Unlock instant, AI-driven research and patent intelligence for your innovation.

A semiconductor process data correction method based on correlation entropy and deep neural network

A deep neural network and process data technology, applied in the field of semiconductor process data correction based on correlation entropy and deep neural network, can solve problems such as inappropriate correction results and failure to consider the relationship, so as to reduce the pressure of parameter adjustment and promote more accurate expressive effect

Active Publication Date: 2021-06-29
ZHEJIANG UNIV
View PDF7 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, this method only considers the statistical knowledge of a single variable, and does not consider the relationship between other variables in the whole process, which may cause inappropriate correction results

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A semiconductor process data correction method based on correlation entropy and deep neural network
  • A semiconductor process data correction method based on correlation entropy and deep neural network
  • A semiconductor process data correction method based on correlation entropy and deep neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0063] The present invention will be further described in detail below with reference to the accompanying drawings and examples, and the following examples are intended to facilitate the understanding of the present invention, but will not afford any limits.

[0064] The following is an example with the previvalence of deposition process results in a factory in China, and the wafer height value via a multi-stage chemical process is virtualized.

[0065] During the production process, the chemical vapor deposition process is similar to that of the surface of the surface to coat the solid thin film coating on the surface of the semiconductor industry. This process is complex because it relates to many chemical reactions, and the reactors in the multi-reactor system are independently controlled to deposit the membrane in the process chamber under various conditions. The chemical vapor deposition equipment is equipped with a significant number of sensors. These measurements include ra...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a semiconductor process data correction method based on correlation entropy and deep neural network, comprising: (1) collecting the output signal of the process variable sensor corresponding to the variable to be corrected; (2) inputting each variable into the established The deep neural network model extracts the correlation information of variables layer by layer, collects the feature output of the last layer of the model to the mapping function, and compares it with the input variables to establish a regression model; (3) saves the parameter weight of the current model and calculates the final objective function value, if the stop condition is not met, update the parameter weights and repeat step (2); (4) change the number of network layers and feature iteration layers, and repeat steps (2) and (3) until the maximum number of layers is reached; (5 ) Select the number of network layers and feature iteration layers that get the best correction results; save the parameter values ​​of each layer, calculate the data to be corrected and obtain the corrected value. By using the invention, the data correction result with lower error is obtained.

Description

Technical field [0001] The present invention belongs to the field of process monitoring in industrial systems, in particular, to a semiconductor process data correction method based on associated entropy and deep neural networks. Background technique [0002] In recent years, data driving methods such as process monitoring, soft measurements have been established as a powerful process control tool in the semiconductor industry. Therefore, the reliability and accuracy of the measurement process data is critical to the efficient, profitable and safe operation of the plant in the chemical industry. [0003] However, due to factors such as the variability of the process and the limitations of the measurement technique, the online measurement data is usually subject to the interference of random errors and significant errors. By improving the original data set, process performance and maintenance efficiency can be significantly improved. Therefore, data correction that can be reduced ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G01D18/00G06N3/04G06N3/08
CPCG01D18/00G06N3/08G06N3/048G06N3/045
Inventor 谢磊吴小菲徐浩杰苏宏业
Owner ZHEJIANG UNIV