Often, a technological breakthrough can lead to new products and even new applications that had previously been considered fantasy. Developments in low-power, low-noise readout ASICs could open new opportunities for microelectromechanical-systems (MEMS) sensors, particularly in medical monitoring, implantable devices, and military/sports monitoring.
MEMS are made of components between 1 and 100 µm, fabricated using modified IC manufacturing techniques. They have been available for more than 20 years. However, advances in manufacturing techniques and requirements for the low-cost systems that these sensors can provide have led to accelerated adoption of the technology over the last few years.
MEMS can be manufactured from silicon, quartz, and other glasses with different additives. They can be used to measure many different parameters, including acceleration, pressure, angular velocity, flow, and force. They even perform chemical analysis.
Many automobiles contain airbags activated using MEMS accelerometers. Many use MEMS inertial sensors to monitor the movement of the car to, for example, detect lane departure. Your GPS system may include a MEMS pressure sensor to improve its accuracy. The physical sensing mechanism for detection can be a change in resistance, piezoelectric voltage, resonant frequency, optical reflection, or capacitance.
Capacitive MEMS-based sensors offer advantages due to reduced cost, a simple structure, and the ability to integrate the sensor close to the readout electronics. Also worth noting is the fact that capacitive sensors are more suited to low-power applications because of their near-zero power requirement (at dc) and their lack of Johnson noise.
However, physically small sensors result in changes in capacitance that are equally small. This places exacting requirements on the readout circuit. For instance, a typical MEMS accelerometer will show a change of around 200 fF per g, where g is the acceleration due to gravity.
For many accelerometer applications, the useful information is at milli-g levels. For example, accelerometers can be used to measure tilt. However, obtaining a resolution of less than a degree requires a resolution of the order of 1milli- g.
Hence, the noise performance should be 1000 times less than the capacitance change, which means a noise of 200 aF or around 30 aF rms. (Note that 1 fF = 1 x 10–15 F and 1 aF = 1 x 10–18 F.) This in itself is a large technical challenge that few companies supplying this sort of component are able to achieve.
If we then consider the requirements for portable battery-powered devices, with a typical battery capacity of 100 mAh, we find that a supply current of 750 µA yields a battery lifetime of six days; a supply current of 25 µA yields a battery lifetime of 166 days; and a supply current of 2 µA yields a battery lifetime of nearly six years.
To have any real use, the device should be autonomous for a reasonable amount of time dependent upon its accessibility. Implantable medical devices must have an autonomous lifetime of at least five years. So, it is clear that low-power consumption is required. Many products demonstrating these properties are now available.
Noise performance as low as 20 aF rms is possible with a continuous operating current of 25 µA. A supply current of 2 µA is even possible in periodic mode. One product, when used with a capacitive sensor, can detect pressure differences equivalent to a 10-cm change in altitude.
Of course, the additional communication and processing technology is also necessary. This is provided using a high-performance, low-power, 32-bit processor core. A 32-bit core is important since linearization of the 22-bit raw data is necessary to achieve the required precision.
One can envision tiny portable vibration detectors containing an RF interface that could be used in a security situation or in a military environment to monitor access and activity. Such systems could be placed easily and operate over long periods of time. Data entry and equipment control might also be performed in the field using movement and gestures detected by accelerometers.
Sports monitors are already becoming available, fitted onto sport shoes or attached to laces. They will now be available for longer operation. For example, a monitoring accelerometer within ski bindings that causes them to release safely when required is now possible. Performance-monitoring wrist watches will be available not only to monitor movement, but also other body parameters such as blood glucose and cholesterol. Clothing will also be manufactured to detect tiny pressure changes useful in virtual reality scenarios. Measurement of body gait and posture could reduce common complaints, such as backache, as well.
In the medical field, implantable accelerometers are already available in pacemaker applications for adjusting the frequency of stimulations. Other measurements, such as ventricular pressure, may become possible.
Internal tracking of tumors using inertial capacitive-based sensors is another real application undergoing trials. Developments of such devices could be inserted into other organs to aid surgery or radiation therapy. Sensors will also be implanted in the surgical and neurological tools to report on a surgeon’s progress and technique.
Capacitive sensors can be designed as chemical sensors by using an absorbent dielectric or absorbent electrode that changes mass and moves when the chemical is absorbed. This opens their use to measurement of chemicals within the body.
Multi-sensor patches that measure several body parameters will become available. In fact, the whole drug delivery system—measurement, control and drug delivery—could be incorporated into a tiny portable instrument.
Furthermore, the integration of capacitive chemical sensors and a low-power ASIC will enable highly portable environmental detectors. They could be placed in any situation to monitor for environmental incidents. Equally, low-power inertial and vibration sensors could be mounted within buildings to monitor structural movements, which will aid and reduce the cost of maintenance and inspection. Their use could be of particular interest in zones of high seismic activity. Low-power vibration sensors could also be used to monitor misuse of equipment to reduce warranty liability.
Finally, when supply requirements are this low, the possibility to completely power the device using energy-scavenging means becomes possible. Energy-harvesting techniques that convert vibration, heat or light into useable energy will become more prevalent. In many cases, the requirement for batteries will disappear.
Andrew Glascott-Jones is the applications engineer for e2v’s mixed-signal ASICs business unit in Grenoble, France. He brings 25 years of experience to his position as an applications engineer within e2v’s mixed-signal ASICs business unit. His specific area of expertise lies in the design of electronic measuring systems, including precision metrology, particle sizing, X-ray imaging and laser spectroscopy.