Energy Harvesting State of the Art - Harb 2010

Renewable Energy 36 (2011) 2641e 2641 e2654

Contents lists available at ScienceDirect

Renewable Renewable Energy Energy j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m/ m/ l o c a t e / r e n e n e

Energy harvesting: State-of-the-art Adnan Harb

*

Department of Electrical Engineering, UAE University, 17555 Al Ain, United Arab Emirates

a r t i c l e

i n f o

Article history:

a b s t r a c t

This paper presents a brief history of energy harvesting for low-power systems followed by a review of the state-of-the-art of energy harvesting techniques, power conversion, power management, and battery charging. The advances in energy harvesting from vibration, thermal, and RF sources are reviewed as well as power power management management techniques. techniques. Examples of discrete discrete form implement implementation ation and integrat integrated ed form implementation using microelectromechanical systems (MEMS) and CMOS microelectronic processes are also given. The comparison between the reviewed works concludes this paper.  2010 Elsevier Ltd. All rights reserved.

Received 5 April 2010 Accepted 9 June 2010 Available online 10 July 2010 Keywords:

Energy harvesting Energy sources Power management Power electronics MEMS CMOS

1. Introduction

Energy Energy (orpower) (orpower) harvest harvestingor ingor (scaven (scavengin ging) g) is withoutany withoutany doubt doubt a very attractive technique for a wide variety of self-powered microsystems. systems. Examples of such systems are wireless wireless sensors, biomedical biomedical implants, military monitoring devices, structure-embedded instrumentation, remote weather station, calculators, watches, Bluetooth headset headsets. s. Recen Recently tly,, Nokia Nokia announ announced ced it is develop developing ing a mobile mobile protot prototype ype that that could could harves harvestt energy energy from from ambient ambient radio radio waves waves emitted emitted from mobile antennas, antennas, TV masts and other sources [1 sources [1e6] 6].. Energy harvesting has become of a growing interest in the last few years and research research report number has kept increasing increasing for the last decade. The scope of this paper is to provide the research community with an update of the state-of-the-art of energy harvesting from vibration, thermal, and RF sources. The principle of energy harvesting approaches can be found in Ref. [7] [7].. In the following, following, we list the energy energy harvestin harvesting g sources sources (Section (Section 2), a brief history of energy scavenging (Section 3), state-of-the-art based based on the revie review w of sever several al recen recently tly publish published ed papers papers (Secti (Section on 4) and Conclusion and Conclusion.. 2. Energy Energy harvesting harvesting sources sources

power power systems, systems, etc., they are not game changers changers for electronic electronic designers whose mission in life is to snip the wires e including powe powerr cords cords and and even even batt batter ery y powe powere red d syst system emss wher where e the the perpetual perpetual device is the ultimate design goal [8] goal [8].. Progress in ultralow-power microelectronic technology with the advance in microenergy energy Harve Harvestin sting g makes makes the number number of batter battery y chargi charging ng cycles cycles the main limit towards the perpetual self-powered device. Towards this mission, and to meet the design community s long march march to ultra-lowultra-low-powe powerr technology technology,, we can identify identify several several micro-energy harvesting sources: ’





 

Even Even though though macromacro-ene energy rgy harves harvestin ting g has been been around around for centuries in the form of windmills, watermills and passive solar

 

* Tel.: þ 971

3 713 606; fax:

þ971 3 762 3156.

E-mail address: adnan.harb@ua [email protected]. eu.ac.ae.

0960-1481/$ e see front matter  2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.renene.2010.06.014 doi:10.1016/j.renene.2010.06.014

 

Motion, Motion, vibrati vibration on or mechan mechanical ical energy energy:: �oors, oors, stairs, stairs, object object s movement, transfer energy from the engine to the battery during braking, braking, etc. The electromech electromechanical anical transducer can be electromagne electromagnetic tic [9] [9],, electrostatic electrostatic [10], [10], or piezoel piezoelect ectric ric [11,12].. [11,12] Electromagnetic (RF): Base stations, wireless internet, internet, satellite lite commun communicat ication, ion, radio, radio, TV, digital digital multim multimedi edia a broadbroadcasting, etc. One must not confuse between electromagnetic energy source and electromagnetic transducer. In some articles, electromagnetic generator is used for electromagnetic transducer. Thermal. Momentum generated generated by radioactive reactions reactions into electrical energy. Pressure gradients. Micro water � ow (e.g. faucet). Solar and light. Biological. ’

2642

A. Harb / Renewable Energy 36 (2011) 2641e 2654

Fig. 3. Three steps to harvest energy: (a) Battery pre-charges the capacitor, (b) vibrations cause capacitance value to decrease and energy is harvested into the battery, and (c) reset [1]. Fig. 1. (a) Piezoelectric generator voltage versus time after recti �cation for a single impact applied to the generator, (b) Schematic diagram of the connections between the piezoelectric � lm and the electrostatic generators. The diodes are Schottky type with forward voltage drop near 0.33 V [24].

Another classi�cation scheme may consider who or what provides the energy for conversion: the �rst kind is called the human energy source. The energy is provided by the activity of human beings or animals. The second kind is the energy harvesting source that gets its energy from the environment [13,14]. The power transferred to a load is limited by the availability of the raw energy, and the ef �ciency of the transducer and the conversion circuit. The discontinuous nature of energy harvesting has consequences on the way the electronic devices powered by energy

harvesting are operated. In principle, we can distinguish two situations: 1. The power consumption of the device is lower than the average harvested power. In this case, the electronic device may operate continuously. 2. The power consumption of the device is greater than the average harvested power. The operation must be discontinuous, and the time between operations depends on the stored energy of the device [13]. 3. History

The �rst observation of harvesting energy in form of current from natural source was in 1826. Thomas Johann Seebeck found that a current would �ow in a closed circuit made of two dissimilar metals when they are maintained at different temperatures [15,16]. For the following three decades, the basic thermoelectric effects were explored and understood macroscopically, and their

Fig. 2. Energy harvesting and battery-charging system proposed by Torres et al. [1].

Fig. 4. A complete harvesting cycle [1].


2643

Fig. 7. A dual polarity boost converter [29].

applicability to thermometry, power generation, and refrigeration was recognized [17]. In 1839, as he was experimenting with an electrolytic cell composed of two metal electrodes, Edmund Becquerel discovered

the photovoltaic effect [18]. The � rst large area solar cell was constructed in 1894 by Charles Fritts who coated a layer of selenium witha thin layer of gold [19]. While the photovoltaic effect was �rst observed by Edmund Becquerel, it became fully comprehensible only after developing the quantum theory of light and solid state physics in the early 1900s [18]. Joseph Henry and Michael Faraday independently discovered the principle of producing electricity from magnetism, known as electromagnetic induction, in 1831 [20]. In October of the same year, Faraday invented the � rst direct-current generator consisting of a copper plate rotating between magnetic poles [21]. The �rst observation of harvesting energy in form of charge was in 1880. Pierre and Jacques Curie successfully predicted and proved experimentally that certain crystals would exhibit a surface charge when subject to mechanical stress. This phenomenon was given the name piezoelectricity [15].

Fig. 6. A plan view of the variable capacitor implemented using MEMS technology [28].

Fig. 8. Spice simulation of the circuit of Fig. 7, positive half cycle, for a source displacement of 25 m m at 322 Hz. Top: Accumulated energy extracted from the coil, output to the reservoir, and dissipated in the three main loss mechanisms. The three loss mechanism lines are of similar magnitude. Bottom Boost inductor instantaneous current [29].

Fig. 5. (a) Capacitor C par added as a hybrid alternative to voltage-constrained and charge-constrained energy harvesting systems. (b) Timing waveforms [28].

2644


Fig. 9. Modi�ed buck converter [29].

4. State-of-the-art review 4.1. Harvesting from vibrations

There has been much recent interest in using MEMS (Microelectromechanical Systems) to scavenge energy from ambient vibration and transfer it to electrical load. Such device is mechanically modeled with the base excitation of an elastically mounted seismic mass moving past a coil [22]. A mathematical model of the transferred energy is developed in Ref. [22]. It takes into consideration the seismic mass amplitude, the magnitude and frequency of the excitation, the electrical analoge of the mechanical damping coef �cient. It is found that the matching condition to transfer maximum power to the load is different from the well known simple electrical matching and is given by:

Rload

¼

Rint þ K 2 =c m

Fig. 11. Magnetic spring generator structure using single moving magnet and 2 � xed magnets [30].

Where Rload is the load resistance, Rint is the coil internal resistance, K is the electromechanical coupling (transducer) coef �cient, and c m is the mechanical damping coef �cient. In order to �nd optimal architectures for maximal power generation under the different operating constraints, analysis and veri�cation by simulation of three classes of MEMS-based vibration-driven microgenerator architectures are presented in Ref. [23]. The three classes are velocity-damped resonant generators (VDRGs), Coulomb-damped resonant generators (CDRGs), and

Fig. 10. Ef �ciencies of an electrostatic micro-generator. The maximum ef �ciency of a buck converter is reached for 10 cells of the unity MOSFET considered in simulation [29]. Percentage scale was not in the original � gure.

Fig.12. Measured no-load voltage during walking and slow running for the generator with higher ef �ciency [30].


2645

Fig. 13. The electromagnetic generator (left) and its cross section (right) [32].

Coulomb-force parametric generator (CFPG). The later is a new class that doesn t operate in a resonant matter. The authors conclude that each class is superior in power generation in different conditions related to the resonant frequency and Z l/Y 0, where Z l is the maximum possible of the amplitude of the mass-to-frame displacement and Y 0 is the source motion amplitude. For ideal implementations, the CFPG produces the most power where ( Z l/ Y 0) < 0.1, the VDRG is superior above the resonant frequency when ( Z l/Y 0) > 0.1, and the CDRG is superior below but near the resonant frequency when ( Z l/Y 0) > 0.1. Rocha et al. [24] described a system integrated in footwear to harvest energy from vibration when people walk. Piezoelectric and electrostatic generators were used to convert pressure variations to energy because they provide the highest density of power comparing to electromagnetic [25,26]. When the person steps with the foot on the �oor, the capacitance (two metallic plates separated by a �exible dielectric material) of the electrostatic generator increases about two times, which means that the voltage decreases to one half. At this time, the piezoelectric generator (polymeric material coated in both sides by a conducting material, which form the electrodes) voltage is higher than the one at the electrostatic generator terminals (Fig. 1a), so, its capacitance will be charged. When the person raises his/her foot, the capacitance of the electrostatic generator decreases and its voltage increases. In this case, the load that is an energy storage device (thin � lm lithium battery of 3 V) will have its voltage lower than the electrostatic generator, the charge is transferred from the generator to the battery and the diode bridge is blocked (Fig. 1b). The power generated ranges from tens to hundreds of milliwatts and the average energy generated in 1 h, by a running person when the generator is coupled to a resistive load doesn t exceed 51 mJ.

Torres and Rincón-Mora [1,27] proposed a voltage-constrained energy harvesting system, i.e. the voltage of the harvested energy storage battery limits the maximum voltage of the capacitor (Fig. 2). It works as follows: The pre-charge control block lets the battery pre-charge the capacitor C VAR to V BAT through inductance L (Energy loss) when its capacitance is maximum. It energizes L by switching S 1 and S 3 ON while S 2 and S 4 are OFF, then it charges C VAR to V BAT by turning S 2 and S 4 ON while S 1 and S 3 are OFF. S 1 through S 4 are then all turned OFF leaving the capacitor under charge-constrained Q C-VAR (open-circuit). C VAR is designed in such a way ambient vibrations cause its capacitance to decrease, converting mechanical energy to electrical (Energy harvesting): V C-VAR increases when C VAR decreases as Q C-VAR ¼ constant and Q VAR ¼ C VAR $V C-VAR ¼ constant. When V C-VAR is high enough, the energy harvested in the capacitor will charge the battery through diode D (current I HARV ) and V C-VAR will remain limited by V BAT þ V D. The capacitor is then reset to its maximum value before getting pre-charged again by the battery. Figs. 3 and 4 show the three harvesting steps, and the associated C VAR , V C-VAR , I Harv and E Harv respectively. The energy net gain depends on the capacitance value excursion and the battery voltage. Thus, other than storing energy, the battery role is to hold constant the voltage across the capacitor. Otherwise a 1e200 pF variation of the capacitor value, ampli �es the initial voltage across it by a factor of 200 which can surpass the breakdown limits of most modern CMOS technologies. The experimental results showed a net energy harvesting gain of 6.9 nJ/cycle at 200 Hz. The harvester circuit was fabricated with the AMI semiconductor 0.5 mm CMOS technology except for the inductor, the

Fig. 14. Output voltages for different resonator shapes at 1 M U load resistance [32].

Fig. 15. Output power for different resonator shapes at resonance [32].

’

’

2646


Table 1

Experimental results for three different resonators [32]. Resonator

Resistance [U]

Electrical resonance [kHz]

Load resistance [U]

Mechanical resonance [Hz]

Quality factor Q

Power at resonance [mW]

Voltage at resonance [mV]

Shape A Shape B Shape C

14.1 27.1 9.8

4910 2420 13950

65 76 47

127 102 98.2

31.7 20.4 27.2

265 290 229

185.7 183.2 148.5

the moving mass at the anchors and the others to each of the stationary combs. This device s wafer will be supported by an identical silicon handle wafer. The two wafers and the separation oxide form the parasitic capacitor C par. An analysis of the capacitance with the comb structure fully closed yields a capacitor maximum value of 260 pF and minimum value of 2 pF with the comb structure fully open. The analysis and design of the control and power electronics is based in these two values with a maximum gap voltage of 8 V. The systemworks as follows (Fig. 5a and b): At startup, the capacitor combination of C par and C MEMS has no voltage across it, so V C ¼ V DD. At the beginning of t 1, a controller trigger the conversion process. During t 1, SW2 is on, SW1 is off, and the inductor current increases. At t 2, SW2 is off, SW1 is on, and the inductor transfers energy to the capacitor. During t 3, both switches are off and the variable capacitor plates move. This time constant is near 400 ms while the resonant on time of the switches is approximately 600 ns. It is therefore a reasonable approximation to say that the MEMS capacitor value is constant during t 1, t 2, t 4 and t 5. During t 3 the plates move from their minimum separation (C max þ C par) to their maximum separation ( C min þ C par). The mechanical energy has moved the plates apart and caused the voltage across the capacitor combination to reach a maximum, and the energy harvesting is performed. During t 4 SW1 is on, SW2 is off, and the capacitor combination transfers energy to the inductor. Note that this LC time constant is smaller than t 2 because the overall capacitor value has decreased. Once the capacitor voltage reaches zero, corresponding to one-quarter of the resonant period of the LC, SW1 is turned off, SW2 is turned on, and the energy put into the inductor is transferred to the reservoir during t 5. This process repeats at the frequency of the mechanical vibration, which corresponds to variations in C MEMS. The system controller has been fabricated in a 0.6 m m CMOS process. In order to achieve the highest possible power density from an inertial energy scavenger, the authors of Ref. [29] found it is necessary to optimize the damping force under a given operation ’

Fig. 16. Step-down dcedc conversion circuitry [33].

state detectors and part of the timer that were kept off chip for experimental �exibility. A 3 V supply emulated a moderately charged Li-Ion battery ( V BAT). Voltage-constrained energy harvesting system provides more energy than the charge-constrained case. However, the major obstacle for this approach is that it would require an additional voltage source to hold the voltage. A hybrid alternative is proposed in Ref. [28] where a second capacitor of constant value C par is added in parallel to the variable energy harvesting capacitor ( C MEMS in Fig. 5a). Analysis shows that as C par approaches in�nity, the charge available approaches that available through voltage-constrained system. The variable capacitor is implemented using MEMS technology (Fig. 6). It consists of three basic parts: a �oating mass, a folded spring (one per side), and two sets of interdigitated combs, one per side. Each spring consists of four spring bars, a free rigid beam, and a rigid anchor. The spring bars are connected to both the anchor and the free beam, limiting the motion of the mass to one dimension (perpendicular to the �gure plan). The interdigitated combs form two variable capacitors by connecting one terminal to

Fig. 17. Theoretical and experimental optimal duty cycles for step-down converter as a function of excitation [33].

Fig. 18. Maximum RMS power against displacement [34].


2647

Fig. 19. (a) the rectifying circuit. (b) The recti �ed output [35].

condition (size of the relative motion between the mass and the frame). In an electromagnetic generator (i.e. magnetic transducer) for which the source vibrates with an amplitude Y o and drives a VDRG with an internal displacement limit of Z l, the optimal damping will be that which just allows the maximum internal displacement ( Z l). The optimal damping factor is given by: z

¼

1 Y 0 2 Z l

As per the power processing circuit, it is advantageous to operate it at relatively high voltage (above 1 V) to reduce power loss in switches. The authors proposed and veri�ed by simulation a compromise between the inductance s number of turns, its size, the resistive loss, the volume of the permanent magnet, and the output voltage as the main challenge in such circuit is the low output voltage of the transducer ( 195 mV). The used dual polarity boost circuit separately process the positive and negative half cycles of the generated voltage (Fig. 7). It is proposed to use it in discontinuous synchronous conduction mode to reduce power loss in the switches (MOSFETs). An optimal ef �ciency of about 50% has ’

w

been reached in several operating conditions for an output power of 50 mW at 1.65 V (Fig. 8). The authors used the converter depicted in Fig. 9 for constantcharge electrostatic mic rogenerators [29]. The trade off in optimizing the overall ef �ciency depends on several parameters: because of the high voltage produced by the transducer, a considerable amount of energy is stored in the depletion layer parasitic capacitance of the blocking junction of the high-side MOSFET. This energy is lost when the MOSFET is turned on. Decreasing the size of the MOSFET will, from one side decrease the parasitic capacitor and hence the losses, and from the other side degrade the ef �ciency of the converter (bad switch). The simulation shows a possibility to locate an optimum switch size that maximizes the ef �ciency of the converter. The authors found that a multiple of 10 of a unity transistor provides the best ef �ciency for the modi �ed buck converter (Fig. 10). A magnetic spring generator is described in Ref. [30]. A free to move permanent magnet is placed inside a tube, and two other magnets are � xed at both ends of the tube in such a way that the facing surfaces of the �xed and moving magnets have the same polarization. Finally, a coil is wrapped around the outside of the

Fig. 20. Illustration of the generator that realizes the frequency upconversion: (upper left) isometric, (upper right) side, and (lower) schematic views [38].

2648


Fig. 21. Simulation and measurement results for a single cantilever of the FupC design [38].

tube (Fig.11). When the tube moves, the middle magnet vibrates up and down and a voltage will be induced in the coil. Two different versions were prototyped and tested while placed inside a rucksack of walking/slow running person. For the higher ef �ciency version (one �xed and one moving magnets), the average measured maximum load powers of the generators were 0.95 and 2.46 mW during walking and slow running condition, respectively. The output voltage is presented in Fig. 12. Using a simple diode capacitor recti�er, the prototype was able to transfer 3.54 J to a rechargeable battery in 1 h. This energy level is enough to power low-power sensor modules without battery [31]. An electromechanical power generator to convert vibrations to electrical energy with electromagnetic transducer is proposed in Ref. [32]. The transducer is shown in Fig. 13. The resonator,

a moveable planar inductor that can move with vibration, is � xed by its external edge at the cylindrical case and it is placed symmetrically between two sets of magnets ( Fig. 13, left). In order to improve the distribution of the magnetic �eld density in the area between the magnets, the two central magnets have opposite magnetization vectors, while the two external magnets have opposite magnetization vectors, but inverted than the other two (Fig. 13, right). Output voltages at resonance in the range of 150e185 mV (Fig. 14) and an output power in the range of 230e290 mW (Fig. 15) have been experimentally obtained for three different resonators A, B, and C (Table 1). An optimized method of harvesting vibrational energy with a piezoelectric element using a step-down dcedc converter is presented in Ref. [33]. Analysis of the converter in discontinuous

Fig. 22. (a) Parametric frequency increased generator (PFIG), (b) Theory of operation e the generator is depicted at three instances of time during an incident displacement [41].


2649

Fig. 23. Full-wave recti�er circuit used to store the charge provided by a pyroelectric cell [43].

current conduction mode results in an expression for the stepdown converter duty cycle-power relationship. For a vibrating piezoelectric element modeled as a sinusoidal current source I psin(ut ) in parallel with its electrode capacitance C p (Fig. 16), it was found that the optimal duty cycle for maximum harvested power is:

Dopt

r ffi4ffi uffi ffiLC ffi ffi ffi f ffi ffi

z

Fig. 25. Experimental data using the circuit depicted in Fig. 23 when consecutive heating/cooling cycles were applied to a PZT cell [43].

p s

p

Where L is the inductance and f s is the switching frequency. The circuit has been implemented with the accompanying control circuitry, i.e. switch gate drive, pulse generator, duty cycle generator and threshold control. Fig. 17 shows the theoretical optimal duty cycle and the experimentally determined optimal duty cycle as a function of the mechanical excitation (characterized by the piezoelectric element s unloaded or open-circuit voltage). Over the range of excitation considered, up to 100 V, the two curves follow a similar trend, both becoming nearly constant above excitations of 45 V. A prototyping of an inertial linear electromagnetic millimeter scale generator is reported in Ref. [34]. Fig. 18 shows the experimental output power at resonance against source vibration amplitude. This device is capable of generating 0.3 m W at an excitation frequency of 4 MHz. Another small scale vibration-induced power harvester with total volume of about 1 cm 3 is presented in Ref. [35]. With the quadrupler rectifying circuit of Fig.19, the loaded generator is able of producing 1.3 V DC from 4 V, 60 e120 Hz input corresponding to 200 m m input vibration. The power output for this system is 100 m W. ’

w

w

Fig. 24. Experimental values of PZT pyroelectric cell I , T and calculated dT /dt over the time when switching on (temperature increases) and off (temperature decreases) an air dryer [43].

More focus on piezoelectric and electromagnetic transducer fabrication to harvest from kinetic energy can be found in Ref. [36,37]. 4.2. Harvesting from low-frequency vibrations

Astheef �ciency of vibration-based harvesters is proportional to excitation frequency, Sari et al. [38] proposed a generator that converts low-frequency environmental vibrations to a higher frequency by employing the frequency upconversion (FupC) technique. The generator is depicted in Fig. 20. It is composed of two mechanical structures: a magnet attached to a diaphragm that resonates with ambient frequency (in the range of 1 e100 Hz), and an array of cantilevers (resonance frequency in the range of 2e3 kHz) located right below the diaphragm. At the tip of each cantilever, nickel is electroplated for interaction with the magnet. As the diaphragm resonates in response to external vibrations, the magnet catches the cantilevers at a certain instance of its movement, pulls them up, and releases them at another point. The released cantilevers start resonating at their high damped natural frequency with the given initial condition, realizing the FupC. The motion of the released cantilevers exponentially decays out, and before it completely dies, the cycle starts again. The authors derived a mathematical model for the output voltage and the harvested power. Fig. 21 shows the simulated and measured output voltages from a single cantilever. In the �gure, the catch and release points of the cantilevers and the peak voltage output are also indicated. Voltage generation is realized right after the release of the cantilevers where they oscillate with their damped natural frequencies. The time in which output voltage is almost zero, depends on the ambient frequency (95 Hz in this case), and damping factor of the cantilevers oscillations.

Fig. 26. A boost converter using digitally controlled FETs [45].

2650


Fig. 27. Left: The principle of the thermoelectric microsystem proposed by [48]. When the heat �ows across the pn junction, an electrical power current is generated by the Seebeck effect. Right: Schematic of the step-up circuit.

Fig. 28. Smart RF energy harvesting system [53].

The experimental result shows that from a natural diaphragm frequency of 113 Hz, the system is able to produce a peak power and peak voltage of 0.25 nW and 0.57 mV respectively per cantilever. This is to compare with the popular traditional large mass coil (LMC)-type [39,40] electromagnetic energy harvester that, scaled for 113 Hz, produces a peak power and peak voltage of 0.23 nW and 0.24 mV respectively.

A non-resonant inertial generator for harvesting low-frequency vibrations is presented in Ref. [41] and its principle is shown in Fig. 22. The Parametric Frequency Increased Generator (PFIG) is designed to accommodate the large amplitudes associated with low-frequency vibrations [42]. Two Frequency Increased Generators (FIGs) are placed on either side of an inertial mass, oriented to face each other (Fig. 22a). Attached to the bottom of the FIG spring

Fig. 29. A 3-stage Schottky diode based Villard voltage multiplier circuit [53].

Fig. 30. Power generated at variable distance from a 13 dBm RF source [53].


2651

Fig. 32. System architecture [56]. Fig. 31. Improved Villard voltage multiplier circuit [53].

is a magnet for power generation, while on top, a smaller magnet (actuation magnet) is used to generate a magnetic force in order to latch the FIG and the inertial mass together. The operation of the PFIG is outlined in Fig. 22b. The generator operates such that the inertial mass snaps back and forth between the two FIG generators, attaching magnetically. As the inertial mass moves, it pulls the FIG spring along. When the inertial mass approaches the opposing FIG, the magnetic force of attraction begins to increase. As the forces on the FIG/inertial mass system overwhelm the holding magnetic force, the inertial mass detaches and is pulled to the opposing FIG. The freed device now resonates at its high natural frequency converting the stored mechanical energy in its spring, to electrical. This process is subsequently repeated in the opposite direction. With a total volume of 3.74 cm3, an inertial mass of 9 g, and an input frequency of 10 Hz, the harvested energy is 288 m W. 4.3. Harvesting from thermal sources

Energy harvesting from pyroelectric cells based on screenprinted PZT (lead zirconate titanate) and PVDF (polyvinylidene �uoride) �lms and subject to temperature variation has been reported in Ref. [43]. The pyroelectric cell is modeled with current source I p, capacitor C p and resistor R p (Fig. 23), with

l p

¼

dT S l dt

where S is the electrode surface of the cell, l is the pyroelectric coef �cient and T is the temperature. Fig. 24 shows the measured characteristic of the generated current when temperature changes. The current changes direction during free cooling off (temperature decrease). As the current is alternative during heating up and cooling down, a diode based full-wave recti�er has been used to store energy (Fig. 23). The experimental results of Fig. 23 for C L ¼ 1 mF, shows how the voltage increases when consecutive heating/cooling cycles were applied to a PZT cell (Fig. 25). As it can be seen from Fig. 25, circuit losses increased with output voltage increase. The output voltage Table 2

Output voltage comparison of traditional work and circuit simulated in Ref. [53]. Input power (mW)

316.23 562.34 1000 1778.28 3162.28

Output power (mW) Traditional

This Work

8.43 45.12 133.85 282.58 495.75

12.39 94.19 347.03 495.75 743.63

% Increase over traditional

47.06% 108.79% 159.26% 75.44% 50%

reached a maximum of 21.5 V for the measured cell. Two parallel cells were able to produce a maximum voltage of 31 V and provide energy of 0.5 mJ. With these results, it is possible to power a lowpower RF transmitter [44]. PVDF cells produce less energy because of their low pyroelectric coef �cient. Carlson et al. presented a switched mode DCeDC boost converter with digital control for thermal source energy harvester [45]. The boostconverter is shown in Fig. 26. The control is based on turning the pFET off when the inductor current falls to zero. The control circuit that is based on the status of V D (high or low) is almost free of analog circuitry and is thus free of static power dissipation. The whole circuit except for an inductance and two �ltering capacitors was fabricated with a 0.13 mm CMOS process. The circuit was then tested with body heat from a human arm. The converter was allowed to reach thermal equilibrium over a period of 10 min. At equilibrium, the generator produced 34 mV (unloaded voltage) and the boost converter was able to deliver 34 m W at 1 V. Using the Seebeck effect [15,16,46,47], a thermoelectric microconverter for energy scavenging systems that can supply individual electroencephalogram (EEG) modules was fabricated using thin �lms of bismuth and antimony tellurides (Fig. 27 left). With a simple step-up circuit (Fig. 27 right), a 1 cm2 thermoelectric microconverter provided a power of about 18 m W [48]. In analogy with electrical matching, Leonov and Fiorini developed equations for thermal matching of thermoelectric energy harvester [49]. Thermal matching is required to maximize the harvested power and serves in the design of TEGs (thermoelectric generators). More details on TEGs design, fabrication, and testing can be found in Ref. [50,51]. In Ref. [50], D.T.S., a German company presented a prototype of its product, the Low-Power Thermoelectric Generator (LPTG). Its manufacturing process is fully compatible with standard microelectronic technologies with power output of a few 10 m W.

4.4. Harvesting from RF sources

An energy harvesting system performance using multiple RF sources was demonstrated in Ref. [52,53]. In Ref. [52], multiple energy harvesting antennas in one area are proposed and studied to increase the power/area ratio. It was shown that an increase of 83% in area resulted in 300% increase in power. Jabbar et al. proposed Schottky diode and CMOS process compatible based version of a 2-antennas RF energy harvesting systems shown in Fig. 28 [53]. The former uses a 3-stage Schottky diode Villard voltage multiplier circuit (Fig. 29). The system was tested with different RF frequencies at different source powers and distances. C tune was tuned with respect to each antenna to get the maximum output power. Fig. 30

2652


Fig. 33. The DC/DC converter architecture [56].

uses an improved version of the NMOS traditional Villard voltage multiplier described in Ref. [54]. The improved version is depicted in Fig. 31. It uses only one self-biasing circuit, consisting of R1, R2 and C 1 for both the transistors as compared to two in Ref. [54]. Simulation using the TSMC (Taiwan Semiconductor Manufacturing Company) 0.25 mm CMOS technology showed a higher output voltage with respect to the traditional circuit. The output power increase range from about 50% (for low and high input power) to about 160% for medium input power (Table 2). As increased input RF energy results in higher output voltage, the transferred power becomes larger than the circuit loss and ef �ciency increases. However, at high input energy, the diode connected MOSFET s leakage current increases and the ef �ciency decreases hence creating an optimal value of the ef �ciency. This is a consequence of larger output voltage resulting in larger bias voltage on the gate preventing the diode connected MOSFET s from ever fully turning off. A study of reception and recti�cation of broad-band statistically time-varying low-power-density microwave radiation is presented ’

Fig. 34. RF converter architecture. Vbandgap is the same as the 570 mV of Fig. 33 [56].

’

shows the output power for a constant source power of 13 dBm and distance between the source and circuit is varied from 1 to 12 cm. The 2.4 GHz circuit response was not in a straight line due to the presence of wireless Access Points in thebuilding. The CMOS version

Fig. 35. Micro-battery voltage monitor. When Act goes low, the comparison between V BAT and 1.5 V is activated, and if V BAT < 1.5 V (at about 11,500 s), soc �ag goes low [56].

Fig. 36. Micro-battery charger test result. When V BAT reaches 2.75 V, soc becomes high at the next Act [56]. “

”


2653

Table 3

Summary of the reviewed energy harvesters. Ref.

Reported power (calculated)

Source

Comment

[24] [1] [29] [30] [30] [30] [32] [34] [35] [38] [41] [43] [45] [48] [49] [52] [52] [55]

51 mJ (14.17 m W) 6.9 nJ/Cycle (1.38 m W) 50 mW 0.95 mW 2.46 mW 3.54 J (983 m W) 230e290 m W 0.3 m W 100 m W 0.25 nW/Cantilever 288 m W 0.5 mJ/2 parallel cells 34 m W 18 m W Few 10 m W 86 dBm 12.39 m W 105e0.1 mW/cm2

Vibrations from running person Vibrations Vibrations Vibrations, walking person Vibrations, running person Vibrations, average walking and running Vibrations 4 MHz vibrations Vibrations Low frequency vibrations Low frequency vibrations Thermal Thermal Thermal Thermal RF, 13 dBm, at 4 cm RF, 316.23 m W RF

1h @ 200 Hz Simulated

in Ref. [55]. A 64-element dual-circularly-polarized spiral rectenna array is designed and characterized over a frequency range of 2e18 GHz with single-tone and multitone incident waves. The integrated design of the antenna and recti �er eliminates matching and �ltering circuits, allowing for a compact element design. The recti�ed dc power and ef �ciency is characterized as a function of dc load and dc circuit topology, RF frequency, polarization, and incidence angle for power densities between 10 5e101 mW/cm2. 4.5. Harvesting power management

Lhermet et al. [56] have fabricated a power generator unit including two micropower sources (radio frequency RF and Thermal) and their management integrated circuit (IC) ( Fig. 32). The RF and thermal microgenerators, the micro-battery and the L eC power � lter of the DC/DC converter are not integrated within the IC. The raw input power from the thermogenerator is processed by a power switch followed by the L eC power �lter, yielding the conditioned output power. The control system varies the switch duty cycle to cause the output voltage to follow a reference voltage of 570 mV from bandgap (Fig. 33). The RF/DC converter is less complicated. It is composed of a limiter, a recti �er and a control loop to provide a stabilized DC output (Fig. 34). As the micro-battery can be charged by either converter, therefore, the power supply manager manages priority between the two sources when they are simultaneously present and activates self-powered micro-battery protection in the case of external power source interruption. The power manager compares the micro-battery voltage to a reference voltage every hour and half (an internal digital signal Act goes low ) and sets an internal digital signal (soc) to low when micro-battery is discharged ( <1.5 V) (Fig. 35). If an internal source is available, the charger starts providing a 27 mA current until the battery voltage reaches 2.75 V (Fig. 36). More in harvesting power management can be found in Ref. [7]. “

”

“

”

5. Conclusion

As the ambient energy sources are much diversi �ed (vibrations, RF, thermal, etc.), and the transducers are available in several types, many techniques to harvest and to convert to DC power supply energy from the available sources, have been presented.

1h millimeter scale Volume 1 cm 3 1 e100 Hz 10 Hz Time NA w

Microelectronic process Schottky diode based Simulated CMOS based

A summary of the energy provided by the reviewed energy harvesters is presented in Table 3. It is hard to make a fair comparison because of the high number of parameters that affect the performance of the generators. Examples of these parameters are the harvested power, the ef �ciency, the output voltage, the application, the size of the harvester i.e. discrete components versus MEMS and microelectronics, the number of unit in parallel or in series to provide more energy and/or voltage, the technology cost, the availability, abundance and vicinity of the energy sources, etc. We can however note that vibrations are the most available and the highest power provider sources. Energy harvesting techniques and their applications are expanding and becoming more attractive especially with advance in microelectronics and MEMS. Power management allows for harvesting from multiple sources, which, depending on the application, can lead to directly power the application circuit without using battery. Acknowledgment

The author would like to acknowledge the support from the United Arab Emirates University. References [1] Erick O. Torres, Gabriel A. Rincón-Mora. Electrostatic energy-harvesting and battery-charging CMOS system prototype. IEEE Transactions on Circuits and Systems I 2009;56(9):1938e48. [2] Want Roy, Farkas Keith I, Narayanaswami Chandra. Energy harvesting and conservation. IEEE Journal of Pervasive Computing; January eMarch, 2005:14e7. [3] Graham-Row. Duncan. Wireless power harvesting for cell phones, www. technologyreview.com ; 9 June, 2009. [4] Walko. John. Nokia working on energy-harvesting handset, www.EETimes. com; 11 June, 2009. [5] Zhang Xiaoyu, Jiang Hanjun, Zhang Lingwei, Zhang Chun, Wang Zhihua, Chen Xinkai. An energy-ef �cient ASIC for wireless body sensor networks in medical applications. IEEE Transactions on Biomedical Circuits and Systems 2010;4(1):11e8. [6] Priya Shashank, Inman Daniel J. Energy harvesting technologies. Springer; 2009. [7] Vullers RJM, Van Schaijk R, Doms I, Van Hoof C, Mertens R. Micropower energy harvesting. Solid-State Electronics 2009;53:684 e93. [8] Murugavel Raju. Energy harvesting, white paper, Texas instruments; November, 2008. [9] El-hami M, Glynne-Jones P, White NM, Hill M, Beeby S, James E, et al. Design and fabrication of a new vibration-based electromechanical power generator. Sensors and Actuators A: Physical 2001;92:335 e42. [10] Miyazaki Masayuki, Tanaka Hidetoshi, Ono Goichi, Nagano Tomohiro, Ohkubo Norio, Kawahara Takayuki, et al. Electric-energy generation using

2654

[11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25]

[26] [27] [28] [29] [30] [31] [32] [33] [34]


variable-capacitive resonator for power-free LSI. Proceedings of International Symposium on Low Power Electronics and Design (ISLPED); 2003:193 e8. Keawboonchuay C, Engel TG. Maximum power generation in a piezoelectric pulse generator. IEEE Transactions on Plasma Science 2003;31(1):123 e8. Yang Jiashi, Chen Ziguang, Hu Yuantai. An exact analysis of a rectangular plate piezoelectric generator. IEEE Transactions on Ultrasonics, Ferroelectrics and Frequency Control 2007;54(1):190e5. Mateu Loreto, Moll Francesc. Review of energy harvesting techniques and applications for microelectronics. Proceedings of the SPIE Microtechnologies for the New Millenium 2005;5837:359 e73. Stevels A, Jansen AJ. Renewable energy in portable radios: an environmental benchmarking study. The Journal of Sustainable Product Design; January, 1998:51e5. Justin R. Farmer. A comparison of power harvesting techniques and related energy storage issues M.S Thesis. Dept. Mech. Eng., Virginia Polytechnic Institute and State Univ. Blacksburg, VA; 2007. Blatt FJ, Schroeder PA, Foiles CL, Greig D. Thermoelectric power of materials. New York: Plenum Press; 1976. Nolas GS, Sharp J, Goldsmid HJ. Thermoelectrics: basic principles and new materials developments. New York: Springer; 2001. Mah O. Fundamentals of photovoltaic materials. National Solar Power Research Institute, Inc.; 1998. Damaschke JM. Design of a low-input-voltage converter for thermoelectric generator. IEEE Transactions on Industry Applications 1997;33(5):1203 e7. Myers RL. The basics of physics. Westport, Conn: Greenwood Press; 2006. Elliot RS. Electromagnetics: history, theory, and applications. Piscataway: IEEE Press; 1993. Stephen NG. On energy harvesting from ambient vibration. Journal of Sound and Vibration 2006;293:409e25. Mitcheson PD, Green TC, Yeatman EM, Holmes AS. Architectures for vibrationdriven micropower generators. Journal of Microelectromechanical Systems 2004;13(3):429e40. Rocha JG, Goncalves LM, Rocha PF, Silva MP, Lanceros-Méndez S. Energy harvesting from piezoelectric materials fully integrated in footwear. IEEE Transactions on Industrial Electronics 2010;57(3):813 e9. Silva HR, Afonso JA, Morim PC, Oliveira PM, Correia JH, Rocha LA. Wireless hydrotherapy smart-suit network for posture monitoring. Proceedings of IEEE International Symposium on Industrial Electronics, ISIE; 4 e7 June, 2007:2713e7. Roundy S, Wright PK, Rabaey J. Energy scavenging for wireless sensor networks with special focus on vibrations. Kluwer Academic Press; 2003. Torres Erick O, Rincón-Mora Gabriel A. A 0.7-mm BiCMOS electrostatic energyharvesting system IC. IEEE Journal of Solid State Circuits 2010;45(2):483 e96. Meninger Scott, Mur-Miranda Jose Oscar, Amirtharajah Rajeevan, Chandrakasan Anantha P, Lang Jeffrey H. Vibration-to-electric energy conversion. IEEE Transactions on VLSI Systems 2001;9(1):64 e76. Mitcheson PD, Green TC, Yeatman EM. Power processing circuits for electromagnetic, electrostatic and piezoelectric inertial energy scavengers. Microsystem Technologies 2007;13:1629 e35. Saha CR, O Donnell T, Wang N, McCloskey P. Electromagnetic generator for harvesting energy from human motion. Sensors and Actuators A: Physical 2008;47:248e53. Bharatula NB, Ossevoort S, Stager M, Troster G. Towards wearable autonomous microsystems, pervasive computing. In: Proceeding of the 2nd International conference, Vienna, Austria; 2004. p. 225 e37. Marioli D, Sardini E, Serpelloni M. Electromagnetic generators employing planar inductors for autonomous sensor applications. Procedia Chemistry 2009;1:469e72. Ottman G, Hofmann H, Lesieutre G. Optimized piezoelectric energy harvesting circuit using step-down converter in discontinuous conduction mode. IEEE Transactions on Power Electronics 2003;18(2):696 e703. Williams C, Shearwood C, Harradine M, Mellor P, Birch T, Yates R. Development of an electromagnetic micro-generator. IEE Proceedings Part G: Circuits, Devices and Systems 2001;148(6):337e42. ’

[35] Li W, Ho T, Chan G, Leong P, Wong HY. Infrared signal transmission by a lasermicromachined vibration-induced power generator. In: Proceedings of the 43rd IEEE Midwest symposium on circuits and systems, vol. 1; 2000. p. 236 e9. [36] Khaligh Alireza, Zeng Peng, Zheng Cong. Kinetic energy harvesting using piezoelectric and electromagnetic technologies dstate of the art. IEEE Transactions on Industrial Electronics 2010;57(3):850 e60. [37] Swallow LM, Luo JK, Siores E, Patel I, Dodds D. A piezoelectric � ber composite based energy harvesting device for potential wearable applications. Smart Materials and Structures 2008;17(2):025017. [38] Sari Ibrahim, Balkan Tuna, Külah Haluk. An electromagnetic micro power generator for low-frequency environmental vibrations based on the frequency upconversion technique. IEEE Journal of Microelectromechanical Systems 2010;19(1):14e27. [39] Williams CB, Yates RB. Analysis of a micro-electric generator for microsystems. Physica A 1996;52(1/3):8 e11. [40] Williams CB, Woods RC, Yates RB. Feasibility study of a vibration powered micro-electric generator. Proceedings of IEE Colloquim on Compact Power Sources 1996;96/107:7/1e7/3. [41] Galchev Tzeno, Kim Hanseup, Naja� Khalil. A parametric frequency increased power generator for scavenging low frequency ambient vibrations,. Procedia Chemistry 2009;1:1439e42. [42] Galchev T, Kim H, Naja� K. Non-resonant bi-stable frequency increased power generator for low-frequency ambient vibration. IEEE Solid-State Sensors, Actuators and Microsystems Conference: Transducers; 2009:632 e5. [43] Cuadras A, Gasulla M, Ferrari V. Thermal energy harvesting through pyroelectricity. Sensors and Actuators A: Physical 2010;158:132 e9. [44] Ferrari M, Ferrari V, Marioli D, Taroni A. Modeling, fabrication and performance measurements of a piezoelectric energy converter for power harvesting in autonomous microsystems. IEEE Transactions on Instrumentation and Measurement 2006;55(6):2096 e101. [45] Carlson Eric J, Strunz Kai, Otis Brian P. A 20 mV input boost converter with ef �cient digital control for thermoelectric energy harvesting,. IEEE Journal of Solid State Circuits 2010;45(4):741 e50. [46] Vining C. Semiconductors are cool. Nature 2001;413(6856):577 e8. [47] Bell L. Cooling, heating, generating power, and recovering waste heat with thermoelectric systems. Science 2008;321(5895):1457 e61. [48] Carmo João Paulo, Gonçalves Luis Miguel, Correia José Higino. Thermoelectric microconverter for energy harvesting systems. IEEE Transactions on Industrial Electronics 2010;57(3):861e7. [49] Leono V, Fiorini. P. Thermal matching of a thermoelectric energy scavenger with the ambient. In: Proceedings of 5th European Conference on thermoelectrics (ECT), Odessa, Ukraine; 2007. [50] Jovanovi V, Ghamaty. S. Design, fabrication and testing of energy harvesting thermoelectric generator. In: Matsuzaki Y, editor. Smart structures and materials 2006: smart structures and integrated systems. Japan: Nagoya Univ; 2006. p. 142e9. [51] Stordeur M, Stark I. Low power thermoelectric generator-self-suf �cient energy supply for micro systems. In: I EEE 16th conference on thermoelectrics. Dresden; 1997. p. 575e7. [52] Mi M, Mickle MH, Capelli C, Swift H. RF energy harvesting with multiple antennas in the same space. IEEE Antennas and Propagation Magazine 2005;47(5):100e6. [53] Jabbar Hamid, Song Young S, Jeong Taikyeong Ted. RF energy harvesting system and circuits for charging of mobile devices. IEEE Transactions Consumer Electronics 2010;56(1):247e53. [54] Salter TS. Low power smartdust receiver with novel applications and improvements of an RF power harvesting circuit. PhD Thesis. Department of Electrical Engineering, University of Maryland, USA; 2009. [55] Hagerty J, Helmbrecht F, McCalpin W, Zane R, Popovi&cacute Z. Recycling ambient microwave energy with broad-band rectenna arrays. I EEE Transactions on Microwave Theory and Techniques 2004;52(3):1014 e24. [56] Lhermet Hélène, Condemine Cyril, Plissonnier Marc, Salot Raphaël, Audebert Patrick, Rosset Marion. Ef �cient power management circuit: from thermal energy harvesting to above-IC microbattery energy storage. IEEE Journal of Solid State Circuits 2008;43(1):246e55.

Energy Harvesting State of the Art - Harb 2010

Recommend Documents