IMEC improves piezoelectric energy harvesters to drive vehicle health monitoring


by Jan Provoost, IMEC, Leuven, Belgium, and Rob van Schaijk, IMEC Holst Center, Eindhoven, The Netherlands

In the not-so-distant future, microsized autonomous wireless sensors may become standard components of our intelligent environment. Their miniature size and autonomy make them suited to continuously measure health parameters of the environment, machines and vehicles, or even the human body. One example where they could already be applied is tire pressure monitoring systems (TPMS) in vehicles, to increase the safety of cars while at the same time decreasing their fuel consumption.

Today, wireless sensor systems that are commercially available have a limited autonomy–they still need batteries. The sensors and wireless electronics become ever smaller and sophisticated, making them applicable in more environments, but the scaling of electrochemical batteries has run into technological restrictions. So, wireless sensors either need large batteries to give them a longer autonomy, which makes the sensor packages too big, or small batteries, which make the sensors less autonomous. For TPMS, large, heavy sensors or frequent battery replacement are not the best options.

One solution for the battery issue is to build microsized energy harvesting into the sensor modules. Energy harvesters take their energy from the environment in the form of vibration, light, or heat, and convert this energy into electricity. Sensors in machines, for example, could harvest energy from the continuous vibrations of the machine’s components. In TPMS, they can tap into the vibrations of the car’s tires.

Within the context of its activities at the Holst Centre, which it formed in 2005 with Dutch research centre TNO, IMEC has recently shown a new piezoelectric energy harvester for microsensors. The new harvester delivers an experimental output power of 60µW, a new record for micromachined energy harvesters, and enough to drive simple sensor systems that intermittently transfer sensor readings to a master, which would be the case for TPMS systems.

Micromachined energy harvesters can make wireless sensors autonomous

The Holst Centre research fits a longstanding ambition of IMEC: to create autonomous micropower solutions for wireless sensing. These micromodules consist of a sensor or actuator, a module to acquire and preprocess the signals, and a wireless radio to send them to a base station. Their autonomy will allow them to operate an indefinite time without battery recharge or connection to a power grid.

For some applications, such as sensing in highly accessible environments, microsensors may run on batteries. But for other applications, such as continuous machine monitoring, or monitoring in an environment that is difficult to access, batteries are not the best option. Depending on the type of battery, the autonomy of a 100µW module would be limited to a few months, or half a year maximum.

IMEC’s solution is to tackle the energy problem from both sides: consumption and generation. To reduce the energy consumption, IMEC is working on micromodules that run on a minimal amount of energy, with a goal of an average of 100µW. As for energy generation, IMEC looks into generating and storing power at the micro-scale to improve the autonomy of wireless autonomous modules. For generating energy, the choice is to develop micromachined energy harvesters and combine these with added energy storage, as backup when the harvester is not active or to handle peak loads when the harvester cannot generate enough power.

Figure 1: Piezoelectric energy harvesters.
Click here to enlarge image

Each form of energy harvesting has its characteristics and application for which it is best suited. Outdoor sensors, for example, may best be combined with photovoltaic cells, which can generate up to 10mW/cm2. Monitoring in machines, on the other hand, will require harvesting the vibrations or heat coming from the machines. Typical for machine components is that they vibrate at constant, predictable frequencies. Tapping into these vibrations could deliver 100µW/cm2, enough to drive the micropower devices that IMEC envisages.

How micromachined vibration harvesters work

Vibrational energy scavengers use electromagnetic, electrostatic, or piezoelectric conversion to generate electrical power. A microsized piezoelectric transducer (Figure 1) is simplest to design, and has so far shown the best result. It consists of a cantilever with one or several piezoelectric layers sandwiched between metallic electrodes forming a capacitor (Figure 2). At the tip of the cantilever, a seismic mass captures the vibrations of the machine to which the scavenger is attached.

Figure 2: Schematic of a piezoelectricenergy scavenger with cantilever and mass.
Click here to enlarge image

The vibrations of the machine cause the mass of the harvester to vibrate, which stretches the piezoelectric layer on the cantilever, generating a voltage across the piezoelectric capacitor. The generated energy is extracted by a resistive load.

These piezoelectric transducers have a resonance frequency that depends on their mass and stiffness of the cantilever. When the machine vibrations cause the transducer to vibrate at this frequency, the transducer will generate its maximum power (Figure 3). For best results, the machine vibrations and the resonance frequency of the harvester should match. This can be done by adapting the mass and cantilever stiffness to the environment in which the harvester will operate.

Autonomous wireless sensors reduce maintenance costs

Wireless microsensors powered by microsized vibration harvesters could, for example, be used to monitor the health of jet engines, train components, windmill propellers, or helicopter blades–monitoring the fatigue and wear of ball-bearings, rotating blades, gears, or stressed surfaces. For such equipment, the unexpected malfunction of rotating elements is a major cause of shutdowns and production losses. However, these components are notoriously difficult to monitor for defects, and in many cases the rotating components are difficult to access, making the use of batteries that have to be replaced regularly a bad option. There is thus a large market opportunity for wireless and autonomous sensors that reliably monitor component health and predict breakdowns.

Figure 3: Resonance curve for the AIN piezoelectric harvester.
Click here to enlarge image

One specific type of application is the tire pressure monitoring system (TPMS). In the near future, these sensors will become standard on new vehicles; they are already obligatory on all new cars in the US. TPMS systems measure and transmit the pressure of the tires of a vehicle a few times per hour. By alerting the drivers if the tire pressure drops, they increase the vehicle safety and decrease the fuel consumption.

There are several issues with current TPMS solutions, including their cost, weight, size, and lifetime. MEMS-based sensors in autonomous micro-power modules would solve many of the issues with current systems–they are ultralight, microsized, autonomous, and inexpensive. But the environment in which TPMS operate brings specific challenges. One is the vibration frequency of the wheels, which is dependent on the speed of the vehicle. A vibration harvester needs to be optimized to a vibration that is available at most speeds. Also, the modules must withstand shocks of several 100g which calls for extremely reliable sensors and components.

State-of-the-art micromachined vibration harvesters

IMEC’s new vibration harvester consists of a piezoelectric capacitor formed by a Pt electrode, an AIN piezoelectric layer, and a top Al electrode. It is fabricated in a silicon-based process using three wafers bonded by SU-8.

The resulting harvester delivers an experimental output power of 60µW. It weights only 34mg. The cantilever beam and mass are 6mm long, and the beam is only 5mm wide. The output power was measured at a resonance frequency of 500Hz and an acceleration of 2g.

Last year, IMEC showcased a piezoelectric harvester with a reported 40µW, but this device had a piezoelectric layer fabricated in PZT. The current AIN layer can be made in a simpler, standard CMOS-compatible deposition process, allowing production at a lower cost. Moreover, the PZT device operated at 1.8kHz; the lower resonance frequency (500Hz) of the new harvester corresponds with vibration frequencies in, for example, industrial equipment or car tires.

These state-of-the-art piezoelectric harvesters can still be improved along several lines. First, the fabrication process can be improved using SOI wafers. Second, a vacuum package should be designed to eliminate the effect of air damping of the cantilever movement. Third, the load should be optimized to have a maximum power output. Further into the future, vibration harvesters will be able to tune their resonance frequency, and optimum power output, to the application. Another option is to make broadband harvesters that generate power at a broad spectrum of vibrations.

With their 60µW output power, IMEC’s harvesters are already powerful enough to drive simple wireless sensors that intermittently transfer sensor readings to a master, which is the case in TPMS systems.

Jan Provoost is scientific editor at IMEC (Leuven, Belgium), Europe’s largest independent research center in microelectronics. Email:
Rob Van Schaijk is principal researcher and activity leader of IMEC’s micropower program at the Holst Centre (Eindhoven, The Netherlands). E-mail: