Efficiency in sealing: A BKM case study
DALIA VERNIKOVSKY, Applied Seals North America Inc., Newark, CA, and BRAD ECKER and SETH URBACH, Microchip Technology, Inc., Chandler, AZ.
This case study describes Microchip's experience with perfluoroelastomers o-ring seals.
There are many reasons to review the productivity of electronics manufacturing equipment, whether processes are being upgraded or efficiencies need to be reviewed for optimal performance. In this case study, a best-known method (BKM) approach was taken to optimize the tool platform and the preventive maintenance (PM) cycles for that platform. The study was the result of a collaboration between a Microchip team of equipment engineers, technicians and process engineers, and an engineering/sales team from Applied Seals North America (ANSA).
We focused on three aspects of equipment optimization:
??? The failure modes of existing seals (o-rings) were evaluated. In this case, six used seals were examined and a failure analysis (FA) conducted for each.
??? The equipment was audited for proper sizing of the seals, as well as chemical compatibility. One of the six seals required special sizing and geometry to assure the optimal life of the sealing element. Figure 1 depicts several different failure mechanisms observed. It's important to note that most of the FFKM family of seals has a high concentration of the polymer base versus fillers. As a result, size and shape rarely influence higher costs (the terminology, FFKM, indicates the TFE backbone polymers that distinguish this material).
Figure 1. Seal failures can be caused by compression (left), powdery contamination from etching (middle), or damage due to poor installation (right).
??? We ensured that the seals were properly installed. Half of the failures of FFKM materials are typically a direct result of overstretching or twisting the material in the groove because most technicians are unaware of the sensitivity of these materials to basic handling issues. For example, white or translucent FFKM materials cannot be stretched on average of more than 3% and must be formed to shape in order to avoid stress points which allow early etching or chemical attack.
Tests took place with the incumbent seals on one chamber and the new seals on another chamber to accurately depict the same conditions when the test began.
Figure 2 depicts the data obtained, indicating a new seal design and new material of construction (a new compound) doubled the life of the products.
Figure 2. SPC chart of weekly tool qualification. On 5/12, the chamber was brought back into control with process response (no maintenance intervention).
The need for standards
Why was this test so important? Newer generation sealing components ??? FFKMs, otherwise known as "perfluoroelastomers" -- are recognized for their chemical resistance and general ability to reduce contamination. However, if the components are not carefully evaluated and set in the equipment properly, new failures will arise. These failures take a great deal of time and money to be identified and resolved, which impacts productivity. This was the case at Microchip, where the seals exhibited a limited life expectancy.
Unfortunately, standards for this new generation of seals have yet to be adequately defined for the semiconductor industry. As a result, there is a significant gap in understanding how these parts must be installed and handled. This lack of understanding can result in significant hours of downtime. However, with the proper training and education for this new generation of product, these failures can be avoided and manufacturers can use "the right seal for the right application," ensuring minimal downtime and reliable production runs. ASNA is heading the effort to generate these much needed standards through SEMI/Sematech, chairing a committee (F51-0200) to create such a document. The goal is to establish some level of performance criteria to guide our industry. More attention and participation by end-users and OEMs will assure that these standards are effective in areas of process, types of performance attributes and some criteria to capture those performance parameters (more information to be found at http://www.appliedsealsna.com">www.appliedsealsna.com)
The lack of understanding can also create a disassociation between the o-rings that are thought to be the answer to resolve advanced processes and their ability to do their job, that is, to seal (advanced processes include all the new hafnium precursors, harsher fluorine cleans and smaller and smaller killer defect parameters predominantly in etch, CVD and 22nm and below nodes). The conventional wisdom is that the "cleaner" the o-rings, the more they will be etched away and, as a result, the less they will mechanically fit. This issue can be avoided if the o-ring is designed properly and handled appropriately from the outset. Without addressing this important issue, failure of seals will continue to impact productivity and their misapplication will continue to cost the industry. Failures could continue to be factors in hardware of all types of advanced technologies and issues of chemical breakdown or mechanical failure will continue to be encountered. If these issues are not addressed, the effects of their misapplications and mishandling could impact the advances of these technologies for years. This is especially true because the cleaning gases (now generally NF3) continue to evolve, and potentially have an even great impact on the life of the seals.
Many families of sealing elements are considered the same, when they are, in fact, very different one from the other and are not interchangeable. A basic knowledge of what they are made of, how they are made, the various parts that include cross-linking agents, curative variations, and fillers must also be considered. Triazine cures (developed for high-temperatures reaching 300??C) are good for diffusion processes but not optimal in etch and CVD as they can become sticky when various chemicals are introduced. Peroxide-cured materials are superior for chemical resistance and cover over 80% of the process requirements in this industry. This is an important factor in the successful application of materials matched to successful production requirements. Along with that, basic fillers such as BaSO4, TiO2 and silicas are used but once again, understanding that there are many variations and forms of these materials will help discern how well they perform under these requirements (i.e., process dependency, cleaning gases, throughput , particles, etc.).
In summary, there are many key issues and challenges our industry faces with these components. As exemplified by a project dedicated to reducing particles on EUV mask blanks ??? a joint effort between SEMATECH and ASNA -- it was ascertained that many of the particles that led to defects were actually traced to the seals used in the deposition tool for the blanks. This program is dedicated to bringing EUV technology into practical use for the 22 nm node and beyond, yet must still deal with what some common hardware issues that remain unresolved.
Moving forward, work should focus efforts required to achieve the levels and types of particles, fillers, materials and designs that will facilitate the advancement of our science and assure that the industry's future aspirations will be met. Seals are only one element, but as more and more is discovered about contamination sources, we must address them so that they do not limit the ability to reach the levels of advancement we can achieve. The conclusion: working with material experts and components will indeed be a path that will enable the successful implementation of advanced technological leaps.
Research and input provided by Brad Ecker, Project Leader and Seth Urbach, Equipment Engineer for the Microchip and the full team that was pulled together to make this project possible.
DALIA VERNIKOVSKY, is the CEO of Applied Seals North America Inc. BRAD ECKER is a senior process engineer at Microchip and SETH URBACH is an equipment engineer at Microchip Technology Inc.
Solid State Technology, Volume 55, Issue 6, July 2012