Download files
Abstract
Miniaturisation has been an ongoing objective of the research community with the promise of low cost, small sample consumption, potential for mass production, and good portability. The development of the materials industry is important to the establishment of MEMS. Presently, increasingly more types of polymeric materials are being used to take advantage of the suitability of their different mechanical, chemical and optical properties. The most critical step in the manufacturing process of polymeric MEMS is the bonding or packaging step. The bond should be strong, durable and non-destructive on
the micro-patterns. Among the various bonding technologies available for polymer bonding, solvent bonding has been shown to be the most robust, durable and simplest method. The solvent/solvents is/are applied to the polymer layers to assist in the bonding. Variables such as the solubility parameter, time, and temperature, as well as process techniques, influence the bond strength. However, current solvent bonding techniques are ad hoc and there is no well-defined guideline for the selection of an appropriate solvent-polymer pair. The available research on solvent bonding relies on
empirical selections of the solvents to make a workable bond. There is a lack of a systematic study or reliable statistical analysis of the relationship between the processing conditions and the final bond strength, which would guide the formation of a robust solvent bonding. This thesis details the statistical methods that were used to design solvent bonding
experiments. Process parameters, such as sealing methods, bonding area, shape of bonded wafers, temperature, time, solubility and molecular volume were studied and related to variations in the measured bond strengths. The bond strengths were measured through tensile tests where the data were obtained from stressdisplacements curves. The experimental results were statistically analysed and existing theories for solvent bonding were applied. The experimental results based on Latin squares and Factorial designs showed that the solvent type was the most significant variable affecting the bond strength. The effects of time and temperature were not as significant as the effect of solvent type, but high temperatures were found to increase the bond strength. The Hildebrand s solubility parameters were found not to represent the solvent bond strength correctly, as it had only one parameter. The Hansen solubility parameter was used in conjunction with the relative energy difference, which related the solubility differences between solvents and polymers. The bond strengths from the experiments of solvent bonding at 25°C showed that the bond strength is influenced not only by the relative energy
difference (RED) but also by the solvent Molar Volume. An empirical model was developed from these experiments to provide a guideline for the solvent selection. The solvent bond strength was also affected by the bonding techniques, such as the type of sealant, shape of bonded wafers, and ratio of perimeter to bonding area. These factors were studied and their effects on optimal solvent bonding outcomes are discussed in the thesis. The strain-stress curves obtained from the tensile test results have been compared with a standard tensile stress-strain curve of Type I PMMA dog bones to
illustrate the physical change of polymer structure during the solvent bonding. The interfacial polymer chains were re-oriented into a side-to-side configuration and slightly entangled, so the bonded samples broke at a relatively low tensile strength without the polymer-chain stretching stage occurring. A promising method to improve the solvent bond strength is to cross-link the polymer chains; this is achieved by in situ polymerisation bonding. Several batches of experiments were performed to establish the relationship between its bonding strength and time, temperature, monomer concentration and initiator.