Numerical Simulation and Experimental Verification of Additive Distribution in Through-Silicon Via during Copper Filling Process

Abstract

Numerical model was built based on a set of PDEs and the curvature-enhanced accelerator coverage (CEAC) mechanism. Arbitrary Lagrange-Eulerian (ALE) method was used to track the moving cathode surface for describing via filling process. The filling profile changed from conformal model to bottom-up model due to the CEAC mechanism. Contour lines were colorized to reflect the local surface additive distribution. Additive surface concentration distribution in through-silicon vias during the copper deposition process was investigated according to the texture coefficient Mhkl of deposits using the micro-area X-ray technique. Surface concentration of accelerator in the via bottom was much larger than that in the top region due to the additive concentration gradient in via and the reduction of bottom surface during via filling process. The V-shaped filling profile demonstrated the rationality of the achieved accelerator surface concentration distribution. Evolution of a submonolayer quantity of a surface-confined accelerator mainly caused the super-conformal filling of TSVs. These provide a novel method to investigate the additive surface concentration distribution in the micrometer scale through-silicon via. Simulation results were compared with the experimental data; the changing tendency of surface concentration of accelerator along the via direction in simulation result is in accordance with the experimental result qualitatively (decrease from via bottom to via top region, and there is a slight increase in the top region).