Defect review, analysis, and classification are vital in semiconductor fabrication, providing the means to monitor and control the quality of individual steps in the manufacturing sequence. Shrinking feature dimensions and the transition to 3D devices make complete and accurate imaging more challenging. High-resolution images of defects to determine their character, analysis of their composition, and examination of root causes are essential in enabling chipmakers to establish accurate statistical process control (SPC), to ramp up production runs rapidly, and achieve consistently high production yields.
Designed for 1x nm technologies and beyond, the Applied SEMVision G6, enhanced by Purity Automatic Defect Classification (ADC), is the most advanced of the SEMVision family of defect review and analysis systems that have been leaders in accuracy, comprehensiveness, and speed of these vital aspects of semiconductor fabrication for the past 15 years. It offers customers the highest quality imaging in the industry and best-in-class defect classification capabilities.
The SEMVision G6 system’s resolution is a 30% improvement over the previous generation, making it the highest available in the industry. Unique beam tilt makes it possible to view defects on the sidewalls and at the base of emerging 3D device structures, such as FinFETs. High dynamic range detection and the collection of back-scattered electrons combined with energy filtering enable high aspect ratio imaging of trenches and holes. High-energy imaging enables “see-through” detection of defects in underlying layers. The system’s advanced detection assembly and sophisticated processing make possible high-quality topographical images of tiny and shallow defects. In addition to superior resolution and image quality, SEMVision G6 delivers the quickest defect review and analysis available through fast navigation, a new stage with nanometric precision, a new automatic defect review algorithm that boosts capture rate, and high-throughput automated defect redetection that filters out “nuisances.”
Purity ADC enables chipmakers for the first time to rely on the automatic review system to correctly identify classes of yield-limiting defects in a production environment. Previously, classification was an entirely manual process. ADC offers significantly superior classification quality and consistency through advanced machine learning algorithms that dynamically “learn” the numerous defect classes in today’s complex processes and a statistical classification engine that achieves benchmark classification accuracy and repeatability. Automatic classification also permits more comprehensive sampling, faster excursion identification, and shorter time to production “go-no go” decisions. Combined, these result in shorter cycle time and higher productivity—especially critical for foundries, where good yields must be achieved rapidly for thousands of new chips each year.