This paper elucidates the principles of semiconductor CMP and provides problem-solving approaches based on these principles. It also reviews various applications, including those beyond semiconductors, and discusses emerging technologies. CMP requires three key functions: material removal (Function 1), surface protection (Function 2), and contamination control (Function 3). Material removal involves both mechanical and chemical actions, which can be effectively combined not only in an additive manner but also in a multiplicative manner to achieve synergistic effects. By understanding the underlying science in relation to the material properties and geometries involved, CMP can be applied to alloys, resins, composites, and complex three-dimensional shapes.

View full abstract

This study introduces a case where a Neural Network Potential (NNP), integrating AI technology with first-principles calculations,was applied to analyze the chemical mechanical polishing (CMP) process of SiO_₂ surfaces using CeO_₂ abrasives. Integrating NNP enhanced computational speed and accuracy, allowing analyses to surpass traditional computational chemistry's capabilities.Specifically, the simulations clarified the dissolution process on the silica surface of the substrate, where ceria abrasives come into contact, proceeding from the formation of Ce-O-Si bonds to the generation of Si(OH)_₄. This approach facilitates the incorporation of environmental factors, such as substrate shape, enabling a comprehensive evaluation of the environmental dependency of the CMP process.

View full abstract

Polishing torque is important for monitoring the chemical mechanical polishing (CMP) process due to its close correlation with material removal. This report presents an innovative model-based technique for estimating the material removal rate (MRR) using only in-process data from CMP machines. The proposed method employs either the sequential least squares method or Kalman filter for real-time state estimation. Real-time estimation of MRR enables material removal control without relying on conventional endpoint detection (EDP) technologies. The accuracy of the proposed approach is validated through oxide CMP experiments, demonstrating precise estimation of the center-slowed MRR profile toward the end of the pad life.

View full abstract

The size of commercially available SiC wafers has increased to 8 inches in diameter with the spread of SiC power semiconductors.Furthermore, the scale of their mass production is also expanding rapidly, and more efficient manufacturing process technology is required. On the other hand, SiC wafers are known to be difficult to process due to their extremely hard crystalline material. This paper reviews the development of technologies to achieve both higher speed and lower cost in future SiC wafer processing.

View full abstract

Hydrophilic and soft polymer sponges are commonly used in post-chemical mechanical planarization cleaning. This paper introduces the friction mechanisms that occur during these sponge sliding. Sponges with high porosity can retain large amounts of liquid, which causes them to exhibit viscoelastic properties when deformed. The combination of the liquid and the surface also influences the frictional behavior. Additionally, this paper presents examples of adhesion force measurements between the sponge and the surface and highlights key considerations during the cleaning process.

View full abstract

Double-sided lapping and polishing are processes in which both the top and bottom surfaces of workpieces are lapped or polished simultaneously. These processes are used in the manufacturing of various products, such as semiconductor wafers, quartz oscillators, block gauges, and bearing rollers, because the workpieces with good flatness on both surfaces can be obtained by the processes. This article presents the behavior and profile change of the workpieces during the processes, which is fundamental in the processes. The rotation speed of the workpieces is demonstrated to vary with the processing time during the process, and the variation depends not only on the process condition but also on the friction between the workpieces, the tools and the carrier. Additionally, elastic deformation of the workpieces is shown to affect the profile change and prevent flattening the workpiece surfaces when the workpieces are thin and large.

View full abstract

High-pressure property of lubricant is important for elucidating the lubrication state of metalworking, bearing and gear in the elastohydrodynamic lubrication (EHL) region. The dimensionless density-pressure-temperature linear equation was derived in a previous report and it was applied to the estimation of high-pressure density. Furthermore, I accumulated the data of various lubricants for a material-specific constant included in this equation, namely dimensionless density constant ε. In this study, a multiple regression equation was derived by conducting the multiple regression analysis on the dimensionless density constant ε as target variables, while the physical properties and the chemical structure of the lubricant were explanatory variables. As a result, it was revealed that the dimensionless density constant ε_mr-eq in the dimensionless density-pressure-temperature linear equation and the high-pressure density ρ_mr-eq could be estimated from this multiple regression equation for unknown lubricants only by analyzing their physical properties and chemical structures, without experimentally measuring density using high-pressure density test apparatus.

View full abstract

[in Japanese]

View full abstract