Sealing better yields

As profit margins decrease so the scope for investing in new, more efficient plant diminishes. It’s at this point that looking for ways of improving manufacturing performance and productivity start to take on a new meaning, and urgency. One way of doing this is to change the seals less frequently, thus extending the preventative maintenance cycle for seals and reducing machine downtime.

Alternative sealing solutions are often overlooked as a way to increase productivity on Deposition, Etch and Ash Tools, for example. Semiconductor engineers will be familiar with the range of sealing materials available to them. In most cases, Fabs will use a perfluoroelastomer (FFKM) or a fluoroelastomer (FKM) material due to exposure to the combination of harsh chemicals and temperatures used in most semiconductor processes. Perfluoroelastomers are often preferred over Fluoroelastomers due to their lower etch rates and higher temperature capabilities.

When reviewing Preventative Maintenance (PM) cycles for seals on the production line, there are a variety of improvements to be considered with a change to a newer technology seal. Aside from the obvious unit cost, one must look at tool down time to install, time between PM’s, wet clean procedures and, of course, yield after installation. From a Management point of view, die yield is the most critical factor. Less workable die per wafer equals less profit. There are many variables attributed to a high or low yield. But from a sealing perspective, particles from an o-ring have a major impact on increasing yield. Especially, when these particles are deposited in a Front End process, but not found until somewhere in the Back End process.

Seal specific solutions
Traditionally, o-rings and seals have used inorganic fillers like Silica (SiO2), Barium Sulphate (BaSO4, or ‘barytes’), Zinc Oxide (ZnO), Carbon (C) or Titanium Dioxide (TiO2) as the base material to ‘fill’ the o-ring to provide rigidity, and to improve the general physical characteristics of the elastomer. The type of filler material used is dependent on the application of the o-ring.

Perlast has developed sealing technologies specifically for the semiconductor market. These rely on organic fillers to deliver increased resistance to chlorinated/fluorinated gases, solvents and cleaning agents. In addition the presence of the organic filler significantly reduces the risk of particulation of the seal, and thus offers far lower particle counts than inorganic filled perfluoroelastomers.

Understanding how long the current seal works versus a newer technology seal is very important in increasing the time between preventative maintenance cycles. During semiconductor processing the seal is exposed to both aggressive gases and solvents and high temperatures. As the temperatures increases, initially, the polymer degrades to a gaseous product releasing particulates which can contaminate etching plasmas and silicon.

The graph on the previous page shows that a regular, inorganically-filled system such as traditional FKM and FFKM, on degradation, leave substantial residues, generally made up of mixed metal-oxide particles. These particles are the cause of ‘particulation’, and subsequently, having to replace an o-ring before one would actually need to. The organic system is shown to leave no residue at all, decomposing entirely to gaseous products.

When moving to an organically filled seal, particles are reduced dramatically. The latest generation seals using organic fillers degrade to form only gaseous species. This is a step-change improvement over the standard filled grades which have extractable particles well above 150nm. Subsequently, this can enhance a Deposition process PM cycle since seals no longer have to be replaced as often due to particle contamination, particles are virtually nonexistent, leading to significant increase in yield.

Improving productivity
To emphasise the huge improvements with an organic filled sealing material, production trials at a large semiconductor fab, compared performance of traditionally-filled material in a gate valve application. The traditionally-filled grade lasted only 6,000 wafers before particulation issues forced a seal change, however, when using Perlast organic filled material, the wafer count reached 20,000 wafers. A significant increase in productivity and quality, as well as reduced preventative maintenance as well as costs.

One clear indication that a semiconductor manufacturer should consider switching to organic-filled seals is when the seals preventative maintenance is less than claimed by the equipment manufacturer, or where there is a recurrent contamination problem. Of course, it may also be that the PM cycles are not fully optimised with service engineers erring on the side of caution.

One method of determining PM cycles is to examine seals using a number of different analytical methods, after various exposure times, this can allow accurate PM cycles to be set. For examples, thermo gravimetric analysis allows the polymer content of an elastomer can be accurately measured.

Reduction in equipment Cost-of-Ownership
From a production perspective, the knock-on effect of using organic filled elastomers is to typically reduce the Cost of Ownership (CoO) of equipment by as much as 25%. For clarification, Cost of Ownership is calculated based on: Variable (Operating) Costs, Depreciation Costs, Indirect Variable Costs and Indirect Depreciation Costs. Seals fall under the Variable Costs and are one of the main factors in determining the Equipment’s Cost of Consumables (CoC). When reducing CoC by utilising better performing, cost effective seals, it brings down the Equipment’s original CoO numbers.

In a competitive world, all companies in the supply chain need to continue to invest in the latest filler technology to push costs down and increase overall productivity and yield. Longer PM intervals can be achieved by using organic based sealing solutions. These materials are already proven to extend preventative maintenance cycles by lasting up to ten times longer than traditional seals and have a lower particle count throughout their service life.

Every incremental improvement in CoO and CoC is required by manufacturers as cost of research compared to cost of manufacturing is pushing economic returns to the tightest margins. Fab managers need to look at the entire process and discover cost effective improvements to their process without negatively impacting on the overall device performance.