C. Cammarata, Scr. Metall.

Mater., 29 (1993) 907.