The Future of Medical Microelectromechanical Systems

Robert S. Seeley

Will microelectromechanical systems (MEMS) for the medical industry reach the potential envisioned for them?

MEMS are micron-sized structures such as beams, cantilevers, diaphragms, valves, plates, and switches that can function as tiny sensors and actuators. They are fabricated by integrated circuit (IC) manufacturing processes; that is, by bulk and surface micromachining. Like ICs, thousands of micromachines can be fabricated on a single wafer with supporting circuits integrated on the chip. They can be mass-produced in the millions at low prices, which is their main appeal.

In bulk micromachining, most of the silicon wafer is etched away to leave the microstructures. The signal-processing circuitry that connects the chip to the larger device is not located on the chip. In surface micromachining, which produces features that are 20 times smaller than those made with bulk micromachining, the MEMS are deposited and etched in layers. This technology, which is also used for fabricating metal oxide semiconductors, allows signal processing to take place on the chip itself.

In the medical industry, micron-sized pressure transducers for disposable blood pressure sensors sell by the millions. For example, about 1 million intrauterine pressure sensors that monitor pressure around an infant's head during delivery and that incorporate micron-sized transducers are sold in a year. About 500,000 disposable angioplasty devices that are used to monitor pressure in balloon catheters and that use the transducers are sold per year. Kidney dialysis pressure sensors measure pressure across the membrane in the dialysis machine; models using the transducers sell at about 10,000 pieces per year.

Medical MEMS applications with sales of 10,000 units or less per year include endoscopy for measuring pressure in the stomach or other organs in which the endoscope is inserted. MEMS pressure sensors also monitor the blockage level in infusion pumps, and appear in sphygmomanometers and other noninvasive blood pressure monitors.

Perhaps the smallest device that incorporates MEMS transducers is Lucas NovaSensor's (Fremont, CA) intracardial catheter-tip blood pressure sensor, which is used for diagnostics during cardiac catheterization, and is only 0.15 x 0.4 x 0.9 mm.

Disposable Blood Pressure Sensors

The strongest MEMS success story to date is for disposable micromachined blood pressure sensors, which sell 17 million units per year. These $10 devices connect to a patient's IV line and monitor blood pressure through the IV solution. Introduced to the market in 1982, their principal manufacturers are Lucas NovaSensor, EG&G IC Sensors (Milpitas, CA), and Motorola (Phoenix, AZ).

Disposable blood pressure sensors replace reusable silicon-beam or quartz-capacitive pressure transducers that can cost as much as $600 and have to be sterilized and recalibrated for reuse. These expensive devices measure blood pressure with a saline-filled tube-and-diaphragm arrangement that has to be connected to an artery with a needle.

In the silicon MEMS blood pressure transducer, pressure corresponds to deflection of a micromachined diaphragm. A resistive element, a strain gauge, is ion-implanted on the thin silicon diaphragm. The piezoresistor changes output voltage with variations in pressure. Temperature compensation and calibration can be integrated in one sensor die.

The MEMS transducer senses blood pressure through a silicon-based dielectric gel between the sensor and the saline solution. The gel isolates the MEMS sensor and circuit from the saline solution, protecting the electric circuitry from the solution. The gel also protects the patient from currents straying back up the IV line. The gel is a nontoxic, nonallergenic polymer that passes pyrogen testing and meets USP requirements.

A close-up view of a Wheatstone bridge, part of a silicon chip used in airflow sensors. The sensors, which can be packaged to measure low flow rates and infer differential pressure measurements, are used in respiratory equipment. Photo courtesy of Honeywell, Micro Switch Div., Freeport, IL.

Packaging Challenges

The disposable blood pressure sensor's gel package illustrates the formidable packaging challenges that MEMS devices pose. Says Kurt Petersen, cofounder of Lucas NovaSensor, "In many real-world applications, the sensor is the easy part of the project; its packaging takes up most of the development resources." The package must enable these micron-sized devices to withstand the body's harsh environment, the potential for corrosion of metallic and ceramic surfaces, and device fouling from body materials.

MEMS devices have the advantage of silicon's toughness. On the micron scale, silicon combines tensile strength that is stronger than steel and has low density and high elasticity and hardness. But, says Dale Gee, Lucas NovaSensor director of new product programs, "In the physiological environment, etched silicon micromachined devices tend to be brittle, and the deposited circuitry tends to be delicate."

Also, because the devices are electronic, they must be unable to shock the patient. Developers must consider how the devices can function safely in the patient's fluid environment.

The issue of biocompatibility is equally daunting. Like all materials that are to be used on or in the human body, the devices must be benign. To ensure that they are benign, they must undergo full toxicity testing.

MEMS devices and the packages around them must also be able to withstand sterilization. "Radiation sterilization for MEMS devices is really tricky," says NovaSensor's Gee. "When you irradiate these silicon chips, you hope they don't shift." He says NovaSensor is developing its disposable blood pressure sensor and package to be radiation tolerant.

These disposable sensors use MEMS transducers to measure changes in blood pressure. Photo courtesy of Motorola, Sensor Products Div., Phoenix, AZ.

Micron-Sized Accelerometers

Pressure sensors are the largest medical MEMS market to date. A distant second medical application is accelerometer MEMS. EG&G IC Sensors fabricates accelerometer MEMS by suspending a micromachined silicon mass on double cantilever beams. Piezoresistors in the beams deliver electrical signals proportional to beam deflection when a motion offsets the mass.

Advanced pacemaker designs include an IC Sensors accelerometer MEMS that measures patient activity. When the patient moves, the accelerometer measures the motion and then signals the pacemaker to adjust its rate.

Most accelerometer applications remain under development. IC Sensors is interested in applying its accelerometer MEMS in patient activity—monitoring systems for sleep disorders and other conditions. Harold Joseph, director of sales and marketing at IC Sensors, envisions building an accelerometer into a data logger module that would fit into a patient's hand and record the patient's movements.

According to Steve Hendry, product marketing manager, Sensor Products Division of Motorola Semiconductor Products Sector, hospital-bed providers are also interested in accelerometers for critical-care hospital beds that would monitor patient position and the rate at which the patient rolls over.

Beyond accelerometers and pressure sensors, there are many ideas for future MEMS applications.

Blood-Analysis System on a Chip

MEMS makers are concentrating on developing micron-sized chemical sensors and fluid-flow devices. Motorola is working on silicon-based chemical sensors. Its first chemical sensor, Hendry says, will be a carbon monoxide detector for home use.

Kurt Petersen, of Lucas NovaSensor, has created a new company, Cepheid (San Jose), which he says is building biomedical test instruments in which micromachined fluidic chips are key components. He says that Cepheid intends to provide micromachining capabilities to large pharmaceutical or medical device companies.

These MEMS transducers are used in intercardial catheter-tip sensors for monitoring blood pressure during cardiac catheterization. Photo courtesy of Lucas NovaSensor, Fremont, CA.

Researchers at Stanford, MIT, other universities, and micromachining companies are working on developing microfluidic instruments small enough to fit on a chip, such as machines for electrophoresis, mass spectrometry, flow cytometry (measuring the properties of cell populations or any collection of small particles), and polymerase chain reaction (amplifying DNA).

A handheld blood chemistry-analysis system based on a chemical sensor chip, a sealed packet of calibration solution, and various fluid channels and chambers molded into plastic has been developed by i-Stat (Princeton, NJ). The instrument contains all the electronics that interface with the MEMS chip and a display and control keypad as well as the actuators to operate the fluidic components, and can eliminate the hours required for traditional laboratory analysis. The instrument automatically performs the calibration measurements, directs the blood sample over the sensor chip, records measurements of the sample, and displays the results on the handheld instrument.

When it comes to microvalves, "The Holy Grail is an implantable pump, which is affordable enough to place in large numbers of patients, and has a long enough life to benefit patient outcome," says Donald VerLee, associate research fellow at Abbott Laboratories, Hospital Products Division (Abbott Park, IL). The largest market for this pump would be among diabetics, for whom it could infuse insulin.

But microchemical sensors and fluid-flow devices again raise the problem of media isolation and packaging. How can the silicon on the device be isolated from the body fluid that the device is measuring? One proposed solution involves coating the fluid-contact surfaces to make them hydrophilic or hydrophobic.

Microfluid Dynamics

VerLee points out, "Normal fluid dynamics for macrostructures don't apply in structures this tiny. As you go smaller and smaller, physical forces come less into play, and chemical attraction becomes the dominant force; capillary pressure becomes the dominating pressure. So if the microchannels and chambers are hydrophilic, the fluid you're trying to push might become stuck in place by capillary action. Similarly, if you make the surfaces hydrophobic, then any dissolved gas that comes out of solution will bubble on these surfaces. You get vapor lock. The fluids won't move.

"The people designing MEMS are wrestling to understand all this, and design their structures so these issues don't become problems," he continues. "There are significant barriers, but smart people in the MEMS community are working on them."

Bright Future

In spite of these obstacles, officials at System Planning Corp. (SPC; Arlington, VA), a market research firm, are optimistic about the potential for developing MEMS microvalves in infusion pumps. The researchers predict that by the year 2000, there will be a $2.6 billion market for MEMS fluid regulation and control devices and 1 million microvalves for insulin infusion pumps will be sold annually.

The officials point out that there are many needed medical products that could include MEMS devices, such as sensors for implantable devices, glucose sensors for insulin drug pumps, physiological sensors for monitoring patients with chronic disease, total artificial heart—control systems, systems for patient monitoring, and rapid and automatic diagnostic blood-testing systems.

The main government promoter of MEMS is the Defense Department's Advanced Research Projects Agency. The agency, which sponsors 50 defense and commercial projects at universities and companies and has a $20 million budget, is currently supporting the development of a device incorporating MEMS that could monitor all major body systems such as blood pressure, temperature, oxygenation, and respiration. The device would expedite diagnosis and treatment, especially right after injury.

In the opinion of many industry observers, the outlook for MEMS is bright despite any technical barriers. SPC staff members predict an overall $14-billion market for MEMS devices by the year 2000, up from $1.4 billion today. The big markets include automotive, industrial, defense, medical, mass data storage, and optical switching.

However, VerLee cautions against some of the more fantastic predictions for this technology. For example, a micromachine cruising through arteries and routing out plaque would unfortunately not be feasible because it would probably shock the patient. "Some of the hype is not realistic," he says. "We need to separate it from reality. The hype can help, but can also hurt when it causes the researchers to lose their credibility."

Yet many in the medical device industry predict that MEMS, while perhaps not destined to achieve all the whimsical potential imagined for them, will dramatically change the manufacture of some medical devices. The tiny machines, which will aid in new medical achievements as well as lower the costs of existing devices and procedures, appear to be destined for a successful future.

Robert S. Seeley is a freelance writer specializing in advanced technology issues.