Thin-Film Contamination Detection via Optical Emission Spectroscopy

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing thin-film deposition techniques face challenges in ensuring proper formation of thin-films, leading to contamination and defects that can result in performance issues and reduced yields in integrated circuits.

Innovation Solution

A thin-film deposition system utilizing a radiation source, emissions sensor, and machine learning-based control system to detect contamination in-situ by analyzing the energy spectrum of emissions from the thin-film, allowing for immediate cessation of the deposition process and preventing further defective films.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If thin-film deposition techniques are used to form very thin-films, then the computing power and integration density are improved, but the reliability and quality of the thin-films deteriorate due to contamination and defects

Engineering Contradiction:

Improveintegration densityVSAvoidthin-film quality

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The system performs preliminary detection during the deposition process by continuously monitoring emission spectra. The machine learning model analyzes spectral data in real-time to identify contamination or defects before they propagate, allowing for early intervention and process correction while maintaining high integration density

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system implements closed-loop feedback by continuously measuring emission spectra during deposition, comparing actual spectra against reference spectra using machine learning, and providing real-time feedback to control the deposition process. This ensures thin-film quality maintenance while achieving high integration density through dynamic process adjustment

Inventive Principle:

Principle #23Feedback

2Device complexity

If traditional detection methods are used, then the device complexity is reduced, but the measurement precision and detection capability deteriorate, failing to detect contamination in real-time

Engineering Contradiction:

Improvedetection system complexityVSAvoidcontamination detection precision

Core Design Contradiction:

Device complexityVSMeasurement precision

Solution Approach 1:

The system replaces complex mechanical detection methods with optical emission spectroscopy combined with machine learning algorithms. This substitution achieves high measurement precision for contamination detection while actually reducing overall system complexity by eliminating mechanical moving parts and complex sampling mechanisms

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system changes the detection parameter from physical inspection or electrical measurement to optical emission spectrum analysis. By monitoring spectral parameters during deposition, the system achieves high detection precision for contamination and defects with a relatively simple optical detection setup

Inventive Principle:

Principle #35Parameter changes

3Reliability

If real-time detection is implemented, then the reliability and wafer yields are improved, but the device complexity and processing time increase

Engineering Contradiction:

Improvewafer yieldVSAvoiddeposition system complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The emission spectroscopy system serves multiple functions simultaneously: it monitors deposition rate, detects contamination, identifies defects, and validates film composition. This multi-functionality improves wafer yield and reliability while avoiding the need for separate detection systems, thereby limiting the increase in overall device complexity

Inventive Principle:

Principle #6Universality (Multi-functionality)

Solution Approach 2:

The system performs continuous detection throughout the entire deposition process without interrupting the deposition flow. The machine learning model continuously analyzes emission spectra in real-time, ensuring high wafer yield through immediate defect detection while maintaining continuous productive action without idle detection cycles

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 enables real-time detection and prevention of contamination, improving wafer yields, reducing the number of scrapped wafers, and ensuring the integrity of thin-films, thereby enhancing the performance and reliability of integrated circuits.

Implementation Method 1

irradiating the thin-film with excitation light and detecting an emission spectrum from the thin film responsive to the excitation light

Methodology Applied

Scientific EffectPhotoluminescence: Photoluminescence

Implementation Method 2

detecting an emission spectrum from the thin-film responsive to the excitation light

Methodology Applied

Scientific EffectSpectrum detection:

Data Source

PatentUS11686683B2System and method for detecting contamination of thin-films

Publication Date: 2023.06.27 TAIWAN SEMICONDUCTOR MANUFACTURING CO LTD

AI summary

A thin-film deposition system deposits a thin-film on a wafer. A radiation source irradiates the wafer with excitation light. An emissions sensor detects an emission spectrum from the wafer responsive to the excitation light. A machine learning based analysis model analyzes the spectrum and detects contamination of the thin-film based on the spectrum.