The wafer-monitoring system

Micron’s global ­production ­network of more than twelve sites worldwide not only ­enables to ­efficiently serve a wide-range of customers. But also fosters a culture of collaboration and innovation. A culture that brings ­together people with diverse ­backgrounds, expansive knowledge and practical ­industry experience. Micron has ­proactively embarked on smart manufacturing projects that use big-data solutions and machine learning. Everything to streamline operations, improve quality and ­reduce costs by more than ­10 million US dollars.

With hundreds of precise steps and sensitive processes, semiconductor manufacturing is one of the most complex industries in the world. This makes it an ideal candidate to benefit from smart manufacturing concepts. Micron recently started an Industrial Internet of Things (IIoT) initiative in its manufacturing facilities worldwide. As an example, Micron deployed an acoustic wafer-monitoring system for quality control using edge servers and dynamic signal-acquisition cards. In doing so, Micron draws on its own IIoT edge solutions for customers in the industrial environment.

IIoT edge solutions listen for defects in chemical mechanical polishing machines

The system uses high-precision acoustic sensors to monitor the condition of the system during the wafer-polishing stage of the process. The wafers are chemically and mechanically polished in special equipment. Dried polishing sludge are particles resulting from tool wear. And diamond-particles released during surface treatment can lead to defects during polishing that are not immediately detected in prior systems. This is because a wafer goes through further processing steps before it is inspected in a real-time defect analysis (RDA). Defects that occur during polishing can affect the subsequent steps in wafer production and even lead to a production stoppage.

Micron’s wafer-monitoring system

Sensors installed in the equipment for manufacturing operations, such as torque- and pressure-sensors, do not detect or monitor these defects. Micron’s acoustic wafer monitoring system makes it possible to detect out-of-band acoustic signals. Generated during polishing and to identify anomalies in real time. To do this, the detected signals are compared with an acoustic fingerprint from an optimally running polishing process.

Analysis using AI

Advanced signal-processing techniques in edge servers and dynamic signal-acquisition cards extract specific characteristics of the noise generated during polishing. The resulting data cluster is analysed using machine-learning and deep-learning algorithms. They compare the acoustic fingerprint of the current polishing process with the stored basic fingerprint. In this way, anomalies can be detected at an early-stage and potentially problematic tool conditions can be identified and fixed.

Selecting the right memory solutions ­early in the design process

In developing this IIoT solution, Micron benefited from 25 years of experience in serving industrial use cases and customers. Micron’s  IIoT-connected wafer monitoring system has reduced defect detection time by more than 90 per cent. And resulted in better quality control with 100 per cent inline monitoring.

Selecting the right components and memory solutions in the early stages of the design process is crucial for long-term success. Micron’s complete industrial portfolio can help customers meet appropriate, functional product requirements while ensuring long-term product quality and reliability.

Micron’s Industrial ­Quotient: The New IQ for Success in IIoT

 

Learn more about Micron and its smart manufacturing projects: www.micron.com.