The recent advances in ultralow power device integration, communication electronics, and microelectromechanical systems (MEMS) technology have fuelled the emerging technology of wireless sensor networks (WSNs). The spatial distributed nature of WSNs often requires that batteries power the individual sensor nodes. One of the major limitations on performance and lifetime of WSNs is the limited capacity of these finite power sources, which must be manually replaced when they are depleted. Moreover, the embedded nature of some of the sensors and hazardous sensing environment make battery replacement very difficult and costly. The process of harnessing and converting ambient energy sources into usable electrical energy is called energy harvesting. Energy harvesting raises the possibility of self-powered systems which are ubiquitous and truly autonomous, and without human intervention for energy replenishment. Among the ambient energy sources such as solar energy, heat, and wind, mechanical vibrations are an attractive ambient source mainly because they are widely available and are ideal for the use of piezoelectric materials, which have the ability to convert mechanical strain energy into electrical energy. This paper presents a concise review of piezoelectric microgenerators and nanogenerators as a renewable energy resource to power wireless sensors.
The advances in low power electronics, and wireless sensor networks (WSNs) in particular, have driven numerous researches in the field of energy harvesting in the past decade [
This paper discusses the recent advances in micro- and nanoscale energy generation using piezoelectric materials for ultra low power sensor applications.
The paper is organised as follows: Section
A wireless sensor node is designed to perform sensing, data acquisition, localized processing, and wireless communication and is usually powered by battery. A power generator which scavenges energy from the immediate environment of the sensor can potentially be used to recharge the battery or independently power the senor node. A typical wireless sensor node is shown in Figure
Wireless sensor node showing the main subsystems.
The subsystems in Figure
Power consumption distribution for a wireless sensor node.
The sensing subsystem consists mainly of the sensors and an analog to digital conversion (ADC) unit and is responsible for converting the physical phenomena of interest into digital signal form. The power consumed in the sensing subsystem is used in sensor sampling, which includes the wake-up and stabilization time associated with the sensor and the data acquisition time. At all other times, the sensors are completely off and consume no power. The power consumption of the ADC is typically proportional to the amount of the samples acquired and the sampling rate (SR) used [
Sensor specifications for wireless module in building management system [
Sensor | Voltage (V) | Current (mA) | Power (mW) | Sampling time (s) | Energy/sample ( |
---|---|---|---|---|---|
Temperature | 3.3 | 0.008 | 0.026 | 0.0002 | 0.00528 |
Light | 3.3 | 0.03 | 0.099 | 0.0002 | 0.0198 |
Humidity | 3.3 | 0.3 | 0.99 | 0.8 | 792 |
Vibration | 3.3 | 0.6 | 1.98 | 0.02 | 39.6 |
Barometric pressure | 5.0 | 7.0 | 35.0 | 0.02 | 0.7 |
The computing subsystem is comprised of a processing unit which is usually a microcontroller unit (MCU) and the supporting electronics. This subsystem controls all sensor node activities and performs some local processing. When not processing data and not controlling the system operation, the processor is in a low power sleep mode. Table
Power parameters of part microprocessors [
Microprocessor | Supply current, | Supply voltage (V) | Run frequency (MHz) | Current at power down mode, |
---|---|---|---|---|
C8051F930 | 4.25 | 0.9 | 25 | 0.05 |
PIC18F4620 | 16 | 4.2 | 40 | 0.1 |
MC9s08GT | 6.5 | 3 | 16 | 2.5 |
AMTEGA 128L | 5.5 | 3 | 4 | <5 |
MSP430CG4618 | 0.4 | 2.2 | 1 | 0.35 |
ML610Q431 | 0.65 | 1.1 | 4 | 0.25 |
The energy required by the computing subsystem to complete a task,
The communications subsystem comprises mainly the radio transceiver (RF transceiver) with the amplifiers and associated electronics. The RF transceiver enables the wireless module to communicate and transmit the processed sensor data. When the RF transceiver is not transmitting or receiving, the transceiver is in a low power sleep mode. As shown in Figure
Power parameters of part microprocessors [
RF module | Reception mode current, | Transmission mode current, | Current at power down mode, | |
---|---|---|---|---|
CC2420 | 2.1~3.6 | 18.8 | 17.4 | 0.9 |
MC13192 | 2.0~3.4 | 42 | 35 | 1 |
UZ2400 | 2.7~3.6 | 18 | 22 | 2 |
xBee | 2.8~3.4 | 50 | 45 | <10 |
xBee-PRO | 2.8~3.4 | 55 | 270 | <10 |
NanoPAN5360 | 2.8~3.6 | 35 | 78 | 1.5 |
NanoPAN5361 | 2.8~3.6 | 35 | 78 | 1.5 |
Table
From Table
The Texas instruments MSP430 family of microcontrollers (see MSP430CG4618 power parameters in Table
Power consumption of some commercial wireless sensor nodes.
Crossbow MICAz [ | Intel Mote 2 [ | Jennie JN5139 [ | |
---|---|---|---|
Radio standard | IEEE802.15.4/ZigBee | IEEE802.15.4 | IEEE802.15.4/ZigBee |
Typical range | 100 m (outdoor), 30 m (indoor) | 30 m | 1 km |
Data rate (kbps) | 250 kbps | 250 kbps | 250 kbps |
Sleep mode (deep sleep) | 15 | 390 | 2.8 |
Processor only | 8 mA active mode | 31–53 mA | 2.7 + 0.325 mA/MHz |
RX | 19.7 mA | 44 mA | 34 mA |
TX | 17.4 mA (+0 dbm) | 44 mA | 34 mA (+3 dBm) |
Supply voltage (minimum) | 2.7 V | 3.2 V | 2.7 V |
Average | 2.8 mW | 12 mW | 3 mW |
At present, batteries still dominate energy source for low power electronics in general. Typical characteristics of Li-ion and thinfilm batteries are shown in Table
Characteristics of Li-ion, thin film batteries [
Characteristic | Battery | Supercapacitor | |
Li-ion | Thin film | ||
Operating voltage (V) | 3–3.70 | 3.70 | 1.25 |
Energy density (W h/l) | 435 | <50 | 6 |
Specific energy (W h/kg) | 211 | <1 | 1.5 |
Self-discharge rate (%/month) at 20°C | 0.1–1 | 0.1–1 | 100 |
Cycle life (cycles) | 2000 | >1000 | >10,000 |
Temperature range (°C) | −20/50 | −20/+70 | −40/+65 |
Batteries, particularly Li-ion and thin films variants, are considerably a cheap and convenient and the best solution available in terms of energy density. Over the past two decades, research and development in battery technology has resulted in an increased battery energy density by a factor of three. Still, battery technology has evolved very slowly compared to electronic technology. For example, while computer disk storage density has increased over 1,200 times since 1990, battery’s energy density has increased only about 3 times [
To provide a reliable source of energy for a wireless sensor system, one can consider extracting energy from the environment in order to complement the battery energy storage or even replace it. The process by which energy from the physical environment is captured and converted into usable electrical energy is called energy harvesting. Table light energy: captured from sunlight or room light via photo sensors, or solar panels, mechanical vibrations: from sources such as car engine compartment, trains, ships, helicopters, bridges, floors (offices, train stations, nightclubs), speakers, window panes, walls, household appliances (fridges, washing machines, microwave ovens), pumps, motors, compressors, chillers, conveyors). Table thermal energy from furnaces, domestic radiators, human skin, vehicle exhausts, and friction sources, radio frequency: microwaves, infrared, cell phones, and high power line emissions.
Power densities of typical ambient energy sources.
Energy source | Characteristics | Efficiency | Power density | Comments/challenges |
---|---|---|---|---|
Light | Outdoor Indoor | 10–25% | 100 mW/cm2 | While solar energy is harvesting is an established technology, aiming for small-scale harvesters is difficult because power output directly linked to surface area. For design of embedded wireless sensor nodes to be deployed indoors or overcast areas such as buildings, and forestry terrains, where access to direct sunlight is often not available, solar energy source may not be a suitable choice. |
Thermal | Human Industrial | 0.1% 3% | 60 | Electric current is generated when there is a temperature difference between two junctions of a conducting material (called the Seebeck effect). Thermal energy harvesting uses temperature differences or gradients to generate electricity. Efficiency of conversion is limited by the Carnot efficiency. The efficiency of thermoelectric generators is typically less than 1% for temperature gradient less than 40°C and it is hard to find such temperature gradient in the normal ambient environment. |
Vibration | Hz-human kHz-machines | 25–50% | 4 | Energy from vibrations can be extracted using a suitable mechanical-to-electrical energy converter or generator. Generators proposed to date use electromagnetic, electrostatic, or piezoelectric principles. Vibration energy harvesting is highly dependent on excitation (power tends to be proportional to the driving frequency and the input displacement). |
Radio frequency | GSM 900 MHz | 50% | 0.1 | Without a dedicated radiating source, ambient levels are very low and are spread over a wide spectrum. There is a limit to the amount of power available for harvesting since the IEEE 802.11 standard prescribes the maximum allowable transmission power allowable (1000 mW in the USA, 100 mW in Europe, and 10 mW/MHz in Japan). |
Some vibration sources and acceleration magnitude and frequency of fundamental vibration from [
Vibration sources | Peak acceleration (m/s2) | Frequency of peak (Hz) |
---|---|---|
Car engine compartment | 12 | 200 |
Small microwave oven | 2.5 | 121 |
HVAC vents in office building | 0.2–1.5 | 60 |
Windows next to a busy road | 0.7 | 100 |
Notebook computer while CD is being read | 0.6 | 75 |
Second story floor of busy office | 0.2 | 100 |
From Table
Piezoelectricity stems from the Greek word “piezo” for pressure and the word “electric” for electricity. When a force or stress is applied to a piezoelectric material, it leads to an electric charge being induced across the material. This is known as the direct piezoelectric effect. Conversely, the application of a charge or electric field to the same material will result in a change in strain or mechanical deformation. This is known as the indirect piezoelectric effect. It is the direct piezoelectric effect that is employed in energy harvesting. Examples of ceramics which exhibit the piezoelectric effect are lead-zirconate-titanate (PZT), lead-titanate (PbTiO2), lead-zirconate (PbZrO3), and barium-titanate (BaTiO3). To date, the most commonly used piezoelectric ceramic is PZT mainly because it has very high electromechanical coupling ability. However, PZT is an extremely brittle material and hence this presents limitations to the strain that it can safely withstand without being damaged [
Research in nanoscience has outputted novel piezoelectric material systems used to fabricate next generation nanogenerators employed in energy harvesting technology. The notable work of Wang and Song introduced piezoelectric nanogeneration using a single zinc oxide (ZnO) nanowire by atomic microscopy [
In piezoelectric energy harvesting from vibration, a mass is suspended by a beam, with a piezoelectric layer on top of the beam. When the mass vibrates, the piezoelectric lever is mechanically deformed and a voltage is generated. The most common energy harvesting systems are cantilever structures that are mainly designed to operate at their resonance frequencies. Such structures (unimorph or bimorph cantilevers) are popular because they enable relatively high stress levels on the piezoelectric material while minimizing the dimensions of the devices [
Summary of piezoelectric MEMS energy harvesting devices.
MEMS device description | Design/dimensions | Resonant frequency | Power output/voltage (reported) | Ref |
---|---|---|---|---|
AIN and PZT MEMS devices. | Piezoelectric generator located on top of a beam, piezoelectric layer sandwiched between top and bottom electrodes | 300; 700 and 1000 Hz | 1–100 | [ |
Energy harvesting MEMS device based on thin film PZT cantilevers, | Cantilever size: length = 13.5 mm, width = 9 mm, thickness = 192 | 3 modes: 13.9; 21.9 and 48.5 kHz | 2.4 V with 5.2 MΩ load, 1.01 | [ |
PZT-based MEMS with interdigital electrodes | Cantilever size: length = 3.000 | 570–575 Hz |
1.127 | [ |
PZT harvesters and MEMS technology | Device packaged using two wafers | 1.8 kHz | 40 | [ |
Thin film PZT-based MEMS power generator array for vibration energy harvesting, operating in d31 mode | Cantilever size: 2.000–3.500 | 226–234 Hz | 3.98 | [ |
Thick film PZT free standing energy harvester operating in d31 mode | Cantilever size: length = 13.5 mm, width = 9 mm, thickness = 192 | 229 Hz | 270 nW at 9.81 m/s2; 130 V | [ |
Two layered PMNZT bender devices for micropower generation | Cantilever size: length = 10 mm, width = 10 mm | 120 Hz | 2.0 | [ |
Laser machined piezoelectric cantilever devices for energy harvesting | 10 cantilevers on both sides of ridge, 5 of them are placed with tip mass alternately:length = 5.74 mm, width = 4 mm | 870 Hz | 1.13 | [ |
Multilayer unimorph PZT cantilever with micromachined Si proof mass based on SOI | Device volume (mass and beam)~0.7690 mm3 | 183.8 Hz | 0.32 | [ |
High performance MEMS PZT thin film energy harvester based on d33 mode | 800 × 100 | 528 Hz | 1.1 | [ |
Piezoelectric energy harvesters based on Aluminum nitride (AIN) | Cantilever devices: length = 1.31 mm–2.10 mm, width = 3.0 mm–7.0 mm | 200–1200 Hz | 60 | [ |
Typical piezoelectric energy harvesting system [
Prototype piezoelectric energy harvesting system powering a pressure sensor-developed at Clarkson University.
The ground breaking work by Zhong L. Wang and his Nano Research Group at the Georgia Institute of Technology, USA, has greatly influenced the current research efforts in the conversion of nanoscale mechanical energy into usable electrical energy using nanogenerators. In their original paper Wang and Song first introduced piezoelectric nanogeneration by examining the piezoelectric properties of a single ZnO nanowire (NW) by atomic force microscopy in 2006 [
The insulating properties of piezoelectric insulator materials do not permit carrier transport from metal electrodes into the insulating active materials. As a result, the nanogenerators fabricated from these materials produce alternating (AC) power. On the other hand, the power generation mechanisms of nanogenerators fabricated from piezoelectric semiconductor materials produce both AC and direct power (DC). The coupled semiconducting and piezoelectric properties are in essence responsible for the DC and AC power, respectively. When piezoelectric semiconducting nanowires are subjected to an external force perpendicular to the nanowires, a piezoelectric potential is generated along the nanowires owing to the relative displacement of cations with respect to anion under uniaxial strain [
The electrical contact plays a crucial in role pumping out charges on the surface of the tips. The effective Schottky contact must be formed between the counter electrode and the tip of the NW because the Ohmic contact will neutralize the electrical field generated at the tips. Owing to formation of the Schottky contact, the electrons will pass to the counter electrode from the surface of the tip when the counter electrode is in contact with the region of the negative potential and the current will be measured, whereas no current will be generated when it is in contact with the regions of a positive potential, resulting in the generation of the DC output [
(a) Schematic definition of an NW and the coordination system. (b) Longitudinal strain
A summary of piezoelectric nanogenerators and their demonstrated capabilities as power sources is presented in Tables
Summary of piezoelectric nanogenerators
Output Performance | ||||||
Key material attributes | Generator type and dimensions (length × diameter) | Voltage | Current | Current density | Power or power density | Reference |
AC type 50 | 2.03 V | 107 nA | — | 11 mW/cm3 | [ | |
DC type 2 | — | — | 2 | — | [ | |
DC type 3 | 20 mV | 0.5 | — | [ | ||
InN by use of VLS. Eg: 0.7–0.9 eV; EA: 5.8 eV | DC type 5 | 1.0 V | — | — | — | [ |
GaN synthesised by CVD. Eg: 3.4 eV; EA: 4.1 eV | DC type 10–20 | 20 mV | — | — | — | [ |
PVDF synthesised by E-SP. Eg: 9.23 eV: EA: −0.53 eV | Ac type 6.5 | 5–30 mV | 0.5–3 nA | [ | ||
PZT synthesised HT process. Eg: 2.4 eV; EA: 2.15 eV | AC type 5 | 0.7 V | — | 4 | 2.8 mW/cm3 | [ |
BaTiO3 synthesised by HTCR growth. Eg: 3.3 eV; EA: 2.90 eV | AC type 15 | 25 mV | — | — | — | [ |
Summary of some promising and demonstrated capabilities of nanogenerators (NGs).
Demonstrated capabilities of nanogenerators | Reference |
---|---|
[ | |
[ | |
[ | |
[ | |
[ | |
[ | |
[ |
A simple nanogenerator is principally a nanowire which is a one-dimensional nanomaterial that has a typical diameter less than 100 nm and a length of 1
Energy harvesting from an oscillating human index finger using ZnO single wire generators [
Concept and power generation of the PZT nanofiber generator [
Ambient mechanical vibrations are harvested and converted to useful electrical energy which is either stored in a storage element or is supplied directly to the load. Energy storage is a key element of the energy harvesting system because it is a bridge of stability between the energy source and the load that provides a constant energy flow from an otherwise variable environmental source. The power interface circuits condition the harvested energy to enable the charging of low capacitor batteries or supercapacitors and also provide compatibility with the load requirements. For a sensor node fully powered by ambient energy, the generated mean power (
The key question is how can the power consumption of the wireless sensor node be reduced so that energy harvesting can handle the supply requirements? The answer to the question is practically realized by what is called
Since WSNs operate on a strict power budget, ultralow power microcontroller units (MCUs) are required for processing and power management. A typical MCU like the Texas instruments MSP430 is ideal for energy harvesting since it has a low standby current of less than 1
As discussed earlier in sections, most embedded sensor systems support sleeping modes, making a direct approach to duty cycling an attractive choice to power management. This direct implementation of the duty cycling technique, though popular in energy harvesting for wireless sensor networks, is not always the best choice [
The operation principle of DVS technique is that increasing a circuit’s voltage allows it to switch faster, but with an increase in energy consumption, and conversely the decrease in circuit voltage causes the circuit to have a low switching time with an accompanied decrease in energy consumption [
The performance of a piezoelectric energy harvesting systems primarily depends on the piezoelectric properties used to fabricate the generators. Generally, thin film piezoelectric materials show better piezoelectric properties compared to bulk piezoelectric materials. The use of single crystals and nanomaterials (nanowires) has, in principle, improved the power density and energy conversion efficiency hence the advance the miniaturization of device size while maintaining a reasonable power output. Despite great research efforts on these nanomaterials, there is lack of fundamental scientific understanding of and experimental research on piezoelectric and flexoelectric effects in single crystalline nanowires. This lag in research at this fundamental level compromises fidelity of the mathematical algorithms used in modeling and predicting the piezoelectric potential, mechanical to electrical energy conversion efficiency and device material optimization.
The other challenge relates to the coupling of piezoelectric and semiconducting effect—resulting in the so-called piezotronic effect. The scientific understanding of the interaction of electron distribution and semiconductor band structures requires additional research efforts. The research will potentially present an opportunity to facilitate in situ rectification of the potential output by making use of the Schottky barrier formed between ZnO and metal electrodes. While single crystal materials offer better piezoelectric performance and give better power density compared to their bulk material counterparts, costs of these materials are still very high and at times very inhibitive. The current fabrication methods and the associated device integration techniques at nanoscale are not yet suited for large scale processing, and research efforts along this line will substantially reduce fabrication costs and help translate piezoelectric energy harvesting from mere experimental curiosity into real engineered device realisations to power wireless sensors.
The design of piezoelectric micropower generators and nanogenerators is in itself a multidisciplinary area with challenges based in fundamental physics, material science, mechanical engineering, and electrical engineering. Different researchers from different discipline and background have reported several researches in the area of piezoelectric energy harvesting. The multidisciplinary approach and a holistic paradigm is perhaps the most promising way of designing piezoelectric energy harvesting device.
As can be observed from the review, there is still a need to improve the power output of piezoelectric generators to match the requirements of wireless sensor devices. This challenge can be addressed by using piezoelectric material with the best piezoelectric properties, the best device geometries, and the best power electronics to condition and manage the power output. This is arguably calls for a holistic design and optimization regime, together with an established international metrology standard of piezoelectric energy harvesting (which currently does not exist). Latest advances in synchronised switching techniques have been reported as the latest achievements in power conditioning interface circuits to date [
Vibration energy harvesting using piezoelectric generators was discussed and its potential as an alternative energy source for wireless sensor devices overviewed. The maturity of piezoelectric energy harvesting as technology entails that WSNs are energy efficient and their dependence on batteries is limited. With advancement in ultralow microelectronics and ultra-low power wireless microcontroller units, power consumption of sensor nodes is getting lower and hence the harvested ambient energy may be sufficient to eliminate batteries completely. In addition, piezoelectric nanogenerators open new avenues for ambient power harvesting through foldable power options and miniaturization of power packages thus enabling implantable medical sensing capabilities. Energy harvesting using piezoelectric generators is an attractive alternative energy source that has the potential to provide energy autonomy to wireless sensor devices.