Investigation of 2-Way Injection Method on Cu CMP Process

Chemical mechanical polishing (CMP) is one of the process for planarization in semiconductor manufacturing. With the decrease of device size, copper (Cu) has been accepted as the material of interconnect for good performance. After electroplating process of Cu, overburden Cu surface is polished for planarization using CMP process. Generally, oxidants, dispersants, abrasive particles, chelating agents, and corrosion inhibitors are used in the Cu CMP slurry, and planarization of Cu is achieved by reaction of slurry components. Hydrogen peroxide is one of the oxidants which is widely used in Cu CMP slurry and react with Cu surface to form a passivation layer. In general, it is known that hydrogen peroxide is easily decomposed in the alkaline state. Therefore, in the alkaline slurry, decomposition of hydrogen peroxide can occur, which leads to deterioration of CMP performance. In this study, a new slurry injection method is proposed as an alternative to prevent the decomposition of hydrogen peroxide that may occur when hydrogen peroxide and slurry are mixed. The concentration of hydrogen peroxide decreased when hydrogen peroxide and slurry are mixed, and accordingly, the removal rate decreased (Fig. 1(a)). It indicates that hydrogen peroxide is decomposed in alkaline state, and this affects the CMP performance. In order to prevent the decomposition of hydrogen peroxide, hydrogen peroxide and slurry were separately injected (2-way injection) during the CMP process. Fig. 1(b) shows the comparison of the removal rate between the general method and the 2-way injection method in the slurry supply system. As shown in the graph, when using the 2-way injection method, the removal rate is almost similar to that of the general injection method. Therefore, the 2-way injection method can replace the general injection method. Figure 1. (a) the removal rate and concentration of hydrogen peroxide with slurry aging time, (b) the removal rate of general and 2-way injection method Figure 1