Issue



Automation is key to getting lean for 450mm manufacturing


04/01/2009







ISMI has been the leader in implementing small-lot manufacturing and single wafer processing for 300mm Prime and next-generation factory (NGF) 450mm manufacturing. By driving and gathering the implementation requirements from chip makers and equipment suppliers, ISMI makes it simpler for chip makers to select NGF MES supplier products and plan for their fab MES deployment. For the advanced 300mm Prime and the NGF 450mm, the main goal is to enhance the productivity improvements necessary to implement fully continuous processing. To achieve this goal, wafer processing interruptions and improving waste reduction in the fab are the keys to success.

Brandon Lee, Dennis Talbert, CIMAC, San Jose, CA USA

As global competition increases, semiconductor chip makers are constantly looking for ways to be more efficient and competitive. At a time when getting more for less has become a common buzzword in any business sector, it is an absolute requirement for everyone to re-evaluate what they have and initiate changes to lower costs and maximize profits. In the semiconductor industry in particular, there exists a continuous effort to raise manufacturing efficiency to a higher level.

Given the current economic environment, one may think it risky to start a new project — on the contrary, it is an ideal time to focus on productivity improvements and cycle time reduction. When the upturn hits, those who seek to get more for less will benefit most from investing in manufacturing efficiency.


MES deployment lifecycle.
Click here to enlarge image

For NGF projects, there are priority items from the 19 Point Guidelines published by ISMI. Many of them, including equipment availability and predictability, are related to improvements for equipment suppliers. This drives the factory manufacturing execution system (MES) to be smarter than the MES that are running in current 300mm fabs. The NGF MES must be able to adapt to predictive and preventive maintenance, wafer delay reduction, continuous processing of materials, and most importantly, provide a single point of control for the factory system.

MES: replacement or upgrade

Manufacturing execution systems are ofTen a leading factor in achieving the efficiencies required for 450mm manufacturing by coordinating the flow of materials, facilitating communication between tools, relaying vital information to operational management, etc. The application of new versions of sofTware (either the MES sofTware itself or supporting sofTware) is ofTen central to achieving these efficiencies and a competitive advantage. In addition, with new manufacturing technologies and increasing business and customer reporting demands, more functionality is required. All of these are reasons that may require upgrade or replacement of existing MES sofTware solutions.

However, the very thought of changing an operational MES can place significant stress on the operational organization of a fab. Do the lean fabs of today have the workforce to evaluate and implement possible alternatives? This article discusses many of the significant considerations involved with the evaluation of MES alternatives and upgrades; the strategies for requirements gathering, management, and tracking; the implementation of MES functionality; and, finally, the deployment of an operationally significant solution.

The decision process

There can be many driving forces to replace an existing MES. Several are discussed below.

Technology obsolescence. The end of life of a previously supported and successful technology can be a necessary driving force for change. An example of this would be the need to migrate the MES from one server platform to a current platform technology.

New technologies. The desire to take advantage of newer technologies is another motivator. Advancements in human/machine interfaces (HMI) have sparked the move to a change in operational sofTware as users sought to make operational personnel more productive and data inputs more accurate. Over the years, we have seen MES sofTware transform from character-based terminal interfaces to Windows-based GUIs to web-enabled operator and engineer user interfaces.

Business consolidation. The business/technical environment response to marketplace conditions has also brought about the need to change an MES at an operational organization. If companies merge, it is probably better in the long run to standardize on a single MES solution. As a result, at least one of the organizations will likely have to change its operational MES.

Tool connectivity. Another driving force for the change of MES sofTware solutions is tool connectivity. As more sophisticated tools and processing technologies are adopted, there is a need to introduce more automation to perform transactions previously done manually, and to reduce input processing errors by human operators. In the drive to reduce recurrent costs and make fabs less labor intensive, the capabilities of new or upgraded MES sofTware are essential for the cost-conscious fab operations of today. This frequently requires MES sofTware upgrades or even migration to a new MES.

The latest versions of MES sofTware have added security capabilities to allow users to access data that is required for them to conduct their tasks but at the same time restricting them to only the data that they have a need to know. This becomes essential in protecting an end user’s intellectual property (IP) and is a requirement for fabs acting as foundries. However, even when the MES is limited to only one end user, access to data can be limited by the role that user performs.

In addition, the introduction or change of manufacturing technology in the fab itself can be a motivation to change MES technology to the most current products. Many of the popular legacy MES products lack the capabilities to support a fab migrating to new product manufacturing sizes; such as the movement from 200 to 300mm technologies, for example.

In CIMAC’s latest two projects to upgrade MES, the major motivators to accomplish the transition were changes of manufacturing technology; the desire to take advantage of technological advancements; tool connectivity for new tools to support advances in manufacturing technology; and taking advantage of additional security features to protect customer’s intellectual property. It is anticipated that these conditions, as well as possible business consolidations sparked by the world’s current dynamic business climate, will continue to be the motivators for future MES upgrades and migrations.

Handling MES upgrades

We have identified the following factors as key to determining whether to manage and implement the changed MES sofTware environment with external assistance, or to manage and implement with internal resources: management of business requirements; availability of internal work force; internal work force’s expertise in MES nomenclature and capabilities; and the company culture of the operational fab. Clearly, it need not be a case of all internal resources or all external. However, the following discussion may provide some insights into what the considerations might be in order to evaluate whether to utilize external or internal resources.

Establish whether internal resources are knowledgeable and experienced with the scope of activities required to bring the fab’s business requirements into a viable, operational MES. This amounts to much more than operational domain or product processing knowledge. It includes critical tasks such as business analysis, requirements collection and management, rendering functional specifications in a way that sofTware can be produced to satisfy them, change management affecting the data, the sofTware development, the operational environment and other relevant processes and artifacts, as well as an understanding of sofTware validation and verification practices.

Fab management must be prepared to assign significant resources to a new MES implementation. Given that an operational fab is still going to require substantial engineering and legacy MES support, will the necessary internal work force be available? Ken Simpson, a long-time MES consultant, indicates that as much as 10% of engineering staff may be involved in the project [1]. In addition, will the fab engineering staff have the necessary expertise in MES nomenclature and capabilities? This is not typically a fab engineer’s core expertise.

CIMAC, for example, recently provided an end user service program that enabled a semiconductor manufacturer to assess the gap between its existing capabilities and what was desired, and evaluate different vendor proposals as solutions to alleviate problem areas. Senior CIMAC MES specialists analyzed the end user’s current manufacturing workflows and manufacturing applications to provide a complete picture of current manufacturing operations. Working in conjunction with internal management, they identified improvement opportunities, analyzed potential technology investments, and provided a recommendation as to the best proposed vendor alternative.

The company culture of the fab is also a determining factor in making the decision. Issues of concern are whether internal documentation is adequate to describe processes and procedures, or whether such processing is performed by rote, or by some collective memory of the operational culture. A lack of clear documentation of operational processes may hinder any vendor from translating perceived functional requirements into an implementation that reflects what fab operations really wants done. Are there areas of special concern that need attention and, if so, does internal staff have the expertise and the time to address these special concerns?

CIMAC’s latest experiences revealed several cases of de facto procedures that superseded all existing documentation of what was done by whom, how, and when. Requirements gathering personnel had to insist upon meeting with operational-level stake holders to clarify perceived operations from the real and the desired. In addition, the management of requirements changes brought about by changes in technology, personnel, or critical reviews of the documented requirements had to be managed and communicated to implementation personnel. These tasks were critical to the success of the implementation migration.

Defining requirements of an MES System

As shown in the figure, selecting a best of breed MES entails requirements collection, analysis, and management — very important to a successful implementation and deployment of a MES whether it be a mere upgrade or a completely new system. When evaluating different solutions it is important to ensure that the request for quotation (RFQ) matches what the operational fab requires. Several methodologies and techniques can be used to do this. However, there will be a need to define the important business processes. In many older fabs, documentation of these processes may be nonexistent or out of date. To avoid costly sofTware revisions in the future, business models and processes need to be rigorously defined and documented.

One must be able to trace these essential business requirements throughout the implementation and deployment process of the MES. To accomplish this, a change management process must be specified for requirements. In fact, not only might the requirements be changed during the selection, implementation, and deployment of the MES, but relevant processes, products, data, personnel, etc., will be changing as well. Successful deployment will depend to a great extent on how the changing environment is managed during this time. Will internal processes and resources have the expertise and bandwidth to accomplish these tasks while maintaining operational throughput?

Another concern is how to put the RFQ into MES language so that requirements are fulfilled as desired. In other words, the business logic and processes need to be translated into “MES speak” to better communicate with the MES sofTware vendors. The RFQ needs sufficient definition to cover details that could otherwise be overlooked. For example, requirements for specifying a state model to reflect processing capability of a tool should also require that the model be configurable with additional states the end user’s business practices may require at some later time. The requirements specification activity may be an area where there is a need for a company with core expertise in MES integration to take part.

Reducing risk during vendor selection

MES sofTware vendor selection can be a highly resource-consuming endeavor. How can fab operations management ensure that it gets the most out of this effort regardless of whether internal or external resources are used? Our experience has been that the following actions provide some valuable input to the MES vendor selection process and substantially enhance the probability of a successful implementation and deployment.

Product demonstration. A demonstration that shows the ability to solve problems that would realistically be encountered is key, along with reference site visits. These sites should use a configured system similar to the one being considered. Additionally, get the vendor product in front of staff that will use it, both operations as well as engineering personnel. This will provide valuable feedback on usability features as well as provide helpful initial insights into how the MES standard product is a match for your fab operations.

We specifically recommended a visit to the vendor’s demonstration lab for operational personnel to view the HMI that would assist with the movement of lots, the collection and analysis of data, and the changing of equipment status. An alternative to this would be the set-up of a functional demonstration in the customer’s environment to garner feedback and validate the operational functionality of the potential solution.

Staff a quality assurance function. A quality assurance staff that does more than merely functionally test the delivered sofTware is important. A successful deployment is dependent upon reviews of the development process from planning through the installation. On recent projects, we have recommended that operational personnel participate in functionally testing sofTware early in the development process, which enables significant feedback on the effectiveness of the HMIs, as well as providing validation of the completeness of the delivered functionality.

We also recommend an organization plan for a realistic performance simulation or characterization before the final deployment. It is important to know well beforehand how the manufacturing sofTware is likely to behave when your organization starts running a full load of product. A well-thought-out and developed performance characterization will protect your organization from future surprises and enable your organization to solve bottlenecks before they actually occur [2].

Staff a change management function. A change management function will be even more important if you are migrating from one MES to another. The production environment is dynamic, with products and manufacturing processes constantly changing and being refined. Business requirements are likewise dynamic and reporting requirements and customer requests continually need to be updated, which will have an impact on the MES sofTware and the system requirements. These changes need to be managed and controlled. It is likely that the fab’s existing infrastructure and processes will not be able to account for all of this new change in the environment. Consequently, a change management function focused on the MES implementation needs to be staffed and set to work from the requirements collection phase through data migration and final system deployment.

Successful deployment execution

First, and most importantly, one must have the highest level of management commitment. The support from the top levels of management is crucial to the success of the plan. At some point in time, whether ample use of internal or external resources is utilized, scarce resources will be requested and the backing of top management will be required.

Secondly, there is a definite need for strong, experienced project management — from the vendor and the end user sides. On one recent upgrade deployment, the vendor had three different project managers during the duration of a one-year project. This change in leadership led to some issues with the interpretation of requirements, as well as some issues in execution to schedule. Indeed, there were not only changes in project management by the vendor but also by the end user. This lack of continuity in management led to communication issues with regard to schedule and the interpretation of requirement commitments. All of these issues were resolved once stable project management was implemented.

Comprehensive schedules with details of all resources needed from the end user are essential to making a successful deployment possible. Our recent experience has indicated that it is necessary but not sufficient for the vendor to present and track a plan only specifying vendor-supplied resources without indicating the essential end-user-supplied resources. For example, an early validation of the HMI is not possible without having the operational personnel interacting with it.

MES upgrades or replacement projects need a risk management structure to identify risk and must have a mitigation plan ready. The risks and mitigation plans must be updated constantly to keep pace with the progress of the sofTware development as well as the constantly changing business and manufacturing process environment.

We recommend that sufficient resources from both the end user and the vendor be definitely committed before the project starts. This includes a clearly defined budget. Once again, periodic schedule and plan reviews will highlight the resource issue and management commitment and the risk and mitigation plans will assist in reacting to resource issues.

Finally, since we have stressed the dynamic environment in which these upgrades and migrations will be taking place, our experience has shown that the need for a change management process that is documented, staffed, and enforced is of tremendous value in ensuring the successful deployment of such a project. We recommend that a specific project resource for this task be assigned if it cannot be integrated into the end user’s existing organizational infrastructure.

Conclusion

Ultimately, it is not an easy task to improve equipment productivity, lower cycle time, and transition into small lot size and single wafer manufacturing. It requires significant factory operation improvements to extract the full benefit from these measures and that is why requirement gathering and careful planning of MES selection is one of the top priorities in the NGF productivity improvement lifecycle.

References

  1. Private discussion.
  2. B. Lee, S. Dash, “Performance Testing Manufacturing Execution Systems (MES),” Semiconductor Fabtech, 2008.

Brandon Lee received his two BS degrees, mathematics and computer and information science, from the U. of Oregon and wrote a post-undergraduate honor’s thesis in fractals and now is the director of marketing and business development at CIMAC, Inc., 1590 The Alameda, suite #100, San Jose, CA 95126 USA; ph.: 409-985-4300; email [email protected].

Dennis Talbert received his BA from Howard U. and is an MES and sofTware quality consultant.