E-Manufacturing Opportunities in Semiconductor Processing
Dave Bloss and Devadas Pillai Intel Corp., Chandler, Ariz. -- Semiconductor International, 7/1/2001
| At a Glance |  | ||
|
| ||
| |||
As these new information technologies and capabilities become available, the challenge for IC manufacturers is how to seamlessly integrate their legacy manufacturing systems with new capabilities to keep up with an increasing rate of change. Effective realization of efficiency improvements needed to survive in a competitive environment will require use of many of these emerging IT capabilities in manufacturing.
Words commonly used to describe this changing paradigm are e-manufacturing or e-operations. The "e" implies the linking of customers and suppliers in the supply chain through a secure Internet-enabled network that, in turn, enables and provides operational flexibility to each element in the supply chain. The desired end result is a reduction of overall manufacturing costs coupled with significant improvements in responsiveness to customer needs amid changing priorities and operational conditions.
In response to these business challenges, IC makers are developing and enhancing key capabilities such as collecting and storing information in a format that can be accessible at the point of need or use, while connecting the elements of the supply chain via secure supplier linkages so that information can be shared, monitored and analyzed. These capabilities allow information to be gathered, shared and manipulated at factory and enterprise levels, while also allowing relevant data to be shared across the supply chain.
A key challenge IC manufacturers face is how to get the relevant information and data from those legacy factory systems that were not architected originally to serve this purpose, but that continue to be used cost-effectively today.
The approach going forward
Up to this point these significant changes in factory systems and operational paradigms have been focused on solving specific issues, and have resulted largely in expensive point solutions. To achieve widespread benefits across the supply chain, IC makers must now define a common roadmap of capabilities, and work together with suppliers to define solutions for IC industry customers.
Because these capabilities are not semiconductor industry-specific, the solutions cannot be semiconductor-specific either. Mainstream computing technologies developed outside the semiconductor industry must be utilized to significantly reduce risk, development cost and implementation costs.
As a first step, IC makers are analyzing these basic building blocks in the area of production equipment productivity enhancement programs. Prioritization in this area is only natural because of its high upfront capital cost and subsequent depreciation. There is a tremendous urgency to achieve benefits in the areas of overall equipment efficiency, increased availability and uptime, process improvement and overall manufacturing cost reduction. The definition and standardization of these building blocks will enable cost-effective implementations to occur at a rapid pace throughout the industry.
Even with industry collaboration on capabilities, IC makers will still face significant challenges and hurdles in the realization of specific implementations. Barriers such as accurate data collection from equipment, cost-effective integration of third-party applications, and enterprise data security concerns must be removed quickly to successfully introduce new e-manufacturing capabilities into IC fabs.
Another concern is whether standards-setting bodies can keep up with cycle-time reduction expectations that users of this capability will demand. At the current time, the process is rigid and not built for speed. Using Internet capabilities for streamlining standards ballot submission, voting and approval loops will be necessary to attain a 10× reduction in cycle time for this process.
A vision for e-manufacturing
From a high level, e-manufacturing can be characterized as providing the right data to the right people at the right time, coupled with decision support systems that act upon this information with or without people intervention. The overall goal of this effort is to synchronize the planning, procurement, ramping and operations of a factory and its support functions at significantly faster (i.e., Internet) speeds and greatly reduced costs. This synchronization is enabled by efficient information flows between the factory, its suppliers, its customers and its internal support groups.
This vision can be achieved in a cost-effective manner, through the use of emerging information technologies applied to specific capabilities built from well-defined and standardized industrywide building blocks. They must be integrated and presented to all relevant users in such a way that decisions can be made proactively and not reactively.
IC maker capabilities and benefits
One possible way of organizing e-manufacturing capabilities from an IC maker's perspective is shown in Figure 1. It consists of 1) information and data sources; 2) consistent data repositories; 3) integrated applications; and 4) high-value solutions aimed at monitoring manufacturing asset performance, analysis and decision support systems, and improving linkages with suppliers.
The data repositories and information sources form the overall enterprise infrastructure for e-manufacturing. The enterprise infrastructure architecture is fundamental in laying the basic foundation for other higher-level layered capabilities related to e-manufacturing. Data comes in many forms, and the infrastructure must have the capabilities for hosting and accessing such data. Reference data, current data, historical data, external data, and raw detailed data are all required for decision making.
The factory's real-time and on-line execution systems such as manufacturing execution systems (MES), automated material handling systems (AMHS), equipment control and factory scheduling are typical sources of data, which must be coupled with information that is originating from other relevant data sources (people, equipment or other systems) as well.
To eliminate data redundancy and inaccuracies, every data element in the infrastructure must have one data source. The data must be harnessed in an integrated fashion to make timely factory and enterprise decisions.
Producing integrated applications that use common data sources and are configurable by the end user is another critical step in making this vast array of data useful for the factory in a cost-effective way. In this light, infrastructure is an e-manufacturing enabler by providing a secure and integrated environment of application databases and Web-enabled user interfaces based on consistent sources of data. On-line reporting and data retrieval capabilities targeted at making this data available in an integrated, flexible manner is the goal of an e-manufacturing infrastructure.
The use of emerging mainstream information technologies, such as on-line analytical processing (OLAP) and portals, allows the infrastructure to be built cost-effectively. Portals are viewed as very desirable because they automatically manage the login/password authentication process (you don't have to log in to each application individually), and provide a configurable front end for users to tailor-make the portal display to match their specific needs based on the set of problems they are interested in at the time.
The ability to easily share information with manufacturing suppliers in a secure manner is a cornerstone capability required in an e-manufacturing strategy. Through the effective sharing of the raw, filtered and analyzed data generated from common sources on the factory floor, IC makers and suppliers can efficiently identify issues and trends and collaborate on resolution plans.
A good example of this data sharing is through one integrated and secure user-interface; equipment suppliers can see how each of their tools is performing (availability, unscheduled downtime, etc.) as well as a history of performance across multiple factories at the same time. They can determine which tools are lagging the others in terms of performance and proactively set roadmaps for recovery and continuous improvement. This ability, labeled supplier linkage, could include capabilities such as e-diagnostics, specifications and on-line technical manual and specs management, support service forecasting, labor forecasting and collaboration.
In particular, the e-diagnostics capability is being defined and is maturing through a joint group of IC makers and suppliers fostered by International SEMATECH. This joint collaboration, which has been active for about a year, is allowing the industry as a whole — rather than just one company — to innovate. The group has developed an e-diagnostics guidebook that defines a set of guidelines, a capability and data taxonomy, security guidelines, industry use case scenarios, and an implementation roadmap. The guidebook, along with associated detailed information, is available on the International SEMATECH Web site, www.sematech.org/public/resources/ediag/guidelines/guidelines.htm.
Of specific interest is the capability taxonomy. It defines four levels of capabilities important to e-diagnostic implementations to reduce development lead time, cost and risk. These include access and remote collaboration, data collection and control, analysis and prediction (Fig. 2).
In general, the e-diagnostic capability gives equipment suppliers and IC makers greater visibility into actual tool performance across the enterprise, and has already demonstrated return on investment via reduction of install and qualification times as well as the average time to repair equipment on the factory floor. As process and metrology equipment continue on the upward trend of complexity and capital cost, these indicators are becoming increasingly important to efficient ramp and reduced maintenance costs of semiconductor fabs.
Another area where supplier linkage is having a positive impact on semiconductor manufacturing is that of forecasting future needs and changes. IC makers and suppliers are sharing enterprise data and have effectively reduced the cycle time required by equipment suppliers to respond to IC maker capital plans. Forecasts are also being shared with equipment suppliers via secure Web applications in terms of experienced equipment technician needs. This is positioning equipment suppliers to quickly respond to changes. Support resources, in turn, are being committed closer to the actual need date, resulting is increased agility and helping to minimize the risk to a quick and steep factory ramp.
The ability to monitor the data generated by the factory is another key element in the e-manufacturing strategy. As massive amounts of data are constantly generated, factory technicians, engineers and managers need agile tools with which to view data critical to the decisions they make as part of their jobs. Specific tools such as equipment and process monitor boards, factory indicator performance to plan, and integrated capacity planning are integral parts of this overall capability. Using Web-enabled technologies, manufacturing portals have been developed that allow specific data to be viewed in an individual, customizable way. This allows massive amounts of data to be filtered and presented in a context that is useful to each individual in the factory, all from common data sources. The result is a drastic reduction in the cycle time required to collect accurate data and the elimination of confusion resulting from multiple data sources, allowing the right decisions to be made at the right times.
Another crucial element in the e-manufacturing suite is the ability to easily perform detailed operations and performance analysis. This analysis is focused on enterprisewide data collection — complemented with flexible, integrated tools — that is composed of parameters, algorithms and models that turn data into information and trends.
While engineers and planners are still required to transform information into decisions, a few critical analysis capabilities are fully automated (optimization techniques and multiple-criteria decision-making methods are gaining momentum) that allow such decisions to be more accurate and to occur much more rapidly then ever before. The ability to calculate equipment performance using common data sources and standardized algorithms allows IC makers and equipment suppliers to focus energy on the identification and resolution of root causes and on improving the indicators as opposed to calculating them. This has resulted in improvements in equipment performance and design at a rapid pace.
On another front, because of the specificity of individual customer needs and diverse product capabilities, the number of individual line items that semiconductor fabs must manage and deliver on schedule continues to increase. Visibility into all products in process at all stages of manufacturing is increasingly important to the effective management of factory output and customer satisfaction. Tools that accurately analyze the health of these products, analyze yield, contain excursions, and present data in a context useful for the task at hand are all enabled by fundamental e-manufacturing capabilities like common data collection, data mining and presentation. These capabilities are allowing visibility into the manufacturing process health and rapid response to changing factory and demand conditions.
Global collaboration, standards
As IC makers develop requirements and implementation strategies for the realization of e-manufacturing capabilities, it becomes evident that the basic building blocks of these requirements and capabilities are common across the industry. In much the same spirit as the 300 mm transition to automated material handling, industry collaboration on these common requirements and capabilities can drastically reduce implementation timelines, cost and risk for both IC makers and suppliers.
Industry consortia such as International SEMATECH, Semiconductor Leading Edge Technologies Ltd. (Selete) and the Japan Electronics and Information Technology Association (JEITA) have begun to investigate this area and produce joint global guidance on common requirements in the form of guidelines. A three-phase approach has been defined, with guideline rollouts targeted to occur throughout the remainder of 2001. The guidance will focus specifically on the area of equipment engineering, and has objectives in the areas of overall equipment effectiveness (OEE) improvement, process capability (CpK) improvement, and factory equipment engineering cost reduction.
Equipment engineering refers to all operations for tool availability improvement and maintenance inside and outside the factory. It contains concepts related to factory throughput and maintenance improvement, tool health monitoring and troubleshooting, tool performance improvements, supplier collaborations, parts management, maintenance management, operational planning and process control.
The guidelines are targeted to include capabilities such as advanced equipment health monitoring, remote diagnostics, spare parts management, preventive and predictive maintenance, recipe management, fault detection and classification and advanced process control (APC). The e-manufacturing infrastructure must enable each one of these capabilities in the shortest possible time.
Implementation challenges
IC makers face many challenges in implementing "e" capabilities into both new and existing factories. At the heart of these challenges is obtaining automated relevant, accurate data from the equipment in a manner that is easily integrated with other factory systems. Currently, the accuracy of equipment health and process performance data is dependent on the particular supplier and model of equipment in question. This leads to inaccurate data being propagated throughout the factory systems, which in turn leads to poor decisions.
Process, metrology and facilities equipment must generate accurate performance data related to both equipment and "process" health to feed supplier links, monitoring and analysis tools. For this to work, IC makers and equipment suppliers must collaborate very quickly and come to a joint agreement on the definitions, content, methodology and collection frequency for each high-priority data element required for APC and process/production monitoring.
The second major challenge IC makers face is the integration of these process, metrology and facilities with factory systems. Currently, equipment and process data is provided in non-standard, rigid implementations that are again dependent upon each specific supplier and tool model. This leads to expensive custom integration with factory systems and often error-prone and inefficient manual data collection in cases where automated integration is not feasible.
The use of outdated, semiconductor-specific computing technology such as SECS-II and GEM exacerbates the issue further by requiring semiconductor-specific solutions. Factory equipment must provide standardized automated interfaces to reduce cost and increase integration efficiency, and must be based on utilizing mainstream computing technologies (not IC industry-specific) to allow solutions developed elsewhere to be easily adopted with less risk within the semiconductor industry.
Another challenge IC makers face in implementing "e" systems is the integration of third-party applications into traditional factory systems. Currently, integration of third-party applications for the purpose of providing new capabilities is cost-prohibitive because of the use of proprietary interfaces and closed architectures.
|
3. The use of open architectures and mainstream computing technology is critical in selecting and integrating applications into factory systems. An example of an open e-diagnostics architecture is shown.
The use of open architectures and mainstream computing technology is critical in selecting and integrating these applications into factory systems to reduce the cost of integration and long-term support. An example of an open e-diagnostics architecture is shown in Figure 3. Designing third-party applications using modular architectures and open interface technologies such as XML and HTTP will greatly reduce the burden of integration on factory systems, and allow the IC maker selections based on best-of-breed capability criteria as opposed to integration considerations.
A final "e" — and one of the most important implementation challenges IC makers face — is security. The philosophy of e-manufacturing encourages wide availability of valuable data to increase factory productivity. This data is a valuable asset to both IC makers and suppliers, and requires creation of new business processes and security measures to protect the intellectual property of all involved.
Classifying this data relative to its intellectual property value to determine the level of protection required is the first step in securing the "e" system. The use of technologies such as encryption, firewalls, proxy servers, hardened operating systems, user identification/authentication and detailed logging are required to ensure that only authorized users have access to specific functions, equipment and data within the "e" systems. Business procedures are equally important and must also be in place to complement technology use, to ensure that security is maintained when user, capability or data changes are made to the systems.
Conclusions
Affordable Internet concepts and technologies are changing the landscape for the IC industry. E-manufacturing concepts such as supplier linkage, monitoring and analysis are key components in realizing the productivity improvement mandates that semiconductor manufacturers and suppliers must be ready for.
Intel is approaching such concepts through active participation in industry consortia and non-proprietary forums to develop technologies and industry roadmaps that allow reduced timeline, cost and risk factory implementations ahead of need. Significant challenges remain in the areas of accurate and standardized equipment data, open architectures and systems and security for e-manufacturing capabilities to be realized in semiconductor fabs.
Suggested additional reading
J. Baliga, "E-Business Enters the Semiconductor Industry," Semiconductor International, March 2001.
D. Bloss, "E-Diagnostics, an IC Maker's Perspective," SEMI Technology Symposium, December 2000, Tokyo.
D.F. Blumberg, "Service Program Diagnostics and Decision Support Technology for Improving Health Technology and Service Efficiency and Productivity," Blumberg Health Technology Service Report, August 1998, p. 370.
"Making Sense of E-Manufacturing: A Roadmap for Manufacturers," an industry white paper by Rockwell Automation, with Allen Bradley, Reliance Electric, Dodge and Rockwell Software, November 2000.
E-Diagnostics Guidelines Version 1.0, Security Guidelines Version 1.1, and E-Diagnostics Guidebook, SEMATECH Web site, www.sematech.org/public/resources/ediag/guidelines/guidelines.htm.
Global Equipment Engineering Service Collaboration — Big Picture Results from the Selete/JEITA/SEMATECH Face-to-Face Meeting, February 2001, Narita, Japan.
D. Pillai, D. Bloss, "Internet Enabled Semiconductor Manufacturing," abstract submitted to the International Symposium for Semiconductor Manufacturing (ISSM) conf. proc., October 2001.
A. Raman, P. Fioravanti, "Infrastructure and Security Architecture for E-Diagnostics," IMEC 2001 abstract, February 2001.
P. Singer, "E-Diagnostics — Monitoring Tool Performance," Semiconductor International, March 2001.
Dave Bloss is automation integration manager in Intel’s Technology and Manufacturing group. He manages programs responsible for the development and delivery of automation capabilities within Intel Fab and Sort manufacturing worldwide. He has participated in and led a variety of industry activities including ITRS, I300I/J300 CIM Global Guidance, International SEMATECH e-diagnostics Working Group, and various AMHS and Production Equipment SEMI International Standards task forces over the last five years.Devadas Pillai is the director of operational decision support technology in Intel’s Technology and Manufacturing group. His team is responsible for developing and utilizing leading-edge operational decision support capabilities for use in factory optimizations, dynamic modeling simulations, and finding ways to exploit the use of e-manufacturing tools across the Intel manufacturing network. His team also coordinates 300 mm factory integration and layout standardization efforts and guides SRC’s Factory Operations research priorities for the company.
ACKNOWLEDGMENTSThe authors are greatly indebted to numerous individuals at Intel Corp. who are delivering the capabilities outlined in this report. Thanks also goes to the team members from IC makers, equipment suppliers and software suppliers participating in the e-diagnostics standardization and guidelines effort channeled via International SEMATECH and Selete.
