Extreme Bottom-up in Through Silicon Vias by Leveler Pre-adsorption

In the TSV copper filling process, combination of several additives, such as suppressor, accelerator and levelers, are used in the electroplating bath. Besides, strong agitation is often applied to achieve good bottom-up feature. From past studies, it is considered that levelers, which show strong suppression effect in the electroplating, have an essential role in the bottom-up feature. It is assumed that the bottom-up TSV filling is achieved based on diffusion-adsorption model, in which less suppression in the via bottom is caused by diffusional limitation of the leveler and preferential deposition happens around via bottom. However, in the practical filling process, in addition to the leveler, other additives are used, and the filling mechanism is not clear yet. In this study, only leveler is added in the plating bath, and the effect of leveler on filling feature is discussed.

Experimental

The leveler used in this study is a commercial product,and the composition is not opened. Besides, the product name is neither provided. At first, in order to grasp the suppression effect of the leveler, chronoamperometry was carried out with addition of only leveler in an acid copper sulfate plating bath ( 1M CuSO₄ + 0.3 M H₂SO₄+ 1mM HCl ). It was assumed that initial adsorption of the leveler on plating surface has some impacts, and pre-immersion time before the plating start was varied. -520mV vs. MSE was applied with a rotating disk electrode at a rotational speed of 400rpm. Leveler concentrations of manufacture’s standard (x1) and 5 times concentrated one (x5) were examined. Figure 1 shows the results. Strong suppression was not obtained with standard leveler concentration, and suppression became stronger along plating time. Less current density at initial plating period was observed with longer pre-immersion. Stronger suppression was observed with 5x leveler concentration. From these results, it was supposed that suppression is limited by mass transfer of leveler. In order to clarify the diffusion-adsorption effect, TSV filling experiment with pre-immersion was performed. Figure 2 shows example results. With short pre-immersion, it is supposed that little leveler adsorbed inside the via, while strong suppression was observed on upper area of the via with 30min pre-immersion, and it was found that the leveler diffused deeper with long pre-immersion. Figure 3 shows an example result after some optimization. Extreme bottom-up was achieved with relatively high aspect ratio vias.