overcome in the technology for producing high performance metal matrix ... Metal
-matrix composites, especially those reinforced with nano-particles, enjoy the ...
TMS2013 Annual Meeting Supplemental Proceedings TMS (The Minerals, Metals & Materials Society), 2013
APPLICATION OF EXTERNAL FIELDS TO TECHNOLOGY OF METALMATRIX COMPOSITE MATERIALS N. Hari Babu1, Zhongyun Fan', and Dmitry G. Eskin1 'Brunei Centre for Advanced Solidification Technology, Brunei University; Kingston Lane, Uxbridge, UB8 3PH United Kingdom
Keywords: external field; shear; ultrasound; cavitation; metal matrix composite Abstract Introduction of nonmetallic particles to the melt of light metals is the main and first bottleneck to overcome in the technology for producing high performance metal matrix composites (MMC). The second bottleneck is the uniform distribution of these particles in the liquid and, eventually, the solid matrix. External fields applied to the liquid/particles mixture are known to facilitate wetting, mixing and distribution of the particles. This paper gives an overview on the application of intensive melt shearing and ultrasonic cavitation to the technology of MMC based on aluminum. The mechanisms are discussed and examples are given. The outlook of a new European program on the application of external fields in liquid metal processing ExoMet (EC grant 280421) is presented. Introduction Metal-matrix composites, especially those reinforced with nano-particles, enjoy the revival of research and industrial interest caused by the new developments in materials processing and the increased demand for light-weight structural and functional materials. Use of high strength lightweight structural materials in transport applications improves fuel efficiency of vehicles thus effectively reducing CO2 emission. Contribution towards reducing carbon emissions plays a crucial role in achieving sustainable development for which a number of different strategies from the transport industry are required. Metal matrix composites (MMC) in which metallic matrices (AI, Mg) are reinforced with high strength and high modulus phases, such as carbides (SiC, B4C), nitrides (S13N4, A1N), oxides (AI2O3, S1O2), as well as elemental materials (C, Si) represent this larger class of light-weight structural materials [1]. The shape of the reinforcement can vary from long microscopic fibers to nano-scale particles and tubes. In this paper we will mostly consider particle-reinforced MMCs. Such MMCs have been shown to offer improvements in strength, rigidity, temperature stability, wear resistance, reliability and control of physical properties such as density and coefficient of thermal expansion, thereby providing improved engineering performance in comparison to the un-reinforced matrix [1-8]. AI & Mg based MMCs posses both high specific strength (Fig. 1) and high wear resistance characteristics, making them desirable for a number of structural and functional applications, e.g. ground transportation, thermal management, aerospace. In all these applications, the principal advantage of MMCs is that their physical and mechanical properties can be enhanced to desired level by tailoring their microstructures for specific engineering needs.
1037
250
MMC
^L· /Qvf^ lXSj \ V $£r
Simultaneous improvement in ductility & strength to be achieved bynanopaniculate dispersion in metal matrix
Al alloys
Figure 1. Specific strength vs elongation for steel, Al and Mg. Mg or Al based paniculate MMCs have higher TS/δ, highlighting their importance over other structural metallic metals [9-15]. MC is metal conditioning.
Mg alloys Steel/ Elongation-to-failure {%)
(a)
Conventional & most commonly used process by the industries
Figure 2. Processing steps involved in (a) most commonly used MMC processing and (b) application of external fields such as intensive shearing and ultrasonic cavitation. MMCs have many features in common with grain refining master alloys that, in the case of Al, contain T1B2 and TiC particles distributed in the Al matrix. There are different ways to produce MMCs, e.g. powder metallurgy (limited in size of parts), reactive formation (limited in choice of phases) and liquid metal processing (versatile but challenging). In this paper we will look at the liquid metal processing route: its challenges and opportunities. The typical liquid metal routes for MMCs are stir casting and infiltration. The latter is the most commonly used industrial method and accounts for largest volume in primary production (-60%) [7]. In this method, a ceramic preform of the desired shape is infiltrated with the liquid metal assisted by artificial pressure difference, e.g. squeeze casting. The processing steps are shown in Fig. 2a. It was first widely used in the automotive industry, but is now also a preferred process in the thermal management industry as a result of the ability to produce high quality near net shape or net shape components. Due to the high volume fraction (Vf) of ceramic reinforcement, liquid metal infiltration is not suitable for components that require thermomechanical deformation or for fracture sensitive applications. The stir casting involves fewer processing steps and usually involves stirring of the melt with an impeller with simultaneous introduction of particles and their distribution by the forced flows. However, agglomeration of particles and their incorporation into the liquid and, eventually, solid matrix remain the biggest challenge. Main challenges in the liquid-metal processing of MMCs can be summarized as follows: (1) introduction of particles into the melt assuring the access of the melt to the individual particle surface; (2) wetting of the particle with the liquid phase (melt); (3) preventing agglomeration of
1038
fine particles; (4) preventing flotation of particles to the surface or their sedimentation to the bottom of the liquid bath; (5) dispersion of particles in the liquid phase with de-agglomeration; and (6) preserving the uniform particles distribution in the solid metal. The problem of agglomeration will be addressed in this paper. Preliminary considerations Current challenge An MMC should consist of fine particulates distributed uniformly in a ductile matrix and with clean interface between particulate and matrix. However, the current processing methods often produce agglomerated particles in the ductile matrix and as a result the MMC exhibits extremely low ductility [16-18]. Agglomeration is more severe when the particulate size is in sub-micron or nano-scale range. Severe agglomeration nature of nano-particles, due to high cohesive energy, combined with lack of dispersive technology for mixing poorly-wettable nano-particles have hindered the progress in fabrication of high performance nano-particulate MMCs with liquid processing routes. The possible solution of the agglomeration problem is in application of external fields such as intensive melt shearing and ultrasonic cavitation. If successful this technology can be very efficient, also in shortening the processing route as illustrated in Fig. 2b. Potential solutions for deagglomeration (a) Intensive shearing: Application of sufficient shear stress (τ) on particulate clusters embedded in liquid metal may overcome the average cohesive force or the tensile strength of the cluster and cause erosion and then deagglomeration of the cluster. In addition, under high shearing and turbulence conditions, liquid can penetrate into the cluster and displace individual particles within the cluster, enabling the wetting. The displacement distance of the particle from the cluster center can be tuned by controlling the applied shear stress on the clusters. Molecular dynamics studies [19-21] suggest that the intensive shearing can even displace atoms that are held together with high-strength bonds. Yet another advantage of intensive shear is short duration of the mixing (at difference with long mixing by stirring) that prevents chemical reaction between the particle and liquid metal. (b) Ultrasonic cavitation: Ultrasonic vibrations with high frequency (17-25 kHz) and high amplitude (10 to 40 μπι null to peak) induce in the liquid phase cavitation and acoustic flows. Cavitation is the formation of pulsating and imploding bubbles that generate upon collapse high local temperature and pressure surges. Cavitation bubbles are known to nucleate in liquids on poorly wetted interfaces so the reinforcing particles represent the ideal nucleation sites for cavitation. The pressure produced by cavitation easies the penetration of the liquid phase through very fine capillaries such as the gaps between the particles in agglomerates. Therefore the ultrasonic cavitation will facilitate the deagglomeration by the action of collapsing bubbles within the agglomerates. Cavitation is also known to decrease the surface tension at the solid/liquid interfaces, so that the wetting could be improved. In addition, acoustic and secondary flows induced by the cavitation region in the melt assist in distribution of particles in the liquid volume. There are promising examples on application of ultrasonic cavitation in lab-scale production of Al- and Mg-based composites reinforced with micron- and sub-micron size particles [3, 4, 22]. Distributive mixing There are two distinctive stages of particles introduction and distribution, i.e. distributive and dispersion mixing.
1039
A stirring process is usually employed to pre-mix the Mg or Al melt with hard ceramic particulates. At this stage, it is essential to optimize the conditions to achieve spatial homogeneity of particulates (similar volume fraction in any part of the mixture). Proper mixing techniques and impeller design must be employed, in order to produce adequate melt circulation and homogeneous distribution of the reinforcement throughout the matrix material. Designing an impeller that can distribute agglomerates/particles homogeneously on macroscopic scale in a molten alloy volume is crucial [23]. As the impeller rotates at moderate speeds (up to 800 rpm), a vortex is generated in the melt that draws the reinforcement particles into the melt from the surface. The desirable situation is when the particles are drawn inside the melt from the surface by a vortex but without turbulence so that there is no excessive oxidation and dross formation. In the area of the impeller proper the turbulence is, on the other hand, expected to be present, assisting in distribution of particles in the liquid volume. Creation of high level of shear is also necessary for liquid metal/particulate mixture because of poor wetting of particulates and liquid metal. In such a system, the pressure required to force the liquid to enter porous agglomerate increases as the pore size decreases. The stirring condition, melt temperature, and the type, amount and nature of the particles are some of the main factors to be investigated. The process for distributive mixing can be optimized by conducting experiments with various types of impellers designed to achieve homogeneous distributive mixing in chemical, pharmaceutical, mining and food processing technologies. The advantages of mixing process both in liquid state and in semi-liquid (above the coherency point) state of alloy can also be explored. Let us look at the example of conventional mechanical stirring of Al alloys while adding reinforcement particles [24, 25]. The distributive mixing equipment used in this study is schematically shown in Fig. 3. The setup consisted of a driving motor to create the torque on the impeller, a control part for the vertical movement of the motor and the attached stirrer assembly and a transfer tube for the introduction of the reinforcement particles into the melt. Properly cleaned A356 metal alloy ingots were melted in cylindrical crucibles using a top loaded resistance furnace set at 650 °C. Accurately measured (in vol. %) reinforcing particles were heated at 400 CC inside an electrical resistance furnace. With the help of the transfer tube, preheated particles were added in the melt through the vortex created by stirring. To ensure a uniform distribution, the impeller was designed to have a d, / £>v ratio equal to 0.4 with the impeller blade width, w„ 12.6 mm [26, 27]. Schematic diagrams of the flat bottom cylindrical clay graphite crucible and the impeller geometry are shown in Fig. 3. The impeller was coated with boron nitride to prevent reaction with the molten aluminum. A controlled argon atmosphere was maintained inside the furnace throughout the whole experiment to prevent excessive melt oxidation. The reinforcement particles were transferred slowly and continuously into the melt which was mechanically stirred at 600 rpm to achieve spatial homogeneity of the reinforcement particles (similar volumefractionin any part of the mixture). Position control
Figure 3. Schematic diagram of (a) the experimental set-up for distributive mixing (b) dimensions of a clay graphite crucible and (c) an impeller used for distributive mixing.
1040
Dispersive mixing with intensive shearing The key technology for application of intensive shearing was the development of a twin-screw slurry maker (shown in Fig. 4a), which has a pair of co-rotating and self-wiping screws rotating inside a barrel [28]. The screws have specially designed profiles to achieve high shear rate and high intensity of turbulence. Liquid/particulate slurry in the twin-screw will undergo a shear deformation with cyclic variation of shear rate. The shear stress (τ) is given by τ = ηπΛί(0/σ-2), (1) where η is the viscosity, N is the rotating speed of the screw, D is the outer diameter of the screw, and G is the gap between screw flight and the barrel surface. In order to break-up the agglomerates, the pre-mixed liquid/particulate slurry is fed into the twin-screw machine operating in the liquid or highly fluid temperature region. The slurry is then sheared by the rotating twin screw mixing in and distributing the particles. The applied shear stress (τ) with twin-screw will be maintained higher than the tensile strength of agglomerate T. Based on the assumption of agglomerates as a collection of spherical particles, Rumpf [29] calculated the tensile strength and suggested that T-F/d2, (2) where F is inter particle force and d is the diameter of the particle. The smaller the particle, the higher the tensile strength. Similarly the shorter the inter particle distance (which is likely the case for fine particles), the higher the tensile strength, and thus the higher the shear stress to be applied to break the clusters
Figure 4. Schematic illustration of (a) twin-screw geometry for applying intensive shearing; (b) new rotor-stator high shear device for dispersing reinforcement particles in liquid metals; and (c) ultrasonic cavitation system. Recently a new high-shear device has been developed (Fig. 4b) [30]. In this design the liquid enters the chamber with rapidly rotating impeller that then forces the liquid to exit through fine holes at the sides if the chamber. Small gap between the impeller and the chamber walls along with the high velocity of the expelled through the holes liquid create shear forces required for efficient mixing and deagglomeration of particles. Dispersive mixing with ultrasonic cavitation The high-intensity ultrasonic oscillations create vast number of microscopic bubbles that are distributed within the volume by acoustic and secondary flows. Under the action of ultrasound wave these bubbles rapidly pulsate and then implode creating new bubbles and also pulses of very high temperature (up to 6000 K), pressure (up to 500 MPa) and supersonic jets [31]. The bubbles preferentially form at the interfaces and gas pockets [32]. Therefore the agglomerates of the particles and particles themselves are ideal nuclei for cavitation. The mechanisms of
1041
deagglomeartion can be represented as follows: the cavitation bubbles are formed at the interfaces particle/gas pocket/liquid. These bubbles pulsate intensely, loosening the agglomerate and then implode. The resultant pressure and momentum pulses literary rip the agglomerates apart. The local pressure generated far exceeds the forces that hold together the particles in agglomerates, i.e. 1 to 15 MPa (capillary and adhesive forces) [29], The flows generated by the cavitation zone, distribute the particles further in the volume. Experimental results Microscopic particles Preliminary experiments (Fig. 5) on Al/SiC composites reveal successful dispersion of SiC within the metal matrix in microscopic scale. The intensive shearing allowed uniform distribution of 1-5 μπι sized SiC particulates without the presence of clusters, but with good distribution (average inter-particle distance of ~ 1-2 μπι), while for the ultrasonic cavitation processing these parameters were 5-7 μπι and 3-8 μηι, respectively [33]. Figure 5. (a) A typical microstructure of A356-5 vol. % SiC MMC with particle size 1-5 μπι produced with mechanical stirring and the HPDC process at 630 °C, revealing the presence of SiC particle clusters; (b) same produced with melt shearing; (c) microstructure of a AA6063-20 vol.% SiC MMC billet with particle size 5-7 μπι produced by mechanical stirring; and d, the same produced by ultrasonic cavitation with electromagnetic stirring.
Nano-sized particles Theoretical calculations [34] supported by some experimental evidence [3, 4] reveal that when compared to large (few microns) inclusions, presence of nano-scale inclusions significantly increases the strength while retaining ductility. The tensile strength of the clusters made of nanosized particles is inversely proportional to the square of the particle diameter [29]. Therefore, dispersion of nano-clusters in the liquid metal is a significant processing problem. So far there were reports on successful application of ultrasonic cavitation in dispersion of nano-sized particles in Al and Mg melts. The main challenges are the way to introduce the particles in the melt and large processing time/volume ratio that at the moment prevents the upscaling of ultrasonic processing. The twin-screw machine enables us to vary the shear stress by controlling the shear rate (rotating speed of screws, N) and if necessary the gap (G) between screw flight and barrel can be also decreased to further increase the shear stress. With the current twin-screw machine the shear stress values of up to 50 MPa can be applied for the slurry of Al melt containing SiC. The initial experiments [35-37] on Al alloy with SiC reinforcement show that intensive shearing solves the most significant problem (i.e, the difficulty of achieving a homogeneous distribution
1042
of reinforcement in the matrix composites) associated with the production of MMCs and the simultaneous improvement in ultimate tensile strength and elongation has been observed. Microstructure of Al/SiC composites produced by applying intensive shearing, reveals successful dispersion of nano SiC within the metal matrix in microscopic scale. This demonstrates that processing of MMCs containing nm scale (
