Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Data-driven surrogate model for etch rate profiles using sensor data from a reactive ion etcher

Abhijit Pranav Pamarty, Robert Neuweiler, Le Quyen Do, Keaton Johnson, James J. Sanchez, Dinesh Koli

Abstract

Reliable predictions of the etch rate profile are desirable in semiconductor manufacturing to prevent etch rate target misses and yield rate excursions. Conventional methods for analyzing etch rate require extensive metrology, which adds considerable costs to manufacturing. We demonstrate a data-driven method to predict the etch rate profiles of a capacitively-coupled plasma RIE etcher from the tool's sensor data. The model employs a hybrid autoencoder-multiquadric interpolation-based approach, with the autoencoder being used to encode the features of the wafers' etch rate profiles into a latent space representation. The tool's sensor data is then used to construct interpolation maps for the latent space variables using multiquadric radial basis functions, which are then used to generate synthetic wafer etch rate profiles using the decoder. The accuracy of the model is determined using experimental data, and the errors are analyzed in interpolation and extrapolation.

Data-driven surrogate model for etch rate profiles using sensor data from a reactive ion etcher

Abstract

Reliable predictions of the etch rate profile are desirable in semiconductor manufacturing to prevent etch rate target misses and yield rate excursions. Conventional methods for analyzing etch rate require extensive metrology, which adds considerable costs to manufacturing. We demonstrate a data-driven method to predict the etch rate profiles of a capacitively-coupled plasma RIE etcher from the tool's sensor data. The model employs a hybrid autoencoder-multiquadric interpolation-based approach, with the autoencoder being used to encode the features of the wafers' etch rate profiles into a latent space representation. The tool's sensor data is then used to construct interpolation maps for the latent space variables using multiquadric radial basis functions, which are then used to generate synthetic wafer etch rate profiles using the decoder. The accuracy of the model is determined using experimental data, and the errors are analyzed in interpolation and extrapolation.
Paper Structure (8 sections, 18 equations, 9 figures, 2 tables)

This paper contains 8 sections, 18 equations, 9 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Methodology
  3. Dataset
  4. Autoencoder
  5. Mapping
  6. Wafer statistics calculation
  7. Results
  8. Conclusions

Figures (9)

  • Figure 1: The measured etch rate against radius, for the case where [$\Delta \overset{\cdot}{m_1}$, $\Delta \overset{\cdot}{m_2}$, $\Delta P$] = [$0$, $0$, $0$].
  • Figure 2: The normalized etch rate heatmap of the interpolated wafer, for the case where [$\Delta \overset{\cdot}{m_1}$, $\Delta \overset{\cdot}{m_2}$, $\Delta P$] = [$0$, $0$, $0$].
  • Figure 3: The convolutional autoencoder's architecture with the input and output heatmaps in blue, convolutional layers in beige, fully connected layers in green and the latent space in lilac.
  • Figure 4: The mean reconstruction error $\mathcal{E}_{r,\mu}$ for the train and test wafers, evaluated at the measurement points.
  • Figure 5: (a,b,c) The measured etch rate heatmap, the reconstructed etch rate heatmap and the local reconstruction error $\mathcal{E}_{r,i}$ for the case where [$\Delta \overset{\cdot}{m_1}$, $\Delta \overset{\cdot}{m_2}$, $\Delta P$] = [$-10$, $0$, $0$] (d,e,f) the same for the case where [$\Delta \overset{\cdot}{m_1}$, $\Delta \overset{\cdot}{m_2}$, $\Delta P$] = [$12$, $0$, $0$]
  • ...and 4 more figures