Hoechst AZ markets 193-nm, CARL products

07/01/1997

Hoechst AZ markets 193-nm, CARL products

Hoechst Celanese, producer of the AZ line of photoresists and ancillary products, is investing $80 million to expand its worldwide electronic materials facilities as part of a broad effort to gain market share over the next few years. In addition, the AZ business unit is beginning to sample 193-nm resists, and is commencing a marketing effort for chemical amplification of resist lines (CARL), a Siemens-developed process that could extend i-line lithography to 0.25 ?m on certain device layers.

Hoechst`s experience with 248-nm precursors has been helpful in developing deep UV products, according to US general manager Friedrich Herold. In the 193-nm line, where cyclic olefins are a likely element of resists, Hoechst has many years of experience with the compounds from its engineering plastics business. "We`re fully back-integrated in resins, catalysts, and polymers" for 193 nm, said Herold. Some samples of 193-nm products have gone to customers, and process development samples will ship soon. Availability of exposure equipment is one problem, he said, adding, "when we talk to customers and equipment makers, not everyone believes 100% in 193 - it`s not 100% defined when and to what extent it will be used."

Hoechst`s 193-nm research is being carried out in the US at a Center of Excellence, which also works on antireflective coatings, advanced i-line processes, and thick-film products. A center in Japan handles deep UV and flat panel display products, while most raw materials work is done in Germany. As products move from R&D into the market, they are customized for particular regions.

Meanwhile, AZ has begun preliminary marketing efforts for the CARL process, developed and now being used in production by Siemens. The procedure, applicable to reflective substrates with heavy topography, has shown an ability to print features of 0.25-?m and less using i-line tools, said Dana Durham, director of technology.

CARL, whose development dates back to the mid-1980s, calls for an initial application of a thick resist layer, which planarizes the topography without the need for CMP. Then, a very thin layer of an imaging resist is applied on top of the first layer, and the wafer is exposed and developed. Next, a silyation agent is puddled on, and the imaged resist absorbs more or less of the material, causing the linewidths to change in accordance with how much time the agent is left on, and what temperature is maintained. Finally, the resulting image is transferred into and through the thick resist layer in an oxygen plasma process. "It`s pretty nifty," said Durham. He cautioned, however, "it`s not free - there are extra process complexities. But you don`t need deep UV. And depending on the application and your cost of ownership modeling, it could be attractive." -P.N.D.