Real-time Hotspot Analyzer Using Process Variability Bands
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current lithographic simulations for integrated circuit manufacturing are computationally expensive and time-consuming, leading to overly conservative rework decisions that increase costs, as they struggle to analyze real-time variations in exposure dose, focus, and alignment effectively.
Innovation Solution
A bifurcated wafer analysis process involving offline hotspot detection and real-time wafer assessment, using process variability bands generated from mask layout data and manufacturing parameters to identify and characterize hotspots, allowing for dynamic disposition of fabricated integrated circuits based on measured parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional lithographic simulations are used to analyze real-time variations in exposure dose, focus, and alignment, then manufacturing precision can be improved, but computation time and cost increase significantly
Solution Approach 1:
The system pre-calculates and stores process variability bands for various manufacturing parameters (exposure dose, focus, alignment) before actual wafer fabrication. These pre-computed bands are stored in a database and can be quickly queried during real-time analysis, avoiding the need to perform computationally expensive simulations at each measurement point.
Solution Approach 2:
The analysis process is divided into two distinct phases: an offline phase where process variability bands are pre-computed and stored, and an online phase where actual wafer measurements are compared against these pre-stored bands. This segmentation allows the computationally intensive work to be done in advance, enabling fast real-time decision-making.
2Measurement precision
If traditional lithographic simulations are performed for every wafer measurement, then hotspot detection accuracy improves, but manufacturing productivity decreases
Solution Approach 1:
Process variability bands covering a range of possible manufacturing conditions are pre-calculated and stored in a database before actual wafer fabrication begins. During real-time analysis, the system simply queries this pre-computed database with actual measurement values to determine if hotspots are present, avoiding repeated expensive simulations.
Solution Approach 2:
Instead of performing original computationally expensive simulations for each wafer, the system creates and uses copies of pre-computed process variability band data stored in a database. These copied data structures can be quickly accessed and compared against actual measurements without repeating the simulation work.
3Manufacturing precision
If conservative rework decisions are made based on limited analysis capability, then pattern fidelity is maintained, but manufacturing cost increases
Solution Approach 1:
The system implements a feedback mechanism where actual wafer measurements (exposure dose, focus, alignment) are compared against pre-computed process variability bands to dynamically determine whether rework is truly necessary. This feedback loop enables more accurate decision-making, avoiding both unnecessary rework and missed defects.
Solution Approach 2:
The system transitions from static, conservative rework thresholds to dynamic, measurement-based decision-making. By comparing actual measured parameters against pre-computed variability bands that account for process variations, the system can dynamically adjust rework decisions based on the specific conditions of each wafer.
Data Source
AI summary
This application discloses a hotspot identification system to generate process variability bands for structures of an integrated circuit capable of being fabricated utilizing at least one lithographic mask based, at least in part, on a mask layout data describing the lithographic mask and a distribution of manufacturing parameters during fabrication. The hotspot identification system can utilize the process variability bands to identify a subset of the structures that correspond to hotspots in the integrated circuit and identify corresponding values for the manufacturing parameters associated with the identified hotspots. A wafer testing system can implement a real-time wafer assessment process by comparing measured manufacturing parameters associated with a fabricated integrated circuit to the values for the manufacturing parameters associated with the identified hotspots, and dynamically identifying a disposition for the fabricated integrated circuit corresponding to one or more structures associated with the identified hotspot based on the comparison.