| Summary: | Porous pseudo-elastic NiTi alloy with a nearly three-dimensionally interconnected pore structure with structural characteristics that show promise for implant applications has been successfully fabricated by the metal injection moulding (MIM) process, followed by transient liquid phase sintering, using a mixture of Ni and Ti elemental powders. Three different powder volume fractions ranging from 65.5 to 69.5 vol% with a nominal Ni-Ti composition of SO.9at%Ni were mixed with a binder system, comprising a mainly water-soluble binder system known as polyethylene-glycol (PEG), in a new technique using a speed mixer, principally incorporating a dual asymmetric centrifuge (DAC). The powder-binder mixture was then analysed using a capillary rheometer at various temperatures and shear rates. It was found that the feedstock exhibited pseudo-plastic behaviour, which is favourable for the MIM process. A temperature range of 120°-130°C was considered as the optimum operating condition for the injection moulding process. The parts were moulded into cylindrical shapes, leached in warm water (60°C for 10 hours), thermally debound in argon and subsequently sintered in a vacuum furnace at different temperatures ranging from 950°C to 1250°C and with different holding times. The physical, thermal and mechanical behaviour of the as-sintered parts, in terms of pore morphologies, phase constituent analysis, phase transformation temperatures and load-unload compression test, were systematically investigated. The binder system used in MIM not only serves as a temporary vehicle to support the metal powder, particularly during mixing, injection and debinding, but also can act as a pore former, particularly during the water leaching and thermal debinding processes which finally facilitate the enlargement of pore channels during the subsequent sintering process. On sintering, the particles bond together first into a network, forming several inter-metallic phases such as NiTiz, NiTi, Ni3 Ti and Ni4 Ti3, with fine scale porosity in that network due to Kirkendall effects. The formation of a transient liquid phase (TLP) close to the eutectic composition leads to rapid bonding. This liquid covers the surface of Ni particles by capillary force and consequently enlarges the pore channels. As the alloy homogenises, this liquid phase disappears, and asymmetric diffusion of the nickel into the titanium particle network results in swelling and macroscopic expansion of the structure. Increasing the sintering temperature and holding time enhances the inter-diffusion between the elemental powders, leading to a major fraction of the desired B2 NiTi phase, whilst minimizing other secondary phases such as Niz Ti, Ni3 Ti and Ni4 Ti3, which are known to be brittle and unresponsive to pseudo-elasticity. For all the processing conditions employed, the variation of porosity, as well as the average pore size, was very small; 3S-4S% and 80-120 ~m, respectively. However, it seems that the use of greater powder loading (69.Svol%) and greater sintering temperature (> 10S0°C) was remarkable in terms of better isotropic dimensional changes after sintering, due to greater inter-particle friction, and low impurity content, due to the smaller amount of binder used and the greater amount of transient liquid, leading to better phase homogenisation. As a result, the ductility of the porous samples was enhanced remarkably (maximum strength and strain are> 400 MPa and> 30 %, respectively), and is slightly higher than with other PM routes such as SHS and HIP techniques using a mixture of elemental Ni and Ti powders. The porous samples also exhibited a quite promising pseudo-elasticity; the elastic deformation for all samples was around 10% strain and> 80% unloading recovery at room temperature for the load-unload compression up to 8% strain, and the recovery was nearly completed (<5% plastic strain) after the samples were heated above Aj temperature. The result is much higher than that of the conventional elastic deformation of ordinary alloys. Further, the average stiffness calculated from the stress-strain curves was around 2-3 GPa during loading and S-8 OPa during unloading, very close to that of cancellous bone(<3 GPa), which makes these alloys attractive as bone implants in biomedical applications.
|
|---|
