Fig. 1. Schematic sketch outlining the major variables (P, Vscan and solder composition) in an experimental setup for laser soldering. The point $~\chi$ is located at the bottom corner corresponding to the start point of the laser beam.

Fig. 2. SEM images of experimental solders of different initial composition and laser processed at different combinations of input power (P) and scan speed (Vscan): (a) scalloped and (b) prismatic morphology of Cu6Sn5 IMC compounds as obtained at the interface, respectively.

Fig. 3. (a) Effects of input power and scan speed on IMC morphology transition from scalloped to prismatic or vice-versa for a Sn-0.7Cu solder and (b) effects of solder types at given laser input power on IMC morphologies.

Fig. 4. Enthalpies of Sn-xAg-yCu solder as a function of temperature.

Fig. 5. Temperature profile corresponding to t = 1.0 s, at the different positions of a pure Sn solder irradiated by different combinations of input power and scan speed.

Fig. 6. Numerically computed values of Tif at point $\chi$ in bottom interface for pure Sn modeled with different combinations of P and Vscan. The T-t curves shown in the figure can be utilized for solder joint design for dynamic heat source conditions.

Fig. 7. Temperature profile corresponding to t = 2.0 s, at the different positions of a Sn-xAg-yCu solders irradiated with input power P = 30 W and scan speed = 60 mm/min: (a) the color bar scaled in the range of 500-600 K for getting meticulous temperature profile within the first 1 mm length from the left end; (b) the complete solder geometries and the color bar is set in the range 298-800 K.

Fig. 8. Simulated values of Tif at point $\chi$ in bottom interface for pure Sn-xAg-yCu solder alloys corresponding to P = 30 W and Vscan = 60 mm/min.

Fig. 9. Result from multi-phase field simulation for grain growth study of two identical IMC grains performed at constant interface temperatures Tif of 453.15 K and 523.15 K.

Fig. 10. Simulated results for sum of squared order parameters (($\underset{i}{\mathop \sum }\,{{({{\eta }_{i}})}^{2}}$) at constant interface temperatures Tif of (a) 453.15 K and (b) 523.15 K.

Fig. 11. A feed forward neural network for regression analysis between input features (solder composition and laser processing parameters) and output feature (Jackson parameter).

Fig. 12. Mean square error (MSE) values from the neural network analysis of training and validation data for a total of 1000 epochs using 9 neurons in the first hidden layer and 5 neurons in the second hidden layer.

Fig. 13. Jackson parameters estimated from machine learning for different solder alloys at (a) P = 30 W, (b) P = 40 W and (c) P = 50 W. The dotted horizontal lines refer to the morphology transition regimes.

Fig. 14. Schematic sketch showing the morphology mapping of IMC in accordance to Jackson parameter. The IMC corresponding to the curve lying in the yellow region has prismatic morphology, whereas Cu6Sn5 within the curve for light pink region has scalloped morphology.

Fig. 15. Material design chart established from this study that shows the optimized values of Vscan,cr as markers of the Cu6Sn5 IMC morphology transition. For a given laser input power and solder alloy type mentioned in the chart, the IMC has scalloped morphology at Vscan > Vscan,cr, whereas at Vscan < Vscan,cr, it has prismatic morphology.