In-Line Part Average Testing for Semiconductor Reliability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current semiconductor device fabrication processes face challenges in detecting reliability defects, particularly latent defects that can lead to early-life failures, as existing inspection methods like electrical testing are insufficient for stringent reliability standards and may result in overkill, where good devices are improperly identified as faulty, impacting throughput and cost.
Innovation Solution
The implementation of advanced In-Line Part Average Testing (I-PAT) systems that use in-line sample analysis tools to identify defects, assign weights based on predicted impact on reliability, and generate defectivity scores to isolate outlier dies, combining defect data with electrical test failures to reduce overkill and improve accuracy in identifying potential failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If electrical testing is used to identify defective devices, then device functionality can be verified, but latent reliability defects may not be detected and good devices may be improperly identified as faulty (overkill)
Solution Approach 1:
The inspection process is segmented into multiple independent analysis tools including optical inspection, electrical testing, and analytical modeling. Each tool analyzes specific aspects of device quality separately, then results are integrated to form a comprehensive assessment. This segmentation allows latent reliability defects to be detected through multiple pathways while maintaining accurate identification of good devices.
Solution Approach 2:
A controller acts as an intermediary that receives data from multiple in-line sample analysis tools, processes the information through analytical models, and generates integrated inspection results. The controller mediates between different testing methodologies, combining their strengths while compensating for individual limitations, thereby improving both reliability defect detection and measurement precision.
2Reliability
If stringent reliability standards are implemented to detect all potential defects, then failure rates can be reduced to parts per billion levels, but production throughput decreases and cost increases
Solution Approach 1:
In-line sample analysis tools perform preliminary defect detection during the fabrication process itself, before devices are completed. By identifying and flagging potential reliability issues early, the system prevents defective devices from proceeding through subsequent production steps, thereby maintaining high throughput while achieving stringent reliability standards through proactive defect management.
Solution Approach 2:
The system dynamically adjusts inspection parameters and weighting factors based on process conditions, device type, and defect patterns. By changing parameters such as inspection sensitivity, sample size, and defect weighting in the analytical model, the system optimizes the balance between detecting latent defects and maintaining production throughput, avoiding unnecessary inspection of low-risk devices.
3Reliability
If comprehensive inspection methods are used to detect latent defects, then reliability control improves, but inspection complexity and cost increase
Solution Approach 1:
The controller and analytical model serve universal functions by processing data from multiple different analysis tools through a unified framework. The same controller hardware and software platform handles optical inspection data, electrical test data, and process parameters, applying consistent analytical logic across diverse input types. This multi-functionality reduces overall system complexity compared to having separate dedicated systems for each inspection method.
Data Source
AI summary
An inspection system may include a controller communicatively coupled to one or more in-line sample analysis tools including, but not limited to, an inspection tool or a metrology tool. The controller may identify defects in a population of dies based on data received from at least one of the one or more in-line sample analysis tools, assign weights to the identified defects indicative of predicted impact of the identified defects on reliability of the dies using a weighted defectivity model, generate defectivity scores for the dies in the population by aggregating the weighted defects in the respective dies in the population, and determine a set of outlier dies based on the defectivity scores for the dies in the population, wherein at least some of the set of outlier dies are isolated from the population.