Maintaining a Clean Environment in DUV Lithography Bays
Laura Peters, Senior Editor -- Semiconductor International, 5/1/1999
Unlike almost every significant trend in the semiconductor industry, the specification for amine contamination levels in deep-UV (DUV) lithography processes does not change with device generation - a surprise in an industry that practically defines change.
According to the 1997 SIA Roadmap, total airborne amine contamination levels must be kept below 1 ppb (molar) over the next five device generations, or as long as chemically-amplified (CA) resists are used. 'The reason the specification does not change is largely due to the continuing success of photoresist manufacturers at formulating resists to be less prone to amine contamination,' explained Devon Kinkead, president of Extraction Systems Inc. (ESI, Franklin, Mass.). ESI supplies airborne amine filtration and real-time monitoring systems for DUV applications. The monitor uses chemiluminescent technology to provide a single-digit output of total amine concentration at up to ten locations in the track system and scanner (see Figure) as shown or, alternatively, several tool clusters. Alarms are set by the user based on the photoresist's sensitivity.
 |
| Fig. 1. Airborne molecular contamination control and monitoring indicates contamination levels at up to 10 sites in the DUV cluster. |
While filters and monitoring techniques have been available in this area for several years, recent progress addresses production issues of filter lifetime, process stability, monitor reliability and cost-effectiveness. 'Initially people buy monitors so they know when to change the filters,' explained Kinkead. 'While that's a viable reason, the reality is there are all sorts of contamination events that cause spikes and valleys in amine levels that have nothing to do with the filtration equipment - like an operator accidentally turning off the air handling equipment or someone leaving a stepper or scanner door open.'
Materials of construction can also provide a source of contamination. ESI recently tested five caulking materials to determine which one(s) was most appropriate for use in a DUV bay. 'Some customers test new materials by exposing a wafer, putting it in a box with the material of interest for a given delay time, taking the wafer out, developing it and taking a SEM (scanning electron microscope) image of it to see whether CD increased or 't-topped' - this is the expensive way to perform such tests.' ESI's monitoring tools and laboratory services lower the cost of such testing dramatically. As it turned out, two materials had no detectable levels of amine off-gassing, three had high levels and one was moderate - proving that seemingly small material decisions can be significant.
Interestingly, ESI got its start in molecular filtration and monitoring in
1991 when it was contracted to control the environment in the new U.S. National
Archives in College Park, Md. 'We had to control four different pollutants in a
two million square foot building down to about the same level required in a
stepper,' explained Kinkead. 'It was a good prerequisite for working in the
semiconductor industry.' 
