Prediction of electrodeposited copper ions in high speed jet electrodeposition for MEMS fabrication using ANN

Abstract

High speed jet electrodeposition offers fabrication of Micro Electro-Mechanical System (MEMS) features. In this technique, high precision and accuracy can be acquired using ultrasonic and controlled environment. This research is an attempt to predict the deposition of copper ion on metallic aluminium substrate. For this, a data set has been generated by performing the experiment multiple times at different variable inputs condition. Consequently, Artificial Neural Network (ANN) has been deployed using MATLAB for prediction. ANN is used to train, test and validate the collected experimental data for development of a function which estimate the electrode deposition rate. The variables include the gap between copper electrode and substrate, voltage applied between electrodes and concentration of CuSO4·5H2O However, no prior data is available to compare the performance because no attempt of ANN implementation has been done specifically in High Speed Jet Electro Deposition (HSJED). The coefficient of correlation R for training, testing and validation proved to be satisfied and came out as 0.94162, 0.98755 and 0.98242 respectively. The performance parameter was Mean Square Error (MSE). The prediction model would help the researcher to know the copper deposition rate and save time and efforts required.