Experimental and numerical study of silicon nitride for powder injection moulding

Poh, Ping Yi, Leslie (2019) Experimental and numerical study of silicon nitride for powder injection moulding. PhD thesis, University of Glasgow.

Abstract

The continued research interest in ceramic injection moulding originates from the pursue to develop new materials and improve mechanical properties. Mould filling simulation has provided the freedom to test out mould designs and new material numerically with minimal material wastage. Therefore, the use of numerical simulations is useful in the development of ceramic injection moulding. However, extensive material data are required to perform these numerical simulations and the results of which need to be validated for any meaningful conclusions to be made.

At present, the simulation of ceramic injection moulding requires extensive material input data of the powder-based feedstock. It is time consuming and costly to characterise a powder-based feedstock for the necessary thermomechanical properties for simulation. This restricts the use of numerical simulations when a new feedstock material or composition is being developed. Moreover, the phenomena of powder-binder separation that occurs during mould filling are not captured by commercial simulation. The defects caused by this phenomenon will affect the quality of the component and cannot be resolved in latter processes. This is the gap this research aims to fill. The objective of the work presented in this thesis is to develop a hybrid material data that can be used in simulating powder-binder separation observed during mould filling. The research work presented in this thesis is structured into three parts.

First, a hybrid material data of the powder-based feedstock is developed. The hybrid material data consists of a combination of estimated and experimental material data. A comparative study was conducted on experimental material properties of powder-polymer mixture found in literature and existing models that estimates elastic modulus, thermal conductivity, coefficient of thermal expansion and specific volume. The comparative study showed that the estimation models selected for this hybrid material data predicts the material properties better than existing models used in past studies. Experiments are conducted to measure the density, specific heat capacity, viscosity and particle size distribution of the silicon nitride feedstock developed in this research work and the elastic modulus, thermal conductivity, coefficient of thermal expansion and specific volume are estimated with the selected estimation models.

Second, powder distribution analysis methods are developed to measure the powder distribution within the injection moulded green bodies which will be used to validate the numerical simulations. The differential scanning calorimetry (DSC) and thermogravimetric analysis (TGA) are developed to measure the volume fraction of ceramic powder in the feedstock and green bodies. An empirical model is developed using the DSC test results and compared with existing rule-of-mixture model at predicting the volume fraction of silicon nitride powder in green bodies. The empirical model showed results closer to the nominal volume fraction of silicon nitride powder and lower variation in measurements at each point of the green body as compared to the rule-of-mixture model. Between the DSC and TGA methods, the TGA tests measured the closest volume fraction of silicon nitride powder to nominal volume fraction and lowest variation.

Third, a suspension balance model and a two-fluid model are developed to simulate the phenomena of powder-binder separation during mould filling of powder-based feedstock. The effects of overestimating elastic modulus, thermal conductivity and coefficient of thermal expansion on injection moulding numerical results are studied with the newly developed hybrid material data and existing hybrid material data. The filling and packing results from the simulation models were used to investigate the filling patterns of the simulation models and the powder distribution within the green body. The numerical results showed that segregation of powder-binder is more prominent in the regions close to the gate and rear of the green body. This is likewise seen in the experimental results from the powder distribution analysis of the injection moulded silicon nitride test-bar.

This research work contributed to a hybrid material data of a powder-based feedstock that can be used as material input data for powder injection moulding simulation. The powder distribution analysis method that is developed is able to measure the distribution of powder within injection moulded green bodies. Furthermore, injection moulding simulations developed are able to simulate the powder-binder separation in silicon nitride green bodies. By developing a hybrid material data that allows several material properties to be estimated and used for numerical simulation, it provides flexibility and reduces cost for numerical investigation of powder-based feedstock in injection moulding.

• Item Type: Thesis (PhD)
• Qualification Level: Doctoral
• Additional Information: Supported by funding from the Economic Development Board of Singapore.
• Keywords: ceramic injection moulding, silicon nitride, thermomechanical properties, powder-binder segregation, numerical simulation, computational fluid dynamic, powder-binder distribution study.
• Subjects: T Technology > TA Engineering (General). Civil engineering (General); T Technology > TJ Mechanical engineering and machinery; T Technology > TS Manufactures
• Colleges/Schools: College of Science and Engineering > School of Engineering
• Supervisor's Name: Goh, Dr. Cindy S.F.
• Date of Award: 2019
• Depositing User: Mr Leslie Poh
• Unique ID: glathesis:2019-41131
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 09 Jul 2019 13:01
• Last Modified: 05 Mar 2020 21:32
• Thesis DOI: 10.5525/gla.thesis.41131
• URI: https://theses.gla.ac.uk/id/eprint/41131

Related URLs

• Article DOI
• Enlighten Publications Record