Making the case for a new MEMS fab model

A new Japanese fab service MEMS Infinity, offering 150-mm and 200-mm wafers, aims to facilitate concept design and evaluation all the way through prototyping and mass production. It has joined hands with MEMS development services provider AMFitzgerald to offer a full-service MEMS silicon wafer foundry and expedite the commercialization of thin-film lead zirconate titanate (PZT) MEMS chip technologies.

Sumitomo Precision Products Co. Ltd (SPP)—a manufacturer of high-precision industrial products with three decades of experience in the MEMS industry—has set up MEMS Infinity with a 20,000 square-foot cleanroom in the industrial and technology hub of Amagasaki, Japan.

Figure 1 MEMS Infinity foundry service features PZT-specific patterning equipment and proprietary high-figure-of-merit epitaxial-PZT (epi-PZT) thin film deposition. Source: Sumitomo Precision Products

Sumitomo Precision Products has already been offering MEMS accelerometers and high-precision MEMS gyros under the name of an affiliated company. Now MEMS Infinity is one of the industry’s few foundries providing a portfolio of PZT thin films, piezoelectric materials that transduce electrical energy to mechanical energy and vice-versa.

These thin-film piezoelectric materials enable new architectures for challenging applications such as high-fidelity and low-power audio, low-cost handheld ultrasound imaging, and instant all-in-focus video. PZT is a versatile piezoelectric material that is sought after for MEMS sensors or actuators. It serves high-growth applications—such as true wireless stereo (TWS), automotive LiDAR, medical ultrasound imaging, AR/VR, and haptics—which require precise performance in a robust, ultra-miniaturized form factor.

“As a material that enables many types of emerging, performance-intensive MEMS devices, thin-film PZT is much in demand,” said Alissa Fitzgerald, founder and CEO of AMFitzgerald. “However, PZT requires specific process tools and expertise that are not widespread.” As a result, it’s been difficult to access commercial-quality material during development stages, especially when wafer volumes are low.

Moreover, using poorly controlled research-grade materials during prototyping slows down product development. The strategic alliance between MEMS Infinity and AMFitzgerald aims to address these challenges by integrating design-to-manufacture solutions while using high-quality PZT from day one.

EDN recently spoke with Fitzgerald to find out more details about thin-film piezoelectric materials and how they meet the needs of next-generation MEMS products.

A new generation of commercial MEMS technology

Sensorization is a mega trend continuing across all markets, and it began with our phones being loaded with sensors. At this technological crossroads, we see a new generation of commercial MEMS devices arising amid the manufacturability of thin-film piezoelectric materials, which have been around for decades in bulk format.

“They were used in many industrial and medical applications in solid bulk format, but now we have tools to deposit them on 1-micron or 2-micron thin films on 200 mm wafers,” Fitzgerald said. “These manufacturing tools are available now, so we are finally seeing piezoelectric MEMS products emerge in the market after more than 10 years of development in academia and elsewhere.”

Piezoelectric MEMS now offer performance and manufacturing gains over traditional capacitive MEMS. Some of the early products include speakers inside the earbuds to provide TWS experience and low-cost ultrasound transducers. “While products like ultrasound transducers have been made from bulk PZT for 40 to 50 years, now we can produce them with thin-film materials, enabling much lower cost transducers, leading to lower cost ultrasound diagnostic systems,” she added.

Figure 2 MEMS applications using PZT thin films include inkjets, autofocus, ultrasound, microphones/speakers, micromirrors, pumps, fluidics, and gyroscopes. Source: AMFitzgerald

It’s important to note that all the high-volume MEMS we see today are based on capacitive deep silicon etch architecture. However, with the advent and maturation of thin film piezoelectric, we see many devices redesigned in piezoelectric architecture. “Many companies are now seeking performance and manufacturing gains compared to the 1990s era traditional capacitive MEMS devices,” Fitzgerald said.

It’s also worth mentioning that while PZT is emerging as the most versatile material, aluminum nitride (AlN) was the first piezoelectric thin film to be commercialized in high volume for RF filters. Companies like Broadcom chose AlN because it’s best for MEMS-CMOS monolithic integration. “It enabled CMOS-MEMS integration on the same piece of silicon and the circuitry surrounding it, and that’s essential for RF filters,” she added.

“We are focused on PZT because it’s a more versatile material that can address many more applications and MEMS device types,” Fitzgerald said. “It spans both sensing and actuating functions and also enables wafer-level or chip-level system electronics integration.”

PZT’s technical merits aside, however, the critical issue is where fabless MEMS companies can access thin film PZT production, especially when the large MEMS fabs are very selective of their customers.

A new MEMS fab model

During the 1980s, the CMOS industry created the fabless business model by forming alliances with design partners to bridge the gap between fabless companies and foundries. However, MEMS foundries have yet to embrace the design partner model. Here, Fitzgerald pointed out that North America currently doesn’t have a MEMS fab for thin film PZT.

“We have been seeking PZT source for quite a while as our clients have been approaching us for this material,” she told EDN. “Large MEMS fabs don’t want to participate in the early-stage development because wafer volumes are small.” As a result, access to PZT has been limited, and there has been an unmet need for small- to medium-level production of high-quality thin films.

Efficient PZT MEMS development requires access to stable, high-performing PZT films, and R&D fabs don’t satisfy this need. The alliance will provide AMFitzgerald access to MEMS Infinity’s thin-film PZT. Furthermore, access to proprietary data on thin-film PZT will improve the accuracy of MEMS design and modeling, minimizing the need for design-build-test cycles.

“When you are developing a new MEMS device, you want to work with the exact material you are using in production,” said Fitzgerald. “So, we aim to lower the barrier for access to this material as well as have foundry-specific MEMS design.” That will inevitably speed up time to market and shrink development costs.

AMFitzgerald will get access to MEMS Infinity’s proprietary data, which will significantly enhance the design and modeling phases of MEMS development. “We can engineer the design specifically for their material, and that helps to reduce design and build test cycles,” she added. “MEMS devices take longer to develop because there are a lot of iterations between the design and foundry stages, and we can reduce that now to help unlock the potential of MEMS devices.”

Figure 3 The alliance will lower the barrier to PZT materials while facilitating design-to-production MEMS service. Source: AMFitzgerald

The alliance between AMFitzgerald and MEMS Infinity is based on a tried-and-tested business model that has been very successful in the CMOS industry, where TSMC is most notable for its design partners’ alliance. “CMOS foundries understood a long time ago that you need design partners to bridge the gap between a fabless company and foundry, but for some reason, this hasn’t taken hold in the MEMS industry yet,” Fitzgerald said. “We think we are going to be the first to have this kind of close collaboration with a foundry.”

Collaboration modus operandi

In MEMS design, prototyping is the third vital step after design and feasibility, and it tends to occur at a research facility. Custom chip designs are validated by prototypes built in-house before transfer to the foundry for production. That’s followed by foundry selection, and then the design must be adjusted for the foundry toolsets. Next, you must move on to ramping up production.

This alliance between AMFitzgerald and MEMS Infinity seeks to provide a shortcut for PZT to facilitate design on the foundry’s material properties and qualities. Moreover, prototyping will occur on the production-ready material from day one at the foundry. In other words, the tie-up aims to shorten the commercialization timeline of MEMS devices.

The two MEMS partners aim to provide the complete PZT development solution. AMFitzgerald will offer MEMS design, modeling and product development services, while MEMS Infinity will provide production for thin films and versatile foundry services. MEMS Infinity will provide production proven PZT processes at its fab.

AMFitzgerald will facilitate global access to MEMS Infinity’s thin-film PZT while using its proprietary thin-film PZT data to improve the accuracy of MEMS design and modelling. That, in turn, will minimize the need for design-build-test cycles and accelerate product development. Consequently, it will unlock the commercial potential of performance-intensive piezoelectric MEMS devices such as ultrasound transceivers, micro-speakers, micromirrors, and microfluidics.

“We aim to provide integrated product development experience, which has been missing in the MEMS industry,” Fitzgerald concluded. “We want to facilitate MEMS Infinity’s high-quality material, which we used for 20 years to produce piezoelectric gyroscopes.”

<!–
googletag.cmd.push(function() { googletag.display(‘div-gpt-ad-native’); });
–>

The post Making the case for a new MEMS fab model appeared first on EDN.

Become a member

Become a member

Become a subscriber

Become a sponsor