Dynamic Reject Thresholds for Interferometric Thickness Measurement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing optical interferometry methods for measuring the thickness of semiconductor slices are inaccurate and require manual adjustment of thresholds, making them unsuitable for production line use due to variability and the need for operator intervention, especially when measuring thin layers or layers with similar foreign values.
Innovation Solution
A method and apparatus that use a processing unit to analyze multiple readings, determine variability intervals, and adjust reject thresholds dynamically based on machining progress, allowing for automatic determination of the thickness by weighting quality factors and integrating frequencies across thickness classes to identify the preponderant group of values, thereby improving accuracy and reducing operator intervention.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If optical interferometry is used to measure thickness, then measurement precision is improved, but device complexity increases due to the need for spectrometers and optical fibres
Solution Approach 1:
The patent applies dynamics by making the reject thresholds dynamic rather than fixed. The processing unit automatically adjusts the thresholds based on the distribution of measured thickness values and their frequencies, allowing the measurement system to adapt to varying conditions without manual intervention or increased hardware complexity
Solution Approach 2:
The system performs self-service through automatic threshold adjustment. The processing unit analyzes the frequency distribution of thickness measurements and autonomously determines optimal reject thresholds, eliminating the need for operator intervention and simplifying operation while maintaining high measurement precision
2Ease of operation
If fixed reject thresholds are used in thickness measurement, then ease of operation is improved, but measurement precision deteriorates due to inability to adapt to varying thickness distributions
Solution Approach 1:
The patent transforms fixed thresholds into dynamic thresholds that automatically adjust based on the measured data distribution. The processing unit calculates frequency distributions and adapts thresholds in real-time, maintaining both ease of operation (no manual adjustment needed) and high measurement precision (adaptive to varying conditions)
Solution Approach 2:
The system implements feedback by continuously analyzing the frequency distribution of thickness measurements and using this information to adjust reject thresholds. This closed-loop approach ensures thresholds remain optimal for the current measurement context without requiring operator intervention
3Adaptability or versatility
If manual threshold adjustment is required, then adaptability is improved, but productivity deteriorates due to operator intervention time
Solution Approach 1:
The system achieves self-service through automatic threshold determination. The processing unit independently analyzes measurement distributions and adjusts thresholds without operator intervention, maintaining full adaptability to different measurement scenarios while maximizing productivity by eliminating manual adjustment time
Solution Approach 2:
The patent applies preliminary action by pre-programming the automatic threshold adjustment algorithm in the processing unit. The system is prepared in advance to autonomously adapt thresholds based on any measurement distribution, eliminating the need for manual setup and enabling immediate high-speed measurements
4Device complexity
If simple frequency counting is used, then device complexity is reduced, but measurement precision deteriorates due to inability to handle parasitic values
Solution Approach 1:
The patent enhances the simple frequency counting approach by making it dynamic and adaptive. The processing unit analyzes the distribution of frequency values and automatically adjusts thresholds based on the observed patterns, maintaining algorithmic simplicity while achieving high precision by adapting to the specific measurement context and filtering parasitic values
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 rapid, accurate, and autonomous thickness measurement with minimal error, suitable for serial production by dynamically adjusting thresholds and focusing on high-quality readings, thus overcoming the limitations of existing methods.
Implementation Method 1
collecting the radiations reflected by the slice of semiconductor material
Implementation Method 2
spectral analysis of the combinations resulting from the interference of the radiations that are reflected
Implementation Method 3
provided with lenses for focusing the radiations emitted by the radiation source on the slice of semiconductor material to be measured and for collecting the radiations reflected by the slice of semiconductor material
Implementation Method 4
A spectral analysis of the combinations resulting from the interference of the radiations that are reflected by the slice of semiconductor material to be measured is carried out by means of the spectrometer
Data Source
AI summary
Methods and apparatuses are used for optically measuring by interferometry the thickness (T) of an object (2) such as a slice of semiconductor material. Readings of the object thickness by optical interferometry are carried out, rough thickness values (RTW) are obtained and frequencies, indicating how often the rough thickness values occur, are evaluated. A limited set of adjacent rough thickness values whose frequency integration or summation represents an absolute maximum is identified, and the actual value of the thickness of the object is determined as a function of the rough thickness values belonging to said limited set of values. The rough thickness values can be divided up into classes (C) with corresponding frequencies (F), and in this case, a preponderant group (Gmax) of thickness classes is identified as the above-mentioned limited set of adjacent rough thickness. A lower reject threshold (Rmin) and a higher reject threshold (Rmax) that define a searching interval including the actual value of the object thickness are also determined, and all the rough thickness values that are outside the searching interval are eliminated from further processing. When measuring the object during a surface machining, the reject thresholds are progressively and automatically updated as a function of a gradual thickness reduction that the object undergoes.