In the fabrication of electronic devices, uniform and good quality conductive printed lines are highly desirable. The goal of the present study is to develop a predictive model for conductive pattern lines produced by the microdispensing system. For this purpose, an artificial neural network (ANN) based on a feed-forward backpropagation algorithm is adopted. Input process parameters are pressure, feed rate, and standoff distance, while the output performance parameter (response) is the width of pattern lines generated through 200 µm and 500 µm nozzles diameter. The dispensing material is carbon paste having a viscosity of 30 Pa s. Best levels of process parameters are identified to achieve lower width of pattern lines based on the Taguchi signal-to-noise ratios. The identified best levels are found valid in the ranges of printing process parameters after training the neural networks. The prediction ability of ANN models is evaluated based on the leave-one-out cross-validation technique. The results showed that the proposed ANN model accomplished better results in predicting the width of pattern lines. In addition, the proposed approach is extendable to different materials with a variety of viscosities as well as to other similar printing techniques.
Read full abstract