Innovators bridge the gap from research to manufacturing.
Their work could spring from the minds of sci-fi writers Octavia Butler or Isaac Asimov, but everything done at R&D institutes focused on electronics engineering is real.
The smattering of innovate R&D institutes EE Times surveyed this fall are compressing the two-year timeline of Moore’s Law to six months: they are lightweighting aircraft by replacing electrical networks with photonics; printing electronics on the wings of drones; amassing data to determine the usable life left in circuit boards; and training robot assemblers for electro-optic devices.
“We’re trying to advance the state of manufacturing of integrated photonics in the United States,” said Nicholas Fahrenkopf, photonics engineering manager at AIM Photonics. “And the biggest way we do that is by bridging that ‘Valley of Death’ between the R&D you get out of academic researchers into high-volume manufacturing.”
Domestically, AIM is part of a network of Manufacturing USA institutes created by the Obama Administration to advance American industry. Among the manufacturing institutes are the ARM Institute for robotics, NextFlex for flexible hybrid electronics (FHEs), and the REMADE Institute that is not only tasked with keeping everything the other institutes make out of landfills, but also hastening a circular economy.
Globally, Imec (Interuniversity Microelectronics Centre) is the cartographer for the semiconductor scaling roadmap. Imec also applies its expertise in advanced semiconductor technology to address the major challenges in climate, health care, mobility, energy, agro-food, and other areas.
“Computing-power needs are exploding due to the rapid rise of digital applications and data processing,” said Katrien Marent, EVP and chief marketing and communications officer at Imec. “With the growing use of artificial intelligence to tackle the major challenges of our time, like climate change or food shortages, the computing need is expected to double every six months from now on. To handle the exponentially growing amounts of data in a sustainable way, we need improved high-performance semiconductor technology.”
Marent and her colleagues used their cartography tools to propose a roadmap extension for their bridge to manufacturing, where they’re tossing over the nanometer for the so-called Ångström Age. Landmarks along the way include high-numerical aperture (NA) extreme ultraviolet (EUV) lithography, new transistor architectures, 3D systems-on-chip (SoC) integration, and a new system-technology co-optimization program and sustainability research effort.
Additive manufacturing coming into play for R&D institutes
In addition to R&D, the Manufacturing USA institutes are fostering a national ecosystem for their particular niche with members and funding both public and private—and developing homegrown talent to do the work.
Regarding talent, Fahrenkopf could envision electronics engineers making the leap into photonics.
“There are a lot of skills that electronics engineers have that are easily transferable to photonics,” he said. “We use the same software when we design chips, we use a lot of the same software when we’re doing simulations. The learning curve is not that steep.”
There is one basic difference: Electronics engineers focus on making things smaller to add more transistors, and, he said, “It’s a completely different paradigm in photonics: You can’t just make a waveguide half as big. That’s not how light works, not how it propagates. So, you have to cram more functionality in differently.”
Cramming in functionality differently happens all the time at NextFlex, where a complete circuit board is 3D printed—an additive manufacturing (AM) technology.
“We think we can dramatically change the look of printed circuit board manufacturing in the U.S.,” said Malcolm Thompson, NextFlex’s executive director. “If you look at the old technology, we replaced six steps by one—and inevitably at lower cost. It turns out to be far less waste because we don’t add materials or create effluent. And so, it’s environmentally friendly and it’s a faster time to market.”
In addition to printing circuits on aircraft and drone wings, the flexible substrates NextFlex and its members have produced or are working on have AM electronics. These are applied to advanced packaging, structural electronics in and on 3D surfaces, and FHEs in mainstream applications like medical diagnostics, wearables, asset monitoring, aerospace, and defense.
Robots get new roles
At the ARM (Advanced Robotics for Manufacturing) Institute, a lot of the work is about enhancing robots with AI so they can perform more manufacturing tasks, such as assembly of electro-optics. To de-risk the adoption of robotics for new and old roles, the institute, as part of the Southwestern Pennsylvania New Economy Collaborative, won a Build Back Better Regional Challenge grant to set up a de-risking center to de-risk the adoption of robotics.
“Large companies, they have the room and facilities to do de-risking,” said Arnie Kravitz, ARM’s chief innovation officer. “A small- or medium-size company that really can’t afford its own de-risking center can come to the ARM Institute with a robotic project that it wants to bring to the factory floor, set it up, run it, get advice, get feedback, enhance it, and do the transition-to-manufacturing phase.”
While all that’s going on, ARM is also collecting data that can then be used with AI and machine learning in training people to use the information to improve manufacturing.
Initially, the de-risking center is available for factories in an 11-county region in southwestern Pennsylvania, where the ARM Institute is located. Kravitz noted, however, “If the model goes well, it’ll spread,” she said.
REMADE revisits electronics’ lifespans
While ARM wants robots to handle assembly in a factory, REMADE sees them as a tool for dis-assembly of electric vehicle batteries and inspection of used circuit boards. REMADE stands for “reducing embodied energy and decreasing emissions.”
“When [circuit boards] come back, they need to be inspected to see what’s wrong with them to see if they can be remanufactured or repaired to be put back into service,” said Mike Thurston, REMADE’s remanufacturing node leader. “Some of the ways they fail are too hard to detect. We’re using vision-based systems and AI to assess the condition of a circuit board and decide if it’s a good candidate for remanufacturing.”
REMADE is also trying to squeeze down to minutes from the hours it currently takes to test battery modules to see if they can be reused, recycled, or remanufactured.
“The time it takes to test modules to see if they’re OK for reuse today is fairly long,” he said, noting that a member company is “looking to reduce that time so they can be more efficiently tested, which will improve the economics around the reuse process.”
This article was originally published on EE Times.