'AI-human approach' reduces chip development costs by 50 percent, according to a Lam Research study in Nature.
FREMONT, CA: "New approaches in innovation are needed to enable the industry to scale fast enough to meet the data-driven world's evolving demand for next-generation chips," said Tim Archer, president, and chief executive officer at Lam Research. "The opportunity for greater collaboration between talented engineers and machines in process engineering highlighted in Lam's study in Nature is a potential game-changer for our customers and our industry at large. This research is a testament to Lam's more than 40-year heritage of industry leadership and semiconductor manufacturing innovation. I congratulate the Lam team on this exciting work."
In a recent study, Lam Research investigated the potential for artificial intelligence (AI) to be used in chip manufacturing process development, which is currently a human-driven process that is necessary for the mass manufacture of every new advanced semiconductor in the world. The study, recently published in the journal Nature, identifies an opportunity to address two major challenges facing the industry: lowering development costs and quickening the pace of innovation to meet the rising demand for next-generation chips. Experts predict the semiconductor market will reach $1 trillion in annual revenue by 20301. According to the study, a "human first, computer last" strategy can achieve process engineering goals significantly faster and for half the price of the current strategy.
Stay ahead of the industry with exclusive feature stories on the top companies, expert insights and the latest news delivered straight to your inbox. Subscribe today.
The complexity of upcoming chips continues to make process development more difficult and expensive. Researchers at Lam pitted skilled process engineers against AI-enabled computer algorithms to find a more effective strategy.
Experienced and talented engineers must first develop a customized recipe that details the precise parameters and permutations required for each process step to construct each chip or transistor that has been built. The construction of these nanometer-sized devices on a silicon wafer involves hundreds of steps. On several occasions, small layers of materials are routinely deposited onto silicon wafers, and the surplus material is then precisely removed via atomic-scale etching. Currently, human engineers carry out this crucial stage of semiconductor development, mostly relying on their intuition and a "trial and error" method. Process development can be time-consuming, expensive, and labor-intensive, with every recipe being specific to the chip design and more than 100 trillion alternatives to consider. As a result, it takes longer and longer to develop new technologies.
A targeted process development recipe was created at the lowest cost based on variables related to test batches, metrology, and overhead costs in the Lam study. In the study, researchers found that humans still excel at dealing with difficult, unconventional challenges, even when using a hybrid human-first, computer-last approach.
"Although critical to the creation of each and every chip produced, the plasma physics of process engineering has been for decades rooted in the same scientific approach that Thomas Edison used: trial and error," said Rick Gottscho, executive vice president and strategic advisor to the CEO – Innovation Ecosystem at Lam Research and co-author of the study. "Our research showed that while engineering talent remains essential to innovation, process engineering costs can be reduced by 50 percent by integrating AI at the right stage and with the right data. The study provides a prescriptive approach for bringing together the best of human-led engineering and what data science and machines offer to create a combination that performs better than one alone. If realized, this hybrid approach can lead to significant savings in both dollars and engineering time for the industry."
Lam is currently incorporating the study's primary findings into its development processes. The Lam study offers preliminary recommendations on effectively combining human expertise, talent, and experience with AI's capacity to quickly evaluate a wide range of potential combinations in process engineering.
"By complementing engineering expertise with AI using the human first, computer last approach, the tedious and laborious aspects of design are alleviated for engineers, freeing them up to focus on the creative areas of development and explore innovations that may have been out of reach either due to bandwidth or cost," said Keren Kanarik, technical managing director of Lam Research, lead author of the research paper and a former process engineer. "While the application of AI in process engineering is still in its infancy and human expertise and domain knowledge is essential for the foreseeable future, the results point us to a path to foundationally change the way processes are developed for manufacturing chips."
