Peak Tracking Algorithm for Flat Object Boundary Layer Inspection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current methods for inspecting flat objects, such as wafers in the semiconductor industry, face challenges in accurately and quickly tracking and assigning peak signals in one-dimensional signal spaces, particularly due to interference and noise, which complicates the measurement of layer thicknesses and profiles.

Innovation Solution

An optical measuring system using spectral interference generates peak signals that are tracked and classified using a real-time capable peak tracking algorithm, employing a Kalman filter to estimate peak signal changes and assign them to specific types, enabling robust movement tracking even in complex situations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If conventional object tracking methods using Kalman filters and multiple sensors are applied, then tracking capability is improved, but device complexity and cost increase significantly

Engineering Contradiction:
Improvepeak tracking reliabilityVSAvoidsystem complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent extracts and utilizes only the essential motion information (peak position changes in one-dimensional signal space) needed for tracking, discarding unnecessary complex image data and multiple sensor requirements. This selective extraction maintains tracking reliability while significantly reducing system complexity.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent introduces a simplified peak tracking algorithm that acts as an intermediary between raw peak signals and classification results. This algorithm uses basic motion models to predict peak positions and assign peak types without requiring complex Kalman filters or multiple sensors, thus reducing system complexity while maintaining reliability.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Measurement precision

If peak tracking is performed without motion prediction, then algorithm simplicity is maintained, but measurement precision and reliability deteriorate due to noise and interference

Engineering Contradiction:
Improvepeak position measurement precisionVSAvoidalgorithm complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies preliminary motion prediction using simple motion models before peak classification. By predicting where peaks should appear based on their previous positions and estimated velocities, the system can accurately identify and classify peaks even in noisy conditions, improving measurement precision with minimal algorithmic complexity.

Inventive Principle:
Principle #10Preliminary action

3Productivity

If rapid inspection of multiple locations is performed, then productivity increases, but measurement precision may deteriorate due to reduced measurement time per location

Engineering Contradiction:
Improveinspection throughputVSAvoidthickness measurement precision
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent implements continuous peak tracking across multiple measurement locations using motion prediction. As the measurement head moves continuously across the wafer, the algorithm continuously predicts peak positions and updates classifications without interruption. This continuous tracking maintains high productivity while preserving measurement precision through predictive correction of peak positions.

Inventive Principle:
Principle #20Continuity of useful action

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

This approach allows for precise and rapid peak tracking and classification, improving measurement accuracy and efficiency by reducing noise sensitivity and enabling continuous updates of peak positions, thus facilitating high-yield chip production and other industrial applications.

Implementation Method 1

The measuring system operates according to the method of spectral interference. It provides peak signals whose position in a one-dimensional signal space corresponds to the distance between the sensor and a boundary layer of the object or the distance between boundary layers of the object.

Methodology Applied
Scientific EffectSpectral interference: Interference

Data Source

PatentEP3795942B1Device and method for inspecting flat objects and for detecting boundary layers of said objects
Publication Date: 2022.04.27 SENTRONICS METROLOGY GMBH
  • EP3795942B1 patent drawingFigure 1a~1b
  • EP3795942B1 patent drawingFigure 2a
  • EP3795942B1 patent drawingFigure 2b~2c

AI summary

A device for inspecting flat objects and detecting their boundary layers comprises an optical measuring system for measuring distances with a sensor. The measuring system provides peak signals whose position in a one-dimensional signal space corresponds to the distance between the sensor and a boundary layer or between boundary layers. The device also includes an evaluation unit configured to evaluate the peak signals in a tracking process and assign a predefined peak type to each peak signal. The evaluation unit assigns peak signals detected during a measurement to different tracks. Each track is updated with the assigned new peak signal. All peak signals within a track are assigned the same peak type. The assignment of a peak signal to a track is based on the change in the peak signal's position within the one-dimensional signal space.The change in a peak signal is estimated for the next measurement using an estimator.