Programming and simulation of densification of ZTA nano-composite

Abstract

Alumina is an important structural ceramics having wide application potential. However, its use in real life systems is limited due to its low fracture toughness resulting in catastrophic failure. Among the several measures that are being taken to overcome the draw- back, the addition of zirconia (platelet, particle, fiber) in the alumina matrix remarkably improves the toughness on account of stress induced transformation toughening. Al2O3–ZrO2 composite powder containing 15 mol% ZrO2 was prepared by sol-gel route using Aluminum nitrate (E-Merck, India) and zirconium oxychloride (E-Merck, India) as precursor. The sol–gel derived powder was properly characterized by X-ray Diffraction study (XRD). The analysis reveal the gel contained pseudoboehmite and amorphous Zr(OH)4 was decomposed in three and two stages respectively. The phase transformation of alumina during calcination followed the sequence of pseudoboehmite → bayerite → boehmite → γ- Al2O3 → θ-Al2O3 → α-Al2O3, while that of ZrO2 follows amorphous ZrO2 → t-ZrO2 → (t + m) ZrO2. Densification behavior of the sol gel precursor was studied through compaction of the powder at uniaxial press and subsequently sintered at muffle furnace. The relative density of the uniaxial pressed specimen depends on several factors; such as pressure, temperature, sintering profile and sample position during firing. An Artificial Neural Network (ANN) can predict the characteristics densification features of sintered specimens at any intermediate processing parameters. To evaluate this, simulation of relative density of ZTA Nano-composite has been carried out through MATLAB program. The simulation was focused on stepwise single layer to multilayered ANN theory. Herein, the project work is emphasized on the prediction of change in densification behavior with respect to change in processing parameters.