MEMS packaging has traditionally been especially challenging because nearly every process and device require a unique approach. We recently spoke with Dr. Ken Gilleo, President of ET-Trends, a consulting and intellectual property firm focused on emerging technologies and device packaging.
MEMS Investor Journal: How has MEMS packaging evolved over time and are we getting closer to having industry standards?
Dr. Ken Gilleo: MEMS is certainly the most difficult device for packaging standardization and the issue will only grow as advanced MEMS evolves. MEMS is a cluster of technologies that is still embryonic even though jetting chips and accelerometers are somewhat mature. Combining electronics with mechanics and optics creates unlimited possibilities. That said, the high-volume areas, like motion sensors, are converging toward similar packages where standards are practical. For example, wafer-level capping, followed by conventional over molding, is now the de facto standard for accelerometers. This high-volume area, can and should, move to standardization. But there will only be “islands” of standardization at best.
Emerging MEMS devices are going to require device-specific packaging far into the foreseeable future. The good news is that Wafer-Level Packaging (WLP) can make it possible for the device to become the package, and this could be the ultimate solution.
MEMS Investor Journal: What are the main approaches for MEMS packaging today?
Dr. Ken Gilleo: Hermetic ceramic cavity packages; classical high-reliability types, but cost-reduced using LTCC (Low Temperature Co-fired Ceramic) materials and processes.
Near-Hermetic Packaging (NHP); high performance plastic cavity packages that fall somewhere between full-hermetic ceramic and non-hermetic plastic. This is a new area where thermoplastics, not the traditional thermoset materials, are being injection molded. They’re “pre-molded” packages, the polymer equivalent to the ceramic cavity.
Wafer-Level Packaging where most, but preferably the entire package, is formed while the MEMS devices are still in wafer format. This is a rapidly evolving area but only wafer capping has been widely commercialized.
MEMS Investor Journal: Which companies do you consider as having the best expertise with MEMS packaging?
Dr. Ken Gilleo: Analog Devices Inc. (ADI) pioneered many elements in MEMS packaging as they commercialized air bag accelerometers; anti-stiction treatments/in-package additives and wafer-level capping processes.
Texas Instruments pioneered and continues to innovate optical-MEMS (MOEMS) packages as the leading producer of digital projector chips; this includes use of getters and efficient glass lid sealing.
Other leaders are Motorola, ST Microdevices, and Bosch - all important MEMS producers. But start-ups, like Akustica, have also done innovative, but device-specific packaging in order to launch products. There are also a few packaging companies that deserve mention, such as Quantum Leap Packaging (QLP), who is promoting injection-molded packaging. Shellcase (Tessera) also bears watching.
MEMS Investor Journal: Are there start-ups or newly commercialized techniques which you find especially interesting?
Dr. Ken Gilleo: As a polymer specialist, I’m unpleasantly surprised that the packaging industry has mostly ignored thermoplastics that are the dominant class of plastics and also have the best properties; superior to the traditional epoxies that were invented about 80-years ago and still the basis for electronic plastic packaging. Hopefully, QLP, RJR Polymer, and ET-Trends, will succeed with low-cost, highly versatile, injection molded packages made with excellent plastics like LCP. Injection molding (IM), one of the most widely used plastic shaping processes, can produce very precise, 3D structures, at low cost, and this is what we are going to need for advanced MEMS, especially fluidic devices. Ironically, the problem may be that the present level of commercial MEMS is too simple to take advantage of IM that can deliver complex structures just as easily as simple ones.
Wafer-Level Packaging is an ideal method for MEMS since it allows fragile, contamination-sensitive device elements to be protected in the ultra-clean environment of the fab. Although the cap-only “pre-package” is in high-volume today, the end game is full WLP - this will streamline and cost-reduce the total process. MEMS processes can be used to make the package especial those with fluidic couplings. WLP patents are issuing from dozens of companies around the world.
Shellcase, recently acquired by CSP-pioneer Tessera, has been developing tiny packages for MEMS and optical chips for several years. The acquisition should boost the rate and scope of development in the future.
MEMS Investor Journal: Packaging of BioMEMS and biosensors has been especially difficult because of bio-fouling and bio-contamination. What are the latest advances in this area and how are the traditional challenges being solved?
Dr. Ken Gilleo: Again, it’s surprising that thermoplastics, with a long and excellent record in biomedical applications, have been mostly ignored. Note that a major break-through in stents was to treat them with drug-bearing polymers. We now use drug-bearing and biocompatible plastics for implantable devices, so why not the same concept for packaging that will contact biological materials, or even be implanted? The human body is mostly polymers, so selecting polymers makes sense for bio-packages. And while the bio-MEMS package might be molded with a common engineering material, like LCP, additives can be used or coatings applied to produce just the right surface. What’s more, micro-channels and ports can be formed by automatic injection molding, making this an ideal technology for Bio-MEMS and other advanced devices. This is so obvious that someone must be working on it.
We may also see the development of “pluggable” packages made by MEMS processes at wafer-level. This seems plausible for bio-MEMS sensors, implants, and analytical products. Patents are starting to issue here.
MEMS Investor Journal: What are currently typical costs for MEMS packaging in various applications? Can you give a few examples?
Dr. Ken Gilleo: Ceramics remains the “gold standard” of packaging, but at a cost premium. However, advancements in materials and processes now make it possible to package MEMS in a full-hermetic ceramic cavity package for less than $0.50. Evolving Chip Scale Packages (CSP) should further reduce costs.
Wafer-level capping enables non-hermetic plastic overmolding to be used, especially for motion sensors. Today’s cheapest package is probably the plastic QFN (Quad Flat-pack No-lead) that comes at just under a nickel for electronics chips. This same package can be used with some, but not all, capped MEMS devices. When the cost of capping is added, the total package cost is probably about $0.15, but this is a guess since capping costs seem to be a well-guarded secret.
Injection-molded cavity packages can be made for $0.08 - $0.15 in volume, and this could be the lowest cost option if the MEMS device doesn’t need capping. But there’s another consideration that will be important in the future. Adding more package features, like ribs, ports, grooves, etc., has little no impact on production cost although it will increase initial tooling. The injection-molding press automatically makes complex parts, in a cycle time of seconds. Micro-scale features can be molded with high precision and nano-scale may be possible. Plastic just may be the best choice for advanced MEMS, especially fluidics. Hermetic barriers seem feasible.
WLP has a long way to evolve, but this should have the lowest cost potential. There are many significant challenges ahead, but some will be solved by the electronics industry that is embracing WLP.
MEMS Investor Journal: In general, what is the percentage of the overall device cost which is attributed to packaging? What can manufacturers do to lower packaging costs?
Dr. Ken Gilleo: The old rule of 80% for the package and 20% for the device has been displaced for most products. We probably hit par at the beginning of this decade. While costs are very product-specific, accelerometer packaging is probably less costly than the device, but this area has a relatively long history. But, MEMS device costs continue to fall making the comparison a moving target. Best guess for a popular product like motion sensors is that packaging is 25% or less of the total. MEMS microphone packages are even a smaller percentage.
Exceptions are optical-MEMS, like the DLP, where the hermetic precision optical package may cost several times more than the device.
**************************************
Dr. Ken Gilleo holds a PhD in Chemistry earned under a full NASA scholarship at the University of Connecticut (UCONN). He has developed new products for 35 years and is an inventor in circuitry, electronics materials and packaging. He holds more than 30 US patents and his products have won 3 R&D 100 Awards.
Ken has authored over 500 papers, presentations and workshops. He writes technical articles and columns for technical magazines. McGraw-Hill recently published his 7th book, MEMS and MOEMS Packaging.
Ken is president of ET-Trends LLC, a consulting and intellectual property firm focused on emerging technologies and device packaging. He is an active member of IEEE and Vice President of Technical Programs for the Surface Mount Technology Association (SMTA).
A lucid article on MEMS packaging. I enjoyed reading it.
Posted by: VITHAL V. ATHANI | September 10, 2006 at 07:08 PM
what is application for tilt sensor?
Posted by: karthik.s | February 13, 2009 at 03:30 AM