Reticle Inspection Sparseness-Weighted Intensity Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Semiconductor reticle inspection systems face challenges in detecting defects and variations in critical dimension (CD) uniformity, especially in sparse pattern regions where feature edges contribute less to the signal, leading to difficulty in detecting CD changes and noise interference.
Innovation Solution
The method involves obtaining patch area images of reticles, determining integrated intensity values, applying a gain based on pattern sparseness metrics, and generating a difference intensity map to correlate feature characteristic variations. A calibration factor is used to convert the difference intensity map into a critical dimension (CD) map, accounting for known CD values from a design database, and adjusting intensity values in flat field areas to minimize noise from feature edges.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional reticle inspection methods are used, then inspection coverage is achieved, but measurement precision deteriorates in sparse pattern regions due to insufficient signal from feature edges
Solution Approach 1:
The patent transforms the inspection parameter from direct intensity measurement to sparseness-weighted intensity measurement. By calculating a sparseness metric for each patch and using it to weight the intensity values, the system adapts the measurement parameters to account for pattern density variations, thereby maintaining measurement precision across both dense and sparse regions
Solution Approach 2:
The patent introduces a sparseness metric as an intermediary parameter between the raw intensity data and the final CD uniformity measurement. This intermediary metric quantifies the pattern density in each patch and serves as a weighting factor, mediating the relationship between intensity signals and measurement results to compensate for insufficient edge signals in sparse regions
2Measurement precision
If intensity values from all patches are used directly, then processing speed is maintained, but measurement precision deteriorates due to noise interference from flat field areas
Solution Approach 1:
The patent applies local quality by treating different patches differently based on their sparseness characteristics. Patches with low sparseness (likely containing flat field areas) receive different weighting or filtering treatment compared to patches with high sparseness, thereby locally optimizing the signal-to-noise ratio without requiring complete reprocessing of all data
3Manufacturing precision
If uniform inspection parameters are applied across all reticle regions, then device complexity is minimized, but manufacturing precision deteriorates due to inability to account for pattern density variations
Solution Approach 1:
The patent implements dynamic inspection parameters by calculating a sparseness metric for each patch and using it to dynamically adjust the weighting of intensity values. This dynamic approach allows the inspection system to adapt to local pattern density variations across the reticle, improving manufacturing precision without requiring complex manual configuration of different parameters for different regions
Data Source
AI summary
Disclosed are methods and apparatus for inspecting a photolithographic reticle. An inspection tool is used to obtain a plurality of patch area images of each patch area of each die of a set of identical dies on a reticle. An integrated intensity value for each patch area image is determined. A gain is applied to the integrated intensity value for each patch area image based on a pattern sparseness metric of such patch area image and its relative value to other patch area images' pattern sparseness metric. A difference between the integrated intensity value of each patch of pairs of the dies, which each pair includes a test die and a reference die, is determined to form a difference intensity map of the reticle. The difference intensity map correlates with a feature characteristic variation that depends on feature edges of the reticle.