Nikkei Electronics Asia -- October 2008
Insights
MEMS Sensors Proliferate in Passive Safety Applications

Every year, globally more than 1.2 million people are killed on the roads. Those numbers are forcing governments to react and take more initiatives, moving from passive to active and predictive safety. The airbag remains the predominant system for saving lives. Most of the new cars produced today are equipped with passenger and front airbags as standard. According to market reports, the total number of airbag sensors in 2010 will reach 205 million units, with new features fueling the growth, such as side bags for head, chest and knee protection, and rollover and smart passenger detection systems. Government regulations or strict independent crash testing organizations like EuroNCAP are forcing car manufacturers to constantly introduce better equipped vehicles to pass their stringent safety tests. To answer this never-ending demand for increased safety, the number of acceleration sensors used to detect a crash has increased to up to six accelerometers in high-end vehicles, and the number still grows.

Designers of such components are forced to use new materials and alternative technologies to keep the overall airbag system cost competitive.

Passive, Active Safety Fusion

One way to keep costs under control is to reduce the number of electronic control units (ECU) in the car. Merging the airbag and electronic stability control (ESC) modules together is one approach. Several approaches are followed by automotive companies: one involves integrating the airbag ECU in the VDC module under the hood close to the hydraulic block; another involves combining ESC sensors (low g and gyro) in the airbag module in the central tunnel of the car. The first cars using one of these approaches should hit the road after 2010 (Fig 1). 

This creates new challenges for the sensor technologies based on micro-electromechanical systems (MEMS).

They have made significant technical steps in recent years, often leveraged by the continuing improvement in semiconductor processing. But the merging of airbag and ESC will impact the sensor characteristics and will require a new level of integration. Multi-axis low g /med g sensing elements will have to be developed, which require new expertise for simulation and modeling, package interaction and processing power.

Transducer Technology

One of the most common methods used to sense acceleration is to measure the displacement of a movable seismic mass which is translated into a variable capacitance measurement. The new generation of accelerometers feature a multichip approach based on a surface micromachined capacitive sensing element and a control ASIC for the signal conditioning (conversion, amplification and filtering). By keeping the MEMS transducer separated from the control IC, putting two die in a plastic package produces significant advantages as it allows the use of technologies that are not "friendly" with each other in terms of processing techniques and development cycle time. Most of the production today uses a poly-silicon sensing element 3um in height and a conventional 1.2um CMOS ASIC.

In safety applications, it is very important for the accelerometer to provide a fast response time. Determining when to fire airbags has to be done quickly, so accelerometers must react instantaneously. Using a transducer design which gets to a steady-state rapidly instead of oscillating improves the device's response time. Furthermore, by definition, an inertial sensor is highly sensitive to acceleration of any origin and the airbag algorithm needs to "recognize" in the data sent from the sensor during an accident only the right crash signature. That is why the sensor output signal is typically cleaned of parasitic high frequencies via electronic low pass filtering (400Hz). A sensor which can eliminate unwanted high frequency acceleration content directly at the sensing element level provides additional benefits.

An innovative transducer technology (Fig 2) has been developed to focus on increasing the thickness of the structural layer to improve performance. Since the height of the movable structure is much larger than the spacing and widths, the technology is known as HARMEMS (high aspect ratio MEMS) where the ratio is between air gap and trench deepness. The technology delivers this high aspect ratio by a combination of a >20um thick SOI layer and <1.5um narrow trenches defined by deep reactive ion etching (DRIE).

The high aspect ratio of the technology, combined with higher-than-vacuum hermetic sealing provides an overdamped mechanical response. In Fig 3, the HARMEMS mechanical response is compared with a 3um underdamped poly-silicon MEMS (Poly-MEMS) device in production today. The Poly-MEMS device is excited to resonance (for this design, above 10kHz). By contrast, the HARMEMS device exhibits no resonance, but rather a cut-off frequency below 1kHz.

(Continue to the next page)

MOST POPULAR ARTICLES

24 Hours | 7 Days