UV Damage Prediction via Visible Spectroscopy in Semiconductor Plasma
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for predicting and controlling damage to semiconductor films due to ultraviolet rays in plasma processing are inadequate, as they either fail to accurately monitor intensity variations or rely on indirect measurements, leading to incomplete prediction and control of damage across different wavelengths.
Innovation Solution
A simulation method and apparatus that calculate particle density and emission intensity at various wavelengths, using an electron energy distribution function and reaction cross-sectional areas, to predict ultraviolet emission spectra and subsequently estimate damage amounts, allowing for precise control of process conditions to minimize damage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If total ultraviolet intensity is monitored using a sensor, then real-time monitoring capability is achieved, but measurement precision is insufficient because wavelength-specific damage effects cannot be distinguished
Solution Approach 1:
The patent segments the continuous ultraviolet spectrum into multiple discrete wavelength bands (e.g., 200-280nm, 280-315nm, 315-400nm). A sensor array with multiple detection elements is configured to measure intensity in each wavelength band separately. This segmentation enables wavelength-specific damage prediction while maintaining real-time monitoring capability, resolving the contradiction between automated monitoring and measurement precision.
Solution Approach 2:
The patent transitions from single-dimensional total intensity measurement to multi-dimensional spectral measurement by adding wavelength as a new dimension. The sensor system measures not only the total ultraviolet intensity but also the distribution across different wavelength bands. This dimensional expansion allows for both real-time monitoring and precise wavelength-specific damage assessment simultaneously.
2Measurement precision
If direct monitoring of 254 nm ultraviolet intensity is performed, then accurate measurement of cleaning wavelength is achieved, but sensor deterioration occurs due to high emission intensity
Solution Approach 1:
The patent applies local quality by assigning different functional properties to different parts of the sensor system. The sensor includes a protective filter or attenuating element specifically positioned for the 254nm wavelength region that reduces the intensity reaching the detection element. This allows accurate measurement of 254nm intensity while protecting the sensor from damage by high emission intensity at this wavelength.
Solution Approach 2:
The patent introduces an intermediary element (such as a neutral density filter, attenuating window, or protective coating) between the ultraviolet source and the sensor. This intermediary selectively attenuates the high-intensity 254nm radiation to safe levels for the sensor while preserving the measurement capability. The intermediary protects the sensor from deterioration while enabling continuous accurate monitoring of the cleaning wavelength intensity.
3Device complexity
If ultraviolet intensity is used as the only index for damage prediction, then simple prediction method is achieved, but manufacturing precision is insufficient because other damage factors are ignored
Solution Approach 1:
The patent merges multiple damage prediction indices into a unified comprehensive evaluation system. In addition to ultraviolet intensity measurement, the system incorporates ion flux measurement, plasma parameter sensing, and process condition monitoring. These multiple indices are combined through a prediction algorithm to provide comprehensive damage assessment, improving manufacturing precision while maintaining reasonable system complexity through integrated measurement.
Solution Approach 2:
The patent creates a universal damage prediction system that can evaluate multiple types of damage mechanisms (ultraviolet damage, ion damage, thermal damage) through a single integrated platform. The system uses multiple sensors and measurement techniques that can detect various damage factors simultaneously, providing a multi-functional damage assessment capability that improves prediction accuracy without requiring separate complex systems for each damage type.
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 precise prediction and control of ultraviolet-induced damage, optimizing semiconductor manufacturing by accurately accounting for emission intensity and damage across different wavelengths, thereby reducing transformation and enhancing semiconductor device quality.
Implementation Method 1
a emission spectrum in an ultraviolet wavelength region is predicted by calculating emission intensity at each wavelength in the ultraviolet wavelength region by using the electron energy distribution function and a reaction cross-sectional area relating to the emission species
Implementation Method 2
a sensor which detects an emission spectrum in a visible wavelength region inside the chamber
Implementation Method 3
In the plasma, collision of particles and interactive reactions with a chamber wall occur, and ions, radicals, and light are generated
Data Source
AI summary
The simulation method is for predicting a damage amount due to ultraviolet rays in manufacturing a semiconductor device. The method includes: calculating particle density by performing simulation based on a differential equation for the particle density; calculating emission intensity at each wavelength in a visible wavelength region based on the calculated particle density; obtaining an electron energy distribution function by comparing the calculated emission intensity at each wavelength in the visible wavelength region with an actually detected emission spectrum in the visible wavelength region with reference to information on emission species and an emission wavelength in a target manufacturing process; predicting an emission spectrum in an ultraviolet wavelength region by using the electron energy distribution function and a reaction cross-sectional area relating to the emission species; and predicting a damage amount due to the ultraviolet rays based on the predicted emission spectrum in the ultraviolet wavelength region.