Newsletter Archives

Commentary, opinions, and insights on all things MEMS.

December 2016: Holiday Cookies

At AMFitzgerald, one of our favorite recreational activities is baking. We think it's no accident that a bunch of MEMS process engineers are also great bakers. After all, the two endeavors have a lot in common. Both require control of sensitive, multi-variable process recipes, attention to detail, and a dogged persistence to get excellent results.

Just as every process engineer knows where the sweet spot is in their LPCVD furnace, every baker knows which rack in their oven delivers the perfectly-baked cookie.

Here's our last process batch of the year:

Our holiday cookie recipes:

What MEMS engineers can do with flour and sugar.

July 2016: MEMS Flaws Are Cryptography Features

MEMS makers spend enormous effort struggling to eliminate process variations which result in imperfect devices, manufacturing yield loss, and lower profit margins.

In a twist of great irony, the very flaws that plague MEMS sensors are in fact features for another application: cryptography.

MEMS gyroscopes, it turns out, are useful as a Physical Unclonable Function (PUF), a device used to create a cryptographic key. When interrogated, a PUF produces a unique response ("physical function") that cannot be predicted by inspecting the physical device, nor can it be duplicated ("unclonable"). Simply put, a PUF is a tamper-proof, copy-proof, black box that spits out a unique code.

Subtle variations in the etch process make each gyro uniquely imperfect in a random, unpredictable way. And even if one could thoroughly measure and inspect it, a gyro could not be perfectly duplicated because silicon etch processes cannot be controlled that finely. All of which are excellent characteristics for a PUF.

A research group at Bosch recently demonstrated how to use a commercially-available gyro as a PUF to generate a 128-bit cryptographic key. Further optimization of a MEMS PUF design could produce a longer, and therefore more secure, key.

Should you think this cryptography application for a gyro is merely an academic curiosity, think again: All data-generating devices in the Internet of Things (IoT) will need encryption, even innocent-seeming wearables, to protect them from hacking. Just last week, some researchers showed that sensor data intercepted from smart watches and fitness trackers can reveal an ATM PIN.

A dual-purpose MEMS device, which in addition to sensing, can produce its own encryption key, is a very exciting emerging technology with huge potential in future IoT markets.

MEMS devices having electrostatic comb fingers (pictured above), such as accelerometers and gyroscopes, have large surface areas which make them vulnerable to silicon etch process variation. (Source: AMFitzgerald)

June 2016: Where has all the silicon gone?

This month, we attended the Solid State Sensors, Actuators and Microsystems Workshop in Hilton Head Island, SC. Since 1984, this venerable conference has been a showcase for the best new MEMS technologies being developed in the Americas. The conference is one of the most competitive in the field, with only about 10% of abstract submissions gaining the privilege of an oral presentation.

The Hilton Head Workshop provides a window onto future commercial products. For example, earlier conferences featured many papers on MEMS accelerometers and gyroscopes, and those products rose to market dominance soon after.

At recent conferences, there's been a trend unsettling for those of us in the silicon business: a decreasing number of papers on silicon technology. This year, only 53% of the oral presentations featured silicon-based devices (generously defined here as a device incorporating silicon, whether as a passive substrate or an active material).

Surveying oral presentation topics over the past 12 years (Figure), it's clear the tide of academic research on microdevices is ebbing away from silicon. Instead, academic researchers are now focusing on developing polymer- and paper-based technology. Devices fabricated from these materials can be made in non-clean environments, at low cost, using simple equipment. They're a much better fit for shrinking research budgets than silicon. Many of these innovations are aimed towards medical applications, where biocompatible and flexible materials are essential.

Paper and plastic microdevices are still relatively primitive in function and capability, and manufacturing infrastructure for this type of device doesn't yet exist. It could take more than 10 years for these new technologies to mature and reach commercial markets.

For those in industry who have been counting on academia to invent exciting new silicon MEMS for their products of tomorrow, here's the bad news: the academic researchers are abandoning silicon.

However, there is certainly plenty more innovating to do on silicon. This means that industry will soon need to pick up the R&D slack or risk stagnation.

Percentage of oral presentations describing silicon microdevices at the biannual Hilton Head Workshop. (Source: AMFitzgerald)

March 2016: Making MEMS Lemonade

In the USA, to motivate optimism when things haven't turned out as expected, we say:

"When life gives you lemons, make lemonade!"

In other words, take the sour stuff, and find a way to make it sweet. In the MEMS business, we make lemonade all the time.

Every fab engineer has experienced the horror of an important wafer emerging from a chamber cracked. It's the last device wafer! It'll take weeks to get new wafers to the same point in the process! What now?

In Dr. Carolyn White's recent presentation at EVG Technology Day, she explained how we make lemonade when things don't go as expected in a new process. In prototyping, often the most valuable output from a wafer run is not the device itself, but the process data and tolerances which inform the next design or process revision. When faced with a broken wafer, we examine each fragment to see what may be salvaged. Since AMFitzgerald works in a development laboratory, we have access to equipment that can accommodate pieces as well as whole wafers. Often, we can continue to process the fragment, or do enough analysis or testing to harvest the valuable data we're seeking.

MEMS manufacturers regularly make lemonade, too. MEMS processing is complex and full of variances: within the wafer, wafer-to-wafer, and lot-to-lot. Although chips may appear to be patterned and processed identically, they never come out all the same. Finished wafers often have beautiful chips in their center and duds at their perimeter.

Long ago, our semiconductor colleagues figured out how to manage the variety of lemons that inevitably come out of a complex silicon process. It's called "binning." All the chips on a finished wafer get probed, and the ones that meet all specs are put in one bin, and sell for the highest price, the ones that are not as good go in a different bin and sell for slightly less, and so on, until all the chips have been sorted. For decades, microprocessors, memory chips, and now MEMS resonators, accelerometers, and gyroscopes have all been sorted, binned, and sold in this way. (Sometimes, these small variances in chip performance may also uniquely identify an individual chip.)

The difference between a "performance" chip and a "consumer grade" chip is often not due to the design, but to where the chip had been located on its wafer.

Lemonade, anyone?

Making lemonade

February 2016: Monkeys and Sharks!

Happy Lunar New Year! The red lanterns are flying in San Francisco and pictures of happy monkeys can be seen all over the city. San Francisco is home to the largest Chinese New Year celebration in the USA. If you happen to be here on February 20th, don't miss the parade! We wish all of our colleagues and partners a wonderful new year.

This even-numbered year also brings the Hilton Head Workshop, where MEMS engineers present their latest research and cavort by the ocean. Sharks will be swimming, not in the warm Atlantic waters, but in the frosty air-conditioning of the Sonesta's ballroom. Technical Program Chair Dr. Tina Lamers has organized the first-ever "MEMS Shark Pup Tank" competition to take place on Wednesday, June 8.

Unlike the TV show Shark Tank, where contestants get shredded by toothy investor Sharks, the "MEMS Shark Pup Tank" will be a supportive and instructive competition for aspiring micro-technology entrepreneurs, from academia or industry, to test their best ideas on industry leaders. Contestants should expect persistent nibbling and ankle-biting from the Shark Pups, instead of a feeding frenzy. Teams must enter by March 31, 2016. Click here for more information and rules.

2016: Year of the Monkey

December 2015: Emerging MEMS & Sensor Technologies to Watch

As the year draws to a close, people naturally start to think about what the next year, and the future, will bring. With growing anxiety over continuing price erosion in high volume MEMS and frenetic M&A activity, the question on many people's minds is:

What will be the next big thing in MEMS and sensors?

We don't have a crystal ball, but we do know from history that most of today's blockbuster MEMS technologies originated in academic laboratories. We expect more to emerge over time.

In this brief report, several noteworthy topics and papers presented at this year's academic conferences are highlighted. The criteria for selection were: commercial relevance, offers a solution to a known or anticipated problem, and technology game-changers.

Nearly all of the research mentioned will need many more years of intensive development and probably more than $100M in investment to bring them all the way to market. Nevertheless, they each hold potential to create new waves of commercial activity in the MEMS industry.

The emerging MEMS and sensor technologies to watch are:

  • Navigation-grade gyroscopes
  • Zero quiescent power devices
  • GaN resonators
  • Graphene FET gas sensors
  • Biodegradable sensors
  • Flexible energy harvesters
  • Paper-based devices
Download the report or a presentation on this same topic that was given at the 2015 MIG Executive Congress in Napa, CA.

Having a crystal ball would make things so much easier.

September 2015: What Navigators Want From MEMS

Earlier this month, we had the opportunity to attend the Institute of Navigation GNSS+ conference in Tampa, FL. This is the world's largest technical meeting and showcase of satellite navigation technology, products, and services. It brought together international academic, industry, military, and government leaders in the positioning, navigation, and timing (PNT) fields. The show attracted over 1,200 attendees and 45 exhibitors.

Overall, the navigation community is enthusiastic about integrating MEMS into navigation systems. They like the idea of getting more data from small, relatively low cost sensors. Recently, U. S. Secretary of Defense Ashton Carter declared his wish to have future MEMS sensors completely eliminate the need for satellite systems like GPS.

What navigators want from MEMS depends on who they are:

The "high integrity" navigators - the people whose systems land airplanes or steer self-driving cars - they would like MEMS sensors with enough performance to enable accurate inertial navigation without GPS for at least ten minutes. If a GPS receiver can't see at least four satellites in the sky, it can't produce accurate navigation data. High integrity navigators are the original developers of sensor fusion systems; they know that no one sensor is perfect, so they design systems to detect loss of a reliable signal, and then adeptly switch between sensor data streams as needed to maintain accurate navigation information. Ten minutes of GPS-independent inertial navigation buys you enough time to get to higher altitudes, out of a tunnel or around a skyscraper, to a position that improves your view of the sky.

The "consumer" navigators - the people who want you to help you find the nearest Starbucks in downtown San Francisco - they would like better low cost MEMS gyroscopes and magnetometers, specifically with improved stability, to improve pedestrian inertial navigation. Although pedestrians are relatively slow-moving compared to vehicles, a key challenge to their accurate navigation is maintaining inertial position fix while their smartphone unexpectedly changes orientation: waving about in the person's hand or sliding around in a purse or pants pocket.

It's clear we MEMS people need to spend more time with these end-users, to first understand how MEMS will integrate with their other sensors and GNSS, and then to derive the essential MEMS sensor specifications for each specific navigation system and use case. The quest for seamless navigation has been, and will continue to be, an exercise in sensor fusion.

For centuries, navigators have had to synthesize information from multiple imperfect sources to find their way

June 2015: The Internet of Dead Things?

Every time I go home to visit my parents, I get "The List" within hours of my arrival. The List consists of computer problems that have to be addressed while I'm home because no one can figure them out. The printer's not working. The computer's too slow. Mysterious pop-ups appear on the desktop.

I'm CEO of a technology company, holder of advanced engineering degrees, and I'm also ... Parental IT Support Technician. My dear colleagues, I bet this is one of your jobs, too.


Two weeks ago, I woke at 06:00 to an odd, fluctuating mechanical noise reverberating through our condo. After some investigating, I discovered the problem: my downstairs neighbor's Nest was malfunctioning. It was cycling his central heater on and off every 5 seconds. He was, of course, in London. I texted him and thankfully he was able to disable it remotely.

And then the next morning at 06:00, the Nest started doing it again.

How many times have you ripped the battery out of your smoke detector because it was making that annoying "low voltage warning" beep? How soon did you replace it?

Never? You know it's the most important safety sensor in your house, right?

I've been attending a lot of talks on the Internet of Things the past few years which focused on Smart Things for the consumer. Smart Homes. Smart Watches. Smart Clothing. I have yet to hear someone address the question, "Who will fix the Internet of Things?" for the average consumer? What will happen when the Things need new batteries, firmware upgrades, or just start mysteriously misbehaving? The Things won't be fail-safe, dog-proof, or toddler-proof; that level of reliability engineering is much too expensive for trillions of low-cost Things. So most certainly, fixing Things will also end up on The List. But what will happen to those consumers who don't have tech-savvy children around to fix the chirping, misbehaving, and dead Things?

We've envisioned and hope for a world of interconnected Smart Things, yet frankly, most people have trouble today keeping their "dumb" things in proper working order. Certainly, a whole new cottage industry of IoT gadget fixers will arise. But there's a good reason plumbers are among the highest paid home contractors - when you need their services, you're happy to pay almost anything, just as long as they come NOW. The toilet must be fixed within hours. But will you care if the Smart Refrigerator stops ordering milk? Will consumers be as willing to pay to have their Smart Homes and gadgets repaired, or will they just ignore dead Things like that smoke detector with its battery ripped out?

When the Internet of Things finally arrives, could we soon after be facing the Internet of Dead Things? (Energy harvesting notwithstanding.)

-Alissa (channeling Andy Rooney)

How most people "fix" a malfunctioning smoke detector

January 2015: RocketMEMS® Update

Most commercial MEMS sensors available for sale have been optimized only for high volume consumer or automotive applications. As a result, companies making products for medical, industrial, or aerospace applications often have trouble sourcing a sensor with the exact specifications they seek.

AMFitzgerald's RocketMEMS® service is designed to address the unmet sensor needs of low volume, high value markets, where sensor customization adds much value to a product. RocketMEMS offers a "semi-custom" solution in which new sensors may be quickly and cost-effectively tailored to a desired specification through use of our verified reference designs and our foundry partner's qualified processes.

Three recently published articles describe the RocketMEMS model and its many benefits:

Achieving System Cost Reduction and Performance Optimization through Semi-Custom MEMS Pressure Sensors                                                  Sensors Magazine, Jan 2015

RocketMEMS®: Tailored MEMS Sensors for Customers Seeking Business Opportunities in the Long Tail Marketplace                                           CMM International, Nov 2014

Choosing MEMS Pressure Sensors for Medical Device Applications        Medical Design Briefs, Nov 2014

RocketMEMS pressure sensor chips

October 2014: Emerging MEMS Technologies to Watch

Report from the Hilton Head Sensors, Actuators and Microsystems Workshop, June 9-12, 2014

Top MEMS researchers from the Americas gathered at Hilton Head Island, South Carolina in early June to present their latest work on novel MEMS devices. The Hilton Head workshop, through the years, has been a reliable predictor of which new MEMS technologies are poised to make the leap from research to commercial products.

In this brief report, we highlight several topics and papers presented at this year's workshop that caught our attention. Our criteria for noteworthiness were commercial relevance, relative maturity and a path towards mass production. Nearly all of the papers mentioned here would need at least three more years of intensive development to bring them to market, but nevertheless, they each hold potential to create new waves of commercial activity in the MEMS industry.

The emerging MEMS technologies to watch are:

  • Navigation-grade gyroscopes
  • Magnetic sensors and materials
  • GaN resonators
You can download the report here.

Ahh...the glorious silicon oxide of Hilton Head Island, SC

August 2014: The Napa earthquake, as detected by fitness trackers

At 3:20am on Sunday August 24, a 6.0 earthquake struck at the heart of California's wine country in Napa and jolted many Californians awake. That Monday morning, a data scientist at Jawbone, maker of the UP fitness bracelet that tracks sleep activity via MEMS accelerometers, sifted through the company's trove of user data. He created this interesting chart (below) that shows what percentage of users woke up as a function of their distance from the quake's epicenter.

If there were any doubt, the Internet of Things is indeed upon us, albeit in cruder form than imagined by most in our industry. In this particular case, it was embodied in the thousands of sensors on the wrists of users, sending data back to the Jawbone servers by way of the users' smartphones.

While this data set is coarse and only confirms the obvious (the closer one was to the epicenter, the more likely one woke up), it provides a glimpse at the power of distributed sensor networks and the insights that may be found in a large data set. The magnitude of shaking in a structure during an earthquake is a function of the stiffness of the structure and the ground below it. Now imagine if each of these UP wearers had had more sensitive accelerometers attached to the foundation and structure of their homes. This data set would have then provided some very useful and actionable information: the geological properties of their land to square meter resolution; the structural integrity of their home; whether the home's prior seismic upgrades were effective, and how much more structural reinforcement might be needed to survive a larger quake.

As IoT evolves, let's hope the industry thinks carefully about how to use it as a tool to get truly useful information. In the meantime, we'll have to just be amused by the fact that even a 6.0 earthquake is not strong enough to wake some Californians!

June 2014: "Engineering Hero" Dr. Morris Chang Speaks at Stanford University

The Stanford Engineering Heroes program, begun in 2010, honors Stanford alumni who have profoundly advanced the course of human, social and economic progress through engineering. On April 23, Dr. Morris Chang, founder of TSMC, received this honor at Stanford, and we had the opportunity to attend the public ceremony.

When introducing Dr. Chang, Jen-Hsun Huang, CEO of Nvidia, strikingly characterized the success and impact of Dr. Chang's career by pointing out the ubiquity of TSMC products: nearly every person in the audience would have a TSMC-made chip in their possession at that very moment, in their mobile phone.

As Dr. Chang described the genesis of TSMC, some of the circumstances of the IC industry at the time reflect the current MEMS industry. In the early 1990's, less than 20 fabless IC companies existed, and the few that did had to work with large companies with captive fabs. Small companies were accepted as customers only if the captive fab found the small company's technology strategically useful to their existing products. Indeed, this situation still exists today in some parts of the MEMS industry.

Dr. Chang described the founding of TSMC as a "solution looking for a problem, and happily, a problem occurred," in that more and more small companies began to seek access to foundry service without strings attached. With TSMC leading the way for the pure-play IC foundry model, in just 10 years, the number of fabless IC companies quickly multiplied to more than 400. In contrast, we have not yet seen similar growth of fabless MEMS companies with high-volume products, despite the fact that MEMS pure-play foundries have existed for over a decade. Only two of the Yole Top 30 MEMS companies are completely fabless. This made us wonder, are we perhaps still missing some important ingredient in the MEMS fabless business model?

Procuring a license to an RCA CMOS process that was two generations behind the then-current CMOS technology enabled the start of TSMC, according to Dr. Chang. Their initial business was built on this obsolete but stable and well-characterized process. As profits accumulated, TSMC was then able to fund R&D work to get to more advanced process nodes, a strategy that helped it to eventually become the world's largest foundry.

This anecdote caught our attention as a potentially useful lesson for the MEMS industry. As we collectively seek to standardize processes, reduce the cost of MEMS development and speed new products to market, we might consider a similar practical starting point. With all the risks the MEMS industry faces with developing new standards and business models, should we be taking a page from the TSMC history book, and finding more ways to leverage older, stable processes, or processes that we can license from others?

Our encounter with Dr. Chang, Engineering Hero, left us pondering these questions and more. We look forward to having lively discussions with you on these topics and more at MEMS industry events throughout the year.

Jen-Hsun Huang, CEO of Nvidia, introduces Dr. Morris Chang (l) and Stanford President Hennessy (r)

December 2013: RocketMEMS™ Evaluation Kits Now Available

If your product development team has been struggling to find just the right MEMS pressure sensor for your new product, this holiday season, give them the gift of RocketMEMS! Your team will be delighted to get the pressure sensor they've always wanted without the headache and expense of custom development.

Most MEMS pressure sensors on the market have been optimized for automotive or consumer electronics applications. OEMs and system integrators in other markets such as medical device, industrial, or aerospace, will find that commercially-available MEMS sensors are the wrong size, range, or sensitivity for their application.

RocketMEMS is a new "semi-custom" MEMS service offered by AMFitzgerald that enables customers to get the exact sensor they need quickly and cost-effectively. AMFitzgerald has developed pressure sensor reference designs tuned for Silex's production-qualified SmartBlocks™. These designs can be further tailored to meet customers' specific needs. Because the designs are already qualified, the risk, time and cost of custom development is eliminated. The customer only has to provide a sensor specification to AMFitzgerald and can expect to receive bare die within six months of run start.

RocketMEMS evaluation kits are available now. The kit (pictured above) contains a packaged pressure sensor, circuit board, cable and software drivers to enable plug and play evaluation of a reference sensor design. The sensor output can be read in I2C digital or unamplified analog formats. Sample bare die are also available. Contact us at to get yours today!

RocketMEMS™ Evaluation Kit

October 2013: MEMS in Emergency Location Devices

Sailors and aviators have long relied on EPIRB (Emergency Position-Indicating Radio Beacon) or ELT (Emergency Locator Transmitter) devices to transmit their location and a distress signal via satellite relay in case of emergency. While highly effective at saving lives, these devices are expensive (thousands of dollars), large (tens of centimeters) and heavy (several kilograms) and designed to be attached to ships, aircraft, or a vehicle. In the past several years, PLBs (personal locator beacons), such as Spot, have become available for hiking and outdoor recreational use. These devices are smaller than ELTs and used by people venturing into the wilder areas of Earth.

Now there's a locator device for those of us having adventures closer to home, such as riding a bicycle through downtown San Francisco during rush hour. A new product, ICEdot, utilizes a MEMS accelerometer and gyroscope to detect crashes. The ICEdot sensor unit is about the size of a pillbox and lightweight enough to mount to a helmet. Like a lot of personal electronics these days, the ICEdot cleverly utilizes the phone's GPS and cell radio to keep its own size and power consumption to a minimum. When a crash is detected, the ICEdot transmits a distress signal via SMS. Of course, it's essential to have one's mobile phone along for the ride.

Curious to know which company's MEMS sensors are inside? Watch this interesting teardown analysis video of the ICEdot.

ICE = In Case of Emergency

July 2013: Emerging MEMS Technologies to Watch

Report from Transducers 2013 - International Conference on Solid-State Sensors, Actuators, and Microsystems. Barcelona, Spain, Park Guell,    June 16-20, 2013

At AMFitzgerald, it's our business to know what's going on in the world of MEMS. To find out more about the latest MEMS technology being developed in the world's top research laboratories, we attended Transducers 2013 in Barcelona, Spain. This high-quality bi-annual international conference always provides a glimpse of the exciting MEMS products yet to come.

In a brief report, we review topics and papers presented at Transducers that caught our attention. Our criteria for noteworthiness were commercial relevance, relative maturity and a path towards mass production. Nearly all of the papers mentioned here would need at least three more years of intensive development to bring them to market, but nevertheless, they each hold potential to create new waves of commercial activity in the MEMS industry.

You can download the report here. For more in-depth analysis of the technologies discussed here and their commercial applications, please contact us.

Dicing and die attach - Gaudi style

June 2013: Contextually aware ... toaster ovens?

Recently, much creative thought has been focused on developing context awareness in mobile phones. If your phone could figure out what you are doing, or perhaps even what you intend to do, then it could better serve you and consume less power. With sensor data and some logic (also known as sensor fusion), a phone could detect the context of your situation and automatically make adjustments without manual prompting. For example, if it senses you are in a meeting, it could silence or soften your ring tones. If it knows you are outdoors walking the dog, it could boost noise cancellation when you take an incoming call. If it has learned that you enjoy an afternoon latte, then at 3pm, the phone would offer a list of nearby cafes and let you know which ones are offering specials.

Why should context awareness be limited only to mobile phones? Many of our less intelligent appliances could benefit from context awareness, too. Already, clothes dryers use humidity sensors to determine how long to dry a load, so you don't have to guess how to set the dryer timer. Breville is now offering a Smart Oven with sensor technology that can dynamically adjust heating power and detect when food is cooked. Finally, an oven that won't burn food! While you still have to manually prompt the Smart Oven that you are cooking a bagel, and not a pizza, it's easy to imagine that if a sensor fusion upgrade were applied, the oven could identify the food type without your input. And wouldn't it be even better, if someday it could learn your cooking preferences, so that it automatically toasts your bagel lightly but makes your spouse's dark and crispy.

If you missed Alissa Fitzgerald's presentation on this topic at Sensors Expo: "Thinking outside the (mobile) box: other high-value applications for sensor fusion," you can download it here.

Did you miss the webinar that introduced RocketMEMS™, our semi-custom pressure sensor offering? You can listen to it here.

The Breville Smart Oven won't burn your food

May 2013: Sensor Fusion - it's not just for mobile phones

The MEMS industry is very excited about sensor fusion, the combination of data from multiple sensors to improve accuracy, make inferences or predict behavior. Sensor fusion will most certainly drive the demand for MEMS sensors to new heights. So far, nearly all of the talk and excitement about sensor fusion has been focused on its application in mobile phones and tablets, and for good reason. The mobile phone and tablet markets are huge. And we all love our gadgets, so of course we want to see them do more cool things.

There has been almost no discussion about everything else, meaning similarly huge markets that were in existence long before the smartphone. Plenty of other products could also benefit from sensor fusion, but they're being overlooked.

Take, for example, the thermostat. In the USA, it is estimated there are 250 million thermostats installed in homes and light commercial buildings, and 10 million new ones purchased annually. (By the way, those numbers are about comparable to the number of smartphone users and annual purchase rate of new smartphones in the USA.) Why aren't we talking more about thermostats? Oh yeah...they don't quite capture the imagination or generate excitement the way our smartphones do.

Nest, of Palo Alto, CA, has already recognized the huge business opportunity in upgrading the lowly thermostat. Founded by a former Apple designer, the company combines sleek product design with sensor fusion. The Nest thermostat detects occupants' behavior patterns, merges that data with local weather information, and then adjusts its heat/cool settings appropriately. The data from multiple MEMS temperature sensors, a humidity sensor and and an IR sensor are used to infer who is home (humans vs. pets) and how well the system is working. It's also an internet-enabled device, so it can be programmed remotely and can update its own software.

On June 4, at Sensors Expo, Alissa Fitzgerald will be presenting more examples of opportunities ripe for sensor fusion in her talk, "Thinking outside the (mobile) box: other high-value applications for sensor fusion."

Nest Thermostat

February 2013: Seeing the Future

Star Trek:TNG fans will recall the blind chief engineer, Geordi La Forge, whose vision was restored and even enhanced by his visor (see photo). Last week, we moved one step closer to science fiction becoming reality with the FDA approval of the Second Sight Argus II. The Argus II vision system can provide 60 pixels to a person whose retina is damaged, but whose optic nerve is intact. A CCD sensor (similar to one found in a digital camera) replaces the retina's function, and the sensor electrical signals travel through micro-machined needle electrodes into the optic nerve, where the brain interprets them as light.

To a seeing person, 60 pixels would seem like a horribly low resolution display. But to a blind person, it provides enough vision to negotiate streets and to avoid obstacles. The Argus II is a life-changing technology and a huge step towards repairing a major disability. MEMS technology, once again, is playing an humble but important supporting role in exciting medical advances.

As medical device technology slowly but surely finds ways to replace faulty human systems (with pacemakers, artificial joints, insulin pumps, cochlear implants, etc.), it seems inevitable that decades from now we will be bionically-enhanced beings, not unlike ... Star Trek's Borg. (Hopefully without the hive mind!)

May we all live long and prosper.

Geordi La Forge (fictional character) and Barbara Campbell, Argus II recipient (NYTimes photo)

January 2013: Got MEMS?

There's no limit to the usefulness of MEMS sensors. There are approximately 265 million dairy cows in the world, and did you know they need MEMS sensors, too? Or rather, their farmers do. A dairy farmer needs to breed his cows to keep milk production up, but the trick is knowing when to breed them. Missed fertility cycles mean lost revenue.

It turns out that cows in heat get restless, usually in the middle of the night when the farmer is not around to observe the tell-tale behavior. By fitting a cow with a MEMS accelerometer collar, her movement patterns can be tracked and correlated to her estrus cycle with great accuracy. Swiss cows can actually text their farmers (in 5 languages) when the moment arrives. And yes, there's an app for that.

Are you laughing? Snickering just a little bit? It's no joke, dairy production is very big business. In the USA, the dairy industry annually generates $140 billion in economic output relating to the production of milk and milk products (e.g. cheese, yogurt, ice cream, butter).

You heard it here first: the MEMS industry should be taking a much closer look at agriculture and its related industries as the potentially next big market opportunity for sensors.

*The title refers to a popular American television advertisement:

The Dairymaster MooMonitor uses a MEMS accelerometer

November 2012: RocketMEMS™ launches!! First payload: Pressure sensors

AMFitzgerald proudly announces the RocketMEMS program, which puts customized MEMS chips in the hands of system integrators faster than ever before. RocketMEMS is a new model for designer-foundry collaboration that enables rapid and cost-effective development of MEMS sensors. The program uses a manufacturing-friendly approach to MEMS development: sensors are designed specifically for existing foundry process flows. This minimizes time, cost, and the risk of qualifying new processes.

AMFitzgerald has selected Silex SmartBlocks™, a suite of standardized process modules, for the first offering from the RocketMEMS program: pressure sensors.

How it works: Customers submit their pressure sensor specifications to AMFitzgerald. Using a sensor architecture we developed for SmartBlocks, we then custom-tailor sensor designs to fit each customers' specific needs. To keep production costs low, we assemble the customers' designs into a multi-project wafer (MPW) run to be fabricated at Silex. Customers receive their bare die about 3 months later.

MEMS solutions for OEMs and system integrators

October 2012: Using MEMS to Lose Weight

Just 10 years ago, the only human performance measuring tool accessible to consumers was the heart rate monitor, which correlated heart beats per minute to athletic performance and calorie consumption. The heart rate monitor strap, worn around the chest, was uncomfortable and awkward, and so was used only by elite athletes during workouts to optimize their training regimens. At the time, no one would have ever bothered to wear it to measure one's athletic performance while doing more mundane things, like walking to the store.

Thanks to MEMS sensors, the measurement of human motion is now easier than ever. So easy, that people can now measure all of their daily motion, not just athletic workouts. Enter the FitBit and Nike+ FuelBand, two of several such health-management products enabled by the MEMS accelerometer. Both products measure steps and estimate calories burned during everyday activities. They are stylishly designed in a way that makes them fun, and not embarrassing, to wear. The brilliant trifecta that drives the adoption of these self-measurement tools is sensors, webpage interfaces and social networking. Both the FitBit and FuelBand upload data to a website (via PC or mobile phone), where you can track your goals and share your progress with friends. A new trend in weight management programs has emerged, based on the simple logic of measure everything: count the calories you eat, subtract the ones you burn, and keep the net result to a reasonable number. Social sharing of the data keeps you accountable to yourself and your peers.

These gadgets get results and are habit forming. If you find yourself suddenly addicted to self-measurement, you're not alone. You can find your fellow enthusiasts at Quantified Self and learn more about how people are using sensors and apps to measure and track things like sleep quality, mood, and spending habits (and possibly discovering some links between the three of those).

Measure your day with a sporty MEMS bracelet by Nike

July 2012: MEMS in Sports Safety Equipment

One could easily argue that the most significant contribution of MEMS technology has not been to consumer electronics (much as we love our iPhones), but to automotive safety systems. MEMS technology has made airbags, anti-rollover stability systems, and tire pressure monitoring ubiquitous in cars. These safety systems literally save thousands of lives every year and no one today would buy a car without them.

MEMS are now finding their way into sports equipment, most significantly into helmets. In the US, there has recently been much reporting on concussion injury in sports and a new awakening that repeated head injuries have much more serious consequences than previously realized. (Consider the tragic story of Derek Boogaard, the NY Rangers hockey player.)

Efforts are underway to incorporate MEMS motion sensors in football helmets and Indy car driver earpieces to more accurately detect concussion injuries. It might not be too long before concussion sensors become mandated for sports helmets, particularly in those for children.

Pioneering a new market in fashionable safety equipment, the Swedish company Hövding has created a clever improvement to the universally-despised bicycle helmet: an airbag that deploys from a collar worn about the neck. Who would have guessed that MEMS technology could be used to save both one's life AND hairdo?

The Hövding airbag bicycle helmet. No need to mess up your hair until you're about to crash

June 2012: Emerging MEMS Technologies to Watch

Report from the Hilton Head Sensors, Actuators and Microsystems Workshop June 3-7, 2012

Top MEMS researchers from the Americas gathered at Hilton Head Island, South Carolina in early June to present their latest work on novel MEMS devices. The Hilton Head workshop, through the years, has been a reliable predictor of which new MEMS technologies are poised to make the leap from research to commercial products.

In this brief report, we highlight several topics and papers presented at this year's workshop that caught our attention. Our criteria for noteworthiness were commercial relevance, relative maturity and a path towards mass production. Nearly all of the papers mentioned here would need at least three more years of intensive development to bring them to market, but nevertheless, they each hold potential to create new waves of commercial activity in the MEMS industry.

The emerging MEMS technologies to watch are:

  • Biodegradable sensors
  • Surface texturing and manipulation
  • Proximity sensors
  • Resonant sensors
You can download the report here.

Ahh...the glorious silicon oxide of Hilton Head Island, SC

May 2012

A Primer for MEMS

This past month, Alissa Fitzgerald gave a five-part webinar series on "MEMS Sensor Technology" hosted by Design News. The webinar series was designed to be an introduction to MEMS for technically-minded people who have only recently heard about the technology and need to come up to speed quickly. Episode 1 begins with "What is MEMS?" and the rest of the series surveys exciting developments in MEMS for consumer, medical, automotive and industrial applications, with one episode focusing entirely on motion sensors. You can download and listen to the audio episodes for free.

Transferring MEMS to Foundry

Alissa also recently gave a lecture at the Berkeley Sensors and Actuators Center on "How to Successfully Transfer MEMS from a University Lab to a Commercial Foundry." This presentation describes what it takes to execute a smooth process transfer based on our experiences doing several such transfers for our customers. Foundry transfer is a critical technical and business step that is too often underestimated. We describe how to select a foundry that's right for your MEMS device, how to get the best quotes from foundries, how to plan and budget effectively for this year-plus effort (yes, it takes that long!), and how to get the best results from your new foundry partner.

The iBeer App: the most frivolous use of MEMS technology we've ever seen.

March 2012: MEMS in Watches

When one thinks of exquisite Swiss mechanical watches, one might imagine a bespectacled watchmaker, assembling tiny components under a magnifying glass. How about a MEMS engineer in a bunny suit? At the recent MIG Executive Congress Europe, held in Zurich, Switzerland on March 20, we learned that the Swiss research institutes EPFL and CSEM have been developing MEMS watch components for years, particularly escapements. Micro-machined single-crystal silicon components, formed by deep reactive ion etch, can be found in mechanical watches by Omega, Patek Philippe, Tourbillon and others.

At the Congress, we also learned that MEMS pressure sensors have been used in Suunto's watches since 1985. The Finnish watchmaker specializes in performance watches for scuba diving and outdoor sports. Suunto's new Ambit watch will also contain a MEMS accelerometer that will be used to improve the accuracy of GPS-calculated speeds, which tend to be inaccurate for human running speeds. The accel will also play a key role in power management by helping to detect, for example, when the watch has been removed and placed on a desk.

According to Prof. Nico de Rooj, the luxury watch industry had its best year ever in 2011 and is expected to reach $20B in revenue this year. The report of the death of wristwatches due to the rise of mobile phones has been greatly exaggerated, apparently.

Patek Philippe's Oscillomax subassembly in the 5550P limited edition watch

February 2012: The MEMS Patent Landscape

Thinking about developing a new MEMS device? Many companies have been inspired by the use of MEMS in consumer electronics to start thinking about how MEMS can improve their products and systems. But before a company jumps into a new MEMS development effort, it should spend some time studying the patent landscape. Although some may think of MEMS as an emerging technology, the US patent database says otherwise. In the database, one can find thousands of patents that reflect decades of global research and commercialization efforts. As an illustration of the challenge posed to new designs, the table on the right shows the number of US patents that contain the words "MEMS" and a specific keyword.

This simple search example is just the tip of the iceberg. More sophisticated search (and search of other countries' databases) will uncover hundreds more patents, especially when one factors in the packaging technology that must accompany a MEMS device. The bottom line is that developing a new MEMS design that has freedom to operate is getting to be very difficult. In certain competitive areas, such as inertial sensors, one may be better off licensing existing patents than attempting de novo design.

At AMFitzgerald, we can help you understand the patent landscape in your technology area and inform the business decision of whether to make vs. license. If you decide to embark on a new MEMS product development, we have strategies for design-around that can help to keep your infringement risk as low as possible.

October 2011: Simulation will save time and money

A MEMS engineer's dilemma: take the time to model a new device, or just build it and see what happens? Too often, engineers don't trust simulation and prefer to go straight to the fab to develop a new design. Simulation tools, after all, are expensive, require a lot of training to use well, and are handicapped by the lack of accurate (and fab-specific) material properties data. But generating one batch of test wafers, for a six-mask process on 150 mm wafers, will cost about $115K and 1-2 months, whether you have a captive fab or outsource. Not to mention the costs of packaging, testing, and analyzing all those test wafers.

Simulation is not just about generating pretty color plots for presentations. In the hands of an experienced engineer, simulation is a powerful tool to examine sensitivity to design variables, explore the effects of process variation, evaluate design rules, and to develop more focused Design of Experiment fab runs. In short, correctly using simulation can save at least one fab run, and usually more - which saves real money and time.

Read more of our thoughts on the usefulness of simulation and learn some practical MEMS simulation tips in this white-paper presentation.

ANSYS model by AMFitzgerald

September 2011: How to Dice Fragile MEMS Devices

After fighting through a multiple mask process and a tricky release of a fragile MEMS device, engineers are often tripped up by the final task of dicing a wafer. Traditional wafer sawing is a violent, messy process, where if the saw vibration and grit don't damage your MEMS, most likely the water jet will. To avoid dicing damage, many MEMS engineers turn to manual tricks (scribe and break, tapes, photoresist) or fancy steps (DRIE die separation) which don't scale well for manufacturing.

Thanks to "stealth" dicing (a term coined by one of the pioneers in the field, Hamamatsu Photonics), the world of MEMS now has a superior option for dicing. In stealth dicing, a laser beam is focused within the depth of the silicon wafer and scanned along the dicing lanes. Localized heating from the laser creates dislocations, or micro-defects, in the silicon crystal. Gentle mechanical tension, usually created by stretching the dicing tape to which the wafer has been mounted, will subsequently cause the silicon to fracture cleanly along the path that was swept by the laser. The result is beautifully cleaved die with zero kerf loss and zero chipping.

At AMFitzgerald, we regularly use stealth dicing for fragile MEMS such as cantilevers, thin membranes and electrostatic actuators. Believe us, we've tried every other dicing trick known to engineers, and we think stealth dicing is always the best option for fragile MEMS.

Learn more about our processing capabilities

Cantilever processed by AMFitzgerald