A Radio-Frequency Energy Harvesting Scheme for ... - Semantic Scholar
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS—II: EXPRESS BRIEFS, VOL. 59, NO. 9, SEPTEMBER 2012
573
A Radio-Frequency Energy Harvesting Scheme for Use in Low-Power Ad Hoc Distributed Networks Wei Zhao, Kwangsik Choi, Scott Bauman, Zeynep Dilli, Thomas Salter, and Martin Peckerar, Fellow, IEEE
Abstract—While RF energy harvesting has proven to be a viable power source for low-power electronics, it is still a challenge to obtain significant amounts of energy fast and efficiently from the ambiance. Available RF power is usually very weak, resulting in a weak voltage applied to a demodulator to drive it into a region of significant nonlinearity. An RF energy harvesting system consisting of a rectenna, a dc–dc voltage converter, and a novel battery cell is proposed. The rectenna is an integration of an antenna and rectifying diodes. In addition, a switched-capacitor dc–dc voltage converter is integrated on a silicon integrated circuit with energy transfer efficiency as high as 40.5%. A new battery system that can be recharged at low voltage (< 1.2 V) is demonstrated. In this brief, all of these elements are tested as a system to achieve RF battery recharging from a commercially available hand-held communication device. The system exhibited an overall harvesting efficiency of 11.6%. Index Terms—DC–DC voltage converter, low-power distributed networks, rectenna, radio-frequency (RF) energy harvesting.
I. I NTRODUCTION
T
HE GOAL OF research as reported here is to develop a power distribution system (PDS) for an ad hoc distributed sensor network. The target network was described elsewhere in the literature [1]. The PDS included a rectenna for RF energy harvesting, a voltage multiplier, and a battery for energy storage. These are the only system subblocks reported here. The system architecture is shown in Fig. 1. Although, the PDS will include boost and buck converters, regulators, and control electronics. The system challenges addressed here are: 1) Matching the antenna to receive as much energy from free-space as possible and efficiently passing that energy to the voltage multiplier; 2) Boosting the resulting signal so that it could charge a battery; 3) Demonstrating battery recharge from low-power RF sources. Most rectifiers “turn on” at voltages approaching a volt, a value hard to achieve from submicrowatt RF sources. However, as shown elsewhere [2], it is only necessary to achieve an Manuscript received November 14, 2011; revised April 22, 2012; accepted June 21, 2012. Date of publication August 10, 2012; date of current version September 11, 2012. This brief was recommended by Associate Editor B. Sahu. W. Zhao, S. Bauman, Z. Dilli, and M. Peckerar are with the Department of Electrical and Computer Engineering, University of Maryland, College Park, MD 20740 USA (e-mail: [email protected]). K. Choi is with Intel Corporation, Hillsboro, OR 97124 USA. T. Salter is with the Laboratory for Physical Sciences, College Park, MD 20740 USA. Color versions of one or more of the figures in this brief are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/TCSII.2012.2206935
Fig. 1. Block diagram for the energy harvesting system.
asymmetric nonlinear current–voltage relationship in the demodulating diode to achieve rectification. As described in [2], metal–insulator–metal tunnel junctions with exceptionally low turn-on voltages are fabricated to meet this challenge. In this brief, Schottky diodes with low turn-on voltage are used for rectification. The second challenge requires boosting a usually weak RF signal up to a rectified voltage in excess of a volt. This problem is addressed in two ways. First is to lower the recharge voltage as much as possible. This excludes using lithium as a battery storage medium, as its recharge voltage exceeds 3 V. A unique flexible cell based on ruthenium (IV) oxide/zinc redox chemistry has been developed. It has a recharging voltage of less than 1.2 V [3]. In addition, a Cockcroft–Walton capacitor boost converter has been designed and fabricated to meet the multiplication needs of the system. During the operation, the capacitors are charged in parallel and then digitally switched to a series configuration, creating an output voltage equal to the sum of the voltages of the charged capacitors and the input. The “stacked” voltage is sufficient to charge a battery. After charging, the battery provides power for the system as required. In addition, an inductive boost converter is an alternate solution to the capacitive charge pump [4]. While individual components of the PDS system have been reported before, this is the first presentation of whole system performance. In this brief, it has been demonstrated that the battery can be successfully recharged through the combination block using RF signals from a commercially available walkietalkie operating at 900 MHz. The maximum output power from this RF source was 1 W [5]. The PDS thus described provides a bridge between the usually weak RF power and various functional electronic blocks that require a considerable power input. II. B LOCKS IN THE RF E NERGY H ARVESTING S YSTEM A. Broad-Band Rectenna The hybrid antenna/Schottky diode assembly is referred to as a “rectenna” in this brief. The rectenna is designed with a combination of full-wave electromagnetic field analysis and
1549-7747/$31.00 © 2012 IEEE
574
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS—II: EXPRESS BRIEFS, VOL. 59, NO. 9, SEPTEMBER 2012
Fig. 2. Broad-band rectenna.
Fig. 4.
Operations of the dc–dc voltage converter.
Fig. 3. Schematic of the dc–dc voltage converter.
harmonic balance nonlinear analysis for broad-band energy harvesting [6]. It is made by a four-element dual circularly polarized spiral antenna and Schottky diodes, featuring input frequencies from 2 to 18 GHz, as shown in Fig. 2. A surfacemounted Schottky diode from Skyworks, Inc. was utilized for voltage rectification. The rectenna covers a low input power density range from 10−5 to 0.1 mW/cm2 , and achieves power conversion efficiency as high as 20%. The details of design and fabrication can be found in [6]. B. Switched-Capacitor DC–DC Converter Fig. 3 shows the schematic of the switched capacitor dc–dc converter with three stages. The converter chip was implemented using the IBM 8RF 0.13-μm CMOS process. The design includes all the switches for a six-stage converter, except external capacitors. The design details can be found in a previous paper [7]. This provides a maximum available output voltage, which is seven times the input voltage through the external capacitors, and a user can select the output multiplication factor from the minimum (1x input) to the maximum (7x input) by connecting an external electrolytic capacitor’s plus pad to the output. In the schematic, the top switches are made from nMOS transistors only, but the middle and bottom switches are made using transmission gates (combinations of nMOS and pMOS transistors) to transfer the accumulated potential on each capacitor to the next stage with the least loss. As shown in Fig. 4, the converter works in two different phases. In phase one [Q = 1, Q = 0, Fig. 4(b)], all capacitors are connected in parallel and are charged up by the input voltage. The output is disconnected from other components. Then, during the battery charging phase [Q = 0, Q = 1, Fig. 4(c)], the parallel configuration of capacitors is changed into series. The output is connected to the battery; therefore, the battery will be recharged by the capacitor arrays.
Fig. 5. Three-stage converter operation for 0.35-V input at 20-Hz switching using 100-μF external capacitors.
Fig. 5 shows the output (∼1.4 V) of a three-stage converter for a 0.35-V input at a 20-Hz clock signal using three 100-μF external capacitors. Note that for tests in this brief, the body contacts of pMOS in converter switches are connected into an external 1.5-V dc power supply. In actual operation, this external source can be replaced with an extra battery for a fully self-powered system. In that case, two batteries power the system while the other one is being recharged. This will keep the system running continuously. All the batteries will be monitored and controlled by a microcontroller. C. Novel Electrochemical Battery A lightweight flexible thin-film battery is developed using monoparticulate films of activated carbon and ruthenium (IV) oxide. The battery offers the highest specific charge storage capacity of any commercially available thin-film cell (> 20 mAh · cm−2 ) [8], [9]. While typical secondary power cells (i.e., those capable of recharging) require relatively high recharge voltages, (Lithium, for example, requires > 3 V to recharge [3], [8]) this cell can be recharged at an exceptionally low voltage of 1.2 V (open-circuit voltage of the battery after being fully charged). Fig. 6 shows two types of battery prepared in the laboratory: (a) rigid and (b) flexible.
ZHAO et al.: RF ENERGY HARVESTING SCHEME FOR USE IN LOW-POWER AD HOC DISTRIBUTED NETWORKS
Fig. 6.
575
(a) Rigid and (b) flexible battery cell.
Fig. 8. Frequency dependence of the 1000-μF capacitor charging with 0.35-V input and three 100-μF external storing capacitors. TABLE I F REQUENCY D EPENDENCE OF THE C ONVERTER IN THE 1000-μF CAPACITOR CHARGING
Fig. 7. First 40 h of a battery’s discharge curve. The RC-discharge profile fit to the capacitive region of the battery operation.
The discharge profiles of the battery cells demonstrate distinct regions of operation. Fig. 7 shows the first 40 h of a battery’s discharge curve over a 1-KΩ load. The initial region of the discharge curve (blue) is a mixture of capacitive behavior and cathode material reduction. The capacitance of the battery can be modeled by fitting an RC-transient curve (red) to this portion of the discharge data; therefore, an estimate of the capacitance could be obtained. This test has been performed across a range of cells to obtain capacitance values ranging from 1.78 to 4.63 F. The flexible battery is designed for ease of reel-to-reel manufacture. We have developed special coating tools capable of depositing particulate layers one nanoparticle in thickness at a time. We can “compound” the particle mixtures in the coating process, minimizing the use of expensive materials and providing precise control of the chemical composition of the layers. This is summarized in [8]. III. C HARGING E XPERIMENTS A. Large-Capacitor Charging In the initial charging tests, a large capacitor (1000 μF) replaced a battery to verify the functionality of the voltage converter. Here, the converter only used three stages. In addition, a number of different clock signal (provided by external signal generator) frequencies with a 50% duty cycle have been tested to find the optimum one. The external capacitors are 100 μF, and the input voltage is 0.35 V. Fig. 8 shows the results of measurement, and the inner graph is a magnified version of 400–700-s section segments of the charging period. All tests are stopped when the capacitor volt-
age reaches 1.38 V. The plots for the 2-, 20-, and 200-Hz tests show similar charging curves, indicating minimal frequency dependence for the range of frequencies used. As shown in Table I, the 0.2-, 2-, and 20-Hz tests have almost the same results for energy transfer efficiency, approximately 40%, for a charging time of about 10 min. The supplied energy is calculated by integrating the measured current from power supply, and the stored energy is calculated using the 0.5CV2 relationship. As shown in Fig. 8, as the clock frequency goes higher, the saturation voltage goes lower. This is the result of the frequency dependence of the dynamic power consumption of CMOS switches [10]. As a result of these tests, a 0.2–20 Hz clock frequency range is used for battery charging tests. B. Battery Charging With the Proposed Converter The battery recharge test includes three steps as follows: 1) initial discharge; 2) recharge; and 3) second discharge. The first discharge consumes part of the battery’s energy before the recharge. From the second discharge, the change in potential after the converter charges the battery can be observed, by comparison with the first discharge. The battery was discharged through a 1-KΩ load for 16 h at first. As shown in Fig. 9(a), the load voltage dropped from 1.171 to 0.735 V. The following charging test was set up with a 0.25-V input from the external power supply and six 100-μF external storing capacitors for a six-stage converter. This gave a 1.75-V open-circuit output. A 50% duty cycle and a 1-Hz clock
576
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS—II: EXPRESS BRIEFS, VOL. 59, NO. 9, SEPTEMBER 2012
Fig. 10. Experimental setup (the voltage converter is an integrated silicon IC in package being tested on the breadboard).
Fig. 9. Results of the battery cycle test. (a) Discharging test results for 16 h (1-KΩ load). (b) Potential change during 24-h charging (0.25-V input, 1-Hz clock signal), and (c) 10-min discharging after charging (1 KΩ load).
signal were implemented to control the charging process. The battery was charged for 24 h. Fig. 9(b) shows the increasing potential during charging. Then the battery was connected to the 1-KΩ load again for 10 min in order to monitor the potential change. As shown in Fig. 9(c), the battery started to provide energy from 1.002 V. Compared with the load voltage before charging, i.e., 0.735 V, the results demonstrated that the dc–dc converter is working properly to charge the battery. C. Battery Charging With the Converter and the Rectenna in Place In this experiment, a walkie-talkie held 0.5 m from the rectenna was used as an RF source generator. It operates in the 900-MHz ISM frequency band, and has a maximum output power of 1 W [5]. Note that the output voltage of the antenna varies with the distance to the source. The distance of 0.5 m is chosen because, at this distance, the rectenna provides an output that is suitable to be boosted. The rectenna receives a power of 0.255 mW, and a Poynting Theorem calculation yields an input voltage of 1.83 V. Measurement on the output end of the
rectenna shows an open-circuit output voltage around 0.35 V with millivolt ripples. In addition, the antenna is able to provide 34-μW power to its 47-Ω load. Next, the components are tested together as a system. The antenna received the signal and coupled that signal to the diodes for demodulation. The dc output from the rectenna was connected to the input of the voltage converter. The voltage converter boosted up the dc signal and charged the battery. The whole system is shown in Fig. 10. In addition, since the output from the rectenna is already a dc signal exhibiting millivolt ripples, no decoupling capacitor is used between the rectenna and the converter. This test procedure also follows the similar three steps in the previous test. The battery was discharged through a 1-KΩ load for 7 min at first. As shown in Fig. 11(a), the load voltage dropped from 0.782 to 0.631 V. The following charging test was set up with a walkie-talkie, a rectenna, and a six-stage converter (implemented by six 100 μF external storing capacitors). A 50% duty cycle and a 2-Hz clock signal was implemented to control the charging process. The battery was charged for 1 h. Fig. 11(b) shows the increasing potential during the charging process. Next, the battery was connected to the 1-KΩ load again for 7 min, in order to observe the change in potential. As shown in Fig. 11(a), the battery started to provide energy from 0.710 V, compared with the load voltage before charging, i.e., 0.631 V. The results show that the battery is successfully charged, and the system is working properly. The charging efficiency of this system can be calculated as follows. Assuming that the walkie-talkie transmits power isotropically, the input power to the system obeys the 1/r2 rule and is also proportional to the antenna area. r is the distance between the walkie-talkie and the rectenna. Therefore, the charging efficiency is the ratio of energy stored in the battery to the total input energy, whereas this battery cell can be modeled as a 2-F supercapacitor, with the method described in Section II-C. A charging efficiency of 11.6% has been achieved, as shown in Table II. The power consumed by the converter Pconverter is also measured with a current meter in series with the power supply. To verify this efficiency, another calculation is performed. The final battery potential of the first discharge, i.e., 0.631 V, is used as a reference. Next, the output power of the second discharge is integrated over the time period from the beginning of discharge, until the voltage drops back to 0.631 V. The integration result is the extra energy that is recharged into the
ZHAO et al.: RF ENERGY HARVESTING SCHEME FOR USE IN LOW-POWER AD HOC DISTRIBUTED NETWORKS
577
(microcontroller and voltage converters), and the third one is recharged from the energy harvester. One point worth mentioning is that lithium batteries may be damaged by overcharging. However, our battery still works normally even if it is overcharged. It has been charged to 1.4 V with no significant damage observed. The control circuit can also prevent possible overcharging. The main challenge ahead is to minimize power consumption in these control peripherals. Some possible methods to improve efficiency may include implementing a different type of a rectenna with higher efficiency, using multiple voltage converters in parallel to keep charging the battery continuously, and upgrading the performance of the battery cell. V. C ONCLUSION An RF energy harvesting system with a rectenna, an integrated dc–dc voltage converter, and a novel battery cell is presented. The functionality of the system is verified by recharging the battery cell with RF signals from a walkietalkie as input. An energy harvesting efficiency of 11.6% has been achieved. While applications centering on RF energy input are highlighted here, the proposed converter and battery combination is also applicable to other harvesting scenarios, where the input power is insufficient to drive the load. These might include vibrational, geothermal, or low-light-level solar energy harvesting. R EFERENCES Fig. 11. (a) Comparison of two discharging processes. (b) Change of battery potential during charging. TABLE II R ELATED VALUES IN C ALCULATION OF C HARGING E FFICIENCY
battery. This yields a stored energy of 118.5 mJ and an overall efficiency of 12.9%. Similar tests have been done recently in the literature, in which 12.39-μW output power has been generated at the 316.23-μW input power level [11]. Compared with the previous work, the system described here harvests from a lower input power (254.6 μW) with a higher efficiency, and it allows for energy storage through the battery recharge process as well. IV. D ISCUSSION To build a “stand-alone” system, the ultralow-power microcontroller EM6819 [12] is used to monitor the batteries’ potential, control the charging process, switch between the two batteries, and provide the clock signal. Tests show that the microcontroller consumes less than 25-μW of power in operation. In addition, a solar cell harvester (currently in design) will also be integrated into the system to improve the power budget, and the system will be able to harvest energy regardless of the time or weather conditions. The complete system will have three batteries. One battery drives the load, another one drives ancillary components
[1] C.-C. Shen, R. Kupershtok, S. Adl, S. S. Bhattacharyya, N. Goldsman, and M. Peckerar, “Sensor support systems for asymmetric threat countermeasures,” IEEE Sens. J., vol. 8, no. 6, pp. 682–692, Jun. 2008. [2] K. Choi, F. Yesilkoy, A. Chryssis, M. Dagenais, and M. Peckerar, “New process development for planar type CIC tunneling diodes,” IEEE Electron Device Lett., vol. 31, no. 8, pp. 809–811, Jul. 2010. [3] Y. T. Ngu, Z. Dilli, M. Peckerar, and N. Goldsman, “High capacitance battery for powering distributed networks node devices,” in Proc. ISDRS, College Park, MD, Dec. 2007, pp. 1–2. [4] T. S. Salter, G. Metze, and N. Goldsman, “Improved RF power harvesting circuit design,” in Proc. Int. Semicond. Device Res. Symp., Dec. 12–14, 2007, pp. 1–2. [5] Owner’s Manual, Model TSX300, TriSquare Electronics Corporation, Kansas City, MO, 2009. [6] J. A. Hagerty, F. B. Helmbrecht, W. H. McCalpin, R. Zane, and Z. B. Popovic, “Recycling ambient microwave energy with broad-band rectenna arrays,” IEEE Trans. Microw. Theory Tech., vol. 52, no. 3, pp. 1014–1024, Mar. 2004. [7] T. Salter, K. Choi, M. Peckerar, G. Metze, and N. Goldsman, “RF energy scavenging system utilizing switched capacitor DC–DC converter,” Electron. Lett., vol. 45, no. 7, pp. 374–376, Mar. 2009. [8] M. Peckerar, M. Dornajafi, Z. Dilli, N. Goldsman, Y. Ngu, B. Boerger, N. V. Wyck, J. Gravelin, B. Grenon, R. B. Proctor, and D. A. Lowy, “Fabrication of flexible ultracapacitor/galvanic cell hybrids using advanced nanoparticle coating technology,” J. Vac. Sci. Technol. B, Microelectron. Nanometer Struct., vol. 29, no. 1, pp. 011008-1–011008-6, Jan./Feb. 2011. [9] M. Peckerar, Z. Dilli, M. Dornajafi, N. Goldsman, Y. Ngu, R. B. Proctor, B. J. Krupsaw, and D. A. Lowy, “A novel high energy density flexible galvanic cell,” Energy Environ. Sci., vol. 4, no. 5, pp. 1807–1812, Apr. 2011. [10] R. J. Baker, H. W. Li, and D. E. Boyce, CMOS, Circuit Design, Layout, and Simulation. New York: IEEE Press, 1998. [11] H. Jabbar, Y. S. Song, and T. T. Jeong, “RF energy harvesting system and circuits for charging of mobile devices,” IEEE Trans. Consum. Electron., vol. 56, no. 1, pp. 247–253, Feb. 2010. [12] EM6819 Datasheet, EM Microelectronic, Marin-Epagnier, Switzerland, 2010.
573
A Radio-Frequency Energy Harvesting Scheme for Use in Low-Power Ad Hoc Distributed Networks Wei Zhao, Kwangsik Choi, Scott Bauman, Zeynep Dilli, Thomas Salter, and Martin Peckerar, Fellow, IEEE
Abstract—While RF energy harvesting has proven to be a viable power source for low-power electronics, it is still a challenge to obtain significant amounts of energy fast and efficiently from the ambiance. Available RF power is usually very weak, resulting in a weak voltage applied to a demodulator to drive it into a region of significant nonlinearity. An RF energy harvesting system consisting of a rectenna, a dc–dc voltage converter, and a novel battery cell is proposed. The rectenna is an integration of an antenna and rectifying diodes. In addition, a switched-capacitor dc–dc voltage converter is integrated on a silicon integrated circuit with energy transfer efficiency as high as 40.5%. A new battery system that can be recharged at low voltage (< 1.2 V) is demonstrated. In this brief, all of these elements are tested as a system to achieve RF battery recharging from a commercially available hand-held communication device. The system exhibited an overall harvesting efficiency of 11.6%. Index Terms—DC–DC voltage converter, low-power distributed networks, rectenna, radio-frequency (RF) energy harvesting.
I. I NTRODUCTION
T
HE GOAL OF research as reported here is to develop a power distribution system (PDS) for an ad hoc distributed sensor network. The target network was described elsewhere in the literature [1]. The PDS included a rectenna for RF energy harvesting, a voltage multiplier, and a battery for energy storage. These are the only system subblocks reported here. The system architecture is shown in Fig. 1. Although, the PDS will include boost and buck converters, regulators, and control electronics. The system challenges addressed here are: 1) Matching the antenna to receive as much energy from free-space as possible and efficiently passing that energy to the voltage multiplier; 2) Boosting the resulting signal so that it could charge a battery; 3) Demonstrating battery recharge from low-power RF sources. Most rectifiers “turn on” at voltages approaching a volt, a value hard to achieve from submicrowatt RF sources. However, as shown elsewhere [2], it is only necessary to achieve an Manuscript received November 14, 2011; revised April 22, 2012; accepted June 21, 2012. Date of publication August 10, 2012; date of current version September 11, 2012. This brief was recommended by Associate Editor B. Sahu. W. Zhao, S. Bauman, Z. Dilli, and M. Peckerar are with the Department of Electrical and Computer Engineering, University of Maryland, College Park, MD 20740 USA (e-mail: [email protected]). K. Choi is with Intel Corporation, Hillsboro, OR 97124 USA. T. Salter is with the Laboratory for Physical Sciences, College Park, MD 20740 USA. Color versions of one or more of the figures in this brief are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/TCSII.2012.2206935
Fig. 1. Block diagram for the energy harvesting system.
asymmetric nonlinear current–voltage relationship in the demodulating diode to achieve rectification. As described in [2], metal–insulator–metal tunnel junctions with exceptionally low turn-on voltages are fabricated to meet this challenge. In this brief, Schottky diodes with low turn-on voltage are used for rectification. The second challenge requires boosting a usually weak RF signal up to a rectified voltage in excess of a volt. This problem is addressed in two ways. First is to lower the recharge voltage as much as possible. This excludes using lithium as a battery storage medium, as its recharge voltage exceeds 3 V. A unique flexible cell based on ruthenium (IV) oxide/zinc redox chemistry has been developed. It has a recharging voltage of less than 1.2 V [3]. In addition, a Cockcroft–Walton capacitor boost converter has been designed and fabricated to meet the multiplication needs of the system. During the operation, the capacitors are charged in parallel and then digitally switched to a series configuration, creating an output voltage equal to the sum of the voltages of the charged capacitors and the input. The “stacked” voltage is sufficient to charge a battery. After charging, the battery provides power for the system as required. In addition, an inductive boost converter is an alternate solution to the capacitive charge pump [4]. While individual components of the PDS system have been reported before, this is the first presentation of whole system performance. In this brief, it has been demonstrated that the battery can be successfully recharged through the combination block using RF signals from a commercially available walkietalkie operating at 900 MHz. The maximum output power from this RF source was 1 W [5]. The PDS thus described provides a bridge between the usually weak RF power and various functional electronic blocks that require a considerable power input. II. B LOCKS IN THE RF E NERGY H ARVESTING S YSTEM A. Broad-Band Rectenna The hybrid antenna/Schottky diode assembly is referred to as a “rectenna” in this brief. The rectenna is designed with a combination of full-wave electromagnetic field analysis and
1549-7747/$31.00 © 2012 IEEE
574
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS—II: EXPRESS BRIEFS, VOL. 59, NO. 9, SEPTEMBER 2012
Fig. 2. Broad-band rectenna.
Fig. 4.
Operations of the dc–dc voltage converter.
Fig. 3. Schematic of the dc–dc voltage converter.
harmonic balance nonlinear analysis for broad-band energy harvesting [6]. It is made by a four-element dual circularly polarized spiral antenna and Schottky diodes, featuring input frequencies from 2 to 18 GHz, as shown in Fig. 2. A surfacemounted Schottky diode from Skyworks, Inc. was utilized for voltage rectification. The rectenna covers a low input power density range from 10−5 to 0.1 mW/cm2 , and achieves power conversion efficiency as high as 20%. The details of design and fabrication can be found in [6]. B. Switched-Capacitor DC–DC Converter Fig. 3 shows the schematic of the switched capacitor dc–dc converter with three stages. The converter chip was implemented using the IBM 8RF 0.13-μm CMOS process. The design includes all the switches for a six-stage converter, except external capacitors. The design details can be found in a previous paper [7]. This provides a maximum available output voltage, which is seven times the input voltage through the external capacitors, and a user can select the output multiplication factor from the minimum (1x input) to the maximum (7x input) by connecting an external electrolytic capacitor’s plus pad to the output. In the schematic, the top switches are made from nMOS transistors only, but the middle and bottom switches are made using transmission gates (combinations of nMOS and pMOS transistors) to transfer the accumulated potential on each capacitor to the next stage with the least loss. As shown in Fig. 4, the converter works in two different phases. In phase one [Q = 1, Q = 0, Fig. 4(b)], all capacitors are connected in parallel and are charged up by the input voltage. The output is disconnected from other components. Then, during the battery charging phase [Q = 0, Q = 1, Fig. 4(c)], the parallel configuration of capacitors is changed into series. The output is connected to the battery; therefore, the battery will be recharged by the capacitor arrays.
Fig. 5. Three-stage converter operation for 0.35-V input at 20-Hz switching using 100-μF external capacitors.
Fig. 5 shows the output (∼1.4 V) of a three-stage converter for a 0.35-V input at a 20-Hz clock signal using three 100-μF external capacitors. Note that for tests in this brief, the body contacts of pMOS in converter switches are connected into an external 1.5-V dc power supply. In actual operation, this external source can be replaced with an extra battery for a fully self-powered system. In that case, two batteries power the system while the other one is being recharged. This will keep the system running continuously. All the batteries will be monitored and controlled by a microcontroller. C. Novel Electrochemical Battery A lightweight flexible thin-film battery is developed using monoparticulate films of activated carbon and ruthenium (IV) oxide. The battery offers the highest specific charge storage capacity of any commercially available thin-film cell (> 20 mAh · cm−2 ) [8], [9]. While typical secondary power cells (i.e., those capable of recharging) require relatively high recharge voltages, (Lithium, for example, requires > 3 V to recharge [3], [8]) this cell can be recharged at an exceptionally low voltage of 1.2 V (open-circuit voltage of the battery after being fully charged). Fig. 6 shows two types of battery prepared in the laboratory: (a) rigid and (b) flexible.
ZHAO et al.: RF ENERGY HARVESTING SCHEME FOR USE IN LOW-POWER AD HOC DISTRIBUTED NETWORKS
Fig. 6.
575
(a) Rigid and (b) flexible battery cell.
Fig. 8. Frequency dependence of the 1000-μF capacitor charging with 0.35-V input and three 100-μF external storing capacitors. TABLE I F REQUENCY D EPENDENCE OF THE C ONVERTER IN THE 1000-μF CAPACITOR CHARGING
Fig. 7. First 40 h of a battery’s discharge curve. The RC-discharge profile fit to the capacitive region of the battery operation.
The discharge profiles of the battery cells demonstrate distinct regions of operation. Fig. 7 shows the first 40 h of a battery’s discharge curve over a 1-KΩ load. The initial region of the discharge curve (blue) is a mixture of capacitive behavior and cathode material reduction. The capacitance of the battery can be modeled by fitting an RC-transient curve (red) to this portion of the discharge data; therefore, an estimate of the capacitance could be obtained. This test has been performed across a range of cells to obtain capacitance values ranging from 1.78 to 4.63 F. The flexible battery is designed for ease of reel-to-reel manufacture. We have developed special coating tools capable of depositing particulate layers one nanoparticle in thickness at a time. We can “compound” the particle mixtures in the coating process, minimizing the use of expensive materials and providing precise control of the chemical composition of the layers. This is summarized in [8]. III. C HARGING E XPERIMENTS A. Large-Capacitor Charging In the initial charging tests, a large capacitor (1000 μF) replaced a battery to verify the functionality of the voltage converter. Here, the converter only used three stages. In addition, a number of different clock signal (provided by external signal generator) frequencies with a 50% duty cycle have been tested to find the optimum one. The external capacitors are 100 μF, and the input voltage is 0.35 V. Fig. 8 shows the results of measurement, and the inner graph is a magnified version of 400–700-s section segments of the charging period. All tests are stopped when the capacitor volt-
age reaches 1.38 V. The plots for the 2-, 20-, and 200-Hz tests show similar charging curves, indicating minimal frequency dependence for the range of frequencies used. As shown in Table I, the 0.2-, 2-, and 20-Hz tests have almost the same results for energy transfer efficiency, approximately 40%, for a charging time of about 10 min. The supplied energy is calculated by integrating the measured current from power supply, and the stored energy is calculated using the 0.5CV2 relationship. As shown in Fig. 8, as the clock frequency goes higher, the saturation voltage goes lower. This is the result of the frequency dependence of the dynamic power consumption of CMOS switches [10]. As a result of these tests, a 0.2–20 Hz clock frequency range is used for battery charging tests. B. Battery Charging With the Proposed Converter The battery recharge test includes three steps as follows: 1) initial discharge; 2) recharge; and 3) second discharge. The first discharge consumes part of the battery’s energy before the recharge. From the second discharge, the change in potential after the converter charges the battery can be observed, by comparison with the first discharge. The battery was discharged through a 1-KΩ load for 16 h at first. As shown in Fig. 9(a), the load voltage dropped from 1.171 to 0.735 V. The following charging test was set up with a 0.25-V input from the external power supply and six 100-μF external storing capacitors for a six-stage converter. This gave a 1.75-V open-circuit output. A 50% duty cycle and a 1-Hz clock
576
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS—II: EXPRESS BRIEFS, VOL. 59, NO. 9, SEPTEMBER 2012
Fig. 10. Experimental setup (the voltage converter is an integrated silicon IC in package being tested on the breadboard).
Fig. 9. Results of the battery cycle test. (a) Discharging test results for 16 h (1-KΩ load). (b) Potential change during 24-h charging (0.25-V input, 1-Hz clock signal), and (c) 10-min discharging after charging (1 KΩ load).
signal were implemented to control the charging process. The battery was charged for 24 h. Fig. 9(b) shows the increasing potential during charging. Then the battery was connected to the 1-KΩ load again for 10 min in order to monitor the potential change. As shown in Fig. 9(c), the battery started to provide energy from 1.002 V. Compared with the load voltage before charging, i.e., 0.735 V, the results demonstrated that the dc–dc converter is working properly to charge the battery. C. Battery Charging With the Converter and the Rectenna in Place In this experiment, a walkie-talkie held 0.5 m from the rectenna was used as an RF source generator. It operates in the 900-MHz ISM frequency band, and has a maximum output power of 1 W [5]. Note that the output voltage of the antenna varies with the distance to the source. The distance of 0.5 m is chosen because, at this distance, the rectenna provides an output that is suitable to be boosted. The rectenna receives a power of 0.255 mW, and a Poynting Theorem calculation yields an input voltage of 1.83 V. Measurement on the output end of the
rectenna shows an open-circuit output voltage around 0.35 V with millivolt ripples. In addition, the antenna is able to provide 34-μW power to its 47-Ω load. Next, the components are tested together as a system. The antenna received the signal and coupled that signal to the diodes for demodulation. The dc output from the rectenna was connected to the input of the voltage converter. The voltage converter boosted up the dc signal and charged the battery. The whole system is shown in Fig. 10. In addition, since the output from the rectenna is already a dc signal exhibiting millivolt ripples, no decoupling capacitor is used between the rectenna and the converter. This test procedure also follows the similar three steps in the previous test. The battery was discharged through a 1-KΩ load for 7 min at first. As shown in Fig. 11(a), the load voltage dropped from 0.782 to 0.631 V. The following charging test was set up with a walkie-talkie, a rectenna, and a six-stage converter (implemented by six 100 μF external storing capacitors). A 50% duty cycle and a 2-Hz clock signal was implemented to control the charging process. The battery was charged for 1 h. Fig. 11(b) shows the increasing potential during the charging process. Next, the battery was connected to the 1-KΩ load again for 7 min, in order to observe the change in potential. As shown in Fig. 11(a), the battery started to provide energy from 0.710 V, compared with the load voltage before charging, i.e., 0.631 V. The results show that the battery is successfully charged, and the system is working properly. The charging efficiency of this system can be calculated as follows. Assuming that the walkie-talkie transmits power isotropically, the input power to the system obeys the 1/r2 rule and is also proportional to the antenna area. r is the distance between the walkie-talkie and the rectenna. Therefore, the charging efficiency is the ratio of energy stored in the battery to the total input energy, whereas this battery cell can be modeled as a 2-F supercapacitor, with the method described in Section II-C. A charging efficiency of 11.6% has been achieved, as shown in Table II. The power consumed by the converter Pconverter is also measured with a current meter in series with the power supply. To verify this efficiency, another calculation is performed. The final battery potential of the first discharge, i.e., 0.631 V, is used as a reference. Next, the output power of the second discharge is integrated over the time period from the beginning of discharge, until the voltage drops back to 0.631 V. The integration result is the extra energy that is recharged into the
ZHAO et al.: RF ENERGY HARVESTING SCHEME FOR USE IN LOW-POWER AD HOC DISTRIBUTED NETWORKS
577
(microcontroller and voltage converters), and the third one is recharged from the energy harvester. One point worth mentioning is that lithium batteries may be damaged by overcharging. However, our battery still works normally even if it is overcharged. It has been charged to 1.4 V with no significant damage observed. The control circuit can also prevent possible overcharging. The main challenge ahead is to minimize power consumption in these control peripherals. Some possible methods to improve efficiency may include implementing a different type of a rectenna with higher efficiency, using multiple voltage converters in parallel to keep charging the battery continuously, and upgrading the performance of the battery cell. V. C ONCLUSION An RF energy harvesting system with a rectenna, an integrated dc–dc voltage converter, and a novel battery cell is presented. The functionality of the system is verified by recharging the battery cell with RF signals from a walkietalkie as input. An energy harvesting efficiency of 11.6% has been achieved. While applications centering on RF energy input are highlighted here, the proposed converter and battery combination is also applicable to other harvesting scenarios, where the input power is insufficient to drive the load. These might include vibrational, geothermal, or low-light-level solar energy harvesting. R EFERENCES Fig. 11. (a) Comparison of two discharging processes. (b) Change of battery potential during charging. TABLE II R ELATED VALUES IN C ALCULATION OF C HARGING E FFICIENCY
battery. This yields a stored energy of 118.5 mJ and an overall efficiency of 12.9%. Similar tests have been done recently in the literature, in which 12.39-μW output power has been generated at the 316.23-μW input power level [11]. Compared with the previous work, the system described here harvests from a lower input power (254.6 μW) with a higher efficiency, and it allows for energy storage through the battery recharge process as well. IV. D ISCUSSION To build a “stand-alone” system, the ultralow-power microcontroller EM6819 [12] is used to monitor the batteries’ potential, control the charging process, switch between the two batteries, and provide the clock signal. Tests show that the microcontroller consumes less than 25-μW of power in operation. In addition, a solar cell harvester (currently in design) will also be integrated into the system to improve the power budget, and the system will be able to harvest energy regardless of the time or weather conditions. The complete system will have three batteries. One battery drives the load, another one drives ancillary components
[1] C.-C. Shen, R. Kupershtok, S. Adl, S. S. Bhattacharyya, N. Goldsman, and M. Peckerar, “Sensor support systems for asymmetric threat countermeasures,” IEEE Sens. J., vol. 8, no. 6, pp. 682–692, Jun. 2008. [2] K. Choi, F. Yesilkoy, A. Chryssis, M. Dagenais, and M. Peckerar, “New process development for planar type CIC tunneling diodes,” IEEE Electron Device Lett., vol. 31, no. 8, pp. 809–811, Jul. 2010. [3] Y. T. Ngu, Z. Dilli, M. Peckerar, and N. Goldsman, “High capacitance battery for powering distributed networks node devices,” in Proc. ISDRS, College Park, MD, Dec. 2007, pp. 1–2. [4] T. S. Salter, G. Metze, and N. Goldsman, “Improved RF power harvesting circuit design,” in Proc. Int. Semicond. Device Res. Symp., Dec. 12–14, 2007, pp. 1–2. [5] Owner’s Manual, Model TSX300, TriSquare Electronics Corporation, Kansas City, MO, 2009. [6] J. A. Hagerty, F. B. Helmbrecht, W. H. McCalpin, R. Zane, and Z. B. Popovic, “Recycling ambient microwave energy with broad-band rectenna arrays,” IEEE Trans. Microw. Theory Tech., vol. 52, no. 3, pp. 1014–1024, Mar. 2004. [7] T. Salter, K. Choi, M. Peckerar, G. Metze, and N. Goldsman, “RF energy scavenging system utilizing switched capacitor DC–DC converter,” Electron. Lett., vol. 45, no. 7, pp. 374–376, Mar. 2009. [8] M. Peckerar, M. Dornajafi, Z. Dilli, N. Goldsman, Y. Ngu, B. Boerger, N. V. Wyck, J. Gravelin, B. Grenon, R. B. Proctor, and D. A. Lowy, “Fabrication of flexible ultracapacitor/galvanic cell hybrids using advanced nanoparticle coating technology,” J. Vac. Sci. Technol. B, Microelectron. Nanometer Struct., vol. 29, no. 1, pp. 011008-1–011008-6, Jan./Feb. 2011. [9] M. Peckerar, Z. Dilli, M. Dornajafi, N. Goldsman, Y. Ngu, R. B. Proctor, B. J. Krupsaw, and D. A. Lowy, “A novel high energy density flexible galvanic cell,” Energy Environ. Sci., vol. 4, no. 5, pp. 1807–1812, Apr. 2011. [10] R. J. Baker, H. W. Li, and D. E. Boyce, CMOS, Circuit Design, Layout, and Simulation. New York: IEEE Press, 1998. [11] H. Jabbar, Y. S. Song, and T. T. Jeong, “RF energy harvesting system and circuits for charging of mobile devices,” IEEE Trans. Consum. Electron., vol. 56, no. 1, pp. 247–253, Feb. 2010. [12] EM6819 Datasheet, EM Microelectronic, Marin-Epagnier, Switzerland, 2010.
