February 2005

Ceramics and nano 'to stiffen MMCs' competitive edge'

Research has revealed the desirable characteristics to be gained from combining PM matrix composites with other materials - and the difficulties involved in manufacturing such products at a reasonable price. Joe Capus reports on a session at last Autumn's Euro PM World Congress in Vienna…

The combination of metallic alloy matrix materials with ceramic particulate reinforcement in order to achieve the property advantages of both in the form of composites has long been the subject of serious research and development efforts.

Although there have been some notable successes over the years, for example in the nuclear energy and aerospace sectors, numerous applications await the development of cost-effective advanced materials, particularly for light-weight automotive components. However, synthesising composite materials for the purpose of developing specific properties is one thing: developing advanced materials for the manufacture of practical components can be quite another challenge.

This aspect was brought out in some of the presentations at a session on metal matrix composites (MMCs) during the PM2004 World Congress in Vienna. Professor Graham Schaffer (University of Queensland, Australia) made the point clearly in a paper with colleague Shuhai Huo on sintered aluminium matrix composites.

While aluminium is readily process by conventional metal-forming techniques, composites reinforced with ceramic particles are difficult to process and expensive to fabricate. That explains the interest in the potential of press-and-sinter PM as a net-shape technology for producing components from MMCs with high- performance properties at competitive cost. Such considerations were highly relevant to potential applications of PM aluminium composites in automotive components, where tighter tolerances and better materials utilisation achievable by PM processing offer a cost advantage over die-casting. With proper alloy design, the increased stiffness and strength of PM MMCs could also open up potential substitution for certain automotive ferrous PM parts, including gears and gerotors.

Professor Schaffer noted that alloy design was like a jigsaw puzzle of properties and processing parameters, with "Cost" in the middle: the job was not finished until one had all the pieces in place. He went on to describe some aspects of the development of PM aluminium composite materials based on a proprietary blended elemental Al-Cu-Mg-Si-Sn powder composition. The air-atomised aluminium powder supplied by the project sponsors AMPAL, a subsidiary of United States Bronze Powders, had a mean particle size of 68.7 microns.

For the purpose of this investigation, the ceramic powder was screened to various narrow particle size fractions. After mixing with 1 per cent Acrawax, the powders were compacted at 200 MPa and sintered in dry nitrogen at 300°C for 20 minutes (for delubrication) and then at 600-630°C for up to one hour. Air-cooled samples were solution-treated at 502°C, water-quenched and aged for 18 hours at 171°C, before submitting to a battery of tests for elastic modulus, fatigue strength and wear testing, as well as microscopic examination.

The presence of inert, rigid particles had a significant influence on the sintering behaviour of the matrix powder. Apart from retarding the sintering response, the ceramic particles affected the sintered density in a way that varied with the relative size of matrix and ceramic particles. Too large a percentage of fine ceramic particles tended to cause expansion rather than shrinkage during sintering, while coarser ceramic particles led to higher sintered densities. When the ceramic particles size averaged 2.5 times the size of the matrix particles, up to 40 per cent (by volume) could be added without causing expansion. The microstructure of a successfully sintered composite with 35 per cent ceramic reinforcement is shown in Figure 1. This degree of reinforcement increased the Young's Modulus by more than 70 per cent (Figure 2). On the other hand, the fatigue strength was drastically improved with only 5 per cent ceramic particles and additional reinforcement was less effective in this regard (Figure 3).

Figure 1: Microstructure of sintered Al-MMC with 35 per cent ceramic reinforcement. Back-scattered SEM; matrix: light grey; ceramic: dark grey; pores: black. After Schaffer and Huo.

Figure 2: Composite stiffness as a function of reinforcement fraction. After Schaffer and Huo.

Figure 3: Composite fatigue strength versus reinforcement fraction. After Schaffer and Huo.

In wear tests using a dry-sand-rubber-wheel apparatus (ASTM G65-94), aluminium-MMC test pieces were compared with an aluminium-silicon die-casting alloy. Wear resistance of the die-casting alloy was superior to the PM aluminium alloy with no reinforcement, but with 5 per cent ceramic, the Al-MMC was equal to the die casting and much superior with 10 per cent.

Some of the challenges facing the practical production of copper-based PM MMCs reinforced with multi-wall carbon nanotubes (MWCNTs) were explored by Professor Jung-Ho Ahn and colleagues at the Andong National University and other institutions in Korea. The very high elastic modulus and high specific strength exhibited by carbon nanotubes make them potentially attractive for use as reinforcing materials in composites. However, as the authors noted, limited success has been achieved so far in MMCs due to the difficulties in dispersing the CNTs in the matrix material.

As-synthesised CNTs tend to agglomerate into entangled aggregates or bundled "nanoropes", preventing uniform distribution within the matrix and also the ability to create a good interface between the CNTs and the matrix metal. Some success had previously been found with hydrogen plasma-evaporated nanopowders and sophisticated processing. However, for the practical application of metal-based CNT composites, Ahn and colleagues have concentrated on exploring low-cost processing methods for making MWCNT-reinforced copper composites. Using two kinds of MWCNTs, synthesised by a CVD method, they tried a variety of procedures to disperse and consolidate the CNTs with minus 325 mesh copper powder. The CNTs were first purified by subjecting them to 10 hours of ultrasonic treatment in a 3:1 mix of concentrated sulphuric and nitric acids at 60°C, followed by washing and drying. The purified CNTs, either in aggregate form or bundled form, were dispersed either by 20 hours treatment in an ultrasonic bath with an aqueous dispersant solution or by ball-milling, dry or in ethanol, for five to 20 hours using hardened steel balls. The resultant ball-milled or ultrasonic treated MWCNTs were then mixed with the copper powder. Three methods of consolidation were investigated:

• Hot-pressing in argon atmosphere at 900°C for one hour;
• Spark-plasma sintering (12 ms pulses of 5000A/10V, followed by heating to 900°C and holding for 5 minutes); and
• Sinter-rolling-sintering, whereby the powder mixtures were pressed at 350 MPa, sintered in vacuum at 900°C for two hours, then rolled 50 per cent, followed by a second sintering at 900°C for one hour.

The resulting MWCNT/Copper composites were characterised by a variety of techniques. When the different types of milling were compared, it was found that best results were obtained by wet milling with moderate energy, using a horizontal ball mill, while dry milling caused a rapid collapse of the CNTs to amorphous carbon. The consolidation methods employed also showed a wide range of results. Sintering alone gave a poor consolidated density of only 50 per cent to 60 per cent of theoretical, due to the presence of CNT aggregates. The other methods showed much improved densities, all close to 90 per cent of theoretical. Spark-plasma sintering resulted in a slightly higher density than the other methods, but combined sintering and rolling showed comparable densities.

The latter processing procedure was said to be easy and suitable for inexpensive mass production of CNT composites. Tests of specific electrical resistance of the MWCNT/copper composites showed that the wet-milled and sinter-rolled-sintered composites exhibited a resistivity of about 2 x 10-8Wm , comparable with that of high-purity copper, despite the density being about 90 per cent of theoretical.

The authors concluded that the dispersion of the nanotubes was one of the most crucial factors to be overcome in the production of high-performance composites. Despite the promising nature of this preliminary study, the results showed there were significant challenges to be overcome in developing low-cost, mass production of CNT composites.

Turning from the exotic world of nanotubes to the more familiar ground of PM stainless steels, David Baxter (QinetiQ, UK) and co-authors A Tarrant (Aerospace Metal Composites, UK) and R Valle (Centro Sviluppo Materiali, Spa, Italy) reported on MMC development work funded under the GROWTH activity within the European Fifth Framework Programme.

The authors described the results of an investigation of the influence of materials and processing parameters on PM MMCs based on stainless steel matrix alloys reinforced with titanium diboride (TiB2) particles. The stainless steels were austenitic 316L, precipitation-hardenable 15-5PH and A286 in the form of minus-150 micron powders. Each of the stainless matrix powders was blended with 10 per cent to 30 per cent of minus-5 micron TiB2, followed by mechanical alloying using a proprietary milling process. After milling, the powders were consolidated by canning and HIPing at 1120°C for four hours under a pressure of 100 MPa. Tensile, fatigue and wear properties of the resultant MMCs were studied.

The mechanical alloying treatment was shown to be a significant factor in the homogeneity of dispersion of the reinforcing TiB2 particles, with tensile strength and ductility as well as fatigue strength improving markedly as the milling time was increased. For 316L + 25 per cent TiB2, tensile strength was increased from 800 to 1100 MPa, while ductility improved from 1.2 per cent to 2 per cent and fatigue strength improved to 700 MPa. However, optimum milling parameters needed to be determined for each matrix/reinforcement particle combination. Even higher strength and ductility combinations were obtained with the heat-treatable 15-5PH and A286 stainless steels containing 10 per cent or 20 per cent TiB2, particularly after ageing of the martensitic 15-5PH (Figure 4). Density, elastic modulus and tensile strength were seen to vary linearly with volume percentage of TiB2 reinforcement (Figures 5 and 6). Ductility fell significantly as the TiB2 content increased from 15 per cent to 30 per cent, but could be improved by further forming operations such as forging and extrusion after HIPing. This is illustrated in Figure 6, where extruded material showed almost double the tensile elongation, versus as-HIPed MMC.

Figure 4: Effect of stainless steel matrix and heat-treatment on tensile strength and ductility of MMCs. After Baxter et al.

Figure 5: Effect of TiB2 volume fraction on density and elastic modulus of MMCs. After Baxter et al.

Figure 6: Effect of TiB2 volume fraction on proof stress, tensile strength and elongation of MMCs. After Baxter et al.

Wear resistance of the stainless steel MMCs was tested by the ASTM G99 pin-on-disc method, where both pin and disc were made of the same MMC material. Wear of 316L was reduced spectacularly by the addition of 25 volume per cent of TiB2, which brought it to a level similar to that of carburised S82 steel. Proportional improvements in wear resistance were found with 15-5PH and A286 containing 19 per cent TiB2 reinforcement.

Microstructural examination of the composites raised some concern about reaction between the steel matrix and the reinforcement particles. For 316L + 25 per cent TiB2, there was evidence of a chromium-rich intermetallic phase that could have detrimental effects on corrosion resistance and strength. However, for the A286 matrix material, formation of the intermetallic precipitates appeared to be suppressed. Although more work was needed to understand these interactions, the finding indicated that the chemistry of the matrix alloy was an important factor in determining the properties of steel MMCs.