Simulation and Prediction of Creep Rate, Activation Energy, and Rupture Time of Sn94Sb5Ag1 Lead-Free Solder Alloy Using Artificial Neural Network Modeling

In the design of modern microelectronic packaging, solder joint reliability is crucial. The assessment of advanced packaging and the adoption of novel solder materials can be sped up by predictions of solder reliability and properties. Sn 94 Sb 5 Ag 1 is one of the solders considered for replacing Pb-free alloys in electronic packaging. In the present study, experimental, simulation, and prediction creep properties of the Sn 94 Sb 5 Ag 1 solder alloy are investigated. An artificial neural network (ANN) model-based nftool was suggested to predict the creep properties. The optimal networks were accomplished using MATLAB. Simulation outcomes were very close to their experimental outcomes. Forecasting applied stresses of 6.23 MPa and 12.48 MPa, the steady-state creep rate values can be predicted. At 12.48 MPa, the activation energy is predicted to be 80.53 kJ/mole, and, at 6.23 MPa, it is predicted to be 99.22 kJ/mole. The creep rupture time can be predicted to be 92 h at 6.23 MPa at room temperature which is difficult to obtain experimentally. Calculations were made for mean, mean squared, root mean squared, and standard division errors. The success of the modeling process is supported by error values. According to this research, the ANN method is regarded as an effective method for predicting creep curve characteristics and confirming the validity of the experimental results. The microstructure features of this alloy were also studied, and it was found that it is characterized by the presence of intermetallic compounds of SbSn and Ag 3 Sn particles within β -Sn matrix.