Wear behaviour of interpenetrating alumina–copper composites

Abstract

The wear behaviour of a variety of alumina–copper interpenetrating composites was tested as a function of copper ligament diameter and volume fraction of copper. The wear mechanisms of pure copper and pure alumina were adhesive and abrasive wear, respectively. In the composites with 1 μm, 5 μm and 15 μm copper ligament diameters, the wear mechanism was a mixture of adhesive and oxidative; in composites with a 30 μm copper ligament diameter a mixture of abrasive and oxidative. Increasing the amount of copper decreased hardness and thus increased wear, except where cyclic tribolayer behaviour occurred or where the alumina grains were weakly bonded. Increasing the copper ligament diameter decreased wear, although this trend was only clear under a load of 20 N. The composites with the highest wear resistance had the highest copper ligament diameter of 30 μm. This was probably due to the higher heat conductivity and fracture toughness caused by the coarse, fibrous copper network replicating the wool felt used to produce it. This was possibly also because these composites had the smallest alumina grain size.

Introduction

Metal ceramic composites are an exciting field of research, enabling a combination of properties not possible in metals or ceramics alone. Alumina–copper composites enable the high heat and electrical conductivity and high toughness of copper to be combined with the high stiffness and hardness of alumina [1].

There are three broad types of ceramic–metal composites. Metal Matrix Composites (MMC) have a continuous metal network with V_metal > V_ceramic. Ceramic Matrix Composites (CMC) have a continuous ceramic network and V_ceramic > V_metal. In Interpenetrating Network Composites (INC) both components are continuous. The only limitation on composition is that there is enough of one phase for it to percolate. Depending on the form of the phase this is around 20–30%, less for fibres and more for particles. The advantages of a continuous ceramic network include a higher Young's modulus, hardness and load bearing capacity than MMCs.

Interpenetrating metal ceramic composites are generally produced via squeeze casting [2] or gas pressure infiltration [3] due to the poor wettability of most metals on oxide ceramics. The wetting angle of copper on alumina is extremely low with a contact angle varying between 124° [4] and 170° [5]. By adding CuO to copper, wettability increases so much that porous alumina can be infiltrated without external pressure [6]. The drawback is that an aluminate forms at the interface of copper and alumina [7] which can decrease the interface strength [8]. Upon cooling after infiltration, large residual stresses build up in the composite (tension in copper and compression in the alumina) due to the difference in thermal expansion coefficient [9], [10]. Scientifically this is an interesting aspect of ceramic–metal composites, since it can affect the mechanical properties and thermal expansion behaviour.

The combination of properties possible in alumina–copper composites makes them particularly interesting for wear applications, for example in automobile, aerospace and even bicycle parts.

Until now relatively few wear studies have been published on alumina–copper composites, with only one found by the authors in the literature. In that study, copper was reinforced with particulate alumina [11]. The authors did not find any articles on the wear of alumina–copper interpenetrating composites, or indeed any copper-based interpenetrating composites. A similar system which has already received quite some attention is alumina–aluminium composites. Aluminium–alumina MMCs have been explored for over 3 decades, interpenetrating composites since the mid-1990s. One drawback of this combination is the relatively low melting point of aluminium (660 °C) limiting the possible temperatures of application. Since copper has a higher melting point (1083 °C), alumina–copper composites have the potential for applications in a wider temperature range. Copper also has a higher thermal conductivity than aluminium (401 W/m K compared with 236 W/m K at 0 °C [12]), allowing heat generated at the wear surface to be more rapidly dispersed. This reduces thermal mismatch stresses and the likelihood of thermal shock.

Wear is not a materials property, but rather a system response [13]. There are many factors influencing wear behaviour, which make it difficult to compare results from different laboratories or different testing methods. Even results from a single laboratory and using the pin-on-disc dry sliding wear test have been shown to have a variation of 28–56% [14]. Despite this, it is interesting to compare trends found in previous studies. Relevant results from the literature have been summarized below.

In aluminium MMCs, increasing alumina content has been found to increase wear resistance [15], [16], [17]. With fibrous alumina reinforcement this effect was only seen up to a content of 20%, above which there was no further improvement [16]. There were no studies found on the effect of the amount of alumina reinforcement in copper MMCs, but in Cu–TiB₂–TiN MMCs, the same trend as above was found [18]. Surprisingly, El-Hadek and Kaytbay found the wear resistance of pure copper to be higher than that of copper–alumina MMCs with 15% alumina and grain sizes of 6 and 20 μm [11]. When reinforced with 100 nm alumina grains the wear resistance of the composite was higher. The larger grains could be more easily removed from the copper matrix, thus increasing weight loss, whereas the 100 nm grains were more strongly bonded to copper. In alumina–aluminium interpenetrating composites, increasing alumina content was also found to increase wear resistance [19], [20]. It was also found that interpenetrating composites have a significantly higher wear resistance than MMCs [21], [22]. It would be expected that the same trend be found in alumina/copper interpenetrating composites.

Increasing metal cell size was recently found to increase wear resistance in alumina–aluminium interpenetrating composites [9], [19]. It was postulated that a larger distance between thicker alumina struts provided a more effective shielding of the composite. There are conflicting results in the literature on the effect of ceramic grain size on the wear resistance of MMCs. Some studies found increasing particle size increases wear resistance in aluminium MMCs [15], [16], stating that increased particle size causes more of the load to be carried by the hard particles thus lowering the wear of the softer matrix. Others found that wear resistance increases with decreasing particle size, both in Al–MMCs [23] and Cu–MMCs [9], [11], [24]. These studies postulated that larger particles are more likely to contain microcracks, therefore fracture more readily and decrease wear resistance. The effect of ceramic grain size on the wear behaviour of interpenetrating composites has not been found by the authors in the literature. This is not surprising, since the (fully sintered) ceramic phase is continuous throughout the composite and one would not expect the grain size to have a significant effect. In the wear of pure alumina, refining grain size was found to slightly increase wear resistance in the mild wear regime, and significantly delay the transition from a mild to a severe wear regime [25]. It is generally agreed that increasing the applied load increases the wear rate [17], [18], [19], [21], [23]. In the mild wear regime this only seems to have a small effect. However, the transition from mild to severe wear, at which point a sudden increase in wear rate of around two orders of magnitude, occurs earlier (in terms of sliding distance) with increasing load [17].

In this study, alumina–copper composites have been produced by gas–pressure infiltration of liquid copper into a porous ceramic perform [26]. The porous ceramics were produced using different sacrificial preforms in order to achieve a wide range of microstructures and metal contents, as in [27]. Using this method, a metal content from 20 to 55% and a metal ligament/cell diameter of 0.5–30 μm has been achieved. This allows for a variation in properties which can be tailored to suit the application. The effect of metal content, ceramic grain size and metal cell or ligament diameter on the wear behaviour of the composites has been explored. Selected samples were also tested at a different load.