ates Today Eneey 7 (2018) 1-8 Contents lists available at ScienceDirect Materials Today Energy journal homepage: www.journals.clsevier.com/materials-today-eneray/ Optical rectification through an Al203 based MIM passive rectenna at 28.3 THz @= G. Jayaswal *, A. Belkadi °, A. Meredov *, B. Pelz °, G. Moddel °, A. Shamim *** * ec! Encering rogram, King Abdullah Univesity of Since and Tcl, Tel <0, Sou Arabia ® Deparment Becca Computr and Ener gine Univers of Caradn, Bauer, CO 80309-0125, USA ‘ace aay Received 17 Jay 2007 Receledin reve form 24 Seperbr 2017 ected 2 November 2017 ‘lable online 21 November 2017 Harevesting energy from waste beat which Mictuates between, approximately, 250 K and 1500 K, ke peaking at 2-11 um. could be a game changer in terms of tapping on to renewable energy sources However, reseatch in tis area has remained elusive due to numerous challenges. We consider waste heat to be an electromagnetic (EM) wave in the mi intared (IR) frequency range, which canbe captured through a resonant antenna and rectified into useful DC through a diode. an arangement typically known as a rectenna. A bowtle antenna has been optimized for IR fleld capture and enhancement though EM simulations. At the overlap of the bowie arms, a meta-insulator-metal (MIM) diode has ‘been realized that can operate at such a high frequency (28.3 THz or 10 ym). The choice of a low permittivity insulator (AO) helps metigae the RC time constant and the diode'scutot frequency, ‘whereas the two different work function metals, Au and, facilitate diode operation through tunneling tno applied bins. A custom optical characterization setup employing 2106 um CO laser has been wed to assess the IR capture and rectification ability ofthe rectenna deve. A polarization dependent voltage ‘ouput whic is well above the noise level and well matched with our ealulations. confims the sue- ‘cessful rectenna operation. Acording to authors est knowledge, this isthe fist demonstration of rectification at 28.3 THz through a MIM diode based rectenna a zero applied bias. ‘© 2017 Elsevier Lt Al rights reserved. ipwort: ‘he enegy harvesting Nano ectenna| etatinsulatorimetal (MIM) dose 1. Introduction nano-antenna and a rectifier typically known asa rectenna device, as shown in Fig. (a) [6-10 Energy harvesting from a renewable source is a promising alternative for sustainable and clean power generation. The con- Vventional photovoltaic (PV) technology harvests energy only from the visible range of the spectrum (400-750 nm). leaving the {nfaved (IR) range completely untapped. About 80% of sola Fadia- tion is absorbed by the atmosphere and the earth surface; these waves are reemitted as mid-IR radiation (515 um) [1-3] The other sources of IR emissions are metal heating, luid heating, steam generation, heat treatment, and agglomeration. The temperature for these processes fluctuates between 250 K and 1500 K, and the cor- responding wavelengths vary in the mid-IR from 2 to 11 ym [4.5 ‘An interesting way of thinking about energy harvesting fromthe IR spectrum is treating the fluctuations as high frequency EM waves, which can be collected and rectified by a combination of a * Conesponding ator mal res [email protected](A, Shain), utp 101016psmener 201731002 2468-00691 2017 Elsevier Al gh reserved Energy harvesting via rectennas in the RFand microwave bands has been demonstrated numerous times [1-13], Antennas in mi- ‘rowave frequencies convert free space electromagnetic waves into guided waves, which are then delivered to the load [14.15]. How- ever, antennas operating at IR frequencies generate an enormous amount of highly localized electric field at the sharp tips of a nanoantenna due to the surface plasmon resonance [5.16.17] In our previous work at 28.3 THz [5]. we have shown in simu lations that by optimizing the gap and geometry of a bowtie an- ‘enna, field enhancements of many orders of magnitude can be Achieved atthe sharp tips (and in the gap between the sharp tips) of the antenna. These fields were quite localized, and therefore it was, necessary to rectify them atthe center (gap between the sharp tips) ‘of the bowtie to benefit from the field enhancement. Since no semiconductor based diode can work at such a high frequency, ‘tunneling diodes, typically known as metal-insulator-metal (MIM) odes [818,19], were realized by overlapping the arms of the bowtie antenna and sandwiching a sub nano-meter thin oxide2 nya a. / Maer Toy nergy 7208) 1-9 ° » Fig Aschemascofrecenaa jc 2) Waves cece hy the antenna are retied bya metal nso metal (MIM) ide. apd etre to te Tod () A aan presenting an MIM God between the two arms. This MIM diode was composed of two diferent metals, namely copper and gold, with a 0.7 nim copper oxide insulator between them to facilitate electron tunneling without any applied bias (5), Though we were able to demonstrate [MIM diode operation experimentally with decent responsivity at DC, we were unable to demonstrate the complete rectenna function at 28.3 THz due to the high permittivity of CuO at that frequency (e,~ 7) (20), which severely effects the cutoff frequency of the de. Vice, and the fragility of the devices due toa sub-nm oxide. MIM diode based rectennas have been demonstrated for RF bands [21-24], however theie real uility is for higher Irequencies Where other semiconductor based diodes cannot work. Even- though a few papers on hgh frequency (-30 THz or so) MIM diode based rectennas have been demonstrated [25~36], there are ‘numerous unresolved issues related to their optical rectification. Some of them utilize symmetric MIM diodes, which means they require a bias for their operation and thus are not suitable for en- ergy harvesting applications [25,26]. For others [27~35], there is some ambiguity as to wether the major contribution to their output a signal from optical rectification or a thermal response due to the Seekbeck effect. Some researchers did attempt to clarily this aspect by comparing the output oftheir MIM diodes to that ofa metal only configuration (without an insulator layer) Results indicated that the main contribution fo the overall output voltage of the MIM diode was the Seebeck [33-35] The issue of thermal contribution to the MIM diode output is an important but icky affair and therefore must be regarded with careful study in order to under- stand the true optical rectification of MIM diodes at these high frequencies, In view of the above and our previous work [5], we chose aluminum oxide (Al.O2) asthe insulator layer for this work. AlaO> has a lower dielectric constant (ranging from 0.5 to 3.5) at higher ‘THz frequencies [20,37] This value is almost half (or even lower) of CuO from our previous work (5, and is expected to help us in enhancing the cutoff frequency of the MIM diode. Au and Ti are used as metal arms, where Au and Ti are expected to have work functions (WE) of 5.1 eV and 433 eV, respectively [20.38]. This higher work function difference between the metals is chosen to facilitate electron tunneling [39,40], The other issue we had in our previous work was the fragility ofthe device due to an extremely thin insulator layer (-0.7 nm), In this work, we have doubled the oxide thickness to =15 nm. The above steps enabled us to demonstrate a 28.3 THz MIM diode based rectenna where the an- tenna can pick up the signal (verified by polarization sensitive ‘output of the rectenna) and the MIM diode can provide optical rectification (verified by the zero bias responsivity and calcula tions). To the authors’ best knowledge, this is the fst demonstra- tion of zero bias rectification of a 28.3 THz signal using an MIM diode based rectenna. The paper presents complete electromag- ‘otc (EM) simulations for the antenna part, quantum mechanical simulations fr the MIM diode part, the nano-fabrication process of the prototype rectenina and complete DC as well a optical char- acterization results. 2, Materials and methods 24. Rectenna design and simulations IMluminating antennas with IR waves generates plasmon oscil- lations on metalfdielectric interface [9 In our previous work, we have compared antennas with different geometries. Bowtie an- tennas showed higher field enhancement, compared to others, and ‘was thus considered for rectenna design [5]. In this work, bowtie antennas were chosen due tothe ease of fabrication and integration ‘simplicity. We performed the simulation studies on the new stack- ‘up to optimize the parameters to have maximum field enhancement. 2uLd. Bowie antenna with the gap ‘A bowtie antenna consisting of Au and Ti arms with a finite gap (30 nm) is shown in Fig. 2 (a). In order to optimize the antenna geometry for highest field enhancement at 28:3 THz, we investigate the main parameters: arm length, tip angle, gap and metal thick- ness to determine the optimum values fr this design. The resultant field enhancement in an electromagnetic solver (CST Microwave Studio) is shown in Fig 2 (b).Itis evident from Fig. 2(b) that there is 4 field enhancement atthe sharp tips of the antenna arms and the sap, which is mainly due to the group and phase velocity of the surface plasmon waves going to zer0 atthe tps, leading to highly localized fields. Electrical properties of materials change at higher frequencies, nd typically the Drude model is used in simulators (1), however, it ignores the electron-electron interaction as well as electron-ion interaction. Therefore, experimentally obtained frequency depen- dent material properties of Au and Ti were inserted into CST [42.43]. A bowtie antenna is defined by its arm lengths and the bow angle. The physical length of the antenna is proportional to the ‘operation wavelength. Thus, operation frequency will change severely withthe change in arm length as shown in Fig. 2 (). From. Fig. 2(4-e), itis evident that bow angle and the metal thickness does not have a significant effect on field enhancement, nonethe- less, a bow angle of 60" and a metal thickness of $0 nm seems 10, provide the best field enhancement, and is thus selected for the final implementation. The field intensity at the tip ofthe antenna increases as the gap between the arms decreases. It is clearly cevient from Fig 24) that there isa higher field enhancement for sinaller gap sizes, and it decreases with the increase in the gap size. The optimized bowtie antenna has 27 um arm length, 60° of bow Angle, and 80 nm of metal thickness. With a one nm gap between ‘the arms and the above-mentioned dimensions, field enhancement Of 8 orders of magnitude can be obtained at the tip of the bowtie 212. Overlapped antenna simulation Since, the final design requires a MIM diode at the center of the bowtie antenna through overlapping of the antenna arms. itis important to also run the simulations with the overlapped configuration ofthe rectenna, as shown in Fig. 3(a). For this study, the optimized antenna dimensions obtained from the electro- ‘magnetic simulations are used but with a very thin insulator layer (Al0s) sandwiched between the two overlapped antenna arm tips. ‘The peak field intensity of the antenna is shifted to larger wave- length when the antenna arms are overlapped when ait is used as6 days ob: Materia oe Energy 72018) 1-9 3 a) Metal thickness) Arm Leng Gap ‘Art leg=2700 1 10 oe eo HR Wavelength (um) 10° | e) | | tal taes=t0 LEtie tieeeaes 10°, se 7 8 8 ww caecmr see mm cntnn Wavelength (um) Wavelength (um) 2 (a) Antenna topology, (b) Feld enancement atthe oe antenn, (6 Ama eng vr, (Bow angle vation) Metal hcknes variation (0 Gap se vratlon, the insulating layer in between them. However, the shift in peak intensity of the antenna introduced by the overlap is compensated by the introduction of a low permittivity dielectric, as shown in Fig, 3(b). Computed antenna efficiency and resistance values are shown in Fig. 3(¢). Our simulation showed that for our particular stack we ate getting a radiation efficiency of around 11% and an antenna resistance (Ra) of -55 0, 22. MIM diode simulation “To assess the tunneling characteristies and DC performance of a MIM diode, itis important to conduct inital simulations of the chosen material set. For this purpose, a quantum mechanical simulator is used, which has been explained in details in Refs, [44.50] The simulator uses the shape ofthe tunnel barrier at diferent voltage values and the transfer-matrix method to calcu- late the tunneling probability 18). The tunneling current is then calculated from the tunneling probability and the Fermi distibu- tion of electrons. The shape of the bartier is determined by the ‘work function of the metals, the electron affinity of the insulator and the applied voltage bias [44], The AlO» electron affinity (waoa) was varied from 15 eV to 35 eV since those are the upper and lower limits reported in literature for this insulator (45-49), ‘The area was taken to be a square with a 200 nm edge length. The starting values for the work function of Au and Ti are taken to be 51 eV and 43 eV (as mentioned in previous section), however. its possible that these values vary due to various parameters involved inthe nano-fabrication process. For example, Au difluses into AlO3 ‘when the oxide is deposited using atomic layer deposition (ALD) method, changing both the work function of Au as well as the electron affinity of AlaO2 [48,49]. Therefore, in simulations, the ‘work function (@) of Au was varied from 5.1 t04 eV and the work function (ihn) of Ti was varied from 4.3 eV to 35 eV. Fig. 4, shows ‘the simulated zero-bias diode resistance (Ry) and zero-bias, responsivity (fy) for the standard values of the work function of ‘Auand Ti using 15 eV as an electron affinity of Al,Os, whereas Fig. 5 ‘shows some results for various combinations. As can be seen in Fg. 5, as the work function of metals decreases and the electron4 {6 Jawa eta acer Ty Energy 7 (2018) 1-8 ® w : * wl wee WR Waveingin (en) ©* i “| "Walt (um) Fig 3 (2) Rectenna stack up ad MIM dade athe enter. () Bel testy fr ovesapped antenna wth land AO) Antenna eficency an input ressance alfinity of the insulator increases, the zero-bias responsivity in- Creases and the resistance decreases, This is expected asthe barrier height decreases with the increase of electron affinity thus facil- tating the tunneling process. Similarly, increasing the work func- tion difference between the metals also facilitates tunneling, chus resulting in lower resistance, 3. Device fabrication ‘After assessing the simulation studies, which highlight the relationship between the device characteristics and the material stack, Itcan be seen that the choice of material, dielectric constant and chosen thickness play a vital role in altering the MIM diode performance. Considering all these aspects, the fabrication of MIM diode was undertaken. The device studied during this experiment is realized on Si(SiO, substrate. A Si (100) wafer with high re- sistivity (p-type; boron doped, resistivity 10000 Q-cm) was used to Improve the performance of the antenna by reducing the electric current leakage from the antenna feed into the substrate. As ‘mentioned in the previous section, in order to prevent the forma- tion of Schottky diode between the antenna and silicon substrate, a 15 jm thick SiO, was thermally grown on Si substrate as shown in Fig, 6(a). Then, SiOz was removed only from the backside of the wafer using BOE (buffered oxide etchant ~ 1 nmjs etching rate), to deposit another conductive layer to enhance the coupling between the antenna and the substrate (Fig. 6(b). For this purpose. a 10 nmj 200 nm Cr/Au metal layer was sputtered on the back side ofthe Si ‘wafet. The first/bottom arm of the rectenna design was then fabricated as shown in the schematics of Fig. 6(e-1). The device was patterned through a lift-off process of the main metal material. To start. positive e-beam resist (PMMA) witha thickness of ~200 nim was spin-coated on the top surface of the Si/SiO, to act as the photosensitive polymer and pattern exposure, as shown in Fig. 6(¢). ‘Then, the bottom arm, having the shape of a bow-tie (2.7 ym arm, length) along with the bias pad (200 im edge square) were patterned using electron beam lithography (EBL) (Model CRESTEC, CABL-9520C). After EBL exposure (curtent | = 500 pA, exposure dose LI uA), the pattern was developed in a solution of methyl isobutyl ketone (MIBK) and isopropyl alcohol (IPA) with the ratio of 1:1 for 60 sees (Fig, 6(€). After development, O> plasma was used to remove the residual resist. Then, a 10 nmy80 nm Cr[Au film was deposited using sputtering and finally patterned through lift-off ‘process and sonication in acetone for 5 mins (Fi. 6(e-) ‘To deposit the oxide and overlap the secondjtop arm, the e- beam resist was again spin-coated on top of the fist atm and exposed by EBL to produce an overlap area of -150 nm(ig.7(a~b)). 1m order to realize this critical sharp overlap, an alignment pro- cedure was carried out using precise steps of global and local alignment mark process. Using the same parameters previously ‘mentioned for EBL exposure, the patterned resist was again developed, as shown in Fig. 7(€) Then a 1.5 nm thin AlgOs film was deposited using ALD in order to ensure uniform and conformal {deposition across the relatively higher aspect ratio structure ofthe arm (Fig. 7). Finally, for top metal was deposition (titanium- 80 nm) as shown in Fig. 7(e). Electron-beam evaporation was chosen instead of the traditional sputtering tool in order to reduce surface damage from the plasma and ion bombardment. The patterning ofthe Ti metal was again done through lift-off process by sonicating the sample in acetone for 5 min (Fig. (0). ‘The SEM image of the final rectenna device is shown in Fig. 8 where the tunneling diode has been fabricated by sandwiching the thin oxide (AljO3) between the two dissimilar metals (Au and Ti). It is clearly evident from the SEM image that the tip-to-tip alignment ‘of antenna arm is successfully realized, and we achieved an overlap area of 200 nm during the fabrication. 4, Results 41. DC measurements and analysis For high frequency applications, MIM diodes ae characterized by the four parameters: differential resistance, responsivity, nonline- arty and cutoff frequency. The diode resistance (Ro) is obtained by differentiating the current with respect to the applied voltage as given by Equation (1) In general, alow value of (Ry) is required 10 Achieve good impedance matching to the antenna [550], R= ryt o6 days ob. / Materia ey Ereey 72018) 1-9 a, «10% b os = z z I | oy 60; 05 55 ozs = 50 io | 45 025 5 4 ‘oa 03-02 01 0 01 02 03 04 Voss Fi. Simultd expected (characteris, restance ad esponsity of an Au ~ ARO (15 nn) ~~. (a) eased UV) dt, eesponding 1 dr, pd the seme HY) (0) Ck 40) 0: 04-03-0201 0 01 02 03 o4 Vous to, 15e¥, t= 15mm, by 43500 lated espns snd esitance rr nated ta Ee eng of 20 8 aecenmanasoman) © [—aenema nssan ease ried 9,2 008A r= 2am nas 4225,01295. 5, 000 am r= 12m marrerenr veri hr re . R= 7OKS | 10° Sos! ad ° dw 05) 108 4 Be eeemrmhrerh ems hh ‘eas Vienas Fig 5. Suto expected characteris, esance and responsi ofa A Diode responsivity (fa) determines the diode’s rectification ability. The (fi), can be expressed as shown in Equation (2). Ifo) 2K) Ahiigher responsivity value enhances the AC to DC conversion efficiency and hence increases the rectification ability. Lastly, the diode cut-off frequency is expressed by Equation (3), and the diode capacitance (C) can be easily calculated through Equation (4). (60) 2 1 te = ame ° C= coer 4) [ALO (15 nm) —TL MIM oe) Caleuated responsivity. b) aed restane ‘where Ra isthe antenna resistance, «is relative permittivity of the ‘oxide Alb0; in our case), and Aand d are the overlap area ancl oxide thickness, respectively [51]. One method of increasing the diode cutoff frequency and its nonlinearity is to decrease the diode ‘capacitance. This can be obtained by decreasing the overlap area (A) andjor increasing the oxide thickness (d). A tradeoff exists as a thicker oxide means an increase in resistance. We have tied to improve this trade-off by incorporating a low dielectric oxide layer in our device The DC characterization ofthe fabricated device was performed using a four point probe setup with a Keithley source meter and HP. voltmeter In order to prevent any damage to the devices, voltage sweep was restricted fom -04 V to 404 V. The two main per- otmance parameters i.e. dynamic resistance and responsivity were extracted from I-V measurements using Equations (1) and (2). To, reduce the effects of the noise, the diodes I-V curves were ‘smoothed and fitted with a 7th order polynomial before calculating. the resistance and the responsivity. As mentioned previously, for6 6 say ot Maer oy nerey 7 (208) 1-9 (sis0, Exposure So») ek elector (As) deposition) Fit am exposure by EB, 6) Remaving the exposed res sing 2 mite of MIBK and IPA developer with aioe) usta spurting OCA, Fst ath ater al haea (0) Bottom arm (0) beam est PNA — (6 Deveopment MIBK PR) nd ebL exposure with aligment Ss SS ss (4 onde ayer (A,0,.S0m by no) (0 T0p rm (ater tit (0) Top Metatizaton (1170 nm Seam Fig 7 Fabrication ofthe al rcenna device a st arm) Second am alignment and exposure by EB, c) Removing the expose resi sing a mice of MIBK and PR laser with a maximum intensity of 3 x 10" Wim. The bbeam is passed through a mechanical chopper to facilitate lock-in detection (Stanford Research Systems SR830) in measuring the ‘open circuit voltage from the rectenna, The beam is also passed through a half-wave plate ta allow the polarization of the beam to be rotated about the antenna axis. as shown in Fig. 11. When the illuminating beam polarization is aligned with the antenna polat- iaation ie. co-pol (0°, 180° and 360°), the output response is maximized. Conversely, when the illumination polarization is perpendicular to the antenna polarization is. crass-pol (90° and 270°), the rectenna response is a minimum. The open circuit voltage dependence on polarization angle of the fabricated Au {Al,03/Ti MIM diode is presented in Fis. 12, with peaks inthe signal's ‘pen circuit voltage at around 0°, 180° and 360°. The dots represent the measured data and the solid line isa sine squared fit to the ‘measurement results Itis evident from Fig. 12, that we have experimentally obtained an output voltage (Voc) of 85 nV which has been calculated from (Wau Veiee)” where the average signal voltage is 100 nV and the average noise signal is 53 nV. ‘laser input power (Pr) of 1020 nW was obtained by mu plying the maximum laser intensity (3 x 10" Wim?) with the Resistance illumination area of the rectenna device, which is 3.410°° mn The DC output of the device can be estimated by (5). Pout = 0.25 V3,/p 6) where V2/ is (85 n¥}(98 KO). The final DC output power (Pau) of 1.79 5° 10-* W is obtained by multiplying the value by afl factor of 025, calculated using equation one in Ref. (3) ‘The final rectenna efcieny is obtained from (6) and is equal to 175 10" Pout Pe © Yectenta We also perform a systematic calculation in order to verify the measured DC output power results and access the reliability of our measurement, In a reetenna, a bowtie antenna isthe first compo: nent to collect the waves. Using our simulated value of 11% for the antenna efficiency, we obtain a power 112 nW (11% of Pj) atthe antenna. Power coming out from the antenna faces impedance mismatch and RC coupling losses. Antenna and diode impedance mismatch efficiency and RC coupling efficiency is calculated from the formula given below (50): 7 ARARo Mmlach ~ TR Re)? 224% 10% a 1 Trccoupng 857103 1+ (80) Using the antenna impedance from the simulation section (Re-55 0), the measured diode resistance (Ry ~ 98 kA) and a Capacitance of 11 fF obtained using (4), we get 025 nW (112 NW > minpstach) after loss due to impedance mismatch and 241 pW (0.25 AW > Mrccouping) after and the addition of RC coupling loss. The final step isto calculate how much AC power going into the diode can be converted to DC power. The upper limit for this is ‘determined by taking the AC power and multiplying by the diode zeto-bias responsivity to get the maximum expected short-circuit ‘current Just as we used the open-circuit voltage from the experi- mental measurement to estimate the DC power, we can use this short-circuit current value to estimate a maximum DC power out from esponavty (M) Fig 9.) characteris, resistance and respons a the ited Aa/A0, MIM diode a) Measuted dee DIV) characteristic andthe 7h rer HY) (0) Messed espns and resistance8 Jawa eta Maer Ty Energy 7 (2018) 1-8 a > @ (o) ra o g Fe p? a 82 i. 5 bases a J a erp non Caen ‘Gren aro ore? oa as = 04 03 2 01 0 01 02 09 04 Vouas (¥) Veuas ExHO,) 44 and WET). Aro lnceased i a edge length of 20 am. Fg The schematic of pia measurement setup used for opal haaterzaton Sonal) Fig. 12. he preiinary otal measurement result. The fact the the response pularzaton dependent (2 ool round O and 180" the signals well stud Fore the ei lve acts teeta gal tu os ou vs, Pout = 0.25 Ie Ro (9) Where fe is the maximum expected short circuit current obtained by multiplying the diode input power with the responsivity (21 pW'0.444/W). Using (9), we have obtained an output power of 21 x 10-® W, which is very close to the measured output DC 2 eV) Messed I) data tscresprtng "onder the smut (0) Caesed respnsvy tnd estas power. Moreover, the rectified voltage from the device is polari- zation dependent and thus confirms that it has been obtained by the rectification of the IR captured by the antenna and that its not random thermal noise. We also calculated the overall efficiency of ‘ourdevice by using Equation (6) resulting in a value of 2.05 x 10". ‘Although the efficiency obtained is smal, it has to be considered that the MIM diode device used here has a larger overlap area (200 nm), due to the alignment challenges. The efficiency can be enhanced by decreasing the overlap area. Furthermore, the diode asymmetry and subsequently its rectification abilty is not the best nd can be further optimized. It is expected that with the optimi- zation of the device structure, such as using, multiple insulator layers, the output DC voltage can be increased. Nonetheless itis encouraging to see that at zero-bias, rectification of 28.3 THz sig- nals is possible. As per the authors’ knowledge, this is the fist demonstration of rectification of a 28.3 THe signal using a MIM diode at zero bias. This result is encouraging and a good indicator of the great potential of using MIM diodes as rectifiers for THz fre- {quencies applications such asa completely passive rectenna design. 5. Conclusion We have demonstrated 28.3 THz (10.6 ym) completely passive rectenna design employing Al,O» as the insulating layer between ‘wo different work function metals, namely Au and T. The bowtie antenna has been optimized to enhance the fields in the gap be- ‘tween the two arms, Since these fields are very localized, the two arms have been overlapped to realize the MIM diode exactly at the ppoint with highest concentration of captured fields. Simulation results for the MIM diode have been presented, which have later been correlated with the measured DC results, A custom optical characterization setup employing a 10.6 jim CO laser has been used to assess the optical confinement and rectification ability of the rectenna device. Detailed calculations of the rectenna output have been carried out and match well with the measured rests, The results, though not the best in terms of efficiency, are prom- ising enough for further investigations in this area ‘Acknowledgement We acknowledge financial support from King Abdullah Univer- sity of Science and Technology (KAUST), Office of SponsoredJawa eo Metra Ty Ener 7 (2018) 1-9 ° Research (OSR) for CRG grant OCRF-2014-CRG-62140281, Authors ‘would also like to thank Shuai Yuan and Shuai Yang for their help with optical set-up and SEM analysis, Rerewencrs [oy P.bosstard, W. Hermann Hung Hunt, A. simon, An assessment of sla ‘nergy conversion technologies and research opportunities, CCE? Ee ‘sess. Anal (2000. [2] Dy. Gorman 5 Viayaraghavan . La G Tamm, New and emerging de ‘elopments slr ene. So ener) 791) (2008 35-83. [3] Bally, Rober eo, A proposed new concep fr 9 slarenergy conver, [Feng Power 94 (2) (1972) 15-77. la]. Grover Medel Metal Single insulator and Mult insulator Didesfor Recenna Soar el in: Reteana Slr Cele Springs, 201, pp. 83-12. 15] MN. Gada, M_AbgetRahman, AE Shamim, Design, optimization and [Bian of a 28.3 THe nano rcenna for infared detection and ety tion, set Rep (20), [6] Br Retna Photvetic Technolgies beyond the Horizon: Optical Ree "enna Sor Cll 2003 Fal Report, NAL) 520-3525. 171 Bloke lssson, Metal instr metal Does for Sole Ener Canerson, Univer of Csirado, 2001 Phi ee Is] Basson. Bake nd Garret Mode. "Meta oxie Becton Tuning Devi for solr Energy Conversion” US. Pate 53474, sued Mar 18, 2003. {9} fo Wang, HS, Gon Tie, Design oa secon howe maoectenna For opcal power and inared detection Front, Phys. 10 (3) (2013) 104101, Wo] NatiMskoeay, PAL Cae, A. Mayes, BL Wet Wil, TE Sulla, Lerner, Natouale devices for retiteaton of bigh frequency radon frm the nffared though the vise: anew approsc, Nentecl, 2012 (2012) ta 12373. lun] Serna rast Kat Chang. Rec Antenas(etenas). Encylopedia UR and Microwave Enginering, 2005 [na] ¥oJen, MC. te K Chang. 35 Gz wetting amen for witless power feansmasson, Heron et 43 (11) (2007) 602-603. [n3] William e Brown, ptnizaon ofthe efcency and ater properties of he ‘ectena clement Microwave Symposium 1JiSIEEE-MITS Intemational, {EEE 1976, pp. 12144. [14]. saher, € Met N. Cldsman,Prasisc aware optimization ofan RF power Scavenging cut with applications to Smarts sensor networks Rao ‘hd Wireless symposium, 2000. S00 IEE, 2000, pp. 332" 335. [u3} 5 agan, WU, High effcleney low-power microwave rete fo wireless ocy harvesting i: Mrowave Symposun igs (IMS) 2013 IEEE MIT'S Imerntina. 2013, pp. 1-4 1W6] Musso Duan. Antonio |. Femandex Dominguez, Michel Bosman. Stefan ‘A Mater Jel KW. Yang. Nanoplasmonis skal own to the nanometer Seale Nao Let. 123) (2012) 1683-1688 1071 Hs Wang, Daniel W. Bron iL, Peter Nonidet Naor Has, Nano- ‘ee hy plasmonic manosrtre Nan Lt (2) (2000) 827622 118] Sachi rower, Garet Movie. Appicabity of metalinsaltrimetal (MIM) ‘ines to sear etonnay IEE] Phoroeaaks 1(1) 2011) 78-83 io) ML Dagens Cho FYesikoy, ANC Chrys MC Peckerar, Sola spectrum feaifeation wing. nanoantenas and tunneling des, in: OPTO. pp. ‘eS0E-I6OSUE ttematonal Society foe Opis and Pts, 2010. [20] £0, Pak, Handbook of Opal Constants of Sos (VoL 3. adem ress [21] KGa Yesoy, Cys SH. Cho, N Goldman, M.Dagemis M. Peck, ‘feed asym etal ienster etal ong dnd ibcaton, ‘Dc ehuractenscs and HF rection ants, Hedtron Deve, IEEE Tras $55 2011) 310-3508. [22] A Kau, Yang Chat, Study of microwave cuts based on Meat ‘ular Seal (MM) dlodes on ex sutstates, In Eecwonie Components nd Technaogy Conference (ECTC) 2074 IEEE Gah, 2018p. 2168-2174 [23] OA Alay, De and RF Characterization of High Frequeney ALD Enhanced Nanostructured Mec insulator metal Diodes 2014 [2A] Hao Gao. Maron K. Mattes Kammerer, Picter Harpe, Mibsevc Dusan Uryhannsen Arthor van Roermund, Peter aus A‘TIGHE RF energy har esting tag with Bx eceney for wireless renperetre sensors in 65am ‘GMOs. in aco Fequeney neat Cres Symposium (RFIC, 2013 TEE, IKKE, 2013, pp. 403-406. 125} C Rumeaux W. Hecmann, PK: Kael H, Rothuizen Nanometr tii NieMO-M dodes for detection and mining of 30 Te acai, Inared ys Teel 39 (3) (1988) 123-183, [Eke Opts W. Hermann. F. Keb. Nanometershi-Slm N-NO: Nidindes or 30 radiation. App Phys R58 (184) 329-24, eo [271 JA. Bean, A. Weeks, GD, Boreman, Peformance optimization of antenna ‘coupled AYAlOxPe tune! lode Infared detectors IEEE J. Quantum Hee- tron 47 2011) 126-135. (ea) Jetey A ean, Ba Tivar, Gary H. Bernstein. Fay. Wolfgang ord ‘Thermal nated detection using dipole antenna coupled meta-oxide- real Aliodes J Vat Se, Technol 8 Nieroeecron, Napom. Stic. res. Meas. enon 27 (1) (2009) 1-14 [29] S¥. Wang, Tawa, TK Custtion. Coupling characteristics of thin Sm ‘priemetal diodes at 1089, Appl Py te. 27 (9) (579) [30] Har iong Tons, SE Schware, Mechanism and properties of point contact ‘tliat metal de detectors at 108 je APL PO bat 25 (10) (974 395-586, {31 Morena etn siya Wang. } Winery, .Gastation, characteristics ‘finerated Mo jncions ade ab atop! equences EEE) Oust. Becton. (3) (1978) 15-10 {32] Yoshizwol Yasuo, Tomar Sdurada, Tye-0. Miyata, Mechanisen and Propertes of point contact W-NI dade detectors at 10 ym, Jn. APPL Page 171) 1978) 171, {33) Gamer Model sae Grover (fds) Rectenna Salar el, vol 4 Spinge, New York 2013 (hapte) (4) Pez, Aelia, C Model, Travelng-wave mea insulator etal diodes for inared rctennas. in: Phitovalaie specs Canferene (PSC), 2016 TEEE #3 IEEE. 2016 jue, p. 1034-1038. (35) Maro ares ecerM Krenz Cargo. Szakmany, Sach N.Tivar, Dane Kable, AletelO ow. Gary H Bersein et lRectennas evs, TREE Trans Narotetol. 12 (6) (2015) 1144-150. (36) M. argo, D. Maso, V_ Rizzo A Costanzo, Design rues fr innovative anoetenas in the nated epon ns Mzowave Conference (MC) TEEE, European, 2013 pp 14 {B71 wow rfactiveindexo (Bal Jost Hotel Fane 1c Scle, Work faction of metals in: Sod Sur Phys. ‘Springer ein Hester, 1978 pp. 1-150, [Bo] NAfmardan, investigation of Metlslstor-metal (MIM) and_ Nano [beat tire Mn funnel Devices Faeted vis Ae Layer Depaiton, PLD. dissertation, Oregon State University. 2013, (40) Perasmy 1 Carey, Al Abdulagaton, PE None} ery. DS, Gin, SSAC George PA Parl Raj Retinal "metal ods: oe a ‘he ur cron herein pra, Mae 25(0) 213) a] Netw. Asberof N.David Mermin, So state Pays. 1976. Sasa ai ne, los A an es A (4s) Abisandar I. Rake, Allsandra 8. Djunsé van. M. Bazar, Maran [Ujewak, Opti properties of metalic fms fr vertia-avy otelee toni devices, App. Ope 37 (22 (1098) 82715285. a Sut Grover Garret Model Engineering the caent-votlag characteris ‘ot meta insltor-metal Gods wing double insular tunel airs Sold'statetlectton. 67 (1) (2012) 94-28, (45) E walla Cowell Sean W. Mut, Douglas A. Kesar, joha Wager, Baier height estimation of asymmettic mecabnslator-metal tunneling des, Abo Ps. 114 (21) (2013) 215703, 6) Asus Hiawa, Dalsuke Masaru, os Kawarads Eft of atomic ayer Gepost temperature on cree conduction im A203 fs formed Sn 20 oxidant. Appl Phys. 120(8) (2010) OBA (a7 Mu tang ¥c Chang CH Chang. Lin wo. Wu. M, Hong Ener ‘and parameters atom yer depostion AIZ03/ ina as heterostructure. ‘op Pe et 89 (1) (03) 012503. 1H] SAC Polack CE: Mert, Bectron tunneling thou asymmetric fms of {hema grown AID. | Appl Ps 3 (5) (1964) 19031512. ay Ka aa ra aay Edun eld, an Ni, Dil Areas, Davi Taner. aj Mitra, Michael Lanagan Broadband dielectric characte. ‘Bun of lumina ide (AOI). J MuoclectrenHeeton Pacing (1) (Gone) 27, {50] Saunt Joshi, Gar Model, Simple gure of merit fr diodes in opi fecensas IEEE} Photvolts G3) (2016) 656-672 ts Bt hr Vint Sy ms hed aa. carom ‘ote optical rece NaL Napotechnt 10 (12) 2015) 1027-1032. {s2] HeLa CL tan DAL shen Xj. Zhang, Vb. Cul 2D i study f the TAl203 tere | Mtr Set 30 (2) (1995) 339-348. {59} Davide Bresagh, ene del Carmen Ptr, lane Kolepert, Steffen Klan, Mathieu Turle. Dieter Neher, Jan Anton Keser. Corpetsion between {combination aad extaton of fee charges determines Ue fll actor af ‘vgn salar ells Nat Commun 6 2015.