Flexible 300-mm MEMS manufacturing is bridging critical gaps and paving the way for broader adoption and integration.
The MEMS and photonics landscapes have undergone a fundamental shift during the past two decades. At the start of the millennium, PICs were the domain of university labs and government grants; a community of scientists and innovators pioneered the R&D necessary to develop device prototypes that seemed to be the core components of the future. Teams worked to produce building blocks — low-loss silicon nitride waveguides, stable couplers, ring modulators that did not drift, and germanium photodetectors — that could be produced with any degree of consistency.
Courtesy of Rogue Valley Microdevices.
MEMS groups were performing similar work with optimized high-aspect-ratio mechanisms and micromirrors that pushed every limit of release and stress control. It was an imaginative and enthralling time. But it was difficult; “wins” were at the subcomponent level, and yield was, let’s just say, in need of improvement.
Today, the dynamic has shifted, and the excitement is palpable. Industry events, including the IEEE Electronic Components and Technology Conference, spotlight the convergence of advanced packaging-, PICs-, and MEMS-derived microfabrication techniques. This convergence is no longer sidebar research — it is the main event. The most valuable companies in the world are integrating PICs directly into their road maps. Integrated photonics, MEMS technology, and their converging microfabrication workflows now headline major AI, data center, and packaging conferences.
The MEMS community’s leap to 300-mm wafer platforms follows the broader semiconductor industry’s transition to 300-mm wafer production. Courtesy of Rogue Valley Microdevices.
Amid this shifting dynamic, industry players and applied researchers are no longer attempting to determine whether the underlying physics and fabrication yield models support functional chips. Now, the race is on to meet the concrete product demands of hyperscalers and AI giants.
Put simply, the sector has graduated from exploring the scientifically feasible to engineering the commercially scalable.
Bridging the manufacturing gap
MEMS and integrated photonics have become the core technologies driving modern sensing, communication, and data processing. These technologies and the systems that they enable are found everywhere, from the sensors in our phones to lidar scanners and optical transceivers.
Now, the evolution of this trend requires major upscaling: The “killer applications” of the next decade, in areas such as AI and the Internet of Things, cannot function without a manufacturing breakthrough. To meet the demands of these applications, PICs and sensors must be manufactured not in the thousands, but in the millions, with semiconductor-grade reliability. Further, progress depends on producing dense, uniform thin films, maintaining precise dimensions during lithography and etching, and combining materials in practical ways that were previously processed separately. These technical limitations already keep many promising devices from reaching stable, high-volume production.
The semiconductor logic and memory industries faced similar scalability challenges in earlier decades. They responded successfully by refining materials control, expanding process automation, and implementing comprehensive metrology. MEMS and photonics stand in this same position, with the opportunity to mature through better process integration and access to 300-mm tool sets that bring higher uniformity and yield.
But a significant disparity remains: While the broader semiconductor industry has standardized 300-mm wafer platforms, much of the MEMS and photonics community continues to operate on smaller substrates using legacy tools. Bridging the gap, both in equipment and process maturity, will determine how quickly these technologies can scale to meet the demanding product road maps of the 2020s and beyond.
The foundation: Thin films, etch control
The performance of both MEMS and PICs begins with the films that define them. Factors such as optical loss, mechanical stress, and electrical isolation all depend on how these films are deposited and patterned. Stoichiometric low-pressure chemical vapor deposition silicon nitride, high-quality silicon dioxide, and plasma-enhanced chemical vapor deposition oxide are common to both MEMS and photonics. In photonics, these films serve as low-loss waveguides or reflective claddings. In MEMS, they provide structural rigidity and electrical isolation. In both cases, film density, quality, and uniformity are essential to the performance and yield of fabricated devices.
In the photonics industry, low optical loss depends on films with smooth interfaces, low wafer total thickness variation, and well-controlled refractive indices. Wafer inspection can discern critical imperfections in a wafer that may influence performance. Courtesy of Rogue Valley Microdevices.
The team at Rogue Valley Microdevices shows a 300-mm wafer. MEMS devices will be manufactured on 300-mm wafers at the company’s newly opened fab on Florida’s Space Coast. Courtesy of Rogue Valley Microdevices.
Another process that photonics and MEMS share — etching — is one of the most critical for achieving high-quality fabrication. Both photonics and MEMS fabrication processes rely on precise silicon and dielectric etches to shape structural, electrical, and optical features. Each approach demands tight control of sidewall quality, surface roughness, and dimensional accuracy. As photonic coupler designs continue to improve and coupling losses drop to <1 dB, manufacturers are introducing more complex material stacks and multistep etch sequences into production. Keeping these processes uniform across a wafer remains one of the central challenges in scaling both MEMS and photonic manufacturing.
Shared platforms and packaging demands
Certain MEMS and PIC devices share a fabrication platform rooted in silicon-on-insulator wafers. The top device layer provides dimensional precision, the buried oxide offers electrical and thermal isolation, and the handle wafer supports alignment and through-silicon structures. Because both mechanical and optical devices already use this material stack, manufacturers can apply process steps such as bonding, chemical-mechanical planarization, and etch release across both types of production with a similar foundation.
When the chips are complete, both MEMS and PIC modules rely on advanced packaging to become part of larger systems. MEMS devices often need hermetic sealing, electrical interconnects, and mechanical protection. PICs require precise optical alignment, electrical interconnects, stable thermal interfaces, and low-reflection surfaces.
As integration continues to grow, packaging has become one of the largest contributors to total cost — as well as one of the hardest problems to scale. The assembly methods developed for electronics do not always meet the optical and mechanical tolerances required for MEMS or photonic devices. Automated alignment, wafer-level bonding, and low-outgassing materials are part of the solution, but their use remains inconsistent across the industry.
Fortunately, compatibility between MEMS, photonics, and electronic packaging workflows will improve naturally as advanced packaging facilities continue to expand into 300-mm production. This in turn will create a clearer path toward high-volume manufacturing.
Maturity gaps in MEMS manufacturing
MEMS production has reached high reliability in categories such as accelerometers, microphones, and pressure sensors. But it remains fragmented in many others. Each new design often brings its own materials and process flow(s), and scaling from prototype to production can require tool recipe changes and process tuning for each device family. In some cases, for example, etch sidewalls are critical. In others, they may not matter. For many experienced MEMS designers, there is a natural hesitancy to design for a specific tool when it is not within the ecosystem of the largest tool vendors. This is often unavoidable. Still, movement toward common 300-mm tools is a critical step toward maturity.
Yield, as is true of performance, is a metric that depends heavily on film quality and etch precision. Small variations in layer thickness or residual stress can shift a resonant frequency, alter a radio frequency spectrum, or lead to stiction during release. For complex MEMS structures, wafer-level bonding and alignment uniformity add even more variables. As a result, process control and efficient metrology are critical. The need for tighter process control continues to grow as MEMS moves into larger-scale and more demanding applications such as adaptive optics, biomedical systems, and microfluidics.
Maturity gaps in photonics manufacturing
Photonics faces a different but equally challenging set of process limits. Achieving low optical loss depends on films with smooth interfaces, low wafer total thickness variation, and well-controlled refractive indices. A few nanometers of variation can change a waveguide’s mode or alter coupling efficiency. These sensitivities make process drift unacceptable. Yet they are difficult to track with inspection tools designed for electronic wafers.
Alignment between fibers and chips remains a bottleneck as well. As optical input and output counts rise, photonic devices must integrate increasingly larger arrays of fibers or lenses that couple directly to on-chip structures. However, assembly is still mostly manual or semiautomated. To reach higher volumes, the industry will need wafer-level alignment and bonding that includes methods used in advanced MEMS packaging.
Finally, thermal stability is still a major limitation to scaling. Many photonic components rely on active thermal tuning to maintain wavelength accuracy, which increases power use and constrains integration density. More uniform silicon nitride and oxide films, combined with better stress control, can improve stability at the material level and reduce reliance on power-hungry compensation later in the system.
Where the fields converge
MEMS and photonics intersect in specific, practical ways. Many PICs now include small mechanical elements that adjust or redirect light using minimal power. Optical switches, micromirrors, and directional couplers built in this way use electrostatically driven motion rather than heat-driven expansion, significantly reducing power draw and enabling faster operation.
The fabrication steps that make these devices possible come from MEMS manufacturing. Sacrificial layer etching, wafer bonding, hard masks, and mechanical release are established techniques that translate well into photonic process flows. Their adoption in photonics is less about merging two fields and more about applying mature, reliable microfabrication and assembly methods to improve device performance. The overlap is still limited, but it is growing as both industries seek new ways to build efficiency and precision into optical systems.
Closing the manufacturing gaps
Improvements in three fundamental areas — material uniformity, precise etching, and reliable packaging — have direct throughlines to progress in both MEMS and photonics.
As it relates to uniform materials, legacy equipment is a persistent bottleneck. Many foundries still run equipment built for 100-, 150-, or 200-mm wafers. Moving to state-of-the-art 300-mm tools within a broader, modern ecosystem improves thermal control, gas flow, and chemical stability. These improvements directly reduce variation in film thickness and stress and lead to better yield and repeatability.
Etching, which defines the core structures that determine device behavior, requires tight depth and profile control — which are essential qualities across both photonics and MEMS fabrication. In MEMS, etching governs parameters such as release and dimensional uniformity. In photonics, it sets optical mode shape and coupling efficiency resulting from waveguide and component definition.
Whether deployed to optimize waveguides and optical components in photonics or MEMS mechanisms and fiber couplers, etch systems can be tailored to the distinct requirements of a given manufacturer. For example, etch systems optimized for larger wafers improve depth accuracy and sidewall consistency across an entire lot, leading to higher device performance, better uniformity, and improved overall yield.
Rogue Valley Microdevices’ 300-mm-capable MEMS fab in Florida. Courtesy of Rogue Valley Microdevices.
Finally, bringing optical, electrical, and mechanical interfaces together at wafer scale remains one of the hardest parts of manufacturing. The bonding and encapsulation protocols used in MEMS manufacturing can improve the reliability of photonic assemblies, while advancements in photonic alignment and metrology can improve MEMS accuracy. Both benefit from stronger collaboration between process and packaging engineers.
The industry already recognizes effective manufacturing practices. The work that remains is to extend that quality through the supply chain with better metrology, process control, and shared infrastructure that can support production at scale.
Scaling as an enabler, not an end
Moving to 300-mm manufacturing brings the process stability and uniformity needed for higher yield and lower cost. Scale matters not only for throughput but also for compatibility with the broader semiconductor ecosystem, where most advanced tools and materials are already optimized for 300-mm operation.
Because of this maturity, processing larger wafers within the ecosystem of 300-mm tools often makes it easier to control film thickness, temperature, and uniformity. This creates a more consistent foundation for both MEMS and photonics production.
Logic and memory fabs reached their maturity through scale and process discipline, and, as mentioned, MEMS and photonics are advancing along a comparable path. The challenges are different, shaped by a wider range of materials and process flows. New fabrication methods will arise, but opportunity lies in meeting the scale of the long-established semiconductor industry.
From road maps to reality
The obstacles slowing MEMS and photonics today almost certainly feel familiar to anyone who has spent time in a fab. These small but stubborn issues have followed these fields from their research roots into commercial production: films that drift across a wafer, etches that behave differently from run to run, and process steps that refuse to scale without constant attention.
After watching these technologies evolve for decades, it is clear that the limiting factors are not inventions we have yet to discover but the manufacturing discipline we must continue to strengthen. The solutions are known. The question is how consistently we apply them.
Larger wafers and more capable tools will unquestionably help to carry momentum forward. The real progress, however, will come from a foundation that behaves the same way every day. MEMS and photonics will be able to scale with the same confidence that reshaped logic and memory — when materials deposit with the stability designers require, when etch profiles no longer raise surprises, and when the path from design to production feels reliable instead of experimental.
With such a foundation in place, these technologies can support the next generation of sensing, communication, and compute systems that are already shaping commercial markets.
Meet the author
Jessica Gomez is founder and CEO of Rogue Valley Microdevices, which operates MEMS foundries in Oregon and Florida. She was named the 2025 Electronics Entrepreneur of the Year by Electronics Weekly; email: jgomez@roguevalleymicro.com.
-Photonics Spectra