Is 3D printing a solution to enhance supply chain resilience for such a core component of the modern tech world?
Conventional semiconductor manufacturing processes limit designers in terms of interconnect architecture, planarity, and substrate shape. In contrast, 3D printed circuit boards are not limited by subtractive manufacturing limitations, saving semiconductor companies a lot of time, effort, and money, and providing designers with increased freedom to design circuit boards with sophisticated architecture and customised designs as required.
Working with a system that is designed for 3D printing circuit boards is an excellent way to complement an existing semiconductor manufacturing process for low-volume, high-complexity boards. The layer-by-layer printing process allows low-volume manufacturing runs of boards with the desired level of complexity, including non-planar circuit boards and high-value boards with very complex shapes.
Why 3D Print Components?
Wafer table thermal management
Better thermal management of critical semiconductor equipment components, such as wafer tables, can improve semiconductor equipment accuracy by 1–2 nm and simultaneously improve speed and throughput. An increased machine speed and uptime leads to more wafers processed and higher overall lifecycle value.
During lithography, keeping temperatures within milliKelvin (mK) ranges is critical as any system disturbance has an impact. Through design for additive manufacturing (DfAM), it’s possible to optimise internal cooling channels and surface patterns, thus dramatically improving surface temperatures and thermal gradients while reducing time constants. A large semiconductor capital equipment manufacturer using AM to produce their wafer tables realised an 83% decrease in ΔT (13.8 to 2.3 mK), and a 5-time reduction in time to wafer stabilisation.
Another benefit of using AM to produce wafer tables is structural optimisation and tables with reduced part counts and assemblies. Producing parts using traditional technologies relies on brazing to join parts together, which is a lengthy, low-yield process with a 50% rejection rate. Replacing multipart assemblies with monolithic additively manufactured parts increases reliability, improves manufacturing yield and reduces labor costs.
Manifold fluid flow optimisation
Using traditional manufacturing processes to produce complex fluid manifolds results in large, heavy parts that have non-optimal fluid flow due to abrupt transitions between components, and channels with sharp angles that lead to disturbance, pressure drops, stagnant zones and leakage.
When AM is employed to produce these same manifolds, engineers can optimise their designs to reduce pressure drop, mechanical disturbances and vibration. A 90% reduction in flow-induced disturbance forces reduces system vibration and realises a 1–2 nm accuracy improvement.
Structural optimisation and advanced flexures
AM gives designers the flexibility to optimise the structural topology of your part (i.e. lightweighting) with a suite of high-strength metal alloys. These designs can more precisely meet the performance requirements of semiconductor capital equipment, improve the strength-to-weight ratio and deliver a faster time to market. Lightweighting semiconductor components and advanced motion mechanisms reduces inertia and improves lithography and wafer processing machine speed and uptime, leading to more wafers processed. In one example, a semiconductor capital equipment manufacturer was able to employ AM to achieve greater than 50% weight reduction in flexures, 23% higher resonant frequency and reduced system vibration.
A likely scenario is that AM will significantly enable newer machines that are either shipping today or will be shipping in the next 1-2 years. With this runway, there is ample time for component and system level redesigns, which will increase productivity and quality. Additionally, the manufacturers will still have enough control over those systems to rigorously test and prove performance gains.
However, while opportunities are indeed emerging, it’s not necessarily a new market for additive nor 3D Systems. At the 3D printing pioneer’s Leuven office in Belgium, major semiconductor equipment manufacturers are said to have been leveraging its direct metal printing for well over a decade. What began as a “secret metal printer” used to print parts as a service has matured to what Scott Green, Principal Solutions Leader at 3D Systems described as “a couple of hundred” successful production projects.
“There’s maybe ten areas in semiconductor capital equipment where we’re contributing regularly,” says Green, citing opportunities in lithography, wafer handling and metrology. Green also pointed to examples of recent large-format EUV (Extreme ultraviolet lithography) machines which can contain well over 100,000 parts.
“The needs and challenges of the semiconductor fabrication industry today are directly aligned with what a direct metal solution offers,” Green tells TCT magazine. “They have challenges where, in order to really push the limits of physics, you’ve got to totally eliminate uncertainty and noise inside of a system and really optimise all the parts of handling, cooling, fluid distribution, light collimation. It’s a very complex machine.”
The design freedoms and part consolidation afforded by additive could offer a solution for parts like heat exchangers, gas manifolds and nozzles. Instead of having tens of components vibrating against each other in an assembly, you could potentially reduce the number of moving parts and links in your supply chain down to one.
The Challenges of AM Adoption Moving Forward
3D printing electronic components is not without its hangups, however. In an interview with TCT Magazine, VELO3D CEO Benny Buller explains that AM is best used for replacing existing parts—not redesigning a system altogether.
“When you are doing legacy parts that you are already producing in one way and just want an identical replacement by additive, the barrier for qualification is much lower,” he said. “But when you start having to redesign the system or the assembly so that you can manufacture, well that’s not fine, because now you’re driving yourself into a lot of risk.”
Buller also notes that AM struggles to deliver the cleanliness and surface control one would find in the cleanroom of a traditional fab. At each layer of the semiconductor fabrication process, wafers are expected to be free of particles that are nanometers in size. This attention-to-detail cannot be replicated in an at-home or in-office 3D-printing environment.
When dealing with the precise chemistries, gases and temperatures expected by the semiconductor industry, those risks simply cannot be afforded. Those same complexities, however, Buller believes suit the capabilities of additive well.
“These are the classical problems additive manufacturing is really good at,” Buller explains. “Control of heat, control of flow, whether it’s flow in gases, form of chemicals, whether it’s forming liquid flow, these are the places where additive manufacturing is really powerful.”
One crucial area where AM does present a challenge, however, is cleanliness, a field Buller is familiar with having spent the early years of his career on the inspection side of the semiconductor space.
“Additive manufacturing, compared to some other manufacturing technologies, has struggled delivering this level of surface cleanliness and this level of surface control,” Buller says of the intense cleanliness levels required at each layer on the semiconductor fabrication process. “When we are doing gas turbines or jet engines, they also care about surface finish but we are talking literally orders of magnitude difference … [The semiconductor industry] cares about particles that are two nanometers in size. It’s a completely different level of cleanliness that they have to deal with.”
Current opportunities for AM lie primarily in semiconductor capital equipment. It’s “the ultimate high volume manufacturing technology” according to Buller, with billions of parts produced every month, but per a recent report in the Harvard Business Review, funding and building out a new semiconductor fab can take at least five years. AM could offer a solution.
“Additive manufacturing has a lot of value to this industry, both in the ability to make better processes and to make equipment that is capable of more uniform, more controllable processes, new ways to make things that were not possible before,” Buller says. “It allows for a more agile supply chain and it helps with shorter lead times.”
There are however also specific opportunities in semiconductor devices themselves as Valentin Storz, General Manager of EMEA at Nano Dimension told TCT. Nano Dimension, a manufacturer of additive electronics systems, known for its DragonFly LDM technology which simultaneously deposits a dielectric polymer and nano-silver for circuitry, is said to operate between the worlds of PCB and semiconductor integrated circuits.
Storz says: “The whole story about IoT, Industry 4.0; everything will have an IP address and communicate. That means every part will become at some place connected and needs some circuitry, some antenna in it and with parts getting smaller and having new form factors, that’s a place for us.”
New opportunities, Storz offers, are those in 3D stacking of chips on top of each other or heterogeneous integration where different components such as circuitry, RF components, optics and potentially even cooling channels are integrated into one package.
Throughout these conversations, Moore’s Law, the notion that the number of transistors on a microchip doubles about every two years, was a common thread. While the trend appears to be flattening in the semiconductor space, innovation continues apace as manufacturers strive to add more complexity to smaller chips and demand for new devices flourishes. It’s here, looking at that five-year roadmap towards next-generation semiconductor fabrication, better geometries and more uniform processes, where AM could find its sweet spot.
“Additive manufacturing allows [manufacturers] to innovate in directions that they couldn’t innovate before,” Buller concludes. “The moment this is demonstrated, that you can get to the cleanliness and you can get to the manufacturing quality that is required to support that, this will be a floodgate.”