Materials Science and Technology

Metal matrix composites (MMCs) will play a significant role in the future of gas turbine aeroengine development. This paper outlines the benefits and some of the potential applications for Al and Ti MMCs and discusses issues involved in the introduction of this relatively new class of composite materials into engine components. The potential for cost savings and performance improvements which may be achieved by the introduction of silicon carbide particulate reinforced Al alloys (Al–SiC_p MMC)are discussed, where the properties of the composite can be tailored to meet the requirements of a specific application. Some of the processing implications for Ti matrix composites (Ti MMCs) are explained for a range of component applications. Illustrations are given of how the manufacturing process can be controlled for a complex component. Finally, the influence of raw material, manufacturing, and component costs on the successful introduction of MMCs into aeroengines is discussed.

There is a growing need in both aerospace and commercial markets for lighter weight, higher stiffness, higher thermal stability materials to solve the design engineers’ problems of reduced mass, higher access speeds, improved mechanical and thermal stability for today's advanced technology. To address those needs, Brush Wellman Inc. has developed, characterised, and put into high volume production a family of beryllium metal matrix composites. There are two classes of materials that have been developed to provide these engineering benefits to the designer in both the commercial and aerospace markets. The first family of materials is aluminium beryllium (AlBeMet, which is a registered tradename of Brush Wellman). This material is a metal matrix composite consisting of pure beryllium and aluminium, with 20–62 vol.-%Be and the remainder aluminium. The material is produced by both powder metallurgy and net shape technologies such as investment casting and semi-solid forming. The materials properties that make it attractive for the design engineer are a density that is 25% less than that of aluminium, a specific stiffness four times those of aluminium, titanium, steel, and magnesium, a higher dampening capacity than aluminium, and a coefficient of thermal expansion almost 50%lower than aluminium. The second family of materials is a beryllium–beryllium oxide metal matrix composite, which are called E materials. This material was developed to address the thermal management needs of the electronic packaging design engineer. The material properties that make this material attractive to the electronic packaging engineer are: a density 20–25% that of Kovar, Invar, and CuMoCu, and 30% less than Al–SiC; a thermal conductivity ranging from 210 to 240 W m^-1 K^-1and a tailorable coefficient of thermal expansion, ranging from 6×10^-6 to 8.7×10^-6 K^-1.

The thematic network ‘Assessment of metal matrix composites for innovations’ (acronym MMC-Assess) has been approved within the Framework Program 4 of the European Commission for a 36 months period starting in October 1998. It is focussing on the assessment of this class of heterogeneous materials for applications in all technical fields. An Internet homepage (http://mmc-assess.tuwien.ac.at/) has been established and will be filled with information on the four MMC categories (particulate, short fibre, continuous fibre, and monofilament reinforced metals) and on the topics of processing, test methods, properties, modelling, specification, application, reliability, and marketing. Twenty-one European partners from universities, research organisations, industry, and national MMC working groups are cooperating with experts outside the consortium and are globally establishing links with MMC activities. The network is open to all interested in MMCs, who are asked to take advantage of the scientific expertise offered and the specific demands of industry.

Metal matrix composites are considered as a possible material to replace the materials actually used in some automotive and railway industrial components. Two different ways of taking advantage of the properties of MMCs at low cost are presented: locally reinforced aluminium components and new castable aluminium MMCs developed by the London & Scandinavian Metallurgy Co. (LSM).

Alumina/nickel composites have been fabricated by hot pressing powder blends of various volume fractions of nickel and alumina. The electrical resistivities and Young's moduli of these composites have been measured and their dependence on the volume fraction of reinforcement has been investigated. Microstructural parameters such as contiguity were measured to quantify the distribution of the phases in these composites, and existing property models based on these data were used to predict the properties of the composites. The percolation threshold of nickel was found to occur at between 7.5 and 15 vol.-%Ni. The Young's modulus decreases as the volume fraction of nickel increases and is dependent on the contiguity of alumina. Composites containing 25, 35, 50, and 65 vol.-%Ni display microstructures with interpenetrating networks of alumina and nickel. The property models were found to fit both the resistivity and modulus data well, although the percolation threshold was predicted at a lower volume fraction than measured experimentally.

Mechanical milling is an established production method for aluminium particulate metal matrix composites (MMCs). There are examples of its use for high performance automotive applications and within the aerospace industry. The production of a titanium particulate MMC is still in the developmental stage. However, compared to conventional titanium alloys such materials offer improvements in stiffness, strength, fatigue and creep properties, high temperature capability, and wear resistance. This paper describes the use of mechanical milling for the production of titanium particulate MMCs with the addition of 10 vol.-%TiB. Gas atomised titanium powders with additions of either boron or TiB₂ were milled in a high purity argon atmosphere to avoid contamination of the powders by oxygen or nitrogen. The distribution of the boron or TiB₂ with increasing milling time is discussed along with the effect of the alloy composition. Gas atomised, hydride dehydride, and sponge fine powder blends are also compared. The powders were subsequently hot isostatically pressed at 500°C for 2 h at 150 MPa followed by 900°C for 2 h at 150 MPa. During this consolidation process TiB was formed by an in situ reaction between either the TiB₂ or boron and the titanium matrix.

From a commercial viewpoint superplastic forming of complex shapes using only a single operation and one surface tool is appealing, especially for metal matrix composites (MMCs) that are hard to form even at elevated temperatures due to low ductility and toughness. Furthermore, secondary machining operations are difficult due to the presence of extremely hard ceramic reinforcements such as SiC. A range of aluminium alloy based MMCs have indeed been shown to exhibit superplastic properties although most of these studies have been concerned with microstructural characterisation using small uniaxial tensile specimens. This paper therefore concentrates on high strain rate biaxial superplastic forming of complex shapes (critical feature) in MMCs where a forming envelope has been defined and post-forming mechanical properties investigated. Particulate reinforced MMCs based on aluminium alloys 7475 and 7178 were superplastically formed in a die with a 45° step at a range of temperatures and pressures. Formed specimens were sectioned to investigate cavitation and cross-sectional thinning. Tensile tests were performed on parent and formed material to investigate the effect of superplastic forming on mechanical properties. The MMCs were successfully formed over the temperature range 450–550°C achieving step angles α of 22–42°. This study has shown that high strain rate superplasticity (∼10^-1 s^-1)can be achieved giving a strain of 70% in only 3.5 s without SiC fracture, reinforcement–matrix decohesion or matrix cavitation making this technique economic and very attractive for commercial exploitation. Cross-sectional thinning was found to be uniform and in the order of ∼25% which could be accounted for at the design stage. The high strain rate superplasticity was found to be grain size dependent (<3 µm) but greater profile definitions were achieved when forming took place just above the matrix solidus. Superplastic forming above the matrix solidus temperature resulted in the achievement of the highest step angles in the complex shapes but had a detrimental effect on mechanical properties. This is thought to be due to the liquid phase present that aids grain boundary and interfacial sliding but has a similar effect to overheating during solution treatment and brittle phases are formed at the grain boundaries.

Continuous carbon fibre preforms (M40), with an average fibre content of 66 vol.-%, were infiltrated with liquid Mg by gas pressure at various temperatures and pressures in order to determine a relationship between manufacturing conditions and the mechanical properties of the material. A saturation curve describes the infiltration behaviour of liquid Mg into the preforms and the relationship between infiltration pressure, pore volume, and mechanical properties. The maximum bending strength of the composites exceeds 1400 MPa and is insensitive to melt temperature between 680 and 720°C. The density of the Mg composite is 1.77 Mg m^-3, 20% lighter than the Al–C composites produced and tested by the authors in an earlier investigation, but yields the same mechanical properties in bending tests at room temperature.

This paper describes the rotating cylinder method for manufacturing, and investment casting for forming, composite slurry. Microstructural features, such as SiC particle distribution and grain refinement of the as cast composites, were investigated. Also the effect of SiC particle fraction and size, and process parameters on the microstructure and the mechanical properties are discussed. Attempts were made to evaluate the thermal stability of oxides against molten AZ91HP magnesium alloy. The oxides examined included CaO, CaZrO₃, and silica bonded Al₂O₃ and zircon flour. Finally, the tensile properties, hardness, and wear resistance of the as cast composites were evaluated and the results are compared with those of the as cast alloy.

Metal foams have acquired popular interest in recent years and are potentially useful for many applications due to their light weight, high specific stiffness, high surface to volume ratio, and adjustable cell structure. Here, current methods of producing metal foams are briefly reviewed. The requirements for high porosity metal foams with open cells are discussed. A novel powder metallurgy route involving a polymeric vehicle is introduced that can readily generate open cell foams with porosity greater than 90%. Coarse Ti powder and fine carbonyl iron powder were tested. Although the resulting polymer metal foam was closed cell, particles were not retained in the windows. Upon pyrolysis to remove the resin, the windows opened and the final sintered metal foam was highly reticulated. Such foams offer a fine reticulated structure with cell diameters in the region of 100–200 µm, and may find applications in the areas of catalysis, biomaterials, and composites.

The influence of the processing conditions on the evolution of the cell morphology of aluminium foams is investigated. The precursor material, a compacted mixture of aluminium and titanium hydride powder, is foamed to different degrees by heating it to different maximum temperatures. The pressure of the surrounding argon atmosphere is varied from low pressure to overpressure. The cell morphology is characterised by the mean cell diameter, its circularity, and the gas/metal interface length per unit area. The cell morphology is found to be primarily determined by the density and cannot be influenced by the outside pressure. The morphology is governed by one simple evolution law: the mean cell wall thickness is constant during expansion.

The impact response of laminated composites consisting of alternate layers of Al alloy foam and Al₂O₃ was studied experimentally in low and intermediate velocity regimes. Low velocity impacts (1.2–2.8 m s^-1) were conducted using an instrumented falling weight apparatus and were compared with static indentation tests (0.2×10^-4 m s^-1). Intermediate velocity impacts were carried out by means of both Hopkinson bar (60 m s -1)and gas gun (200 m s^-1) tests. Post­impact damage was assessed using X-ray radiography and microscopy. It was found that there is good correlation between low velocity impact and quasi­static responses. In both cases, penetration of the layered targets resulted in the formation of a distinctive plug. Increasing impact velocity (intermediate velocity range) switched the penetration mode from plugging to fragmentation, giving rise to an increase in the absorbed energy. In this range, impacts led to localisation of damage in the region under the projectile. Furthermore, a comparison has been made between the penetration response of foam laminates and dense metal laminates of equivalent areal density. Preliminary results suggest that the dense metal laminates are superseded by the foam laminates on an energy absorption basis.

The deformation behaviour of different types of closed cell aluminium foam (Alulight, Alporas) was studied. Compression tests indicate that inhomogeneities in the density distribution might be the key factor in determining the mechanical behaviour of foams. Strengthening and softening of the foam can be related to the formation of deformation bands. Depending on the composition and the microstructure of the cell wall material, cells undergo either ductile or brittle collapse. A three dimensional finite element analysis, which has the capability to simulate the initial deformation of foam samples, is presented. Continuum mechanics methods are used to describe the behaviour of foam. Each foam sample is divided into subregions, according to their density distribution. Scaling laws are used to simulate the mechanical properties of the subregions. Experimental data are compared with results of the presented model. Very good agreement between experiments and modelling was found for the rather inhomogeneous Alulight material.

A series of finite­size particle distributions were simulated to investigate the effects of particle size, shape, orientation, and area fraction on the quantification of homogeneity in structural particulate metal matrix composites (MMCs). It is found that, for nominally random distributions, the values of conventional centre-to-centre nearest-neighbour spacing parameters are influenced by particle morphology, and, as such, are unsuitable for characterising distributions of finite-size particles. However, the coefficient of variation of the mean near-neighbour distance COV(d _mean), derived from particle interfaces using finite-body tessellation, appears independent of particle shape, size distribution, orientation, and area fraction, while showing great sensitivity to particle clustering. In the range of particle morphological characteristics studied, the random distributions were found to exhibit a consistent value of COV(d _mean) equal to 0.36±0.02. The degree of inhomogeneity of any given distribution may then be evaluated by simply comparing the measured COV(d _mean) with this value.

Systematic empirical investigations on the relationship between microstructural features and mechanical performance of unidirectionally reinforced continuous fibre Al matrix composites (CFAMCs) carried out by the present authors in recent years are summarised. The employment of a high strength matrix alloy and the development of a strong fibre/matrix interface are beneficial to maximise the strengthening effect of the fibre reinforcement. Processing defects, such as second brittle phases in the matrix, non-infiltration defects, matrix solidification shrinkage voids, excessive interfacial reactions, the presence of reaction products on the interface, weak interfacial binding, and excessively high fibre volume fraction reduce composite strength to different extents via a number of different mechanisms. Criteria for the microstructure design of CFAMCs for optimum fibre strengthening efficiency are proposed.

The mechanical response of an in situ copper–chromium composite was investigated over a range of temperatures by means of tensile and isothermal creep tests. Scanning electron microscopy was used to characterise the extent, type, and distribution of damage. It was found that the failure mechanisms fell into distinct regimes. At cryogenic temperatures damage tended to occur in the form of reinforcement fracture. Around room temperature, very little damage was observed in the composite. At temperatures of about 400°C, extensive damage was again observed in the form of reinforcement failure and cavitation. Further increase in the test temperature resulted in a transition from a local to a global load sharing, with damage distribution becoming more homogeneous. These experimental observations were rationalised by considering the relative extent of deformation within the two phases as a function of temperature.

A miniaturised test system was used to investigate how the thermal and electrical properties of a unidirectionally reinforced titanium alloy (Ti–6Al–4V)/SiC (SM1140+)metal matrix composite change with mechanical damage at elevated temperature. Thermal conductivity and expansion measurements were obtained in the longitudinal and transverse direction both before and after short term strength and creep tests and at intervals during tests to assess changes in interface characteristics as functions of mechanical or thermal damage. The mechanical tests included monotonic stress–strain and ramp creep at temperatures between 500 and 650°C. The changes in thermal properties were compared with model predictions for the dependence of thermal properties on interface characteristics. The agreement was good for thermal expansion changes but not for thermal conductivity. This was ascribed to the nature of the damage at the interface that probably still allowed thermal transport but not mechanical load transfer.

The low cycle fatigue response of aluminium alloy 2124 reinforced with silicon carbide particulates produced by a commercial powder metallurgy route has been evaluated. This assessment includes measurements of fatigue crack growth rates and ‘total life to failure’ experiments. The relative performance under aqueous saline environments, namely total immersion in a 3.5% salt solution and exposure to a salt fog of similar concentration, were then characterised. While the alloy essentially demonstrates an insensitivity to saline environments in terms of its crack growth characteristics, a process of surface corrosion and pitting is highlighted and deemed responsible for significant reductions in the total fatigue life. The implications for engineering designs attempting to make use of this class of metal matrix composite are discussed.

Alumina (N610) reinforced pure Al (A9) or Al–2 wt-%Cu alloy (AU2) unidirectional (UD) composites which combine a highly ductile matrix with a strong interface bonding, present high static and dynamic mechanical performances. In the present paper, the fatigue behaviour of quasi­UD N610+A9 and N610+AU2 composites is investigated. Notwithstanding the presence of the transverse bundles, the longitudinal behaviour in tension and fatigue of these quasi-UD composites is nearly equivalent to that of the pure UD. Only in the case of the N610+AU2 quasi-UD composite, the fatigue limit is about 30% lower than that of the pure UD. Acoustic emission (AE) monitoring correlated with microfractography and microstuctural examinations has enabled identification of different stages in the evolution of damage and the association of these with damage and failure mechanisms. In fatigue, three main damage mechanisms are activated in a sequential and/or superposed mode during a three stage evolution.

This paper examines the use of a two­dimensional digital image based finite element method to predict the global behaviour of multiphase material systems. Micrographic images are digitised and meshed for implementation into the general purpose finite element code ADINA. The global cyclic response of the composite can be effectively modelled by using an appropriate constitutive relationship to describe the cyclic elastic–plastic behaviour of the matrix phase. The main advantage of the digital image based method is that the actual microstructural details including particle size, shape, and distribution are inherently captured in the analysis. The predicted global stress–strain responses of aluminium alloy 6061-T0/Al₂O₃particulate metal matrix composites under uniaxial and biaxial loading conditions (monotonic and cyclic) are found to correlate accurately with experimental results. When compared with predictions based on existing unit cell models, a noticeable improvement is observed. The effect of the representative length­scale (field of view) used in the analysis was found to be quite important in determining an accurate global response. A statistical analysis using uniformly derived lineal fraction measurements was also performed to demonstrate the correlation between the particle morphology in a particular field of view and the measured global response. Preliminary results indicate that this analysis technique may provide a possible method for determining the appropriate length­scale for which global analysis applies.

A slurry–powder metallurgy method is currently under development as a low cost production route for Ti/SiC composites. One of the critical steps in the process is the complete removal of a fugitive binder used to form the powder slurry. A diffusion-controlled, shrinking-core model has been developed for the prediction of the binder burnout kinetics in a metal powder compact with relatively large particles. This model is suitable for a polymeric binder that decomposes mainly to a monomer. The model considers the degradation of the polymer and the diffusion of the monomer in the core that contains a powder structure filled with a liquid polymer–monomer solution. It is found that the dominant mechanism is the liquid diffusion of monomer and the burnout time increases significantly with the compact size. The model also predicts a longer burnout time for a larger particle volume fraction.

The mechanical behaviour of layered systems, consisting of layers of monolithic titanium alloy, (IMI834) and titanium metal matrix composite, (Ti-MMC) is examined. A finite element based micromechanical model was developed to investigate the mechanical response of the systems subjected to four point bending. Two types of layered systems were examined: the first designated as C type with Ti-MMC in the centre of the beam, and IMI834 on the outside, and an S type which consists of a Ti-MMC layer on the outside and IMI834 on the inside. A representative volume element (RVE)was defined and boundary conditions imposed to simulate bending of the layered systems. While the S type system exhibited higher stiffness and linear behaviour up to fracture, the C type system exhibited more ductile behaviour prior to fracture. The overall predicted behaviour of the two systems was in good agreement with experimental results.

Composite materials based on aluminium are used in different fields where weight, thermal expansion, and thermal stability are key requirements. The aim of the present study was to develop a universal method and scientific approach for evaluating the design of lightweight, Al matrix composites with low coefficients of thermal expansion (CTE) and high dimensional stability, and to produce such composites using the vacuum plasma spray (VPS)process. The methodology is general and could be applied to other composite systems. The VPS-produced Al and Al alloy 6061 based composites were reinforced with a variety of ceramic particles including Si₃N₄, B₄C, TiB₂, and 3Al₂O₃.2SiO₂. These composites have low CTE values ((12–13)×10^-6 K^-1), similar to that of steel, and high dimensional stability (capable of keeping dimensions stable with changes in temperature). They have low porosity (98–99%dense) and a uniform distribution of the strengthening particles. Hot rolling of the VPS-formed composites, followed by heat treatment, resulted in a significant improvement in the mechanical properties. Deformed and heat treated 6061 based composites, containing 20 wt-%TiB₂ and 40 wt-%3Al₂O₃⋅2SiO₂, showed excellent mechanical properties (ultimate tensile strength 210–250 MPa, elongation >4%).

As a critical material for next generation aeroengines, fibre reinforced composites such as silicon carbide reinforced titanium continue to attract strong attention from both industrial and academic sectors. Reducing the processing costs and increasing component processing flexibility remain the priorities of current research. This paper presents a novel powder coated fibre pre-processing technique to meet such industrial requirements. The proposed technique is based on slurry powder metallurgy and presents itself as a cost effective alternative to current processing methods. It involves firstly, mixing matrix powder particles with an appropriate organic binder and solvent to form a slurry, drawing a continuous silicon carbide fibre through the slurry, drying the coated fibre and finally laying up the fibres into a composite preform for subsequent consolidation. The organic component is removed from the preform matrix via a binder burnout phase prior to composite consolidation.

Metal matrix composites have been produced by pressureless infiltration of pure Al into Mg doped SiC preforms after 1 h at 900°C. Aluminium has been found to infiltrate preforms containing between 2 and 14 wt-%Mg, however Al did not infiltrate a preform containing 1 wt-%Mg. Preforms doped with 1 wt-%Mg and Si did result in infiltration. Increasing the Mg content or increasing the Si content in the preform resulted in more extensive infiltration. The effect of Mg and Mg mixed with Si on pressureless infiltration of pure Al, microstructure of MMC as well as mechanical properties are discussed. Although the dopant was uniformly distributed throughout the preform microstructural analysis and hardness measurements indicate that the resultant composite may not be uniform due to infiltration inwards from the edge to the centre of preform.

This paper investigates the behaviour of aluminium based composites, reinforced with aluminium oxide particulates, manufactured using a powder processing route and various sintering times. It is found that a softening effect, compared with the monolithic material, coincides with migration of magnesium from the matrix to the matrix/reinforcement interface and to the specimen surface. The addition of 5 wt-%Si to the matrix alloy results in a retention of hardness in the specimens given short sinter times (0.5 and 1 h). This coincides with reduced migration of magnesium from the matrix. It is postulated that the excess silicon in the matrix results in sufficient magnesium being retained in the matrix during short sinter times to take part in the formation of Mg₂Si precipitates necessary for the hardening of the Al alloy 6061 matrix.

Rotating cylinder viscosity and spiral fluidity experiments were carried out to investigate the effects of SiC particle fraction and size and casting temperature on the viscosity and the fluidity of Mg–8.2Al–0.7Zn (wt-%)alloy containing up to 10 vol.-%SiC particles. The rotating cylinder viscosity (process based viscosity) was deduced from the data measured by a power measurement system during the actual composite manufacturing by the rotating cylinder method, and the fluidity was measured using a specially designed spiral fluidity mould. The spiral fluidity, which is the reciprocal of viscosity, decreased with increasing SiC particle fraction (for a given size) and with decreasing SiC particle size (for a given volume fraction). Although the decrease in the fluidity of the composite slurry was observed due to the expected increase in the viscosity by the presence of SiC particles, the results showed that the fluidity of the composite slurry with 10 vol.-% 50 µm SiC particles was adequate to make a variety of castings at 680°C.

Charpy impact behaviour of SiC fibre reinforced titanium composites (TMCs)has been examined at a range of temperatures on specimens with fibres orientated longitudinally [0]₈ and transversely [90]₈ to the impact direction. Corresponding monolithic alloys have also been studied. Although monolithic Ti–6Al–4V has the superior ‘bend’ strength at room temperature, Timetal 834 is superior at the operating temperatures of gas-turbine components. The TMCs follow the same trends as the monolithic alloys, but their impact strengths are inferior to the monolithic alloys by an order of magnitude. This is attributed to the low impact strength of the SiC fibres, which dominates in [0]₈ specimens, and the low strength of the fibre/matrix interface, which dominates in [90]₈ specimens. For [0]₈ specimens, Ti-6-4 reinforced with carbon coated SiC fibres (Sigma SM1140+) has consistently superior impact strength compared with the other three TMCs. For [90]₈ specimens, Timetal 834 reinforced with TiB_x/carbon coated SiC fibres (SM1240) exhibits the best impact strength at intermediate temperatures.

The aim of this paper is to report the experimental results concerning the microstructural evolution and formability of semisolid forming SiC particulate reinforced AZ91HP Mg composites. The composite materials were produced in the form of cylindrical billets by a rotating cylinder method. The microstructural evolution was characterised by conventional parameters (liquid fraction, average size of globules, and number of globules per unit area) as functions of temperature and isothermal holding time in the semisolid state. In addition, special attention was paid to the shapes of the globules and the liquid droplets entrained within the globules and to the variation of hardness with regard to microstructural evolution. Finally, the formability of the composites was evaluated through thixocasting trials, and then correlated with the effective liquid fraction of the composites.

The thermal cycling creep characteristics of aluminium matrix composites AA 6061+22 vol.-%Al₂O₃particles and A359+20 vol.-%SiC particles were analysed with respect to the dependence of the behaviour on the applied external stress and on different heat treatment conditions. The specimens were thermally cycled using a triangular waveform between 50 and 300°C with dwell times of 3 s at the minimum and maximum temperatures. An enhanced lifetime was found for reinforced aluminium alloy 6061 compared with A359+20 vol.-%SiC and it was found that overaging of the matrix prior to the thermal cycling experiments increases the creep rates significantly in both composites. The determination of the true, thermally compensated strain revealed a much higher in cycle strain evolution in A359+20 vol.-%SiC. Reverse creep during the cooling sequence was observed in both composites where it depends on the applied stress.

The hot deformation behaviour of a 20 vol.-%SiC particle reinforced Al alloy 6061 composite and Al alloy 6061, made via a powder metallurgy (PM) route and treated by a T4 temper, were studied by compression testing over a range of temperatures (300–500°C)and at strain rates of 0.005, 0.05, and 0.09 s^-1. It was observed that the flow stresses of the composites were significantly higher than those of the alloy at lower deformation temperatures. However, the stress–strain curves of both the composite and the alloy showed significant softening during deformation at the lowest strain rate, the softening for the composite being faster than that of the alloy. The activation energy for hot deformation was determined for different strains, using a power law equation, and was found to change significantly with strain for both the alloy and the composite. This phenomenon was explained by the occurrence of dynamic precipitation and coarsening during deformation.

Alumina/aluminium composites with an interpenetrating microstructure were made by infiltrating an alumina preform which had the structure of a reticulated ceramic foam. Low density preforms (6% relative density)with porous struts were prepared from a polyurethane suspension of alumina powder which was pyrolysed and partially sintered after foaming. Higher density preforms consisted of ceramic foams with open cells and dense struts. All these preforms were infiltrated with 6061 aluminium alloy using a modified squeeze caster fitted with a vacuum system and fine control of speed and pressure. Such MMCs had a microstructure in which two 3D networks were interpenetrating. One was either a ceramic reinforced alloy (50 vol.-%reinforcement) or a dense ceramic phase while the other was the unreinforced alloy. The microstructure of the preform fitted an established relationship between the ratio of window diameter to cell diameter k and void volume fraction V _p. In addition, a short fibre foam was made to compare with the traditional fibre preforms.

By employing spherical agglomerate of submicrometre Al₂O₃ particles, a novel composite with controlled non-uniform reinforcement distribution was produced by the squeeze casting technique. The resulting composite has a microstructure in which composite spheres, consisting of Al reinforced by small Al₂O₃ particles, are uniformly distributed in a particle free Al matrix. The response of mechanical properties to the controlled non-uniform distribution of the reinforcement was investigated and the fracture mechanism was discussed in order to understand experimentally the influence of particle clustering on the deformation characteristics. It is demonstrated that a composite with such a controlled microstructure exhibits significant increase in ultimate tensile strength (UTS), elastic modulus, and also a slight increase in ductility compared to composite with a homogeneous microstructure at a given volume fraction. It is proposed that the composite spheres act as units of reinforcement when subjected to deformation and contribute positively to the modulus, UTS, and hardness. The higher ductility is attributed to the extensive ductile deformation of the Al matrix. In addition, the highly restrained deformation observed in the interagglomerate region can also be beneficial to the high strength.

In situ metal matrix composites (MMCs) with Mg₂Si particulate reinforcement have been developed recently as ultralight materials. In this paper, a brief overview of the physical and mechanical properties of Mg₂Si and the current status of research on Mg₂Si reinforced MMCs is presented, followed by more detailed information on recent progress in the research group of the present authors. The effects of element additions and processing parameters on the microstructure of the composites obtained by gravity casting are discussed, together with some mechanical property data.

Aerospace and industrial gas turbine engine components such as discs, blades, and propeller shafts for which long fibre reinforced metal matrix composites are being made, operate under complex mechanical and thermal loading conditions. A major aim of functionally graded materials application is to optimise component response through appropriately tailored microstructures. This paper explores the influences of property gradation, centrifugal body force loading, and thermal loading on stresses in rotating discs. The discs were modelled as non-homogeneous orthotropic materials such as those obtained through non-uniform reinforcement of a metal matrix by long fibres. The results show how different temperature change distribution patterns and property gradation types correlate with hoop stresses developed in the disc.

The incorporation of low density, high modulus ceramic particles into a steel matrix is a potential route to improve the mechanical performance of steels. A powder metallurgy, mechanical blending route has been adopted to produce a homogeneous distribution of TiB₂ particles in both pure Fe and 316L stainless steel matrices. This approach gave large increases in both the static and fatigue strength with increasing TiB₂ volume fractions, in comparison with the matrix material. Additions of TiB₂ also resulted in reduced density and increased stiffness in the composite or lightweight steel materials, giving a specific stiffness increase of 52%with a fraction of 30 vol.-%TiB₂ in Fe, compared with the matrix.

The use of sandwich structures is a well documented and accepted engineering approach. Composite sandwich structures, consisting of a fibre reinforcement impregnated with resin surrounding a foam core, can be used to improve bending stiffness, thermal insulation, and specific mechanical properties at relatively low cost. Such materials can be produced by the resin transfer moulding (RTM)process by incorporating the core within the fibre preform before resin injection. However, this process may result in a number of defects including core shifting, deformation, and delamination, which will have a detrimental effect on mechanical performance. The occurrences of these defects have been studied as a function of the fibre volume fraction, resin pressure, and skin thickness. Delamination between the skin and the core is also a major concern, although this can be controlled by abrading the core surface before moulding. The effect of process parameters on moulding quality and, in particular, the response of the core during moulding and the subsequent delamination resistance of the resulting sandwich structure are presented. This is applied to a range of commercially available core materials.

Metallic foams show some potential for being produced with controlled spatial variations in their density. This suggests employing them as graded materials in space filling lightweight structures designed in analogy to cortical bone, a natural cellular material, that displays increased density in regions of high loading. In the present study the influence of the mechanical properties of aluminium foams on the results of an optimisation of the foam density distribution with regard to structural strength and stiffness was examined. Regression formulae for the relationships between stiffness and strength of metallic foams on one hand and effective density on the other hand can be fitted to the results of uniaxial compression tests of a certain brand of metallic foam. These results and additional assumptions such as overall isotropy and a yield surface suitable for cellular materials can be implemented into a finite element program adapted for performing stiffness or strength optimisation on the basis of a density adaptation similar to the remodelling of bone. Some applications are presented that show how foams with gradients in the apparent density may be employed to obtain optimal structural behaviour for classical design problems.