Effect of Degraded Additives on Under-Fill TSV

Abstract

In the past few years the 3D IC integration (3DI) stacking has emerged as a potentially novel approach for extending computing performance without the high cost of CMOS scaling. The 3DI stacking technology is thought to be the essential technology of the next generation high-end semiconductors which will provides improved performance, and reduced form factor for applications such as logic-memory integration, image sensors, MEMS, and LED. The through silicon via (TSV) is one of the main components for 3DI stacking. Copper plating has been used for TSV metallization due to its low resistivity, low cost, excellent plating efficiency, easy to integrate, excellent reliability and a unique bottom up fill capability in presence of additives in the plating bath. One of the major challenges in the TSV metallization by plating is embedded voids and underfill TSV. Among the many parameters in TSV plating; bath composition especially suppressor to accelerator ratio (S/A) significantly affects TSV under-fill. It has been seen that the additives also degrades over the shelf life aging depends on their storage condition. Since the degradation of both suppressor and accelerator is not happening in the same rate, the S/A ratio vary and thus found to be caused TSV under-fill. Fig. 1(a) shows a cross-section SEM image of a TSV plated with BKM bath additives for 1600s. In contrast, Fig. 1(b) and 1(c) shows cross-section SEM images those were plated with same amount of BKM accelerator but with 10% and 25% less suppressor than the BKM respectively. It can be seen that the growth of copper inside the TSV increased~ 33% (with 25% lower suppressor) than the POR without increase of Cu film thickness on side wall, Chamfers and field area. Similarly, for a given suppressor, the copper fill rate can also varied by varying the accelerator concentration. In this paper, we will describe how additives affect the TSV under-fill and correlate the TSV under-fill with degradation of additives during their shelf life aging. Fig. 1(a) Cross-section SEM image of a TSV deposited with BKM plating additives of accelerator and suppressor and plating process for 1600s. Fig.1 (b) and (c) shows cross –section SEM images of a TSV those were plated with same amount of BKM accelerator but 10 and 25% less suppressor respectively than the BKM condition but exactly same plating process and time as Fig 1a Figure 1