Abstract

This paper will focus on current and future automotive applications of MEMS (Microelectro-mechanical Systems), and MST (Microsystem Technologies). Current product developments in pressure sensors, accelerometers, angular rate sensors, and other MEMS/MST devices are presented. Market figures for automotive electronics and automotive sensors from 2000-2005 are given. The migration from traditional electromechanical and discrete sensor solutions to those embodying MEMS/MST is addressed.

Introduction

Microelectromechanical Systems (MEMS) and Microsystem Technologies (MST) have played a role in automotive engine control in the form of MAP (Manifold Absolute Pressure) sensors since 1979[1]. Today, many automobiles use one of these devices in their electronic engine control system (EECS).

Early 1990's vehicles saw the first silicon accelerometer for an airbag crash sensor application. These devices have obtained extensive coverage in the popular and trade press. However, with the exception of these two applications, many major production vehicle systems do not yet include MEMS/MST. The next five-to-seven years will provide significant opportunities for large volume production of MEMS/MST devices - whether it be in entirely new applications or in the replacement of traditional technologies. This movement is being driven by the factors including:

Longer warranty periods (up to 10 years/150,000 miles) require increased reliability components

Continuously developing federal and regional mandated fuel efficiency, emission and safety standards.

Higher vehicle performance and comfort

Availability of low-cost microcontrollers, memory and displays

Enhanced vehicle diagnostics

Market Figures

Worldwide vehicle production has grown from 47.5 million passenger vehicles, SUVs and light trucks in 2000 to a projected 54.0 million vehicles by the year 2005. The 2000 market for automotive electronics was $22.7 Billion (US) growing to an expected $30.9 Billion (US) in 2005, representing an average growth rate of 6.3%. This equates to an average electronic content per vehicle of $477 (US) in 2000 to $572 (US) in 2005.

It has been reported that the automotive sensor market to be worth $6.17 Billion (US) (904 million units) in 2000 and will grow at a compounded annual growth rate (CAGR) of 6.5% in dollars and 7.0% in units to $8.45 Billion (US) (1,268 billion units) by 2005 [2]. North America accounts for 47% of the year 2000 market for automotive sensors, followed by Europe (26%), Japan, (22%), and then South Korea (5%).

Speed and position sensors accounted for 38% of the 2000 dollar value of total automotive sensors followed by oxygen sensors (20%), mass airflow (13%), acceleration (11%), pressure (10%), temperature (5%), and others (3%).

The major growth areas for sensors in the 2000-2005 time frame is expected in accelerometers for vehicle dynamic control and airbags, pressure sensors for transmission, brake, diesel fuel rail, refrigeration, tire, fuel evap; yaw rate for vehicle dynamic control, roll-over and GPS backup; position sensors for wheel speed, camshaft, crankshaft, pedal position; humidity sensors for cabin comfort control; sunlight and rain/moisture sensing; and distance sensors for near obstacle detection and collision avoidance.

MEMS/MST technologies are currently satisfying or have a great opportunity to fill many of these applications and are expected to constitute a greater share of the automotive sensor market in 2005 versus its current value. A recent Roger Grace Associates/Nexus study has estimated that the 2000 sales of Automotive MEMS/MST to grow from $1.75 Billion (US) to $2.27 Billion (US) by 2005, which constitutes a 16.9% compound annual growth rate. The total MEMS/MST market is estimated to grow from $14.11 Billion (US) in 2000 to $36.22 Billion (US) by 2005, constituting a CAGR of 20.1% [3].

AUTOMOTIVE REQUIREMENTS

A number of stringent performance environmental, reliability, and cost requirements are imposed on automotive components. The harsh underhood automotive environment includes extreme temperature, shock, vibration, humidity, corrosive media, EMI, RFI and a host of other environments.

In addition, automotive components must be able to be produced in extremely large volumes, typically one million or more units per year. This is necessary not only from a vehicle demand point of view, but from the necessity to recoup the large investment associated with fixed design and manufacturing costs. Operating lifetimes of up to 10 years/ 150,000 miles, and very low unit prices are also required. Essentially, we can consider automotive components to require the ruggedness of military parts with the price of consumer products. These qualities are inherent in MEMS/ MST.

Component cost is a significant factor in the selection criteria of automotive system designers. The total cost of a sensor/actuator frequently includes the MEMS/MST device, signal conditioning electronics (e.g., temperature compensation, filtering, amplification), package, connector/cable harness, and testing. As a result, the cost of the MEMS device itself constitutes a one third or less of the total delivered component cost. Therefore, a significant challenge to achieve highly efficient design for manufacturability and testing is imposed on all suppliers who wish to be successful in the automotive sector.

UNIQUE MEMS/MST FEATURES

MEMS/MST are well suited for a wide variety of automotive applications. Due to their batch process manufacturing, large volumes of highly uniform devices can be created at relatively low unit cost. Since MEMS/MST have virtually no moving parts to wear out, they are extremely reliable. Silicon has provided itself as a material for sensors in many applications over the last 20 years in applications including military, consumer and automotive.

With the advent of microprocessor compatibility imposed on many automotive sensor/actuator applications, silicon is uniquely qualified to provide high levels of monolithic vertical functional integration using popular semiconductor and classical micromachining processing.

Migration From Electromechanical Technology to MEMS/MST

MEMS/MST offer the major advantages of cost and performance to automotive electronic systems. There is ample history of the migration of electromechanical sensors and discrete switches to MEMS/MST based sensors. Figure 1 is a summary of this activity.

The early MAP sensors have been virtually replaced by a MEMS/MST approach starting in 1979. Mass Airflow sensors, previously implemented in discrete anemometer wire technology by Hitachi, are being threatened by MEMS/MST technology as demonstrated recently by Bosch's approach. We can cite yet another migration, this one being extremely large-scale, in the transition from variable reluctance (VR) wheel speed sensing for antilock breaking systems (ABS) and vehicle dynamic control (VDC) to Hall-Effect (HE) and possibly to anisotropic magneto resistive ratio (AMR) and giant magneto resistive ratio (GMR) implementations.

Air bag accelerometers have migrated from the Breed "Ball and Tube" and TRW Teknar "Rollamite" switch solutions to the MEMS/MST solution currently being provided by Analog Devices, Motorola (Figure 2), SensoNor, and Nippondenso.

Figure 2: Courtesy of Motorola

Pressure sensors, especially those dealing with harsh media e.g. engine oil, radiator coolant, which have previously been implemented using a ceramic capacitive, discrete pressure switch or other electromechanical approaches are in the process of being replaced by MEMS/MST approaches provided by Keller, Measurement Specialties, SSI Technologies, Fasco, and Integrated Sensor Solutions (ISS)/Texas Instruments. Here, either bonded/fused silicon strain gauges affixed to low cost and rugged packages or piezoresistive chips housed in silicone oil-filled reservoirs capped by a stainless steel diaphragm are expected to take over these existing and new opportunities.

System Applications

We have reviewed the use of MEMS/MST in the following automotive systems:

Safety

Engine/Drive Train

Comfort, Convenience, Security, Engine/ Drive Train

Vehicle Diagnostics/Monitoring

For each system, specific applications are noted (e.g., Digital Engine Control Fuel Level) and the status of the specific application is given (i.e., future, limited production, major production). Also noted is the opportunity afforded to a MEMS/MST solution versus that using another technology (e.g., piezoelectric, hall effect). The appropriateness criteria for MEMS/MST selection was based on detailed market research initially conducted by the author and reported in Reference [1], and updated periodically and as recently as October 2000 during the semi-annual Convergence conference in Detroit, Michigan. The most significant application opportunities will be addressed.

Safety

A summary of MEMS/MST safety application opportunities in automotive safety systems is given in Figure 3.

Airbag actuation is currently and will continue to be a major application of MEMS. Silicon accelerometers in the 50g range are currently being supplied to U.S. vehicles primarily by Analog Devices, GM/Delco, and Motorola. Sensonor (Norway) has historically provided a significant amount of silicon accelerometers to the European market while Nippondenso is the major provider of accelerometers to the Japanese market. With the exception of the Analog Devices approach, all of these accelerometers currently use a multiple-chip solution to sense and provide the appropriate signal conditioning. The Analog Devices unit is a fully integrated monolithic device. A number of the other manufacturers are currently evaluating this approach.

A number of manufacturers have investigated the use of compressed gas as a means to replace/ supplement the sodium hazide explosive approach to airbag deployment. The use of a pressure sensor to monitor gas cylinder pressure was being actively investigated. However, this approach has not been adopted.

Suspension systems have been configured to provide the driver with optimum vehicle performance in high speed cornering, rough roads, sudden braking and acceleration. Numerous systems have been configured using total closed loop control of the suspension system. The fully active systems are extremely expensive ($2500-$4000), consume significant horsepower to operate the hydraulic pump and add considerable weight to the vehicle. The enhanced performance attained by these systems has been marginal as compared to their cost. As a result, their implementation has been and will continue to be extremely limited in large scale production vehicles. However, numerous suppliers have introduced "semi-active" systems. Some of these systems use displacement sensors in the shock absorbers and use a number of linear accelerometers. This application is ideal for MEMS/MST and companies including Analog Devices and Motorola are aggressively pursuing this opportunity with ±2g designs.

Silicon pressure sensors are currently being used in master cylinder brake pressure starting with the 1995 S-class Mercedes. In addition to an angular rate sensor, accelerometers, steering wheel angle and wheel speed sensors are being used for vehicle dynamic control. Currently, most other vehicles only use wheel speed sensors for their vehicle dynamic control (ABS, traction control). Since the unit selling price of the existing angular rate gyros similar to Bosch's surface micromachined [4], Systron Donner's or Matsushita's tuning fork [5] are approximately $25, its implementation is relegated to the top of the line model vehicles, e.g. Mercedes S Class, BMW, Cadillac. Currently, much work is being undertaken by vehicle manufacturers and first tier suppliers (e.g., Bosch, Lucas, Temic, Siemens) to configure systems that are more cost effective. The availability of a low-cost, MEMS/MST -based angular rate sensor similar to that developed by C.S. Draper Lab [6], G.M. [7], or Analog Devices, Bosch, and Systron Donner (Figure 4) is expected to propel the adoption of these enhanced systems into less expensive vehicles over a period of time.

Figure 4: Courtesy of Analog Devices

Current navigation system designs use a combination of global positioning satellites (GPS) and CD ROM maps in addition to wheel rotation sensors and rate gyros, or magnetic compasses. Here, MEMS/MST devices can be put to good uses.

Again, the current cost of these systems pre-empts their widespread use. These systems are currently offered as an option and cost in the $1800 range. System manufacturers have a cost target of $900 in the near future and $500 as a long-term target.

Comfort, Convenience and Security

A summary of MEMS/MST application opportunities in automotive comfort, convenience and security systems is given in Figure 5.

The measurement of compressor pressure in the vehicle air conditioning system offers a major opportunity for MEMS/MST. Currently, other technologies (e.g., Texas Instrument ceramic capacative pressure sensor) are being used. Major developments by a number of MEMS/MST companies are actively pursuing this very large opportunity, including Measurement Specialties which has recently formed a strategic alliance with Texas Instruments. Keller, Fasco, and ISS/Texas Instruments are also pursuing this opportunity.

Engine/Drive Train

A summary of MEMS/MST application opportunities in automotive engine/drive train systems is given in Figure 6.

 Electronic engine control has historically been and is expected to be a major application area of MEMS/MST in automotive applications. Silicon manifold absolute pressure (MAP) sensors are produced by the millions by Delco, Motorola, and Bosch (Figure 7). These devices provide an inferred value of air-to-fuel ratio by measuring intake manifold pressure. A great deal of effort has been undertaken to replace these devices with mass airflow (MAF) devices. Currently available on the market are discrete hot wire anemometer devices (e.g., Hitachi). Because of their construction, they tend to be large and expensive. A thin film equivalent of this device was introduced by Bosch in 1995. A MEMS/MST version of this device is currently under evaluation by a number of organizations. In addition to the MAP/MAF devices, barometric pressure values are needed to provide the engine controller with altitude information to compensate a rich/lean fuel-to-air mixture. MEMS/MST devices are well suited for this application.

Figure 7: Courtesy of Bosch

Cylinder pressure values are of great importance to optimize engine performance; however, due to the extreme high temperature levels, piezoelectric and fiber-optic techniques such as that developed by Optran provide a much more pragmatic solution to this application; however, at this time cost issues preempt their introduction. Exhaust gas recirculation (EGR) applications exist in Ford and Chrysler systems. Ceramic capacative pressure sensorare being replaced by silicon piezoresistive solutions.

Continuously variable transmission (CVT) applications require pressure measurements in hydraulic fluids. MEMS/MST devices which are isolated from the media using various techniques (e.g., isolated diaphragms) could find widespread application. Fasco, Measurement Specialties, Integrated Sensor Solutions (recently acquired by Texas Instruments), and SSI Technologies are developing possible solutions to this application. All of these approaches use a sensor plus silicon CMOS ASIC hybrid, typically electronically programmed using EEPROMS.

The only known application of a MEMS device in a mechanical structure is in fuel injector nozzles. Here, Ford has used micromachined silicon to create highly uniform and rectangular orifices for fuel injection systems. Over 3 million of these devices were manufactured; however, they are not currently in production.

Vehicle Diagnostics/Monitoring

A summary of MEMS application opportunities in automotive vehicle diagnostic/ monitoring is given in Figure 8.

One of the more interesting applications for MEMS/MST is in tire pressure monitoring. For both safety and optimized fuel performance, proper tire inflation is necessary. A number of systems are currently being offered that provide real time measurement of tire pressure. MEMS devices are ideally suited and are being considered by a number of their manufacturers for this purpose. With the favorable acceptance of run-flat tires, these systems became very popular by the model year 2000 vehicles. Run-flat tires (e.g., Michelin 60-series) eliminate the cost and weight of a spare and jack. Lucas NovaSensor, Motorola, and Sensonor are currently pursuing this application, which has become a major opportunity as a result of the recent Clinton Administration's edict that all passenger vehicles operating in the US be equippped with these devices by 2004.

Engine oil monitoring is a huge opportunity for MEMS/MST. The greatest barrier to the adoption of these systems is price. These pressure sensors must be able to survive the elevated temperature requirements of engine oil and isolate the silicon chip from the media. Price target for this application is in the $5-$7 range for a fully signal conditioned, packaged device. Numerous sensor manufacturers are aggressively pursuing this significant application opportunity.

Recent legislation has created a major opportunity for pressure sensors in the evaporative fuel system. In this application, a pressure sensor is used to monitor the pressure level in the fuel tank and insure that no fuel vapor escapes.

Conclusions

Automotive applications of MEMS/MST are expected to continue to constitute a significant part of the MEMS/MST market by the year 2005. It is apparent that there has been and continues to be a major migration of technology solutions from the electromechanical to MEMS/MST in many automotive applications. Today, MAP and airbag acceleration applications are almost entirely being served by MEMS/MST. There are a number of applications that are currently being considered or are in early production as candidates for MEMS/MST, including wheel speed sensing, refrigeration pressure, engine oil pressure, and brake pressure.

In addition, many new applications are being considered for sensor integration into newly developed systems. The cost, reliability, and size of MEMS/MST based solutions makes them the technology of choice. MEMS/MST devices have accumulatively logged many millions of operating hours in automotive applications. Their reliability has been proven.

We expect the introduction of MEMS/MST based devices to proliforate in automotive as well as other applicatoins in the next five to seven years. The major barrier to the introduciton of these devices has been cost - not the cost of the device, but rather the total cost of the device based solution, i.e. package, connector, testing. As low cost packaging developments and high-volume, low-cost testing matures, the success of MEMS/MST based solutions will be insured.

Acknowledgment

The author would like to thank Mr. Joe Giachino, formerly of the Ford Motor Company, and Mr. Bob Sulouff of Analog Devices, and Mr. Steve Hendry of Motorla for their helpful comments; and to Mr. Chris Webber of Strategy Analytics for access to his automotive sensor market data.

References

R.H. Grace, "Semiconductor Sensors and Microstructures in Automotive Applications, Sensors and Actuators", Society of Automotive Engineers International Conference Proceedings, 1991, pp.245-260

C. Webber, "Automotive Sensor Market Trends", Presented at 1999 Sensors Marketing Conference, Strategy Analytics, San Diego, CA, Feb. 1999

R.H. Grace, "Commercialization Issues of MEMS/MST/Micromachines", Sensors Expo Proceedings, Detroit, Michigan, September 2000.

R. Schelling, "A Low-Cost Angular Rate Sensor for Automotive Applications in Surface Micromachining Technology", Third Annual International Conference on Advanced Microsystems for Automotive Applications Proceedings, March 1999.

I. Toshihiko, J. Terada, "Angular Rate Sensor for Automotive Application," Sensors and Actuators, Society of Automotive Engineers International Conference Proceedings, 1995, pp. 49-56

N. Weinberg et. al., "Micromachined Comb Drive Tuning Fork Gyroscope for Commercial Applications," Sensors Expo Proceedings, Sept., 1994, pp. 187-194

J. Johnson, S. Zarabaldi, D. Sparks, "Surface Micromachined Angular Rate Sensor, Sensors and Actuators", Society of Automotive Engineers International Conference Proceedings, 1995, pp. 77-83

M. Mattes, J. Seefeldt, "A One Chip, Polysilicon, Surface Micromachined Pressure Sensor with Integrated CMOS Signal Conditioning Electronics," Society of Automotive Engineers International Conference Proceedings, 1996, pp. 29-34.