As a young man fresh from engineering school, I began my career on a project team designing a 500 MW electric power generating station. Because the contracting public utility specified a pressurized water reactor as the plant's energy source, myriad safety regulations governed the design process—from procurement of pumps, motors, and valves, to the location of control panels and access areas.
These strict requirements complicated an already complex job. "Why not build half a million power plants, one kW each?" joked a colleague. "That way, if there's a meltdown, who cares?"
My irreverent co-worker had a point—one that years later is becoming partly realized thanks to MEMS technology. Building small allows decentralized sensing and control, fast local response, and low power requirements. Semiconductor fabrication techniques permit batch manufacturing, in turn making many new products—from automotive airbags to ink jet printers to disposable blood pressure sensors—cost effective. The following examples represent just a few of the many new MEMS designs looking for a home.
Non-contact angle measurement
Unterpremstätten, Austria—Able to resolve as many as 64 points around the face of a diametrically magnetized cylindrical magnet rotating at speeds reaching 30,000 rpm, this contactless position encoder measures a mere 4 × 5 mm—the size of a standard SOIC-8 package.
Developed by austriamicrosystems AG, the AS5020 Magnetic Angular Position Encoder places an array of integrated Hall effect sensors, signal conditioning circuits, and post processing capability on a single ASIC. The ASIC, fabricated from low-cost, sub-micron CMOS technology, is located above or below the tiny two-pole permanent magnet which functions as the coding source.
During operation, the Hall sensor array detects orientation of the magnetic field. Signal conditioning algorithms provide high linearity for the full temperature range of -40F up to 185F to achieve a position accuracy within ±1.5 degrees. Post-processing generates a 6-bit binary output compatible with any microcontroller-based system. There is no limit, furthermore, on the number of encoder units attached to a single microcontroller.
Absolute positional information is ac-cessed via the 3-wire synchronous serial interface of the ASIC. Multiple devices may be cascaded using a standard microcontroller daisy chain to eliminate wiring complexity and to reduce cost. Self-calibrated, the magnetic an-gular position encoder operates from a supply voltage of 4.5 to 5.5V dc, with an input current of typically 17 mA during active measurement.
Especially desirable in harsh environments such as chemicals or explosive gas, austriamicrosystems AS5020 represents the first generation of MST/MEMS-based magnetic encoder products designed to replace mechanical rotary switches and potentiometers. Future generations will accommodate other interfaces (parallel, USMB, CANbus, ASIbus), will have higher resolutions, and will also replace slide switches.
Applications include motor-speed and stop control, micro-motors, trackballs and joy sticks, conveyor belts, robotics, positioning controls for video and motorized toys, and program switches for household appliances.
Kyoto, Japan—Engineers at Omron Corp. have turned to MEMS techniques for development of a micromachined relay offering subminiature size and high-frequency signal-switching performance. Its operating principle is simple: a spring holds a movable electrode and contact above a fixed electrode with fixed contacts. Applying a voltage across the electrodes pulls the movable contact onto the fixed contacts to close the signal circuit. When voltage is cut, the spring pulls the movable contact away.
Above, an entire magnetic rotary position
encoder fits on a silicon die packaged into a standard SOIC-8.
brushless motors require rotor position feedback to commutate the coil
currents. Contactless MEMS-based devices can provide absolute position of
the rotor or the gear connected to it for motion control applications.
Distorting a single crystal silicon film provides the spring action. Etched into an H shape with broad legs, the silicon film helps form the movable electrode. Its contact mounts on the cross-piece. Two anchors, positioned at diagonally opposed corners of the H, sit on the assembly's glass base substrate, maintaining a gap between the movable and fixed electrodes. Slits etched part way around the anchors give the movable electrode flexibility.
To balance gap size between electrodes and the electrostatic voltage, while achieving the required restoration force, Omron engineers designed a special non-linear spring that develops greater restorative force than possible with a linear spring. Non-linearity is achieved by adding separators, or bumps, on the silicon substrate. These land on bonding pads on the fixed substrate before the contacts meet, causing additional distortion in the silicon layer and the non-linear effect.
For the actual formation of the device, Omron starts with a silicon-on-insulator (SOI) wafer and forms the movable electrode, including the basic H-shape and the anchors. An oxidized film for insulation is put down in the area of the movable contact and then the movable contact itself is formed of gold film. Separately, the parts of the fixed electrode are patterned on a glass substrate. The two parts are fit together and the remaining base of the SOI wafer is removed. The slits are then formed in the silicon layer to complete the formation of the actuator.
The micromachined devices have been successfully tested through 100 million cycles with acceptable insertion and isolation losses at up to 2 GHz. Kazuyuki Hayamizu, head of the MMR project, says the company is developing mass production techniques that will maintain the functionality of the MMR.
Fiber-optic pressure transducer
Sainte Foy, Quebec, Canada—The prescription for medical pressure sensors can be tough to fill. They must be ultra-small to fit within the tight confines of a catheter. And given the life-saving importance of some physiological pressure readings, they need to be accurate even when plagued by electrical interference from nearby surgical or diagnostic equipment. These disposable gauges also have to be cheap enough to throw away. Because of these design considerations, a fiber-optic pressure transducer developed by Fiso Technologies Inc. could be just what the doctor ordered.
Called POP-M, this new transducer targets a variety of invasive pressure-sensing applications—including intrauterine-, intracranial-, and blood-pressure monitoring. With respective tip and optical cable diameters of 0.55 mm and 0.25 mm, the FOP-M sensor fits easily within typical catheter housings. It provides pressure readings accurate to within 1 mmHg, reports Fiso President Claude Belleville. And when supplied in OEM quantities, the sensor costs about $15.
Distortion of a single crystal silicon
film provides spring action for Omron's micromachined relay. Voltage
applied across the electrodes pulls contacts together on Omron's
Like Fiso's earlier fiber-optic sensors for pressure and strain, the FOP-M operates on a "white light interferometric interrogation" principle. As Belleville explains, Fiso builds its sensors around a Fabry-Perot interferometer, a device that consists of two opposing mirrors separated by a cavity whose length changes in response to pressure. These pressure-induced variations in the cavity length modulate the resonant wavelength of broad-band light that bounces between the mirrors. Fiso applies patented technology to correlate the wavelengths to absolute cavity lengths, which in turn translate to specific pressures. The company creates these microelectromechanical sensors through a silicon micromachining process that can produce features as small as 15 microns.
The fiber optics and the interferometric operating principle could add up to several advantages over liquid-filled, piezoelectric, or intensity-based optical sensors. These include interchangeable probes and protection from EMI and RFI generated by electric scalpels, MRI machines, and other medical equipment.
San Mateo, California—Development of micro-scale biochemical processing devices, often called "biochips," promises rapid and inexpensive analysis while simultaneously improving the quality of many test results. There are two reasons for such optimism: One, many discrete operations can be integrated on a single chip; two, some biochemical and bioanalytic processes occur more rapidly and efficiently on the microscale.
Biochips promise improved clinical
Consider a biochip designed to determine the level of theophylline, a common Asthma drug, in blood serum. Developed by Coventor Inc., an independent provider of MEMS development software and product design services, the chip first combines theophylline with another stream of theophylline that has been fluorescently tagged. This mixture is then combined with an antibody specific to theophylline.
Applying a high voltage across the biochip draws a tiny amount of the mixture along a channel on the chip. Each of the chemicals in the mixture moves at a different rate in the channel, providing a means of separation. At the channel's end, the fluorescent signature of the marked theophylline is recorded.
Based on the intensity of the fluorescent signal, the amount of theophylline that has reacted with the antibody can be detected. This, in turn, indicates how much of the unmarked theophylline is present in the blood serum sample.
Such devices could allow similar tests to be performed closer to the point of care, as in a doctor's office rather than clinical laboratory.
Thermodynamic flow sensor
Delft, The Netherlands—Implementing multiple heaters and temperature sensors on a chip using MEMS technologies imparts high dynamic performance as well as an extended measurement range to this tiny sensor. Designed by TNO TPD to measure flow rate and direction for gases and liquids, the sensor works as follows:
Flow rate is determined by "inducing" a known amount of thermal energy into the fluid of interest. Surface temperature indicates flow rate. At higher flow rates, the fluid "absorbs" more thermal energy; at lower flow rates, it absorbs less. Temperature, consequently, is inversely proportional to flow rate.
Flow direction is measured by analyzing temperature distribution across the chip's surface. Model-based heat balance calculations give an accurate reading of both flow rate and direction. Compared to conventional thermodynamic flow sensors, the MEMS-based unit offers:
Extended measurement range. The unit measures flow rates from 0 to 100 m/sec, compared to the typical 0 to 25 m/sec.
Improved dynamic response. Because the MEMS flow sensor can monitor flow rate changes up to 80 Hz, it can be used in a pulsating flow or vibrating environment.
Compact design. No separate reference temperature sensor is needed. In addition, the sensor mounts flush in the wall of a pipe to minimize flow disturbance.
Longer life. Direct exposure to hazardous environments is permitted.
Finally, integration with a MEMS pressure sensor permits measurement of mass and volume flow. Potential applications include compressors, pumps, and turbines. Compliance with hygiene regulations makes it particularly suitable for the food and pharmaceutical industries.