Tackling Advanced Chip Manufacturing Challenges

Intel and PDF Solutions are deepening their partnership to address the growing complexity of semiconductor manufacturing at advanced nodes, according to a recent discussion between Intel CEO Lip-Bu Tan and PDF Solutions CEO John Kibarian at the Direct Connect Intel Foundry event in April.

During the presentation, Kibarian highlighted how the two companies have been collaborating for approximately four years to optimize chip design and manufacturing processes with a particular focus on yield improvement and accelerating time to market. “To maximize performance, profitability, and yields below seven nanometer, you really have to co-optimize,” Kibarian explained. “That co-optimization can’t begin after the fact. The R&D teams need to be working with the product portfolios in mind from the beginning.”

The partnership comes at a crucial time as Intel pushes forward with its foundry services and advanced manufacturing nodes, including the 18A and 14A processes. According to the discussion, these technologies present significant design challenges that require early engagement between chipmakers, designers, and manufacturing partners. Kibarian noted that PDF Solutions along with Intel “initiated a way of working with design information brought into the manufacturing flow, enabled by e-beam inspection and created new ways of looking at characterization data.”

One key innovation highlighted was Intel’s implementation of backside power delivery, a technological advancement that continues through its 14A node. This technology, along with advanced packaging solutions, requires designers to adapt their approaches and understand how to effectively leverage these capabilities.

“When new architectures come out, it usually takes a little while just to make your code work well on it,” Kibarian noted. “The same thing is when you introduce things like backside power. The innovation that Intel has really brought forward with 18A and extending with 14A and advanced packaging as well, it’s a lot of capabilities that designers need to know how to take advantage of.”

The collaboration focuses on leveraging data and machine learning to bridge the gap between design and manufacturing. PDF Solutions’ systems help build Process Design Kits (PDKs) and datasets that enable R&D engineers to work more closely with design and production teams.

“Electrical characterization data is the lingua franca in our industry,” Kibarian said, explaining how PDF’s analytics help connect different engineering teams that might otherwise use different technical languages and approaches.

Both Tan and Kibarian emphasized the importance of handling increasing complexity in semiconductor manufacturing, particularly regarding power management, temperature control and testing procedures. They noted that AI-driven analytics are becoming essential to manage multiple test insertions and complex production flows.

Kibarian added that “18A and 14A are designed with advanced packaging in mind and it is built into the PDK.  It is not a stand-alone die. That means that you have to start thinking about your test flows with multiple test insertions. Most of the engineering teams we see are using AI to manage that complexity. That means more data feed forward and more data analytics, an area where PDF is contributing and where we are partnering with Intel.”

The partnership also aims to create secure communication networks and data exchange systems that allow Intel to work more efficiently with both suppliers and customers. This collaborative approach is described as necessary for tackling the “unbelievably daunting” problems faced by semiconductor engineers.

As chip designs continue to shrink in size and increase in complexity, the executives stressed that early engagement and integrated analytics will be critical for success in future semiconductor development.

“Innovation drives benefit, but innovation needs the know-how to use the innovation,” Kibarian concluded, highlighting the need for continued investment in tools, characterization data, and engineering expertise to fully leverage advanced manufacturing technologies.

Intel’s Tan expressed enthusiasm for deepening the collaboration with PDF Solutions as both companies work to address the semiconductor industry’s most significant challenges. For more information, please visit pdf.com.