Wide-scale applications for Micro-Electromechanical Systems (MEMS) became practical in the 1980s when they started being fabricated with the same silicon wafer processes as semiconductor chips. Once they could be mass produced, they found a major market in automobile safety systems as inertial sensors for airbag deployment. Then came the introduction of smart phones in the late 2000s. According to Dr. Alissa M. Fitzgerald, CEO, A.M. Fitzgerald & Associates, a MEMS and microtechnology product development company, every smart phone these days contains dozens of MEMS chips. And now there are many more safety applications in automobiles beyond just air bag sensing.
All-in-all MEMS manufacturing is a $20 billion industry, according to Yole Développement, a French technology marketing firm that closely tracks the MEMS industry. Fitzgerald views the MEMS industry as bifurcated between mass market producers, who are making billions of chips and the companies like hers who serve specialty markets.
MEMS Now
By far the most common usage of MEMS devices these days is as motion sensors, pressure sensors, and microphones in cell phones and automobiles — in particular internal combustion automobiles. Also, personal devices like fitness trackers.
Automobiles
In automotive combustion engines, MEMS pressure sensors are used in the combustion and exhaust systems to optimize performance and emissions. In the cars’ safety systems, MEMS accelerometers are used to release air bags in the event of a crash, and gyroscopes enable electronic stability control. Some companies use MEMS microphones in the car to detect crash events, even locating where on the car the impact has occurred.
EVs are coming and they are obviously quite different from internal combustion automobiles — there will be many fewer moving parts and emissions will no longer be a problem. So, all those pressure sensors, the combustion management, those go away, said Fitzgerald. But in an EV, there is a large battery pack, which is prone to thermal runaway under certain circumstances. “So, you’d want to have sensors in your battery pack to detect if you’re having a thermal event — a temperature sensor, or a pressure sensor. And you still have the same need for MEMS motion sensors to detect crash events,” she said. “EVs will still use similar pressure sensors as internal combustion vehicles, just in a different way.”
Smartphones
Almost all smartphones use MEMS microphones. But most of the MEMS in phones are used for other functions. For example, accelerometers and gyros for motion sensing — to determine the orientation of the phone, to stabilize the camera imaging, to count steps.
According to Fitzgerald, “A big use of MEMS that a lot of lay people don’t know about, also in smartphones, is for RF filters. For your phone to be able to communicate with a cell tower here or when you get off a plane in Germany, there are a couple of dozen RF filters that selectively tune your phone to the different cell tower frequencies.”
These MEMS devices are acoustic wave filters. They are each designed to resonate at a specific frequency and operate by generating acoustic waves inside or on top of a thin film piezoelectric layer when an RF signal is applied. Only frequencies that match the resonant conditions of the layered structure are allowed to pass, while all other frequencies are attenuated. They are very effective at precisely isolating specific frequencies and rejecting unwanted adjacent signals — critical for crowded RF environments.
What’s Next — Active Noise Cancellation Earbuds
According to Fitzgerald, an emerging high-volume application for MEMS technology is speakers and microphones in earbuds. MEMS speakers are much thinner than electrodynamic coil and magnet speakers and require much less power. Because of that, they can be readily integrated into the small space in an earbud, leaving room for a MEMS microphone and an electronics package for high-quality active noise-cancellation. The next step will be to add artificial intelligence to enable earbuds to incorporate a variety of advanced features. And it will be suited to large-scale production because the MEMS are manufactured on silicon wafers with same techniques as regular electronic chips.
What’s Next — Ultrasonics
Ultrasound has been around for a long time, but MEMS is slowly but surely replacing the traditional technology. With traditional medical ultrasound, the wand that the doctor puts against your body uses a chunk of a piezoelectric crystal that has been machined to form the head. However, MEMS can do far better, said Fitzgerald.
You can think of the MEMS transducers as little diaphragms pushing out ultrasonic waves. Because the MEMS emitters are so tiny, in the range of microns, you can organize them into elements of a phased array, which can be tuned electronically to a range of frequencies.
So, instead of a large cart costing upwards of $100,000, “I think one thing that’s very exciting about MEMS-based ultrasound technology is that you will be able to use your phone as the video display and the transducer head will be small enough to be carried in a medical provider’s pocket,” said Fitzgerald. “And they might eventually cost 100 times less than the big cart. With the advent of AI for image interpretation, the dream is also that instead of needing a trained sonographer, you could have an EMT or ultimately in the future, perhaps a home health device, just like the thermometer or blood pressure cuff.
“A couple of companies are now selling them, so I think we’ll start to see the handheld ones becoming more common in medical settings. The AI version — that’s probably still 5 or more years away.”
And the uses of ultrasound are not just medical. For example, MEMS-based ultrasound is used in industry for precision distance and position monitoring, for detecting internal faults in structural materials, and for gesture recognition.
There is some work being done now on using ultrasound as a data link. You could use ultrasound to do point-to-point line of sight communication in place of some Wi-Fi functions. It would be harder to eavesdrop. Radio emissions are omnidirectional, while ultrasound would be focused point to point, so it would be more secure.
Also, for people who wear electronic devices for their health, for example pacemakers or insulin pumps, instead of sending the signals to your phone and then to the internet and then back to the phone and then back to the body, you could have an ultrasonic network, which is more secure, communicating between your pacemaker or your insulin pump and sensors in your body.
What’s Next — Biotech
Another place where MEMS have been really useful is in biotech. People doing protein or DNA analysis and drug discovery need disposable chips that are a combination of electronics and microfluidics for them to do molecular-scale analysis in diagnostic as well as research settings.
“The consumer market interacted with some of this technology during the COVID pandemic. Pharmacy PCR detectors could take your nasal swab and within 10 minutes determine not only if you had COVID, but which type. Underlying that was MEMS technology. That technology is also core to what’s going on inside hospital labs and biotech labs. A huge number of consumable chips are needed to do that kind of analysis,” said Fitzgerald.
With MEMS microfluidics, you push nanoliter, even picoliter, droplets of liquids through micro scale channels that have little reaction sites. The results of the reaction can then be measured either optically or electronically. CMOS electronics can be integrated on the same chip for electronic signal readout.
According to Fitzgerald, “The three most likely new commercial growth areas for MEMS in the next five years, are the ones I described: microspeakers, ultrasonic imaging, and biotech chips.
Specialty and Emerging Markets
There are MEMS applications with smaller markets, but important uses. For example, medical and surgical instruments, scientific equipment, laser systems, and orbiting satellites.
“An example would be a precision pressure sensor used in a cardiology guidewire for a specific cardiac procedure. There might be 200,000 of those instruments made per year. which might seem like a large number, but in the electronics business, that’s minuscule,” said Fitzgerald.
As the number and size of data centers expands at a high rate, the usage of electrical power is growing at an alarming rate. One way to mitigate the problem is to reduce the power consumption. Using fiber optic interconnects between racks instead of copper wiring will eliminate the joule losses, which waste a significant amount of power. And switching the fiber-optic connections can be done by optical circuit switches. The ability of MEMS to produce nanometer- or micron-scale motion can enable them to do effective optical switching. Higher bandwidth and optical wavelength-division multiplexing will enable massive data throughput, which will improve efficiency.
A potential turning point in the manufacturing of low volume, specialty MEMS is on the horizon: 3D printing. Until recently, 3D printers were producing millimeter-scale dimensions, but now they are pushing down into the micron range. Previously you could only make things that tiny using photolithography. That brings 3D printing into the realm of MEMS dimensions. Being able to directly print MEMS devices would bring down manufacturing costs of specialty products significantly. But 3D printing could never compete with silicon wafer manufacturing for high volume applications.
“The companies making billions of MEMS chips need to stay with silicon wafers because they are a mass production technology. However, companies like ours who are serving specialty markets, will probably, in 10 years, be 3D printing MEMS for customers who only need 1000 or fewer MEMS units per year. A silicon wafer process is very, very expensive to bring up for someone who only needs 1000 chips a year. So micro-scale 3D printing could serve the specialty MEMS markets very well,” said Fitzgerald.
“MEMS, which started off on wafer technology, is now splitting into three different manufacturing pathways. Silicon wafer-based manufacturing is going to continue to evolve incrementally. But now two side branches are forming. One will be 3D printed MEMS. And then the other I think will be MEMS that I would call the goodenough sensors. They will be done on paper, plastic, maybe even fabrics, and might be biodegradable. They will be ultra-low cost because they’re made on very cheap substrates. These two side branches, the 3D printing and the paper and plastic, are still in very early stages, but I think we’re going to start to see commercial products from them in 10 plus years.”
For more information, visit here .
