LEUVEN, Belgium — As imec develops the next level of integrated circuit chips, Cédric Rolin, the institute’s program manager for sustainable semiconductor technologies and systems, and members of his team are focusing on making the global chipmaking process more environmentally sustainable, he told EE Times. Imec is one of the world’s largest research centers for semiconductor technology.

“At imec, our bread and butter is developing the next generation of chips,” Rolin said. “We have this enormous ecosystem within the semiconductor industry. We are leveraging that ecosystem and our knowledge of the next generation of processes, bringing together everyone, and seeing how we can collectively look at every single aspect of this whole process. Then (we will) go and change not just the performance and the cost, but also that environmental aspect. It’s one additional element that we pull into our core program but, since it’s so important, it’s a separate program that we’ve started that we’re trying to get everyone to join.” 

The sustainability program is currently organized under two pillars: assess and improve, Rolin said. 

The first pillar is to assess the level of the environmental impact of integrated circuit chip manufacturing, such as CO₂ emissions and energy use. That assessment also includes identifying high-impact processes. 

The second pillar is to improve specific processes to reduce their environmental impacts. “We do that by conducting focused process development in the imec 300mm R&D Fab,” he said. 

In terms of climate change, a semiconductor fab has the highest impact through the generation of the electricity it consumes and through its direct emissions of greenhouse gasses to the atmosphere.  

“You want to decrease the electricity consumption of the fabs,” Rolin said. “To do that, you want to optimize those high energy consuming processes and you want to decrease the direct emissions from the fab, direct emissions of greenhouse gases, which are an array of fluorinated gases used for dry etching and for chamber cleaning, as well as the nitrous oxide heavily used in CVD (chemical vapor deposition).”

Imec is creating a generic model based on a 2020 IEDM paper that generated a lot of interest and triggered the birth of the new program, Rolin said. 

“The paper was a big success and imec’s partners said, ‘We want to see more of that—it’s important’,” he said. “That’s how the program was created.”

Today, partners include system companies Apple, Amazon and Microsoft, as well as suppliers ASM, ASML, KURITA, SCREEN and Tokyo Electron.

The first interested partners were the fabless companies because they have ambitious net zero targets for 2030 to 2050, Rolin said. Original equipment manufacturers also have shown interest because they want to reduce their overall environmental impact. 

More recently, foundries have joined the effort, and they are expected to add valuable insight.

“What we will expect from the foundries is to help us benchmark the numbers that we have,” he said. “We build models … from the bottom up … based on data collected from different sources. When you build something bottom up, there are always some parts that you miss. Whereas the foundries have the top-down view. They can basically look at the yearly production of their fab, look also from procurement at how much electricity, how much water, how much of all the resources have been consumed. So, they get a clear idea of what it takes to make a wafer in terms of resources.”

Imec’s bottom-up approach adds a complementary set of insight, Rolin said:

“What the foundries miss is that when you do a top-down analysis based on fab inventory, it must be done on an existing fab, and you cannot project that in the future. Also, you cannot drill down into the details of the actual impact of a process in their fab. That’s the strength of the bottom-up approach. The two are very complementary. You want the bottom-up approach to have this high granularity of results, to be able to drill down, to identify high impact and project in the future. You want to have the top-down approach for the benchmarking, to stick to reality. And by having foundry partners joining, the first exercise that we are going to do with them is going to benchmark our results.”

In the immediate future, imec is looking to expand its portfolio in the assess pillar to include radio frequency (RF), advanced packaging, 3D technologies and photonics, he said. With a broader focus, more chipmakers and more types of foundries may become interested in getting involved.

Under the improve pillar, imec is running several small projects in the fab, working to maintain a pipeline of such projects with a five-year goal of having active work on edge gases with a low global warming potential.

Finding qualified workers remains a challenge, Rolin said. 

“We need broad experience,” he added. “We need fab knowledge. We need life-cycle assessment. We need data science and electronics system knowledge. We need environmental science expertise. We are looking for people who can juggle all of that and move from one technology to the next to the next—and then to understand and be able to talk to all the actors. It’s difficult to find people. On the other hand, sustainability is appealing to a lot of people.”

By EETimes