RTA Temperature Compensation Using Matched Concentric Measurement Grids

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In semiconductor manufacturing, temperature deviations in rapid thermal annealing devices are challenging to control accurately due to sensor offset, contamination, and aging of heating lamps, leading to inaccuracies in temperature compensation, especially when measuring points in different machines have inconsistent distributions, affecting production efficiency.

Innovation Solution

A temperature control method is implemented using concentric circle-shaped dot matrices with unequal radial spacings in both the process device and detection device, allowing for precise temperature compensation by determining corresponding relationships between process temperature and target parameters, such as square resistivity and oxide film thickness, to maintain temperature uniformity across the semiconductor workpiece.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional measurement point distributions are used in process devices and detection devices, then device complexity is reduced, but temperature compensation accuracy deteriorates due to inconsistent measurement point distributions between different machines

Engineering Contradiction:
Improvetemperature compensation accuracyVSAvoidmeasurement point distribution complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent changes the spatial distribution parameters of measurement points from traditional uniform grids to concentric circle patterns with specifically designed radial and tangential spacings. This parameter transformation enables consistent measurement point distributions across different machines, thereby improving temperature compensation accuracy without significantly increasing device complexity

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The measurement area is segmented into concentric circular zones with different measurement point densities. The radial spacing decreases from center to edge, creating zones that capture temperature variations more effectively. This segmentation approach improves measurement precision by adapting point density to the thermal gradient distribution pattern

Inventive Principle:
Principle #1Segmentation

2Manufacturing precision

If measurement points are uniformly distributed, then device operation is simplified, but temperature uniformity measurement accuracy deteriorates due to inability to capture radial temperature variations

Engineering Contradiction:
Improvetemperature uniformityVSAvoidmeasurement point setting
Core Design Contradiction:
Manufacturing precisionVSEase of operation

Solution Approach 1:

Different regions of the measurement area are assigned different measurement point densities based on local temperature variation characteristics. The concentric circle pattern with unequal radial spacings provides higher measurement density in regions where temperature gradients are more significant, improving temperature uniformity measurement accuracy while maintaining operational simplicity through a regular geometric pattern

Inventive Principle:
Principle #3Local quality

Data Source

PatentUS20250022720A1Temperature control method for semiconductor process
Publication Date: 2025.01.16 BEIJING E TOWN SEMICON TECH CO LTD
  • US20250022720A1 patent drawing
  • US20250022720A1 patent drawing
  • US20250022720A1 patent drawing

AI summary

Provided is a temperature control method for semiconductor process, comprising: setting a first measurement point of a first point distribution in a first process device to measure a temperature in the first process device; and setting a second measurement point of a second point distribution in a second detection device to measure a target parameter; wherein the first point distribution comprises a concentric circle-shaped dot matrix with unequal radial spacings; and the second point distribution has a concentric circle-shaped dot matrix corresponding to the first point distribution.