Air Products' Electronics Technology Customer Applications (ETCA) Group was established more than 10 years ago. It has been actively involved in assessing the impact of PFCs on global warming and creatively establishing ways to mitigate PFC emissions from semiconductor manufacturing facilities' processes. In a recent interview, three of the original group members—Dr. Pete Maroulis, Dr. Andy Johnson and Dr. Bob Ridgeway—discussed the group's work that led to the winning of a 2002 EPA Climate Protection Award.

The three have more than 40 years of combined experience in working with and for Air Products' global electronics customers. Maroulis, the ETCA group manager, is a senior research associate, while Johnson and Ridgeway are lead research chemists. Maroulis and Ridgeway excel in analytical chemistry with a focus on atmospheric chemistry. Johnson's specialty is physical chemistry as it relates to semiconductor manufacturing.

Rick Pearce is another original team member. Other team members include Tom Strencosky; Dr. Steve Rogers and Dr. Moo-Sung Kim, both working in Korea; Dr. Chris Hartz; and Dr. Haifeng Zhang.

Question: Why was your group established?

Dr. Maroulis: The major impetus was that we wanted to help our customers. Fab emissions were becoming more of an issue and the EPA had challenged the aluminum industry to reduce its emissions of C₂F₆ and CF₄. Well, those same gases were being used by the semiconductor industry, mostly for chamber cleaning. And, even though the EPA eventually worked with the semiconductor industry to develop the Memorandum of Understanding for emissions, no one was exactly sure what those emissions were and what effect they might have on our atmosphere.


Question: How long did it take for the group to achieve success?

Dr. Maroulis: Virtually, from the start. In 1993 we worked with AMD and conducted our first effluent project at Sunnyvale, California. The result was the first-ever program to fully characterize fab emissions.

Dr. Johnson: Before that, I don't think many people really considered what was coming out; they were only concerned about what was going in. Chemistry in the semiconductor process produces other byproducts that have environmental impact. Process analysis is needed to identify these additional emissions. You need to measure something before you can improve it.


Question: What was your next step?

Dr. Maroulis: Once we had developed a working method for characterizing emissions, we turned toward finding ways to minimize emissions by optimizing the use of C₂F₆ and introducing NF₃ as a chamber cleaning gas option.

Dr. Ridgeway: Once we could characterize emissions, the industry started demanding better abatement technology, so Air Products and the group became pretty involved in the whole process. We found that sometimes the cure was worse than the disease; abatement systems emitted even worse byproducts, including NO and NO₂, precursors to smog. Working with SEMATECH and Motorola, we carried out the first full characterization of an abatement system. We learned that higher temperatures, for example, produced greater Knox emissions, so we adjusted accordingly.

Dr. Maroulis: It's a pretty simple formula. First, at every step of your process you do what's best for your particular case to minimize emissions. This entails developing a process with good performance and good cleans. This can be done by optimizing the current process or changing the chamber clean gas.

Air Products' senior research chemist, Dr. Christopher Hartz, checks data during field trials for optimized C2F6 PECVD chamber cleaning.

Dr. Johnson: Less effort had been made developing chamber clean recipes because it didn't impact the silicon wafer. We showed that substantial reductions in emissions could be achieved by adjusting the C₂F₆ process parameters. Further reductions and faster cleans are possible by switching to an NF₃-based process.


Question: How well did the industry appreciate your findings?

Dr. Maroulis: The feedback has been tremendous and extremely positive. That's why winning this EPA Climate Protection Award means so much to us. We took advantage of a lack of data to undertake some much needed research and we kept extending our findings. For example, we worked with Motorola and Applied Materials to develop a retrofit remote NF₃ clean to replace the in situ C₂F₆ chamber cleaning process. This was a complement to the existing AMAT remote NF₃ clean process used on their 300 mm tools. NF₃ not only eliminates PFC emissions, but it also cleans faster. During the Motorola study the clean times were cut from 220 seconds using C₂F₆ to 78 seconds with NF₃ on the AMAT P5000 Dell chamber for a TEOS process. Since switching to this process, Motorola subsequently has also seen significant yield improvement and defect reductions due to a reduction in particles.

Dr. Johnson: At one point, Air Products had made about 90 percent of the PFC effluent characterization studies. That work resulted in the semiconductor industry using our analytical protocols for quantitatively determining emissions from semiconductor processes. It's important to us to be able to see our methods working. At the same time, we are benefiting Air Products and the environment.

Dr. Ridgeway: This has been a very good opportunity to work with our customers to help them reduce their emissions and improve their processes. Everyone in the group sees this as an opportunity, too, to use their skills to the utmost. It's gratifying to be able to say our work has resulted in some 40 publications, 10 U.S. patents and about a dozen SEMATECH technology transfers.


Question: Did the group's work ever produce any R&D "eureka" moments?

Dr. Maroulis: Not in the strictest sense. Our research followed a pretty predictable path and we kept building on what we had, starting out with model calculations on how various gas emissions affected the atmosphere. One of the most surprising things for us, and those we shared our results with, was discovering what byproducts were coming out of the early abatement systems. At times they were more detrimental to the environment than the PFCs used for cleaning. It was really gratifying to learn, too, the impact that our work could have on optimizing and cutting costs for customers, as well as benefiting the environment.

Dr. Johnson: Yes, it has been a "win-win" situation for everyone. We've reduced PFC emissions, lowered costs and improved process performance. Sometimes we tend to think environmental improvements will only cost us money. We've shown that isn't always true.


Question: Is there a limit to optimizing chamber cleaning and reducing emissions?

Dr. Johnson: While improvements can be made in both clean times and emissions from the unoptimized process, there is tradeoff when selecting the optimal conditions. We have covered a broad cross section of the world's installed tool base and have published recipes for some specific tools. We can dramatically reduce customer cleans and emissions recognizing that throughput may be the most important goal. Yet, for some customers, this tool is not the bottleneck. That allows them to do longer cleans, which further reduces their overall costs.


Question: Where does your group's research go from here?

Dr. Ridgeway: One of the things we are doing is moving to HAPs and VOCs, trying to characterize these emissions and then discover how we can reduce them, just as we did with PFCs.

Dr. Maroulis: There is definitely more work for our group to do to help customers further optimize chamber cleans. There are still many C₂F₆ chamber cleaning systems for us to evaluate. We are expanding our work with NF₃ chamber cleaning applications because NF₃ is the dominant chamber cleaning gas for 300 mm CVD processes. Currently, about 75 percent of fabs use C₂F₆ for cleaning, while about 25 percent clean with NF₃. The percentage of NF₃ cleans will significantly increase as additional 300 mm tools reach the marketplace.

Return to EPA Award page