联系我们
CMPEXP
An Efficient CMP Modeling and Simulation Tool

Overview

Chemical Mechanical Planarization (CMP) plays a vital role in the fabrication of integrated circuits. It combines chemical reactions and mechanical polishing techniques to achieve an even and smooth surface on silicon wafers.

With the rapid advances in IC manufacturing processes, the flatness and uniformity of the chip's topography during various manufacturing stages are critical for the yield and product performance. This impact is further magnified when the microchip's process layers are stacked up and process-sensitive patterns exist on the chip's layout. One of the significant challenges in securing high chip yield is to achieve accurate simulation, modeling, and optimization of the CMP process and its effect on the chip's topography. To overcome this challenge, Semitronix offers CMPEXP, a tool that can ensure the manufacturability of chips and yield.

Benefits

Various Types of CMP Models
▪ Supports physical models based on contact mechanics
▪ Supports emerging machine learning models of neural networks
▪ Supports model customization
Efficient Model Tuning
Introduces advanced parameter tuning algorithms to provide efficient model tuning experiences
Main Features

 

  • Extraction of geometry features from a grid-based collection
  • Simulation and analysis of CMP models
  • Detection and localization of CMP hotspots
  • Support for physical and machine learning models
  • Definition of process recipes and automatic model calibration (Etching, ECD, and CMP)
  • Adoption of an efficient distributed parallel computing architecture, which significantly improves model calibration and simulation efficiency. Usage of various visualizations such as data tables, scatter plots, line charts, bar charts, heatmaps, and 3D plots to display model calibration, simulation, and hotspot detection results
  • Generation of CMP hotspot report

 

Workflow

 

88888.png

Advantages

 

  • Optimizes the method to extract layout geometry features to improve model simulation accuracy
  • Improves model generalization to broaden the range of applicable processes
  • Optimizes the algorithm for automatic model calibration to reduce the number of calibration iterations and improve efficiency
  • Optimizes the model simulator and uses a variable step-size simulation iterator to enhance simulation efficiency
  • Uses machine learning techniques to improve simulation accuracy and calibration efficiency of models