Integrated micro product/process quality assurance in micro injection moulding production

The present thesis is dedicated to the advancement of the current state-of-the-art of process/product control of polymer micro components manufactured by micro injection moulding (μIM). The main focus was the achievement of in-line quality assurance and optimization in μIM by means of a combination of product micro/nano metrology and real-time process monitoring. To achieve this, the newest technologies available in the moulding scenario were employed. Multiple important aspects of μIM, as the use of unconventional materials, process simulations and indirect metrology for mould assessment were also investigated.
A thorough experimental investigation was carried out to study the behaviour of thermoplastic elastomers (TPEs) when moulding a commercial micro ring component having a mass of 2.2 mg. The quantification of part defects such as weld lines and air traps allowed drawing clear recommendations on how to set up the process window. The geometrical precision and accuracy of the produced parts were investigated using optical measurements over a broad range of process variables in order to characterize the impact of process variations on the final output. A novel shrinkage behaviour, most probably induced by residual stresses, was observed and studied. Finally, the produced parts were also functionally tested, allowing to define a clear relationship between process settings and functionality.
A novel method based on the definition of product and process fingerprints was applied to both a 3D micro component and a nano-structured one to define an in-line quality assurance strategy based on process monitoring. As for the 3D component, which was designed for medical applications and had a mass of 0.1 mg, geometrical focus variation measurements and in-line monitored pressure and velocity data were combined to determine the most suitable control strategy. It was found that the integral of the injection pressure during filling was the process indicator that was most correlated to the quality of the moulded part and to the size of its defects. Therefore, clear indications could be given on how to monitor the process in-line to obtain the most valuable information with respect to conformity to design specifications.
The nano-structured component was moulded by replicating a laser-structured surface having a texture composed by ripples with an average height of 50 nm and pitch of 900 nm. The replication behaviour was investigated over a wide range of process variables and the simultaneous use of two areal surface roughness parameters was identified as the best way to characterize the produced surfaces. High-speed infrared temperature data were obtained and considered as a tool for in-line quality monitoring.
The results showed that the in-cavity temperature during the initial heating of the cavity, which was caused by the entrance of the polymer melt, had a strong impact on the replication quality, being highly correlated to the variation of the surface roughness parameters. Clear recommendations were drawn on how to in-line monitor the nano scale replication by using the value of this temperature indicator. The use of process simulations as a tool for dimensional quality prediction and virtual process optimization was investigated in this thesis. A commercially available injection moulding simulation software was employed to build models and run simulations for two of the case studies of the project, namely the TPE micro ring and the 3D micro component. Model calibration proved a task of primary importance in order to match experimental measurements with numerical data. In regard to the micro ring simulations, the calibrated model was capable of predicting the final dimensions of the moulded parts with a 1.6 μm accuracy. The effects of the μIM parameters on the part dimensions were also well forecasted, validating the model as a tool for virtual optimization. Process simulations were applied to the 3D micro component to predict the size of the flash that affected the part quality. The results showed that the model accurately captured the variations induced by the different process variables on the defect.
The use of replica technology as a method for indirect measurements of inaccessible micro milled features, which are often present in micro moulds, was assessed. Two benchmark samples were designed and micro milled in order to quantify the degree of fidelity of a commercial two-component silicone rubber when applied to the measurement of both surface texture and geometry. The comparison between direct and indirect measurements allowed defining the applicability of the method. The surface topography measurements showed that the indirect approach led to an overestimation of the real roughness because of the stretch of the replication media generated during its manual removal from the master. On the other hand, indirect measurements of geometry revealed that the silicone shrunk linearly with respect to the original dimension. The shrinkage factor was calculated to provide a method to extrapolate the mould dimensions in the industrial situation in which only indirect data are available.