As an emerging additive manufacturing technology, inkjet printing has been increasingly applied in microelectronics field. However, due to the impacting and rebounding behaviors of conductive ink droplets impinging onto flat substrates, it is challenging to fabricate conductive lines with desired quality, such as suitable line width and line thickness, and matching resistance when it is used for interconnecting multifarious electronic components if there is not a proper configuration of operating parameters. To address this research gap, this article aims to investigate the effect of process parameters on the quality of conductive lines, including the platform temperature, printing speed, number of layers, and delay time (droplet interarrival time), are selected to conduct a full factorial experiment. First, the approximate parameter ranges for ensuring the continuity of conductive lines are determined. Second, this study analyzes the interactive effect among process parameters on line quality. Third, an artificial neural network (ANN) is constructed to predict the quality of printed lines. Results show that the line width does not increase with an increased number of layers, while the line thickness shows an increasing trend. The low resistance and high aspect ratio of printed line are achieved by printing 5 layers with the platform temperature of 70°C, the delay time of 12.2 ms, and the printing speed of 1139.39 mm/min. Moreover, the ANN model can be used to predict line width and line thickness with excellent performance, except for the resistance due to the irregular line edge. This study provides a useful guide for the selection of appropriate printing parameters to realize a diverse range of quality properties for 3D printed conductive lines in integrated circuits.
Keywords: artificial neural network; inkjet printing; joint analysis; line quality; process parameters.
Copyright 2024, Mary Ann Liebert, Inc., publishers.