Materials Case Study
the fab operators, governed by their approach to tightly control proprietary IP and technical data about a given node, geometry or process. This has led to a marked tendency for new
tools to be created, or existing tools to be modified/substantially updated, in isolation from and without a significant amount of consultation with chemistry suppliers. It has inhibited the chance to integrate innovations in tool designs with advances in the performance characteristics of engineered chemistries to optimize the results of a given process step.
The Standard Model: Parallel Development Can Hinder In response to technical direction from the semiconductor manufacturer, a chemistry supplier may develop new performance chemistries with characteristics engineered to solve specific challenges: improved viscosity, improved recyclability, formulations that reduce environmental impact, and most recently, chemistries selective to a specific layer. At the same time, in parallel, tool builders
22 Wafer bath
frequently develop systems that are not compatible with these new chemistries, missing the opportunity to design tools that leverage the performance of these newly developed chemistries, such as broader process temperature latitude or sensitivity to certain materials of construction. After testing offerings from multiple suppliers, the semiconductor manufacturer generally chooses a new chemistry for its process of record, based on results in lab testing or pilot runs. But when this chemistry is utilized in the process tool, costly modifications are often required to optimize the process step:
chemistry temperatures, circulation rates, bath times and additives have to be added or adjusted, and can typically require some level of tool redesign to reach desired yield and performance targets.
A classic example is the introduction of high-dose ion implantation of photoresists. The exposed photoresists are transformed into carbonized-hardened crusts, which necessitated new stripping techniques and chemistries as standard mechanical or chemical processes were not fully effective. To remove these hardened resists, a next-
generation process was needed. The industry- accepted solution called for a process that utilized special, high-temperature chemistries first to loosen and soften hardened photoresist, and then precisely attuned mechanical motion to lift the residue without impacting very small geometries to which they were adhered. From the chemistry supplier’s perspective, certain performance characteristics were crucial. The strippers needed to be: Very stable and not interact with the substrate
Compatible with high-k dielectrics Able to chemically modify and break down the densified resist without affecting the morphology of the surrounding structures.
Multiple companies submitted similar aggressive chemistries designed to break-up and attack the “crusty” post implant resist. Commensurate with these chemistries, a major modification of the tools was needed, such as high temperature mixing tanks and segregation of the active components prior to addition to the mixing tanks. Non-standard metallic and plastic components to minimize corrosion needed to be fitted as well.
These complexities, which arose out of the unique characteristics of the newer chemistries, led to tool delivery delays and in some cases a withdrawal of some tool developers from participation. For some manufacturers, an unworkable solution was introduced into the process, producing delays in ramping to full yield.
Joint Efforts A better approach would have been for tool builders and chemistry suppliers to be teamed by the semiconductor manufacturer to solve the hardened photoresist removal challenge as a combined effort. All three participants will need
www.euroasiasemiconductor.com Issue III 2011
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32