One of the most exciting and rapidly growing set of applications in our industry are those involving RF MEMS. We recently spoke with Dr. Dan Hyman, the founder and CEO of XCOM Wireless, about RF MEMS history, current challenges and future trends.
In this thoroughly comprehensive interview, Dr. Hyman shares his expertise and insight, and offers intriguing predictions about the future potential of RF MEMS devices.
MEMS Investor Journal: Why is there a need for RF MEMS? Why are MEMS devices for RF applications better than existing non-MEMS devices?
Dr. Dan Hyman: Developing RF circuits and subsystems requires a series of engineering trade-offs that are limited by the technology you are using. This is true at the device, component, and circuit level, and this is a part of RF engineering life that has been true for many decades. This is “easy” to deal with for single-mode systems like an old-fashioned cell phone, or modern Bluetooth circuit, but this gets harder and harder to do as frequencies get higher, data bandwidth gets larger, and, most of all, when multiple broadband signals have be handled in the same device. This is a defining trend in the wireless industry, and one that is taxing the limits of conventional technologies and “old-school” radio architectures. Wireless engineers in the broadband, multi-standard world need a better way of doing things, and better technologies that can easily handle these widely varying signals. Enter RF MEMS.
Essentially, RF MEMS devices offer “best of breed” in a host of performance and usage parameters out of all possible technologies you would reasonably consider to be your alternatives in a very wide variety of RF applications. The most widely recognized advantages are low loss, high isolation, near-perfect linearity, and unbelievably large instantaneous bandwidth that conventional mechanical and semiconductor technologies simply can’t even compete with. On top of this are a host of usage parameters like cost, size, speed, ruggedness, reliability, repeatability, and lifetime, that each range from fantastic to poor depending specifically on what you are comparing them to and depending on whether or not a customer values that particular specification. A good RF MEMS Ohmic relay, for example, has fantastic repeatability and ruggedness compared to any semiconductor technology, and has a good (or great) price, size, speed, reliability, and lifetime compared to any conventional mechanical technology.
There are a staggering number of RF applications where RF MEMS are superior in every measurable technical way to incumbent technologies, and it is this variety that allows for a substantial amount of non-competitiveness in the marketplace. Business concerns of supplier networks, fulfillment, qualification & supplier chain, etc. are still real challenges for many applications and customers, however, so niche applications are still the best bet for near-term product releases. Also note that there are many applications where RF MEMS devices are not a good choice. The ubiquitous and over-lauded “T/R switch in a GSM handset,” for example, was never a good idea to consider for RF MEMS, and probably never will be.
MEMS Investor Journal: When were RF MEMS first introduced and what were the original applications?
Dr. Dan Hyman: RF MEMS came into their own as a nascent industry in the early 90’s, which is when most of the classic devices were invented by researchers (myself included) working for major Defense-oriented companies. These device-level development efforts were largely funded by DARPA and the Air Force, with a system-level view of how these nifty gizmos could actually improve radar and communications systems. Soon thereafter, other Federal services and agencies recognized their own needs for improved RF systems, and followed suit with programs directed towards broadband, highly multi-band, and/or millimeter-wave applications such as tactical radio programs (JTRS, WIN-T, etc.), satellite communications, and terrain-penetrating imaging radar (FOPEN, URPEN, etc.). RF MEMS enable switching and tuning of front-end circuits for applications of mode switching, antenna tuning, and antenna beam steering with phase shifters.
The IP space started getting crowded in the late 90’s as universities and research organizations began to get involved. The “small company” interest is a fairly new phenomenon, really only starting in 2000 and focusing on a variety of niche applications in Defense, test equipment, and consumer electronics. Critical service first-responder radio interoperability, homeland security communications and radar became another rapidly growing niche since 9/11. Fortunately, so many lessons were already learned about MEMS devices and processes that XCOM and other venture and government-backed small businesses were able to hit the ground running with technology and process licensing, comparatively fast development times (3-4 years), and product releases in these niche applications.
MEMS Investor Journal: How have these original applications been accepted by the market? Which applications have survived and are on the market today?
Dr. Dan Hyman: The original Defense sponsors have so far been very patient and respectful of RF MEMS advances and product offerings over the years. However, they are still skeptical of adopting RF MEMS products for their specific applications because they have heard every promise in the book from researchers, trade rags, analysts, or companies out in the field (sometimes even their own). There are still significant obstacles for RF MEMS to enter conservative application areas with long histories of well-established technologies and very strict screening requirements. RF MEMS received substantial federal money in part because it truly is “DARPA hard” to get these things into production with the ruggedness, reliability, lifetime, repeatability, and performance demanded by the systems in which they would be used. I would even dare say that “ability to deliver” is still a significant concern for potential customers that require hundreds of thousands or millions of components, as there have been a number of aborted product releases in the industry’s short history.
I would estimate that more than half of total efforts put forth by RF MEMS developers with Defense product goals are dedicated to product testing and also improvement to overcome failures observed in the course of that testing. This is a natural part of product release, iteration, re-release, and ultimate customer qualification that should come as no surprise to any business. The delay to product adoption is normal, and just happens to be fairly long for Defense applications. Fortunately, so much time and energy has been spent characterizing and experimenting with RF MEMS, which Defense customers are highly educated about “what RF MEMS can do for you”. They are a generous and patient, though exacting, group of customers, and they are often willing to help subsidize the cost of development, iteration, iteration, testing, qualification, and manufacturing ramp if they need your component in their system badly enough.
I would assert that the vast majority of the original applications (though not the original programs or envisioned insertion points) for RF MEMS are still valid and interested in the technologies today after 15 years. The subsystem and system engineers are still skeptical with each new company or product release, but they are still eager and willing to try them out, and willing to comment on what they do or don’t like with each new RF MEMS component they get their hands on.
MEMS Investor Journal: How would you categorize the main kinds of RF MEMS devices today?
Dr. Dan Hyman: This is an important question for RF MEMS developers and users, because differences in component architectures have huge implications on how they can be used. Each kind of RF MEMS device is different than another, and even devices within the same category can vary wildly in terms of capabilities, performance, cost, size, etc. The best way I like to discuss them is in terms of their circuit configuration, because that is how the customer will be looking at the part relative to semiconductor options they may already be using (or considering). The circuit configurations I like to think about are as follows:
- Ohmic switch – The simplest three-terminal device architecture, directly comparable to a transistor switch and usable in the same way. It is possible to use switches in more complex circuits, but you have to pay careful attention to biasing, especially in very wide-band applications. The device is called “Ohmic” because it makes a genuine Ohmic metal-to-metal electrical and physical contact when closed.
- Ohmic relay – A four-terminal device that separates the control signal from the load signal within the body of the component package or even at the individual MEMS device itself. Many customers prefer relays to switches because of the ease of control design. Ohmic is as defined above.
- Capacitive switch – A three-terminal device architecture, with control problems as with Ohmic switches. These devices form a capacitor with a gap that can be varied with applied voltage. Changing the gap between the electrodes in a well-defined manner allows for a change in capacitance. This is not useful for direct current applications such as telephone line switching, but it is extremely useful in switching or tuning RF circuits.
- Capacitive relay – A four-terminal device that separates the control signal from the load signal, and changes capacitance as with the less capable switch version above. This is a configuration with a simple control, and is attractive for complex RF integrated circuits such as matching networks and phase shifters.
- Mechanical resonator – A micromachined three-terminal device that has a mechanical resonance in the RF regime. The resonance is useful as either a very exact time reference (better than quartz, and suitable for arraying or integration) or as a very sharp filter (that is very difficult to consistently implement).
- Bulk acoustical resonator – A micromachined two-terminal device that is better defined as a micro-system technology (MST) rather than a MEMS device. The best example is the FBAR, already found in many millions of handsets. Think of them as competitors to SAW filters, and you’re most of the way there.
MEMS Investor Journal: Which of these are already commercialized? Which are likely to be commercialized over the next 1-2 years? Which are going to be commercialized in the longer term?
Dr. Dan Hyman: Well, the two resonator types are already commercialized and shipping in high volumes at this moment. The mechanical resonators have not yet ramped up to the level of “Dick Tracy watch” market, but they are quite aggressively penetrating the reference oscillator market, replacing quartz crystal technologies at a truly impressive rate. The FBARs, of course, are also shipping high volumes and still growing, so the question of “can a MEMS gizmo make it?” has a clear answer of “yes.”
For more traditional RF MEMS switches and relays, the definition of the word “commercialized” is a topic of discussion. A typical definition would mean that a product has been created and released, to one or more customers, or to the open market. The part would be manufactured in a process that has been “qualified” with a proven reproducibility in manufacturing and performance, with that qualification being performed by the company itself (least desirable), by one of its customers (better), or by a third-party incumbent in the industry (best yet). With this reasonable definition, it can be said that four products have been commercialized in the test, instrumentation, and Defense communities: XCOM and Panasonic/Matsushita have commercial Ohmic relay products, and Radant MEMS and TeraVicta have Ohmic switch products.
In the next 1-2 years, other Ohmic switches and relays will be commercialized, as well as other niche variants of existing products re-focused for a particular application area or operational requirement. You will probably also see the first few capacitive switches or relays released inside larger circuit products such as tunable matching networks and phase shifters. In the longer term, you will see variants of these devices, and the actual manufacturing ramp of these devices. The analysts are a bit aggressive with predictions of adoption of RF MEMS by the marketplace, but I do not think they will be far off.
MEMS Investor Journal: In terms of taking RF MEMS devices from lab prototypes to mass production, what do you see as the main challenges?
Dr. Dan Hyman: There are a number of challenges that face anyone who wishes to take an RF MEMS device or circuit prototype into component production. The most significant challenges are those associated with packaging. MEMS actuators are intrinsically challenging devices to package, because you have to have a tiny, fragile device actually affect the environment or signal in some way. The purpose of the package is to interfere with this task in a minimal way, and protect the device from all other aspects of the environment. For RF MEMS devices, this means the package needs to be invisible to RF in the desired signal path (low insertion loss), be opaque to RF in every other possible signal path (good isolation, low cross-talk), defend the device from internal and external heat problems (including solderability, temperature cycling, power handling, etc.), defend it from environmental problems (fatigue, shock, vibration resilience), and defend the device from internal and external contamination (hermeticity, gettering, etc.). The package needs to do these things to an acceptable level, and also cost as little as possible.
These packaging aspects are designed to enable the high RF performance of the MEMS devices inside to shine, yet still provide customer-demanded levels of four critical operational characteristics: lifetime, repeatability, reliability, and ruggedness. These four separate characteristics have widely varying requirements (many orders of magnitude different) for different applications and customers. Lifetime is the number of times you can carry operational loads in service before a particular likelihood of a defined failure event. Repeatability is how similarly each part operates each time it operates, and sometimes how each part operates similarly to other parts throughout their lifetime. Reliability is how likely the part is going to operate properly throughout its lifetime given particular environments or usage. Ruggedness is how well the part survives harsh environments and conditions of storage or operation. All of these things are important at different levels to different customers, and the fact of the matter is that RF MEMS developers tend to focus on the most challenging specifications posed to them by their highest paying customers!
MEMS Investor Journal: What are the key challenges with marketing and selling RF MEMS devices?
Dr. Dan Hyman: Fortunately, the early marketing and sales problems of “customer outreach” and “user education” are not really the significant impediments they used to be. RF MEMS has, fortunately, passed through and survived the “nascent hype” phase of the industry growth cycle. There are no longer announcements of “wiggling beams” and vaporware products “sampling soon.” As for education, there are not very many RF MEMS suppliers in the industry compared to the large number of customers who want them, and it is interesting to continuously learn that every single customer is intimately familiar with every product.
Customer adoption for any of the “high-end” application areas still has taken a long time, and will continue to take a long time. The present marketing and sales efforts in our industry is simply about products or test results that address specific customer concerns and impediments to adoption. All of the developers in the field, whether small business or large, are facing the same questions; we need to address them to continue to make traction, and get our sales victories as we go.
There is, however, one downside to all of the progress RF MEMS developers make, which is that as an industry we continue to sell next years’ parts before we sell this years’ parts. Customers are learning that next year’s model really will handle more power, switch faster, last longer, and cost half as much! While personally satisfying to receive this vote of confidence, I’d rather have a second big order right now. . .
MEMS Investor Journal: In general, could you comment on the lessons learned and future trends for RF MEMS?
Dr. Dan Hyman: Lessons learned:
- Hermeticity – You can’t get too clean, too hermetic, too “well-sealed”. A perfect seal with a well-controlled environment inside is needed for contacting MEMS devices. Know your seal and perfect it, or use a highly trusted service provider to do it for you, because it can very easily make or break your product.
- Don’t over-design – Everything is an engineering tradeoff, so improving something that isn’t asked for by a customer is probably sacrificing something that they care about. Focus on product adoption.
- Don’t over-test – There are millions of possible tests and test conditions you could put a part through, but your customer probably only cares about ten. Perform them wisely, and don’t agonize over failure to meet an esoteric salt spray test, or an irrelevant burn-in test designed for semiconductor devices, or any other test that isn’t specifically asked for by a customer. Invest time and money carefully.
- Shipping quantities - More parts are going to be shipped; that’s a trend that we have seen in 2006 that we are certain is going to continue. It only takes two flagship customers to make a product line successful.
- Invisible usage – More devices are being developed out of existing processes, and I expect RF MEMS will become a “secret sauce” for circuit and subsystem products that will simply not advertise the technology. I will consider this to be an important milestone for the RF MEMS industry. . . “Oh, this thing here? It just works better, it doesn’t matter what’s inside!”
Dr. Hyman is an internationally recognized expert in RF MEMS packaging and contact reliability, with numerous papers, invited lectures, and patents in RF MEMS devices and applications. He holds a M.S. and Ph.D. in Electrical Engineering and Applied Physics from Case Western Reserve University, and a B.S. in Engineering from Harvey Mudd College. He is an active member of the IEEE MTT-S Subcommittee for RF MEMS since his appointment in 2001, and has served as a reliability and commercialization advisor to the NSF and Department of Defense. Dr. Hyman entered the RF MEMS industry in 1995 as a Hughes Fellow for HRL Laboratories, where he designed RF MEMS for U.S. military communications programs. He left HRL to co-found XCOM Wireless and address broader RF MEMS commercialization opportunities.
XCOM Wireless is a small business that specializes in RF MEMS devices, packaging, and circuitry. Most MEMS devices in the industry are expensive, fragile, environmentally-sensitive, and must be treated with great care in laboratory environments. XCOM RF MEMS circuits, by comparison, can be soldered, baked, frozen, drop-kicked, and shaken, which is an absolute requirement for rugged industrial, military, aerospace, and automotive electronics applications.