Efficient and Sensitive Electrically Small Rectenna for Ultra-Low Power RF Energy Harvesting

Transcription

1 Effcent and Senstve Electrcally Small Rectenna for Ultra-Low Power RF Energy Harvestng Assmons, S. D., Fusco, V., Georgads, A., & Samaras, T. (8). Effcent and Senstve Electrcally Small Rectenna for Ultra-Low Power RF Energy Harvestng. Nature Scentfc Reports, 8, [58]. Publshed n: Nature Scentfc Reports Document Verson: Publsher's PDF, also known as Verson of record Queen's Unversty Belfast - Research Portal: Lnk to publcaton record n Queen's Unversty Belfast Research Portal Publsher rghts Copyrght 8 the authors. Ths s an open access artcle publshed under a Creatve Commons Attrbuton Lcense ( whch permts unrestrcted use, dstrbuton and reproducton n any medum, provded the author and source are cted. General rghts Copyrght for the publcatons made accessble va the Queen's Unversty Belfast Research Portal s retaned by the author(s) and / or other copyrght owners and t s a condton of accessng these publcatons that users recognse and abde by the legal requrements assocated wth these rghts. Take down polcy The Research Portal s Queen's nsttutonal repostory that provdes access to Queen's research output. Every effort has been made to ensure that content n the Research Portal does not nfrnge any person's rghts, or applcable UK laws. If you dscover content n the Research Portal that you beleve breaches copyrght or volates any law, please contact openaccess@qub.ac.uk. Download date:. Jan. 9

2 Receved: May 8 Accepted: September 8 Publshed: xx xx xxxx OPEN Effcent and Senstve Electrcally Small Rectenna for Ultra-Low Power RF Energy Harvestng Stylanos D. Assmons, Vncent Fusco, Apostolos Georgads & Theodoros Samaras A new electrcally small antenna wth sze ka =.45 s presented, fabrcated and measured n ths work. Ths s ntrnscally matched to 5 Ω, has omn-drectonal and lnear-polarzed radaton pattern n the horzontal plane wth maxmum drectvty of.75 db and smulated radaton effcency of 9%. The antenna n combnaton wth a low-complex and co-planar rectfer wth one sngle dode forms a hgh effcent and senstve electrcally small rectenna wth ka =.5 at 868 MHz (UHF RFID-band n Europe). The latter has measured effcency.5% for 9 dbm power nput and senstvty of 44 dbm (or equvalently.8 μw/cm power densty), whle at.5 μw/cm s able to supply contnuously,.e., wthout a boost converter or use of any energy tank, a small electrcal devce wth 8 μw. In order to ncrease the dc output voltage and the delvered dc power to the load for lower power densty levels, rectenna-array confguraton s exploted. Applcaton to batteryless, backscatter wreless sensor node powerng s dscussed. Specfcally, for a power densty of.7 μw/cm the RF energy harvestng system delvers 7 μw at.85 V every.5 s. The development of new deployed wreless moble technologes such as Mult-Input Mult-Output (MIMO) systems, Wreless Local Area Networks (WLAN), Rado Frequency Identfcaton (RFID) and Wreless Sensor Networks (WSN) leads to the need for antennas wth relatvely small sze, but wthout antenna performance degradaton. Based on the latter, electrcally small antennas have ganed ncreasng attenton n recent years 9 snce Wheeler s study. There are many defntons for the electrcally small antennas: based on Wheeler s crteron an antenna s electrcally small when t can be crcumscrbed by a radan sphere of radus λ/π, whle based on Kng s crteron, the accepted electrcal sze lmt for an electrcally small antenna s ka 5,. where k = πλ / s the wavenumber and a s the radus of a sphere, whch encloses the antenna. Electrcally small antenna characterstcs consst of relatvely small sze (.e., a small fracton of the wavelength), low nput resstance (.e., hghly capactve/nductve nput mpedance), and characterstc omndrectonal radaton pattern. Usually, small antenna mpedance matchng to 5 Ω load s a dffcult procedure, whle the radaton pattern tends to approach the typcal dpole omndrectonal radaton pattern, wth maxmum drectvty equal to.5. Energy harvestng through RF-to-dc rectfcaton, has been ganng ground over the last decade 7 9,4 7. The number of RF emtters has been rapdly ncreasng due to the development of new wreless technologes and t remans an engneerng challenge how to capture unused ambent RF energy and use ths to supply small electrcal devces, such as backscatter rado sensor networks 8. In typcal RF-to-dc rectfcaton systems, an antenna s combned wth a rectfer, whch manly conssts of one or more dodes n specfc confguratons, formng a rectenna. The man desgn lmtaton n a rectenna s the relatvely low avalable ambent power densty level as well as the senstvty (.e., the ablty to harvest energy and operate at low power densty). Based on long-term RF EMF measurements n the European Unon (EU), power densty vares from.7 to.8594 μw/cm. Hence, t s clear that, n order to fully explot and harvest the RF energy avalable n the envronment, rectennas should be desgned to operate at ultra-low power denstes. In order to ncrease the senstvty for a gven power densty, losses nserted by the rectfer dodes, the matchng School of Electroncs, Electrcal Engneerng and Computer Scence, Queen s Unversty Belfast, Belfast, BT 9DT, Unted Kngdom. School of Engneerng and Physcal Scences, Herot-Watt Unversty, Ednburgh, EH4 4AS, Unted Kngdom. School of Physcs, Arstotle Unversty of Thessalonk, Thessalonk, 544, Greece. Correspondence and requests for materals should be addressed to S.D.A. (emal: s.assmons@qub.ac.uk) Scentfc REPOrts (8) 8:58 DOI:.8/s w

3 ka gan (db) RE (%) radaton pattern drectonal nearly sotropc * 8.88* drectonal * 9.86* nearly omn-drectonal * 96.8* omn-drectonal * 66.5* omn-drectonal * omn-drectonal 9.69.* 9.* drectonal ths work.45.4* 9* omn-drectonal Table. Electrcally Small Antenna Comparsons. + Measured, *Smulated. ka P n (dbm) η (%) matchng network no yes no ths work no Table. Electrcally Small Rectenna Comparsons: rectennas have the same radaton pattern wth the relevant antennas (Table ). + Measured. network and the delectrc losses n the substrate should be lmted. Fnally, snce the locaton of the source s not a-pror known, the rectenna should have an omn-drectonal radaton pattern. The use of electrcally small antennas n rectennas 7 9, has two man advantages: a) a matchng network between the antenna and the rectfer can be avoded due to the ntrnsc hghly capactve/nductve nput mpedance of these antennas, and b) the rectennas have relatvely compact sze. The goals and the contrbuton of ths work are not only ) to desgn an electrcally small antenna, whch ) s ntrnscally matched to 5 Ω, ) has omn-drectonal radaton pattern n the horzontal plane and 4) has hgh radaton effcency (RE), but also, 5) to desgn an electrcally small rectenna, whch has 6) hgh RF-to-dc effcency for low power nput (.e., less than dbm), 7) hgh senstvty (.e., equal or above.7 μw/cm ), and 8) omn-drectonal radaton pattern approprate for ambent energy harvestng. Based to our knowledge and accordng to the relevant lterature (an overvew s gven n Tables and ), the presented electrcally small rectenna has better performance n terms of effcency for low power nput and senstvty compared to pror-art desgns. It s noted that the comparson, n order to be far, relates to only electrcally small antennas and rectennas. Results The small antenna s frst presented n terms of geometry, performance (.e., reflecton coeffcent and radaton pattern), antenna radaton mechansm and parametrc analyss. Next, the coplanar rectfer s presented and ts nput mpedance and RF-to-dc effcency s estmated va smulaton and measurement. The RF-to-dc and the dc output voltage of the electrcally small rectenna versus power nput and power densty s then presented, whle fnally backscatter, wreless sensor node powerng measurements take place. Small Antennas. Antenna Desgn. The proposed antenna conssts of two electrcally small square splt rng resonators (SSRRs) wth maxmum edge dmenson d and gap g, whch are located at dstance h. The two SSRRs are electrcally connected, whle antenna feedng was appled across the gap of one SSRR. Two types of these antennas where fabrcated: the frst ( wre-ssrr antenna) s made of wre wth radus r (Fg. a), whle the second ( strp-ssrr antenna) s made of strp lne wth wdth w (Fg. b). The antennas were desgned n order to operate (.e., reflecton coeffcent relatve to 5 Ω less than db) at 868 MHz, takng nto account the allocated UHF RFID frequences n Europe. It s noted that n our smulatons, an SMA connector wth maxmum dmenson (.e., heght) of 5 mm was placed and taken nto account. The latter was crucal for the desgn procedure because a) n our measurements an SMA was used for feedng, b) the connector s dmensons are relatvely large compared wth the dmensons of the proposed antennas, and c) t was found that the SMA presence nfluences the antenna performance n terms of both reflecton coeffcent and radaton pattern. In the strp-ssrr case, foam materal was used for mechancal support wth ε r =.. After optmzaton wth the Quas Newton algorthm (Ansys HFSS), the optmal obtaned dmensons for both antennas are gven n the capton of Fg. : for the wre-ssrr ka =. 45, whle for the strp-ssrr ka =. 459, and thus both meet Kng s crteron ( ka 5).. The proposed antennas smulated reflecton coeffcent s depcted n Fg.. Both antennas operate at 868 MHz. Specfcally, the wre- and strp-ssrr antenna has a frequency mpedance bandwdth relatve to 5 Ω of.56% (.e., MHz) and.64% (.e., MHz), respectvely. Fgure also depcts the measured Scentfc REPOrts (8) 8:58 DOI:.8/s w

4 d d w g g h z y x h z y x (a) (b) Fgure. Proposed antenna geometry: (a) wre-ssrr wth d = mm (λ/4. ), h = mm (λ/. 6), g =. 4 mm, r =. mm and (b) strp-ssrr wth d = mm (λ/. ), h =. 87 mm (λ/. ), g =. 566 mm, w = 4. mm. (a) (b) Fgure. Wre-SSRR (a) and strp-ssrr (b) antenna smulated (sml) and measured (msr) reflecton coeffcent. reflecton coeffcent, where good agreement s observed, and the smulated D drectvty as nset for the both antennas. It s noted that the reflecton coeffcent estmaton was based only on full-electromagnetc analyss. Estmaton through the equvalent crcut model e.g. 6, could be used, as well. However, the full-electromagnetc study was adopted because, n general, s more accurate snce takes nto account the real shape of the geometry, the total couplng between the antenna parts and the antenna radaton. The total smulated normalzed gan (sold lnes) of the proposed antennas s depcted n Fg.. The wre- and strp-ssrr antennas both have maxmum half-power beamwdth of 96 n the vertcal plane (.e., xz-plane). They also have omndrectonal radaton pattern wth the maxmum gan occurrng n the xz-plane at θ = 78, (wre-ssrr) and at θ = 8, (strp-ssrr). The maxmum smulated drectvty was.75 and.7 db, for the wre- and strp-ssrr antenna, respectvely whle the smulated gan s.4 (wre-ssrr) and.8 db (strp-ssrr), resultng n a smulated radaton effcency of 9% for both antenna types. In Fg. s also depcted the smulated co- (dashed lnes) and cross-polar (dotted lnes) normalzed gan of the proposed antennas: n the xy-plane (blue lnes) and for both the antennas, the cross-polar component,.e., component whch les n the xy-plane (dotted lnes) s always lower than db, whch reveals that the antennas n the horzontal plane are lnear vertcal polarzed. Hence, based on Table, the proposed desgn and especally the wre-ssrr, s the smallest antenna wth the hghest RE and smultaneously has omn-drectonal radaton pattern n the horzontal plane, reported to date. These characterstcs make t approprate for RF energy harvestng and RFID applcatons. Radaton Mechansm. The normalzed magntude and the drecton (vectors) of the wre-ssrr antenna s current dstrbuton s depcted n Fg. 4a,b, respectvely. Fgure 4c depcts the normalzed magntude of the current dstrbuton along the geometry s ponts. It s observed that currents flow n-phase n the vertcal and upper loop, whle n the bottom loop there are two separated regons: the currents flow n-phase from P 4 to P and from P 4 to P 7, respectvely. The maxmum current dstrbuton s located n the vertcal wre, as well as n part of the upper loop, and thus, power s radated manly through these parts, resultng vertcal polarzaton, as mentoned. The bottom loop, where the feed was appled, radates less and operates as a magnetc resonator, whch delvers power to the rest of the structure; at pont P 7 (Fg. 4b) current s added and flows upwards, towards P 8 and for ths reason there s a dscontnuty n the current dstrbuton,.e., current magntude at P 7, (from P 6 to P 7 ) dffers from that at P 7,+ (from P 7 to P 8 ), as depcted n Fg. 4c. Scentfc REPOrts (8) 8:58 DOI:.8/s w

5 Normalzed Gan (db) Angle (deg.) (a) Normalzed Gan (db) Angle (deg.) (b) Fgure. The total (sold lnes), co-polar (dashed lnes) and cross-polar (dotted lnes) smulated normalzed gan for the wre-ssrr (a) and the strp-ssrr (b) antenna n the xy (blue lnes) xz (red lnes) and yz plane (yellow lnes). Current P P 9 P P 8 P P z y P 5 x P s 4 P P 6 P 7 P P Current Dstrbuton P P 7,+ P P 7,- P P P 6. P 5 P P 4 Geometry's Ponts (a) (b) (c) P 8 P 9 P P Fgure 4. The smulated normalzed magntude (a) and the drecton/vectors (b) of the wre-ssrr antenna s current dstrbuton. The normalzed magntude of the current dstrbuton along the geometry s ponts: the contnuous lne represents the numercally estmated values, whle the dashed lne depcts the lnear approxmaton, whch wll be used through the analytcal soluton of the electrcal component n the far-feld regon (c). In order to analytcally estmate the electrcal feld of the proposed antenna n the far-feld regon and to study n-depth the antenna radaton mechansm, t s assumed that the antenna conssts of twelve de-coupled, thn wres, each of these located between ponts P j P j+, where j =,,,, as t s depcted n Fg. 4b, wth lnear dstrbuton, whch n the -th edge s gven by, = = I ˆ + u Il L J J z z z I ( ) u zˆ, L where I( Pu ), I( P l ) s the current value at the upper, lower pont of the -th edge wth =,,,, respectvely, algned wth the current drecton, whle the L s the edge length. Hence, the equvalent current dstrbuton of the antenna s depcted n Fg. 4c wth the dashed lne. The electrcal feld of a fnte dpole antenna, whch les n the z -axs, whch s n parallel wth the -th edge, has length L, deally zero thckness and current dstrbuton J = J ( z ) zˆ s gven by, n far-feld regon, where, () E = E θ ( r, θ, φ) θ^ () jkr ηke L/ jkz θ Eθ ( r, θ, φ) = j sn θ π J( z ) e cos dz, 4 r L / where k = πλ / and η π s the free-space wave-number and mpedance, respectvely, and the ntegral represents the antenna s space factor. After calculaton, Eq. () va Eq. () s wrtten as, () Scentfc REPOrts (8) 8:58 DOI:.8/s w 4

6 Numercal Analytcal Numercal Analytcal (a) 4 7 (b) Fgure 5. Wre-SSRR drectvty n the xz (a) and yz (b) -plane as t s estmated va the analytcal and the numercal procedure. jkr η e kl cosθ kl cosθ 4πkL r { u l cosθ u l u l } = θ E tan kl ( I I)cos ( I I) jkl( I I ) sn. Hence, the total electrcal feld s gven by, (4) E = E. n= Fgure 5 depcts the magntude of the electrcal component n the far-feld regon, as estmated va Eq. (5). It shows agreement between the two methods (.e., numercal and analytcal), whle the shft n angle and the beam-wdth varaton s the result of the followng assumptons: n the analytcal soluton a) the nter-edge mutual couplng was not taken nto consderaton, b) the radus of the wre was consdered as zero (. mm n practce) c) lnear varaton of the current dstrbuton along the edges of the antenna s used (Fg. 4c). It s noted that the analytcal study s presented a) to study n-depth the antenna radaton mechansm and b) to obtan an analytcal formula, whch depends only on the desgn parameters. Parametrc Analyss. Fgure 6 depcts the frequency at mnmum reflecton coeffcent, the bandwdth (BW) relatve to 5 Ω, the maxmum drectvty at 868 MHz and the RE (%) of the wre-ssrr versus all the desgn parameters: the varaton of each parameter s from 9% to % of each ntal value. In all the latter cases, the antenna resonates,.e., the reflecton coeffcent s below db. It s evdent that the resonance frequency s mostly affected by the d parameter, e.g., % ncrement of d shfts the resonance frequency to 8 MHz. BW s affected by both d and h n an opposte way: as the d and h ncrease, the BW decreases and ncreases, respectvely. Drectvty s agan mostly affected by parameter d and ncreases wth the d ncrement. RE s affected by all the parameters, except for the gap g. Specfcally, RE ncreases as h and r ncrease, however, there s an optmum value for the d parameter. Thus, t s possble to enhance the BW and the RE by ncreasng the parameter h,.e., the heght of the antenna, wthout affectng too much the operatng frequency and the maxmum drectvty. Coplanar Rectfer. A rectfer was desgned to operate at 868 MHz for a low-power nput of dbm. Its geometry s coplanar (.e., there s no ground plane) and t was chosen n order to facltate easy connecton wth the antennas SSRR gap. For mechancal stablty and fabrcaton ease, Taconc TLY-5 substrate (ε r =., tanδ =. 9) wth thckness of.5 mm was used. The desgn, the zoomed fabrcated geometry and the equvalent crcut schematc are depcted n Fg. 7: a sngle seres crcut, whch utlzes one low-cost Schottky dode HSMS85B (Avago Technologes, Inc.), one capactor and a load, performs half-wave RF-to-dc rectfcaton. There s no matchng network snce the antenna wll be drectly mpedance matched to the rectfer at a next step. The rectfer s nput mpedance s a functon of the power nput, the operatng frequency and the rectfer s output load, and thus, the goal of the smulaton was to estmate the rectfer s nput mpedance and output load when the RF-to-dc effcency (.e., the rato of the dc power output to the RF power nput) s maxmzed for power nput of dbm at 868 MHz: the maxmum RF-to-dc effcency came to % wth rectfer s nput mpedance of Zr = 5 j57ω and optmum output load of 9 Ω. Fgure 8a depcts the smulated nput mpedance n terms of real (Re) and magnary (Im) part versus frequency for varous power nput levels: t s evdent that the rectfer s capactve and non-lnear, whle at 868 MHz the Re and Im part of the nput mpedance vares from 74 to 84 Ω and from 58.5 to Ω, respectvely, from to dbm power nput, respectvely. Next, the nput mpedance of the generator was fxed at Zs = 5 + j57ω, n order to be matched wth the rectfer, and the RF-to-dc effcency versus power nput at 868 MHz was estmated. Fgure 8b shows the smulated results: the effcency ncreases from % to 9.% for power nput from 4 to dbm, whle for dbm s %. Fgure 8c,d depcts the RF-to-dc effcency versus (5) Scentfc REPOrts (8) 8:58 DOI:.8/s w 5

7 Reflecton Coeffcent (GHz) d h g r BW (%).4. d h g r Drectvty (db) d h g r (a) (b) (c) (d) RE (%) d h g r Fgure 6. Smulated wre-ssrr RE (a) and mpedance BW relatve to 5 Ω (b) versus desgn parameters d, h, g, and r: the varaton of each parameter s from 9% to % of each ntal value. Za dode 7.7 Za dode Pn C R Pout.65 Pn C R Zn (a) (b) (c) Zn Fgure 7. Coplanar rectfer topology, where there s no ground plane (a), pcture of the fabrcated geometry (zoomed by mcroscope) (b) and equvalent crcut schematc (sngle-seres crcut wth a sngle dode performng half-wave rectfcaton) (c). Also, depcted the dmensons of the desgn n mm, whle the footprnt of the dode, the capactor and the resstor s SOT-, 4 and 6, respectvely. frequency (load was fxed at 9 Ω) and load (at 868 MHz), respectvely, for varous power nput levels, when the rectfer s powered by an mpedance matched source (.e., agan Zs = 5 + j57ω): hgher power nput results hgher effcency, as expected, but n slghtly lower frequency (Fg. 8c) and load (Fg. 8d). Small Rectenna. The strp-ssrr antenna was redesgned a) n order to be mpedance matched wth the rectfer, and b) for drect connecton to the latter, to form a rectenna. The obtaned dmensons were: d = 9. mm (λ/8), h = 5. 6 mm (λ/6. 6), g = 45. mm and w = 89. mm (.e., ka = 5),. whle the antenna mpedance was equal to Za = 7. + j55. 6Ω at 868 MHz, resultng n a reflecton coeffcent Γ = Za Zr Z + Z (6) a of. db, where Zr = 5 j57 s the nput mpedance of the rectfer at 868 MHz and for dbm power nput, as mentoned. The antenna s slghtly electrcally larger than the ntal strp-ssrr antenna wth mpedance of 5 Ω, (.e., for the latter was ka =. 459) manly because now the antenna mpedance s dfferent. Addtonally, the antenna sze s affected by the delectrc permttvty of the used substrate, whch s now ε r =. (.e., Taconc TLY-5), nstead of ε r =. (.e., foam), whch s used n the ntal strp-ssrr antenna. However, the rectenna can be crcumscrbed by a radan sphere of radus λ/ π, and thus meets Wheeler s crteron. The antenna s mpedance n terms of Re and Im part versus frequency s depcted n Fg. 9a: the real part s maxmzed at 9 MHz, whle the antenna s manly nductve except the frequency regon from about 9 to 96 MHz. In Fg. 9b the gan n the horzontal plane (.e., xy-plane) s presented for a sngle antenna and an antenna-array of two elements n dstance λ/: n the frst case the smulated gan s omndrectonal wth maxmum ampltude of.8 db, whle for the antenna-array the gan s drectonal wth maxmum of 4.8 db, as expected based on theory. Next, the antenna was connected to the rectfer, formng a rectenna and the reflecton coeffcent between these two parts was estmated. Frst, the power nput was fxed and the reflecton coeffcent versus frequency s depcted n Fg. 9c: n all cases, antenna and rectfer are well mpedance matched, whle specfcally for, and dbm the operatng frequency bandwdth (.e., Γ < db) s , and MHz, respectvely. Then, the frequency was fxed at 868 MHz and the reflecton coeffcent versus power nput s presented n Fg. 9d: t s evdent that the antenna s mpedance matched wth the rectfer at 868 MHz for power nput from.8 to above dbm, whch makes the rectenna able to harvest RF energy from low up to hgher power levels. It s noted that, n both cases, the rectenna s output load was fxed at 9 Ω. The antenna and the rectfer were fabrcated formng the rectenna (Fg. ). At 868 MHz the smulated (sml) rectfer s effcency and the measured (msr) rectenna s effcency versus power nput and power densty for output load of 9 Ω s depcted n Fg. a,b: for 9 dbm power nput the measured RF-to-dc effcency s.5%, r Scentfc REPOrts (8) 8:58 DOI:.8/s w 6

8 Impedance ( ) Re, - dbm Im, - dbm Re, - dbm Im, - dbm Re, - dbm Im, - dbm Frequency (GHz) Effcency (%) Power Input (dbm) Effcency (%) 4 - dbm - dbm - dbm Frequency (GHz) Effcency (%) 4 - dbm - dbm - dbm.. Load (k ) (a) (b) (c) (d) Fgure 8. Rectfer s nput mpedance (real (Re) and magnary (Im) part) versus frequency for varous levels of power nput (a), rectfer s RF-to-dc effcency versus power nput at 868 MHz (b), versus frequency (c) and versus load at 868 MHz (d) for varous levels of power nput. At (a c) the output load was fxed at 9 Ω, whle at (b d) source was mpedance matched to the rectfer (.e., Zs = 5 + j57). All the results are based on smulatons. Impedance (Ω) Re Im Frequency (GHz) 7 4 sngle array Reflecton Coeffcent (db) dbm - dbm - dbm Frequency (GHz) Reflecton Coeffcent (db) Power Input (dbm) (a) (b) (c) (d) Fgure 9. Antenna s mpedance n terms of Re and Im part versus frequency (a), gan n the horzontal plane (.e., xy-plane) for a sngle antenna and an antenna-array of two elements placed sde-by-sde at dstance λ/ (b), rectenna s reflecton coeffcent versus frequency for varous levels of nput power (c) and versus power nput at 868 MHz (d). At (c,d) the rectenna s load was 9 Ω. All the results are based on smulatons. whle for the hgher power nput level of 4 dbm the effcency s measured at 8%. Based on the latter results, t s evdent that the rectenna presents superor performance n terms of RF-to-dc effcency for low power nput compared wth other publshed electroncally small rectennas (Table ). The measured effcency for the lowest power nput of 44 dbm, where the rectenna operates, s.%, and thus, the proposed harvester also presents hgh senstvty (.e., ablty to harvest energy and operate at low power nput or densty) snce based on the lterature harvestng systems rarely operate lower than dbm. The maxmum measured effcency s 6.6% for 4.9 dbm power nput. The smulated rectfer s and the measured rectenna s RF-to-dc effcency at 868 MHz, but now versus power densty s depcted n Fg. b: based on Eqs () and () (as t wll be next explaned n the measurement setup secton), power nput of 4.9, 4, 9 and 44 dbm corresponds to power densty of.5,.77,.88 and.8 μw/cm, respectvely, and once agan t s observed that the rectenna s senstvty s hgh snce operates from.8 μw/cm, whch makes t a good canddate for RF energy harvestng for envronments wth ultra-low power densty. The measured rectenna s voltage output across the optmum load of 9 Ω versus power nput and power densty s depcted n Fg. c,d, respectvely. For 4.9, 4 and 9 dbm power nput, the voltage s 59.5, 57 and 79 mv, respectvely, whle for the power nput of only 44 dbm the dc output voltage s mv. For all the above cases, t s evdent that there s a good agreement between measured and smulated results. Accordng to, the values of (rms) electrc feld n the telecommuncatons spectrum of MHz 6 GHz for 99% of outdoor measurements are below V/m n EU. Specfcally, based on long-term RF EMF measurements, the mean electrc feld strengths were between.8 V/m and.8 V/m, or equvalently power densty was between rms.7 and.8594 μw/cm, gven that S = E /π. The latter ndcates that the senstvty of harvesters should be as low as.7 μw/cm. Based on the smulated and measured results presented n Fg., the effcency of the rectenna vares from about.5% to.5% as the power densty ncreases from.7 to.8594 μw/cm. Addtonally, n 7 t was shown that the power densty levels from a GSM-9 base staton at a dstance from 5 to m vares from. to. μw/cm, and thus, agan, the proposed rectfer covers the latter power densty regon. Supplyng Backscatter Sensor Tags. In ths secton, the ablty of the presented rectenna n supplyng of small electrcal devces, such as backscatter sensor tags, s dscussed. The dc output voltage versus power densty when the rectenna s opened-crcuted was measured through a specfc procedure, whch wll be next explaned n the measurement setup secton. Fgure depcts the results: Scentfc REPOrts (8) 8:58 DOI:.8/s w 7

9 Fgure. Fabrcated and measured rectenna before (left) and after (rght) s assembled n D shape: s also depcted the co-planar rectfer (upper SSRR). as the measured power densty vares from.4 to.5 μw/cm, the voltage vares from 98 to 95 mv. The fttng b curve, whch s also depcted, s gven by V = a S, wth a =. 587, b =. 554 and V n V and S n μw/cm. Next, n order to ncrease not only the open crcuted dc output voltage, but also the total, delvered to the load dc power, two of the proposed rectennas were placed n the horzontal plane (.e., sde-by-sde) at dstance λ/ and were also dc sde connected, n seres confguraton (.e., voltage summng). In the latter harvestng system, at the RF sde each antenna element s termnated to a hypothetcal load, whch represents the nput mpedance of the co-planar rectfer, and hence, the harvester acts as a rectenna-array of two mpedance matched antenna elements, despte the fact that the rectennas are connected after the rectfcaton, at the dc sde,. Equvalently, the two rectennas are loadng each other at the near-feld leadng to mutual couplng whch results drectve radaton patter at the far-feld. Consequently, the nter-element dstance has mpact to the far-feld of the harvestng system, and thus, to the dc added power. The latter was the reason the nter-element dstance of λ/ was used. The radaton pattern of the rectenna-array s dentcal wth the antenna-array pattern, depcted n Fg. 9b. The measured dc output voltage versus power densty for the open-crcuted rectenna-array s also depcted n Fg.. Now, the voltage s enhanced, compared to the sngle rectenna: for.4 and.5 μw/cm the voltage s mv and.85 V, respectvely. The fttng curve for the dc open crcuted voltage versus power densty for the rectenna-array. s now gven by V =. 85S 55 and s depcted n Fg., as well. All the above results are presented n Table, where s also ncluded the RF power nput P n nto the harvester as estmated by Eq. () when G equals.8 and 4.8 db for the sngle and the rectenna-array, respectvely and G cal =.8 db. From a practcal aspect of vew, the sngle rectenna for 4.9 dbm RF power nput presents 6.6% effcency, and thus, the rectenna drectly delvers to the optmum load dc power of 8 μw at.5 V (Fg. ). Ths means that the harvester s able to contnuously supply a small electrcal devce such as a dgtal thermometer or a smoke detector wth power consumpton of and 57 μw, respectvely 8. In ths work, t wll be tested the ablty of the proposed harvester n supplyng wth power the backscatter sensor node presented n 9. The latter has power consumpton of the order of μw and.6 V voltage operaton threshold. Based on Fg., the sngle and the rectenna-array presents an open-crcut dc output voltage greater than.6 V when the power densty s hgher than 6. and.74 μw/cm, respectvely. Hence, theoretcally, t s possble the rectenna or rectenna-array to be drectly connected wth the sensor node and contnuously supply the latter, wthout the use of any boost converter. However, the goal of ths work s, among others, the desgn of a hgh senstve harvester, and thus the latter should delver dc power equal or hgher than μw wth voltage threshold of.6 V 9 for low power densty, equal or lower than μw. For the above reasons, the use of a boost converter s necessary. The low power boost converter bq554 from Texas Instruments was used n ths work 9. The latter s an ntegrated crcut and performs power management and maxmum power pont trackng (MPPT) technque. Specfcally, t has ultra-low quescent current lower than na, cold start voltage of mv, and once started, s able to harvest RF energy from voltage sources, greater than 8 mv. MPPT technque ensures the maxmum extracton and transfer of energy from the source (e.g., rectenna) to the load (e.g., backscatter sensor node), by adjustng the nput mpedance of the boost converter 9. In order to estmate the end-to-end performance of the proposed harvestng system n a realstc scenaro, the rectenna s connected through the boost converter crcut wth the load (the schematc s depcted n Fg. a: t s smlar to ths one presented n ) and the rectenna s voltage output (V IN_DC ) and the voltage across the capactor (V BAT ) was measured va the data acquston DAQ NI USB-656 nstrument. The confguraton of the commercal evaluaton board of the bq554 9 was used. Specfcally, once the voltage of the rectenna V IN_ DC > mv, cold start mode starts and energy flows through the boost converter to the μf capactor, whch s ntally uncharged (.e., VBAT = mv at the begnnng). The V BAT_ OK pn produces a dgtal sgnal when the V BAT reaches.85 V and swtches off when V BAT drops back to.4 V. The PMOS was used n combnaton wth the bq554 as a swtcher snce solates the load from the supply system untl the capactor wll be charged, n order to reduce the energy leakage. Analytcally, as VBAT <. 85V, V BAT_ OK = V (.e., pn swtched off), the PMOS BSH7 stays off and thus, no energy flows to the load (all energy flows to capactor at ths tme). But when VBAT =. 85V and untl t drops back to.4 V, V BAT_ OK, pn swtches on and the nverted through the open dran NMOS Scentfc REPOrts (8) 8:58 DOI:.8/s w 8

10 Effcency (%) 4 rectfer, sml rectenna, msr Power Input (dbm) Effcency (%) 4 rectfer, sml rectenna, msr -4 - Power Densty ( W/cm ) Volt (V) - -4 rectfer, sml rectenna, msr Power Input (dbm) Volt (V) - rectfer, sml rectenna, msr Power Densty ( W/cm ) (a) (b) (c) (d) Fgure. Smulated (sml) rectfer s and measured (msr) rectenna s RF-to-dc effcency versus power nput (a) and power densty (b). Also depcted the smulated rectfer s and measured (msr) rectenna s dc output voltage versus power nput (c) and power densty (d). At all cases frequency was fxed at 868 MHz and load was 9 Ω. Volt (V) 4 sngle, msr array, msr sngle, ft array, ft - - Power Densty ( W/cm ) Fgure. The measured, open-crcut dc output voltage versus power densty of the sngle and the rectennaarray: two sngle rectennas were placed n the horzontal plane (.e., sde-by-sde) at dstance λ/ and were also electroncally connected n seres confguraton (.e., voltage summng). The fttng curves for the both cases are. also depcted: for the sngle rectenna s V =. 587S 554., whle for the rectenna-array V =. 85S 55, where V and S s n V and μw/cm, respectvely. RF Harvestng Delvered Power system S (μw/cm ) P n (dbm) V o.c. (V) Δτ a (s) Δτ b (s) Δτ c (s) ηtot, Δτa (%) ηtot, Δτc (%) to node (μw) sngle rectenna rectenna-array Table. RF Harvestng System Supplyng Batteryless, Backscatter Wreless Sensor Node 9. BSH5 sgnal drves the PMOS: now the latter turns on and current flows from the capactor to the load. Next, when VBAT =. 4 V, PMOS agan turns off and capactor s chargng agan untl VBAT =. 85V: the above procedure s perodcally repeated. For the sake of generalty, a resstor of 47 kω was used as load n ths work: the resstance s not arbtrary, snce t was chosen n order to result n more than μw for voltage threshold hgher than.4 V. All the above are n accordance wth the goal of ths work, whch s to supply the backscatter sensor node n 9, wth power consumpton of μw at.6 V. The choce of wde tme perod of RF harvestng and short tme perod of operaton was the way was used n order to supply the backscatter sensor node 9 wth enough power to operate n a low power densty envronment as t wll explaned below, and thus, the PMOS-swtcher was necessary n order to enhance the end-to-end effcency by reducng the energy leakage. Two scenaros were tested: frst the harvester was the sngle rectenna and next the rectenna array. In each case, the senstvty and the total (.e., end-to-end) effcency of the total-harvester (.e., rectenna(s), boost converter and load) was nvestgated. The latter s gven by, c η = W tot W, (7) n Scentfc REPOrts (8) 8:58 DOI:.8/s w 9

11 rectenna bq554 boost converter VBAT VIN_DC VBAT_OK μf MΩ PMOS BSH7 NMOS BSH5 load 47 kω Voltage (V) V IN_DC V BAT Δτ α 6 9 Tme (s) Voltage (V) V IN_DC 6 9 Tme (s) (a) (b) (c) V BAT Δτ α Fgure. The total-harvester (.e., rectenna(s) and the boost converter) schematc (a), the measured voltage dc across the rectenna output (V IN_ DC ) and the boost converter output (V BAT ) for power densty of.9 (b) and.7 (c) μw/cm. where, the W c s the output dc energy, delvered to the load, as follows, W = c C Vb Va (8) and V b, V a denotes the hgh and low voltage across the capactor, respectvely, whle, τ Wn = Pndt s the nput energy to the total-harvester wthn the tme slot τ, where the capactor s dscharged (charged) from V b to V a (V a to V b ). It s noted that the latter effcency takes nto account all the sub-effcences,.e., the antenna RE, the rectfer and the boost-converter effcency and any other possble losses. In ths work, several measurements took place n order to fnd the total harvester senstvty: for the sngle rectenna, senstvty s S = 9. μw/cm, whch corresponds to Pn =. 5 dbm power nput and to Voc.. = 5. V open-crcut voltage, whle for the rectenna-array, senstvty s S =. 7 μw/cm, Pn = 4. 5dBm (va Eq. () for Pcal = 7. 5dBm and G = 4.8 db) and Voc.. = 7. V (Table ). Fgures b,c and 4 depcts the measurement results usng the DAQ NI USB-656 nstrument: the boost converter s dc voltage nput (V IN_ DC ) and output (V BAT ) for the sngle rectenna and the rectenna-array, when the power densty s 9. and. 7 μw/cm, respectvely, s presented. It s noted that, V IN_ DC also represents the rectenna s output, whle V BAT also represents the dc voltage across the load, snce the voltage drop across the MΩ s nsgnfcant. In both cases, three dstnct tme perods should be consdered. Frst, the cold start perod ( τ a ), n whch V IN_ DC goes from to. 85 V. Second, the dschargng perod ( τb ), when V IN_ DC goes from. 8 to. 4 V, and thrd, the chargng perod ( τ c ), when V IN_ DC changes agan from. 4 to. 85 V. The last two operatons, are perodcally repeated. For the sngle rectenna (Fgs b and 4a), where S = 9. μw/cm ( Pn =. 5dBm), the three latter perods are 7,. 75 and 9 s. Thus, for the frst perod ( τa = 7 s) t s Wc =. 46mJ and η =. 59% : the tot end-to-end effcency s low durng the cold-start perod, snce no external power was used, however, ths perod took place only once, at the begnnng. For the thrd perod ( τc = 9s) t s Wc =. 8mJ and η =. 4%. tot Interest presents the second perod: here, agan Wc =. 8mJ, however, because τb =. 75s s very short, t s possble to estmate the delvered power to the load as W c / τa = 6. 7 μw, and thus, the presented system s able to delver more than 6 μw every 9 s, whch s suffcent for the supply of the sensor tag, presented n 9. For the rectenna-array (Fgs c and 4b), where S =. 7 μw/cm ( Pn = 4. 5dBm), t s τa = 88s, τb =. 685 s and τc =. 5 s. Hence, for the cold start and the chargng perod t s agan Wc =. 46 and. 8 mj, respectvely, as expected, but now the end-to-end effcency s. % and 4. 8%, respectvely. Durng the dschargng perod, the harvester system delvers to the load. 8 mj wthn. 685 s, or equvalently, the rectenna-array harvester delvers to the load more than 7 μw every.5 s. Table summarzes all the above measurement results. Fnally, based on the latter measurement results and on,7, the sensor tag presented n 9 s able to be battery-less and to be suppled usng entrely ambent RF energy, whle, ths s real, although no attempt has been made to optmze the performance of the boost dc-dc converter (was used wth the ntal, commercal evaluaton board s confguratons 9 ). It s noted also, that despte the measurements presented n,7 are based on non-contnuous sgnals, as cellular sgnals are, and thus, the ambent mentoned power levels are the average value, based on 4, under certan load condtons non-contnuous sgnals wth tme varyng envelope may lead to hgher RF-to-dc effcency n comparson to contnuous-wave sgnals. Addtonally, n our case, snce the proposed system s ntended to supply the RF backscatter sensor node presented n 9, and snce the latter s used n a wreless sensor network, whch adopts b-statc archtecture, where an emtter llumnates the sensor wth contnuous sgnal at 868 MHz, the presented analyss responds to a realstc scenaro. (9) Scentfc REPOrts (8) 8:58 DOI:.8/s w

12 Voltage (V) Δτ b V IN_DC V BAT Δτ c Voltage (V) Δτ b V IN_DC V BAT Δτ c Tme (s) (a) Tme (s) (b) Fgure 4. The measured voltage dc across the rectenna output (V _ IN DC ) and the boost converter output (V BAT ) for power densty of.9 (a) and.7 (b) μw/cm : zoom-n of the Fg. b,c, respectvely, after the cold start perod. Dscusson Ths work presented the desgn and the measurement of an electrcally small antenna, wth omn-drectonal radaton pattern and hgh radaton effcency, whch s ntrnscally matched to 5 Ω. It also presented the mplementaton and measurement of an electrcally small rectenna wth hgh RF-to-dc effcency for low-power nput and hgh senstvty. In the rectenna, the antenna was drectly mpedance matched to the RF-to-dc rectfer, requrng no other matchng network desgn. The harvester s able to supply contnuously or through a boost converter battery-less small electrcal devces, e.g. sensor-nodes, for extra low-power densty levels: the latter was tested and shown through measurements. Based to our knowledge and accordng to the relevant lterature, the electrcally small rectenna presents better performance n terms of effcency for low power nput and senstvty compared to pror-art desgns: t s noted that the comparson, n order to be far, relates to only electrcally small antennas and rectennas Fnally, future work wll examne the use of the proposed rectenna n dense rectenna-arrays, wth unt-cell sze smaller enough than the wavelength. Methods Smulaton setup. The antennas were smulated n terms of reflecton coeffcent, radaton pattern and current dstrbuton va Ansys HFSS (ANSYS Inc., Canonsburg, PA, USA) wth the Integral Equaton (IE) method. For smulaton for the rectfer the ADS software (Keysght Technologes) was used. Intally, full electromagnetc analyss wth the MoM method was appled to the mcrostrp trace of the rectfer only, n order to estmate the frngng felds and the electromagnetc couplng between ports. Next, harmonc-balance was employed, takng nto account the non-lnear behavour of the rectfer due to the dode. The latter, was modelled through ts Spce model 4. Measurement setup. The measurement setup s depcted n Fg. 5. A sgnal generator was connected to an antenna of known characterstcs, whch was placed at a specfc pont, transmttng power. A strp-ssrr antenna matched at 5 Ω and wth gan Gcal = 8. db was also fabrcated and placed at far-feld dstance away from the transmtter. Ths antenna, whch hereafter wll be called as calbraton antenna, was connected to a spectrum analyser and the receved power P cal was measured. Hence, the power densty s, 4π Pcal S =. λ G It s noted that the calbraton antenna s co-polarzed and co-algned wth the transmtter, whle t has dentcal radaton pattern wth the rectenna, snce t has the same desgn wth the antenna of the rectenna, but dfferent mpedance (metallc parts of the rectfer are too small compared wth the antenna geometry, and thus, they do not affect the radaton pattern of the rectenna). Next, the calbraton antenna and the spectrum analyser were removed and the proposed rectenna was placed at exactly the same pont. The dc voltage output across the load was measured. The rectenna RF-to-dc effcency was estmated by, cal () Pout V / R η = =, P SA where, Pn = SAeff and A eff s the rectenna s effectve area whch s gven by, A eff n R eff λ = G, 4π where G s the rectenna s gan. Alternatvely, based on Eqs ( ) power nput s gven by, () () Scentfc REPOrts (8) 8:58 DOI:.8/s w

13 harvester transmtter Fgure 5. The measurement setup: rectenna (harvester) s placed n far-feld regon and the voltage across the load s measured. whle the RF-to-dc effcency s gven by, P n G = G P cal, () cal Pout V Gcal η = =. P R G P n R For the sngle rectenna at 868 MHz G = Gcal = 8. db n the horzontal plane (.e., xy plane), as explaned, and thus P = P n cal. References. Volaks, J., Chen, C.-C. & Fujmoto, K. Small antennas: mnaturzaton technques & applcatons (McGraw Hll Professonal, 9).. Tang, M. C., Wang, H. & Zolkowsk, R. W. Desgn and Testng of Smple, Electrcally Small, Low-Profle, Huygens Source Antennas Wth Broadsde Radaton Performance. IEEE Transactons on Antennas and Propagaton 64, (6).. Krues, C. M., Vyas, R. J. & Tentzers, M. M. Desgn and Development of a Novel -D Cubc Antenna for Wreless Sensor Networks (WSNs) and RFID Applcatons. IEEE Transactons on Antennas and Propagaton 57, 9 99 (9). 4. Tang, M. C. & Zolkowsk, R. W. A Study of Low-Profle, Broadsde Radaton, Effcent, Electrcally Small Antennas Based on Complementary Splt Rng Resonators. IEEE Transactons on Antennas and Propagaton 6, (). 5. Tang, M. C. & Zolkowsk, R. W. Effcent, Hgh Drectvty, Large Front-to-Back-Rato, Electrcally Small, Near-Feld-Resonant- Parastc Antenna. IEEE Access, 6 8 (). 6. Zolkowsk, R. W. Low Profle, Broadsde Radatng, Electrcally Small Huygens Source Antennas. IEEE Access, (5). 7. Zhu, N., Zolkowsk, R. W. & Xn, H. A metamateral-nspred, electrcally small rectenna for hgh-effcency, low power harvestng and scavengng at the global postonng system L frequency. Appled Physcs Letters 99, 4 (). 8. Notak, K. et al. A compact dual-band rectenna usng slot-loaded dual band folded dpole antenna. IEEE Antennas and Wreless Propagaton Letters, (). 9. Zhu, N., Zolkowsk, R. W. & Xn, H. Electrcally Small GPS L Rectennas. IEEE Antennas and Wreless Propagaton Letters, ().. Wheeler, H. Fundamental Lmtatons of Small Antennas. Proceedngs of the IRE 5, (947).. Kng, R. W. P. The Theory of Lnear Antennas (Harvard Unversty Press, 956).. Fujmoto, K. & Morshta, H. Modern Small Antennas (Cambrdge Unversty Press, 4).. Brown, W. C. The Hstory of Power Transmsson by Rado Waves. IEEE Transactons on Mcrowave Theory and Technques, 4 (984). 4. Vyas, R. J., Cook, B. B., Kawahara, Y. & Tentzers, M. M. E-WEHP: A Batteryless Embedded Sensor-Platform Wrelessly Powered From Ambent Dgtal-TV Sgnals. IEEE Transactons on Mcrowave Theory and Technques 6, (). 5. Pñuela, M., Mtcheson, P. D. & Lucyszyn, S. Ambent RF Energy Harvestng n Urban and Sem-Urban Envronments. IEEE Transactons on Mcrowave Theory and Technques 6, (). 6. Popovć, Z. et al. Scalable RF Energy Harvestng. IEEE Transactons on Mcrowave Theory and Technques 6, (4). 7. Masott, D., Costanzo, A., Franca, P., Flpp, M. & Roman, A. A Load-Modulated Rectfer for RF Mcropower Harvestng Wth Start-Up Strateges. IEEE Transactons on Mcrowave Theory and Technques 6, (4). 8. Vera, G. A., Georgads, A., Collado, A. & Va, S. Desgn of a.45 GHz rectenna for electromagnetc (EM) energy scavengng. In IEEE Rado and Wreless Symposum (RWS), 6 64 (). 9. Assmons, S. D. & Bletsas, A. Energy harvestng wth a low-cost and hgh effcency rectenna for low-power nput. In 4 IEEE Rado and Wreless Symposum (RWS), 9 (4).. Assmons, S. D., Daskalaks, S. N. & Bletsas, A. Effcent RF harvestng for low-power nput wth low-cost lossy substrate rectenna grd. In 4 IEEE RFID Technology and Applcatons Conference (RFID-TA), 6 (4).. Assmons, S. D. et al. Hgh effcency and trple-band metamateral electromagnetc energy hervester. In 5 9th Internatonal Conference on Electrcal and Electroncs Engneerng (ELECO), (5).. Assmons, S. D., Daskalaks, S. N. & Bletsas, A. Senstve and Effcent RF Harvestng Supply for Batteryless Backscatter Sensor Networks. IEEE Transactons on Mcrowave Theory and Technques 64, 7 8 (6). cal (4) Scentfc REPOrts (8) 8:58 DOI:.8/s w

14 . Bto, J. et al. A Novel Solar and Electromagnetc Energy Harvestng System Wth a -D Prnted Package for Energy Effcent Internetof-Thngs Wreless Sensors. IEEE Transactons on Mcrowave Theory and Technques 65, 8 84 (7). 4. Sharat, N., Rowe, W. S. T., Scott, J. R. & Ghorban, K. Mult-Servce Hghly Senstve Rectfer for Enhanced RF Energy Scavengng. Scentfc Reports 5, 9655 (5). 5. Xe, Y. et al. A Unversal Electromagnetc Energy Converson Adapter Based on a Metamateral Absorber. Scentfc reports 4, 6 (4). 6. Almoneef, T. S., Erkmen, F. & Ramah, O. M. Harvestng the Energy of Mult-Polarzed Electromagnetc Waves. Scentfc reports 7, 4656 (7). 7. Palazz, V. et al. A novel ultra-lghtweght multband rectenna on paper for rf energy harvestng n the next generaton lte bands. IEEE Transactons on Mcrowave Theory and Technques 66, (8). 8. Rensch, H. et al. An Electro-Magnetc Energy Harvestng System Wth 9 nw Idle Mode Power Consumpton for a BAW Based Wreless Sensor Node. IEEE Journal of Sold-State Crcuts 46, (). 9. Daskalaks, S. N., Assmons, S. D., Kampanaks, E. & Bletsas, A. Sol mosture wreless sensng wth analog scatter rado, low power, ultra-low cost and extended communcaton ranges. In IEEE SENSORS 4 Proceedngs, 5 (4).. Daskalaks, S. N., Assmons, S. D., Kampanaks, E. & Bletsas, A. Sol Mosture Scatter Rado Networkng Wth Low Power. IEEE Transactons on Mcrowave Theory and Technques 64, 8 46 (6).. Gajšek, P. et al. Electromagnetc feld exposure assessment n Europe radofrequency felds ( MHz 6 GHz). Journal of exposure scence and envronmental epdemology 5, 7 44 (5).. ERC Recommendaton, 7-, Relatng to the use of Short Range Devces (SRD). REC7E.PDF.. Balans, C. A. Antenna theory: analyss and desgn (John Wley & Sons, ). 4. Baena, J. D. et al. Equvalent-crcut models for splt-rng resonators and complementary splt-rng resonators coupled to planar transmsson lnes. IEEE Transactons on Mcrowave Theory and Technques 5, (5). 5. Saha, C. & Sddqu, J. Y. A comparatve analys for splt rng resonators of dfferent geometrcal shapes. In IEEE Appled Electromagnetcs Conference (AEMC), 4 (). 6. Saha, C. & Sddqu, J. Y. Versatle cad formulaton for estmaton of the resonant frequency and magnetc polarzablty of crcular splt rng resonators. Internatonal Journal of RF and Mcrowave Computer-Aded Engneerng, 4 48 (). 7. Vsser, H. J., Reners, A. C. F. & Theeuwes, J. A. C. Ambent RF Energy Scavengng: GSM and WLAN Power Densty Measurements. In 8 8th European Mcrowave Conference, 7 74 (8). 8. Mms, K., Gbbns, D., Dumanl, S. & Watkns, G. T. Ambent RF energy harvestng tral n domestc settngs. IET Mcrowaves, Antennas Propagaton 9, (5). 9. bq554. Ultra Low-Power Boost Converter Wth Battery Management for Energy Harvester Applcatons. In Texas Instruments SLUSAHB Datasheet, Revsed Dec. 4 (). 4. Bolos, F., Blanco, J., Collado, A. & Georgads, A. Rf energy harvestng from mult-tone and dgtally modulated sgnals. IEEE Transactons on Mcrowave Theory and Technques 64, (6). 4. Avago Technologes. HSMS-85x Seres. In Surface Mount Zero Bas Schottky Detector Dodes (9). Acknowledgements Ths work was supported by the UK Engneerng and Physcal Scences Research Councl (EPSRC) grant EP/ N9/. Author Contrbutons S.D.A. and A.G. conceved the dea. S.D.A. desgned and smulated the electrcally small antennas, rectfer and the rectenna, performed the measurements, nterpreted results and wrote the paper. V.F., A.G. and T.S. supervsed the research and contrbuted to the general concept and nterpretaton of the results. All authors revewed the manuscrpt. Addtonal Informaton Competng Interests: The authors declare no competng nterests. Publsher's note: Sprnger Nature remans neutral wth regard to jursdctonal clams n publshed maps and nsttutonal afflatons. Open Access Ths artcle s lcensed under a Creatve Commons Attrbuton 4. Internatonal Lcense, whch permts use, sharng, adaptaton, dstrbuton and reproducton n any medum or format, as long as you gve approprate credt to the orgnal author(s) and the source, provde a lnk to the Creatve Commons lcense, and ndcate f changes were made. The mages or other thrd party materal n ths artcle are ncluded n the artcle s Creatve Commons lcense, unless ndcated otherwse n a credt lne to the materal. If materal s not ncluded n the artcle s Creatve Commons lcense and your ntended use s not permtted by statutory regulaton or exceeds the permtted use, you wll need to obtan permsson drectly from the copyrght holder. To vew a copy of ths lcense, vst The Author(s) 8 Scentfc REPOrts (8) 8:58 DOI:.8/s w