|
|
|
Integration of machine learning with phase field method to model the electromigration induced Cu6Sn5 IMC growth at anode side Cu/Sn interface
|
|
Anil Kunwar, Yuri Amorim Coutinho, Johan Hektor, Haitao Ma, Nele Moelans
|
|
| |
Table 3 Values of the neural network optimized effective charge numbers at different values of current density. The predicted and experimental growth rate constants are denoted as kem,pred and kem,expt respectively.
|
|
| j (A/m2) | ZCufcc | ZSnfcc | ZCuimc | ZSnimc | ZCuliq | ZSnliq | kem,pred
(μm/min) | kem,expt
(μm/min) | | 1 × 106 | 2.1 | 2.15 | 2.1 | 2.25 | 2.02 | 2.03 | 0.11 | 0.098 | | 2 × 106 | 3.7 | 3.8 | 8.7 | 12.2 | 2.8 | 2.9 | 0.194 | 0.194 | | 3 × 106 | 4.3 | 4.9 | 15.0 | 23.0 | 3.3 | 3.4 | 0.3 | 0.294 | | 4 × 106 | 5.4 | 5.9 | 26.0 | 36.0 | 3.85 | 4.2 | 0.398 | 0.392 | | 5 × 106 | 6.6 | 6.9 | 30.0 | 39.0 | 4.3 | 4.7 | 0.47 | 0.49 | | 6 × 106 | 8.0 | 8.5 | 36.0 | 46.0 | 4.5 | 4.9 | 0.57 | 0.588 | | 7 × 106 | 9.75 | 10 | 42.0 | 52.0 | 5.1 | 5.2 | 0.677 | 0.686 | | 8 × 106 | 9.9 | 10.1 | 48.0 | 56.0 | 5.2 | 5.2 | 0.791 | 0.784 | | 9 × 106 | 10.0 | 10.25 | 52.0 | 62.0 | 5.2 | 5.3 | 0.897 | 0.882 | | 10 × 106 | 11.5 | 11.75 | 58.0 | 68.0 | 5.8 | 5.9 | 1.0 | 0.98 |
|
|
|