the Standard Fall 2008
Data-Driven Tool Architectures: The Gateway to Quality Data
by Dave Faulkner, Executive Vice President, Cimetrix, Inc. and Larry Bourget, Director of Product Development, Axcelis Technologies
Data collection and reporting requirements for semiconductor manufacturing equipment have increased dramatically over the past decade. Factories seem to have an insatiable need for higher quality and larger quantities of data. Investment and guidelines provided by industry organizations, such as the International SEMATECH Manufacturing Initiative (ISMI) and SEMI, have also focused on improvement in yield and data management in an effort to optimize factory productivity and efficiency. Additionally, a need is emerging for on-tool and off-tool data storage and retrieval. All of these initiatives require higher quality data at a greater frequency from manufacturing equipment.
Tool Data Requirements
| 200mm |
|
 |
| 300mm |
|
 |
| Today |
|
 |
Equipment control data (e.g. equipment parameters, process parameters, process performance measurements) is collected from many sources and must be distributed to a number of systems (clients). A data-driven architecture enables delivery of the highest possible quality data, in terms of freshness and frequency, to clients that perform real-time decision making. Diagnostic and processing data can be fed to factory interfaces (e.g SECS/GEM, EDA clients), on-tool data storage, on-tool data analysis, and the operator interface (GUI) to enable optimized equipment and process control performance. This data speed and quality is required for Equipment Data Acquisition (EDA), Advanced Process Control (APC), Fault Detection & Classification (FDC), Run-to-Run Control (R2R), Predictive & Preventative Maintenance (PPM), and equipment "health" monitoring. With a data-driven architecture, process and equipment engineers have access to the information they need to accurately monitor and quickly troubleshoot equipment and process performance.
Typical Data Distribution
Ideal Data Distribution
Unfortunately, many tools in the factory today were deployed with equipment control software that was not designed to meet these new demands. As a result, installed equipment may need to be redesigned internally to provide the data access that is necessary to improve factory yield. In addition, equipment suppliers face many integration challenges including compatibility between older and newer technologies, distribution of information across multiple computers and operating systems, and collecting data from a wide variety of sources. With a competitive marketplace forcing constant improvement of process and metrology technology, equipment suppliers may also find it difficult to invest the effort necessary to create a data-driven tool control architecture.
Axcelis Technologies faced this exact problem when they began searching for a commercial software package for their new multi-chamber wafer cleaning system, the Integra RS™. Customers were demanding high-speed, dependable information from tool suppliers in order to improve yield, productivity, and reliability. However, the available tool architectures did not allow for high frequency data collection at the lowest levels of the tool. Cimetrix shared the vision of Axcelis to provide improved quality and quantity data to not only the equipment supplier but the end user. In 2006, the two companies began working together on a joint development project to bring this vision to life. Axcelis and Cimetrix developed Cimetrix’s CIMControlFramework™ for the Integra RS tool to deliver optimum data collection, storage, management, and analysis.
With the data-driven architecture at the core of CIMControlFramework, and thus the Integra RS, the ability to meet the stringent tool performance and reliability requirements of the future is available today.
Related Links
- CIMControlFramework Web page
- Axcelis Technologies Web site
- Press Release for Related Presentation at AEC/APC 2008
