CMP removal rate uniformity

Before considering CMP removal rate uniformity, this section describes Preston s law, which is the law that the CMP removal rate obeys and is proportional to the loading pressure (P) and to the relative speed (V) between the wafer and the table. 

For an ILD (interlayer dielectric, a part (initial step height)) is poUshed to reduce the unevenness of the height. The remaining step height shows planarity performance. Since W-Plug, STI (shallow trench isolation), and Cu interconnect are so-called recess processes, the remaining oxide erosion and dishing show planarity performance. 

The above types of planarity performance are affected by the above-mentioned WIW nonuniformity and WTW nonuniformity. 

---

Chemical mechanical polishing (CMP) removal rate uniformity and role of carrier parameters... 

Before CMP removal rate uniformity is discussed, the types of CMP, which are shown in Figure 16.1, are introduced. The carriCT type differs depending on the equipment. This means that the controllability and results of CMP removal rate uniformity also differ depending on the equipment. Several types of polisher are shown in Figure 16.1 [1]. The following is a brief explanation of the features of each polisher type, including... 

Carrier B [3] has two major mechanisms. One mechanism is the adoption of a membrane instead of the above-mentioned backing film. Through the membrane, pressure is applied to the back of the wafer. The membrane is concentrically divided into multiple areas. The pressure applied to each area is adjusted to freely control the CMP removal rate uniformity. The other mechanism is the retainer ring, which is used for controlling the rebound of the pad. These two mechanisms are currently used for almost all of the CMP for semiconductors. 

Carrier C [4] does not employ a membrane. High-pressure fluid is directly poured onto the back of the wafer to improve the CMP removal rate uniformity. 

The second main feature of Carrier B is Membrane (3). Membrane (3) is made of an elastic body-Uke rubber. As shown in Figure 16.9, Membrane (3) is divided into several areas. Through Shaft (8) and Holes (7), a different load can be applied to each divided area of Membrane (3). This means that the CMP removal rate of each divided area can be controlled. As mentioned earlier, deposition cannot be evenly formed on an in-coming wafer. By measuring the deposition distribution and controlling the load applied to each divided area during polishing, the CMP removal rate uniformity can be controlled. This is an important measure for CLC (closed loop control). 

Described below are various methods for controlling the profiles of the removal rate that are required for improving the CMP removal rate uniformity. First, profile control by the carrier for the removal rate, which is most important, is described. 

Figure 16.11 shows profile control methods for the removal rate of all parts, including the carrier. Next, Figure 16.12 shows several types of monitors used for CLC that is absolutely required for improving the CMP removal rate uniformity. 

All the examples shown in Figure 16.11 are related to the control of the CMP removal rate uniformity. The examples also lead to the improvement of the above-mentioned WIW nonuniformity and WTW nonuniformity. The following explains the mechanisms and features. 

The removal rate and its profile of the slurry change depending on the abrasive size, the abrasive concentration (wt%), or the material of the abrasive or liquid. Grasp the features of the slurry to improve the CMP removal rate uniformity. For example, changing the method for supplying the slurry may improve the removal rate uniformity. 

Figure 16.12 shows an example of a method where the polish profiles are measured and CLC is applied to improve the CMP removal rate uniformity [6]. 

Now consider an experiment for improving the CMP removal rate uniformity by the carrier. The analysis example is also shown. Conditions used for the experiment and analysis are given in Table 16.1. In the experiment, complicated methods, such as CLC where a membrane is used, are not adopted. Instead, backing film is used to simplify the experiment. 

In order to improve the CMP removal rate uniformity, it is important to use Preston s law and apply methods such as FEM to analyze the removal rate. The analysis has greatly contributed to the development of CMP. 

Figure 16.17 shows a method to improve the CMP removal rate uniformity by carrier speed. It contains three vector figures showing relative speed on a wafer. Each figure illustrates the relative speed vector when the table rotation speed (Nt) and the carrier rotation speed (A h) are changed. The figure in the center represents the relative speed vector in the case of Nt = carrier rotation speed Nu, the figure at the left represents the relative speed vector in the case of Nt > carrier rotation speed Nn, and the... 

---