1

nerarlee

eutawteafvmmo

N

mE

J

DownD. K. Kotter

S. D. Novack

Idaho National Laboratory,2025 Fremont Avenue,Idaho Falls, ID 83415

W. D. SlaferMicroContinuum, Inc.,

57 Smith Place,Cambridge, MA 02138

P. J. Pinhero1Department of Chemical Engineering,

University of Missouri,Columbia, MO 65211

e-mail: pinherop@missouri.edu

Theory and ManufacturingProcesses of Solar NanoantennaElectromagnetic CollectorsThe research described in this paper explores a new and efficient approach for producingelectricity from the abundant energy of the sun, using nanoantenna (nantenna) electro-magnetic collectors (NECs). NEC devices target midinfrared wavelengths, where conven-tional photovoltaic (PV) solar cells are inefficient and where there is an abundance ofsolar energy. The initial concept of designing NECs was based on scaling of radiofrequency antenna theory to the infrared and visible regions. This approach initiallyproved unsuccessful because the optical behavior of materials in the terahertz (THz)region was overlooked and, in addition, economical nanofabrication methods were notpreviously available to produce the optical antenna elements. This paper demonstratesprogress in addressing significant technological barriers including: (1) development offrequency-dependent modeling of double-feedpoint square spiral nantenna elements, (2)selection of materials with proper THz properties, and (3) development of novel manu-facturing methods that could potentially enable economical large-scale manufacturing.We have shown that nantennas can collect infrared energy and induce THz currents andwe have also developed cost-effective proof-of-concept fabrication techniques for thelarge-scale manufacture of simple square-loop nantenna arrays. Future work is plannedto embed rectifiers into the double-feedpoint antenna structures. This work represents animportant first step toward the ultimate realization of a low-cost device that will collectas well as convert this radiation into electricity. This could lead to a broadband, highconversion efficiency low-cost solution to complement conventional PV devices.DOI: 10.1115/1.4000577

Keywords: nantenna, frequency selective surfaces, nanoscale modeling, nanofabrication,nanoimprinting, roll-to-roll manufacturing, rectennaIntroduction

Full spectrum incident and reflective reemitted electromag-etic EM radiation originating from the sun provides a constantnergy source to the earth. Approximately 30% of this energy iseflected back to space from the atmosphere, atmospheric gasesbsorb 19% and reradiated to the earths surface in the mid-IRange 714 m and 51% is absorbed by the surface or organicife and reradiated at around 10 m 1. The energy reaching thearth in both the visible and IR regions and the reradiated IRnergy are under-utilized by current technology.

Several approaches have been used to successfully harvest en-rgy from the sun and conversion of solar energy to electricitysing photovoltaic PV cells is the most common. An alternativeo photovoltaics is the optical rectenna, which is a combination of

rectifier and a receiving antenna. The initial rectenna conceptas demonstrated for microwave power transmission by Ray-

heon Co. in 1964 2. This work illustrated the ability to capturelectromagnetic energy and convert it to dc power at efficienciespproaching 84% 3. Since then, much research has been per-ormed to extend the concept of rectennas to the infrared andisible regime for solar power conversion, and progress has beenade in the fabrication and characterization of metal-insulator-etal diodes for use in infrared rectennas 4,5. It has been dem-

nstrated that optical antennas can couple electromagnetic radia-

1Corresponding author.Contributed by the Solar Energy Division of ASME for publication in the JOUR-

AL OF SOLAR ENERGY ENGINEERING. Manuscript received September 9, 2008; finalanuscript received September 20, 2009; published online January 5, 2010. Assoc.

ditor: Robert Palumbo.

ournal of Solar Energy Engineering Copyright 20

loaded 26 Sep 2012 to 219.223.192.230. Redistribution subject to ASMtion in the visible in the same way as radio antennas do at theircorresponding wavelengths 6.

A major technical challenge continues to be the development ofeconomical manufacturing methods for large-scale fabrication ofnanoantenna-based solar collectors. In addition, research is re-quired to improve the efficiency of rectification of antenna-induced terahertz THz currents to a usable dc signal, and mate-rial properties and behavior of antennas/circuits in the THz solarregions need to be further characterized.

1.1 Limitations of Photovoltaic Technology. Traditional p-njunction solar cells are the most mature of the solar energy har-vesting technologies. The basic physics of energy absorption andcarrier generation is a function of the material characteristics andcorresponding electrical properties i.e., bandgap. A photon needonly have greater energy than that of the band gap in order toexcite an electron from the valence band into the conduction band.However, the solar frequency spectrum approximates a blackbody spectrum at 6000 K, and as such, much of the solar ra-diation reaching the earth is composed of photons with energiesgreater than the band gap of silicon. These higher energy photonswill be absorbed by the solar cell but the difference in energybetween these photons and the silicon bandgap is converted intoheat via lattice vibrations, i.e., phonons rather than into usableelectrical energy. For a single-junction cell this sets an upper ef-ficiency of 31% 7.

The current research path of implementing complex, multijunc-tion PV designs does not appear to be a cost-effective solution toovercoming PV efficiency limitations. Even optimized PV mate-rials require direct perpendicular to the surface sunlight for op-

timum efficiency and are only operational during daylight hours.

FEBRUARY 2010, Vol. 132 / 011014-110 by ASME

E license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

aeaop

usaitacl

2

nVlfti

tcifreatrt

sdpe

t

Fpw

0

Down1.2 Economical Alternative to PV. We have developed anlternative energy harvesting approach based on the use of nant-nnas to absorb incident solar radiation. In contrast to PVs, whichre quantum devices limited by material bandgaps, antennas relyn natural resonance. The bandwidth of operation is a function ofhysical antenna geometries.

Nanoantenna electromagnetic collectors NECs can be config-red as frequency selective surfaces to efficiently absorb the entireolar spectrum. Rather than generating single electron-hole pairs,s in the PV case, the incoming electromagnetic field from the sunnduces a time-changing current in the antenna. Efficient collec-ion of the incident radiation is dependent upon proper design ofntenna resonance and impedance matching of the antenna. Re-ent advances in nanotechnology are now providing a pathway forarge-scale fabrication of nantennas.

Theory of OperationWe have designed nantenna elements that capture electromag-

etic energy from both solar radiation and thermal earth radiation.arious self-complementary antennas, such as dipoles, spirals,

oops, etc., are candidates due to their inherit wide bandwidth andeedpoint configurations for concentrating energy collection andhe antenna element size is related to the wavelength of light wentend to harvest.

The basic theory of operation is as follows: the incident elec-romagnetic radiation flux produces a standing-wave electricalurrent in the finite antenna array structure and absorption of thencoming EM radiation energy occurs at the designed resonantrequency of the antenna 8. When an antenna is excited into aesonance mode, it induces a cyclic plasma movement of freelectrons in the metal antenna. The electrons freely flow along thentenna, generating alternating current at the same frequency ashe resonance. Electromagnetic modeling illustrates that the cur-ent flow is toward the antenna feedpoint. In a balanced antenna,he feedpoint is located at the point of lowest impedance.

Our initial proof-of-concept was performed using spiral antennatructures and Fig. 1 shows a modeled thermal energy profile in-icating that the e-field is clearly concentrated at the center feed-oint. This provides a convenient location from which to collectnergy and transport it to other circuitry for conversion.

Antennas have electromagnetic radiation patterns, which allow

ig. 1 Calculated flow of THz currents to the antenna feed-oint. Red represents highest concentrated E field. Modeledith Ansoft HFSS 21.hem to exhibit gain and directionality and effectively collect and

11014-2 / Vol. 132, FEBRUARY 2010

loaded 26 Sep 2012 to 219.223.192.230. Redistribution subject to ASMconcentrate energy. A radiation pattern for our prototype spiralantenna, configured onto a groundplane reflector, is shown in Fig.2. Other antenna types i.e., dipoles, loops, etc. have similar ra-diation patterns but have not yet been characterized in this re-search. The ground-plane reflector layer effectively provides ahigh gain, forward-looking lobe, which enhances the ability tooptimize collection of radiated power over a predefined radialbeam area, such as a heat plume from an industrial thermal heatsource.

The nantenna radiation pattern displays angular reception char-acteristics that result in a wider angle of incidence exposure toradiation than a typical PV device. Any flux from the sun that fallswithin the radial beam pattern of the antenna is collected. Thisproperty is a critical antenna characteristic that optimizes energycollection from the sun as it moves throughout the sky, thus it maybe possible to reduce the need for mechanical solar trackingmechanisms. It also provides designers another mechanism forincreasing the efficiency of antenna arrays through the expansionof the radial field. Antennas by themselves do not provide a meansof converting the collected energy, so this will need to be accom-plished by associated circuitry, such as rectifiers. As illustrated inFig. 2, the electrical size of the antenna comprised of the radia-tion beam pattern is much larger than the physical size of theantenna. The virtual large surface area antenna focuses the elec-tromagnetic energy onto the nanosized energy conversion materialfabricated at the antenna feedpoint. Theoretical efficiency is im-proved by the enhanced radiation capture area of the antenna.

3 Energy Conversion MethodsThis research has demonstrated that infrared rays are capable of

creating an alternating current in the nantenna at THz frequenciesbut commercial grade electronic components cannot operate atthat switching rate without significant loss. Further research isplanned to explore ways to perform high frequency rectification,which requires embedding a rectifier diode element into the an-tenna structure. One possible embodiment uses metal-insulator-metal MIM tunneling-diodes. The MIM device consists of a thinbarrier dielectric oxide layer sandwiched between two metalelectrodes with different work functions. The device works whena large enough field causes tunneling of electrons across the bar-rier layers. A difference in work functions between the metal junc-tions produces nonlinear effects, resulting in high-speed rectifica-

Fig. 2 Typical electromagnetic radiation patterns of ground-plane spiral antenna. Antenna relative field strength db isplotted on a polar graph, where 0 deg=direct incidence. Thephysical size of the antenna is represented by the red ring andthe effective electrical size of the antenna is the radiation pat-tern approximate relative comparison only.tion. As the insulated layers become thinner, the probability

Transactions of the ASME

E license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

isMatit

ipclu

4

aams

ttgdirlp

ccifi

bsmstmto

Fien

J

Downncreases such that the tunneling effect can be maintained at THzwitching frequencies. There are also other permutations of the

IM diode, such as the MIIM with two insulating layers, whichlthough having higher efficiency, requires a higher bias to movehe electrons out of the quantum well created by the two insulat-ng layers making this less attractive for energy conversion withhe nantenna.

The outputs of the rectifiers can be dc-coupled together, allow-ng arrays of antennas to be networked to further increase outputower. Researchers in Ref. 9 demonstrated that microantenna-oupled infrared detectors Fig. 3 could be ganged together inarge arrays for imaging applications. A similar process can besed to provide large-area solar nantennas arrays.

Proof of Concept Through ModelingThe importance of a working model for nantenna development

nd manufacturing cannot be overemphasized. The cost and timessociated with fabricating complex interdependent structures andaterials can be significantly reduced by applying tolerances re-

ulting from an accurate model.One major goal of this research was to develop novel manufac-

uring methods that would enable economical large-scale fabrica-ion of nanoantenna structures. We determined that a simplifiedeometry, such as a loop antenna, would be a better candidate foreveloping the initial nanoantenna replication processes. Model-ng and manufacturing proof-of-concept trials were therefore car-ied out using a square-loop antenna design. In future research, theessons learned from this work will be applied to other more com-lex antenna designs.

Frequency selective surface FSS structures have been suc-essfully designed and implemented for use in radio frequencyRF and microwave frequency applications 8 and the laws oflassical physics apply for adapting microwave applications to thenfrared. The novel work carried out by INL for this program hasurther optimized FSS designs for operation in the terahertz andnfrared IR spectral regions.

It is recognized that several numerical analysis techniques cane employed for electromagnetic analysis. The basis for our re-earch and development is the adaptation of the method of mo-ents technique, which is a numerical computational method of

olving linear partial differential equations associated with elec-romagnetic fields. Ohio State University has implemented a

ethod of moments-based algorithm in a software product,ermed periodic method of moments PMM, which was devel-

ig. 3 Array of antenna-coupled microbolometersUCFmageRef. 9 The bolometer element will be replaced withmbedded rectifiers. A flexible panel of interconnected nanten-as may one day replace heavy, expensive solar panels.ped for the design of military RF frequency selective surfaces.

ournal of Solar Energy Engineering

loaded 26 Sep 2012 to 219.223.192.230. Redistribution subject to ASMThe details of this code are discussed in Ref. 10. Ohio StatesPMM model serves as the analysis engine for the INL design andmodeling methods.

4.1 Analytical ModelRLC Circuit. To model NEC struc-tures, it is first necessary to understand the electrical equivalentcircuit and basic theory of operation. The primary antenna struc-ture studied in the initial design of an NEC is a periodic array ofsquare-loop antennas, whose RLC circuit analog is shown in Fig.4. The electrical behavior of the structure is described as follows:thermally excited radiation induces current in the metal loops pro-viding inductance to the NEC. The gaps between the metallicloops and the gap within the loop comprise capacitors with adielectric fill. A resistance is present because the antenna is com-posed of lossy metallic elements on a dielectric substrate. Theresulting RLC circuit has a resonance tuned filter behavior 8.

It is evident that the proper selection of element and substratematerial is important in that they both contribute significantly tothe RLC parameters. The electro-optical characteristics of theNEC circuit and the surrounding media have the effect of shapingand optimizing the spectral response.

Our current designs incorporate a metal antenna layer, a dielec-tric standoff layer and a ground-plane see Fig. 5, where thestandoff spacing serves to create an optical resonance cavity. TheNEC-to-ground plane separation acts as a transmission line thatenhances resonance. The thickness of the standoff layer is selectedto be a 1/4 wavelength to ensure proper phasing of the electro-magnetic energy.

Fig. 4 Square FSS element and its RLC circuit analog

Fig. 5 Cross-sectional schematic of NEC structure showing

path of incident wave

FEBRUARY 2010, Vol. 132 / 011014-3

E license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

tnfcmtwlwkvvcrTw

pis

agawttcdrwsor

lcwtoar

5

l

0

Down4.2 Frequency-Dependent Modeling. The material proper-ies of circuits in the infrared are not well characterized. It wasecessary to adapt the PMM modeling software to account for therequency-dependent optical properties of dielectric and antennaonductor materials. The optical functions n and k of variousaterials that comprise NEC devices were determined using spec-

roscopic ellipsometry. These values characterize the way inhich materials respond to excitation by light of a given wave-

ength. One representation is the complex index of refraction n,here the real part n is the refractive index and the imaginary partis the extinction coefficient. The index n describes the phase

elocity of light in a material compared with propagation in aacuum, and the absorption of light is governed by the extinctionoefficient k. These quantities also determine the amount of lighteflected and transmitted at an interface between two materials.his allows accurate simulation of antenna behavior at thermalavelengths.It was demonstrated that the models accurately predict the

hysics of NEC energy absorption. This provides visualization ofnfrared thermal behavior that cannot be directly measured andupports prototyping of nanostructure devices.

4.3 Validation Of NEC Concept. A periodic array of loopntenna elements was designed for resonance at 10 m, the re-ion of maximum thermal reradiated emission from the earthsbsorption. All required antenna geometric and material propertiesere entered into the PMM model, which plots the reflection and

ransmission spectra for the electric field and the power spectra ofhe NEC antenna. The PMM tool was validated previously 11 byomparing experimental and modeled data sets. However, largerata sets for specific applications are required before deriving cor-elation and error bounds. It is anticipated that the research teamill collect future data sets for the purpose of estimating model-

pecific applications. The emissivity plot in Fig. 6 was acquired inrder to estimate NEC efficiencies in the 812 m wavelengthegion supporting the reradiated heat energy application.

The output of the model is in units of emissivity versus wave-ength and is referenced to a blackbody curve. A theoretical effi-iency of 92% was demonstrated at a peak resonance of 10 m,here efficiency is defined as the ratio of the power accepted by

he antenna to the power emitted by the sun at the earths surfacever a defined frequency range. The half-power bandwidth of thentenna is from 9.2 m to 12 m, easily covering the energyegion of interest.

Proof of Concept Through Small Scale Prototype

5.1 Laboratory-Scale Silicon Wafer Prototypes. A

Fig. 6 Modeled spectral output of aantenna structure with a 10 m resoaboratory-scale device was prototyped using standard semicon-

11014-4 / Vol. 132, FEBRUARY 2010

loaded 26 Sep 2012 to 219.223.192.230. Redistribution subject to ASMductor integrated circuit fabrication techniques 12, as outlinedschematically in Fig. 7. Prototype fabrication was performed un-der subcontract at the University of Central Florida. Direct-writeelectron-beam lithography EBL was employed for the fabrica-tion of metallic loop antenna arrays on dielectric and semiconduc-tor substrates. The completed structure is shown in Fig. 8.

E-beam lithography provided a convenient way to systemati-cally investigate dimensions, spacing, and geometrical effects in acontrolled manner and allowed for systematic tradeoff studies.However, the throughput and cost limitations of EBL do not sup-port large-scale manufacturing.

EC. Proof of concept based on loopnce.

Fig. 7 Fabrication process flow for a square-loop IR FSS

Fig. 8 SEM of a small portion of the completed IR NEC madeby e-beam lithography. The entire IR FSS covers 40 mm2 and is

8

n Nnacomposed of 9.7610 elements.

Transactions of the ASME

E license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

6m

w1Tw

pfswsMe

iitad

swpsNln

7

cussnNeac

ehpht

Fu

J

DownProof of Concept Through Experimental Measure-ents

6.1 Silicon Wafer Prototype. Experimental measurementsere performed on the prototype device Fig. 8, consisting of a.0 cm2 array of loop nantennas fabricated on a silicon substrate.his rigid substrate prototype served as the precursor to currentork on large-scale flexible substrate nantennas.The IR nantennas were designed to operate as a reflective band-

ass filter centered at a wavelength of 6.5 m. The spectral sur-ace characteristics from 3 m to 15 m were studied usingpectral radiometer and FTIR analysis methods. The prototypeas elevated to a temperature of 200C and its spectral radiance

pectrum was compared with blackbody emission at 200C.aximum contrast is over 90% between emission near 4 m and

mission at resonance 11.A further metric of IR FSS performance is its spectral emissiv-

ty. The ratio of IR FSS radiance to blackbody radiance is shownn Fig. 9. The emissivity approaches unity at resonance, indicatinghe IR FSS radiates like a blackbody at a specific wavelength with

steep dropoff toward shorter wavelengths and a more gradualropoff in emission toward longer wavelengths.

The experimental spectrum closely correlates to the modeledpectrum, with peak resonance being achieved at the 6.5 mavelength. Resonant wavelengths are obtained by varying FSSarameters, such as the standoff layer thickness and FSS elementize and spacing. This basic design can be adapted to specificEC implementations. Furthermore, out-of-band emission is 90%

ess than emission at resonance, making this IR FSS an excellentarrowband emissive energy concentrator or reflective filter.

Precursors to Roll-to-Roll Manufacturing

7.1 Nanoforming Process. Well-established FSS theory indi-ates that nanoscale antennas could potentially be fabricated forse in any region of the EM spectrum. However, even the featureizes for an IR NEC require the use of EBL, X-ray lithography orome other ultrahigh-resolution fabrication means. These tech-iques are generally very slow and expensive and have limitedECs to small sizes, severely compromising their utility. How-

ver, due to recent advancements in high-resolution lithographynd roll-to-roll R2R processing of flexible polymer films, low-ost manufacturing now appears achievable.

Pioneering work by Polaroid 1316 in continuous R2R micro-mbossing and more recent work by MicroContinum 17,18 inigh-resolution R2R patterning has shown the possibility of massroduction of complex multilayer patterned film structures. Weave extended these techniques to the fabrication of IR FSS struc-

ig. 9 Demonstrated success in energy collection. Validatedsing a spectral radiometer and FTIR.ures on flexible polymeric substrates, using a process that we

ournal of Solar Energy Engineering

loaded 26 Sep 2012 to 219.223.192.230. Redistribution subject to ASMrefer to as nanoforming. A key feature of the manufacturing-scalable design of this process is the use of a master template, arelief pattern made by e-beam, laser, or other precision patterngeneration method from which durable high-resolution productiontemplates tools are produced for continuously replicating therequired pattern in polymer substrates. A single master templatedescribed in the next section is used to make numerous inexpen-sive but highly accurate second and third generation productiontools used for mass replication of the original template pattern inthe polymeric substrate, thereby greatly reducing overall manufac-turing costs. Further processing of the polymeric structure resultsin formation of the completed NEC device.

It is recognized that continuous large-scale production of com-plex multilayer functional devices possessing nanoscale featuresposes a significant problem. Coupling this with a requirement forproducing these devices on flexible, shape-conforming substratescomplicates this problem further. We have addressed these con-cerns by developing a multistep fabrication process by which wehave produced multiple large-area coupons and this comprises thefirst demonstration of feasibility of large-scale manufacture ofmultilayer functional nanoantenna devices on flexible substrates.

7.2 Development and Replication of Master Template. Thefirst step in producing large-area NECs is forming a large-areatemplate with the requisite antenna geometry. One of the mostimportant aspects of NEC design is an understanding of the toler-ances required to make structures behave like antennas at infraredwavelengths. Tolerances were derived through extensive modelingstudies to evaluate the impact of changing geometric parameters,including: antenna wire width/depth, antenna periodicity, gap sizebetween adjacent antennas, etc. Many parameters were codepen-dent and required use of optimization algorithms.

Based on this modeling work, the resulting NEC geometry wasconverted into a master template on a 6 in. 152.4 mm roundsilicon wafer for large-scale device fabrication. The template pro-cess involved a series of steps, including direct-write e-beam li-thography, sequential anisotropic oxide etching, and deep reactiveion etching DRIE into silicon MEMS and Nanotechnology Ex-change, Reston, VA. An SEM of the oxide mask with dimensionsis shown in Figs. 10a and 10b top and side views, respec-tively and 10c tilted view. The results after the second etch stepDRIE are shown in Fig. 11 prior to removal of top SiO2 layer.The template consists of approximately 10109 antennaelements.

The large-area silicon template formed by direct-write EBL andmultiple etch steps is not only very expensive to produce also veryfragile and is thus not well suited for direct use in manufacturing.Instead, a technique similar to that developed for the productionof CDs and DVDs 19,20 is used to fabricate a durable moldingtool capable of repeated use in replicating the master pattern.

This utilizes Ni electroforming, an ultralow-stress electroplat-ing process in which a thin conductive metal layer is first confor-mally coated over the master template surface followed by thedeposition of a thick typically 200 m layer of Ni onto theconductive layer. Separation of the plated layer yields a durableNi surface having a pattern complementary to that of the mastertemplate. In contrast to CD and DVD template patterns compris-ing preformat and data bit patterns, which tend to be shallow witha low aspect ratio 20, the nantenna template is very deep with ahigh aspect ratio, thus direct Ni plating of the silicon pattern wasfound to be very problematic. Ultimately, an intermediate polymertemplate made from the silicon wafer was instead used as thetemplate for the Ni electroforming process. Additional Ni copieswith the same symmetry as the master template were made fromthis first Ni copy with subsequent plating from these copies alsobeing possible. It should be noted that the sequence of formingsuccessive generational copies of an original template has severaladvantages: it allows either positive or negative relief patterns tobe formed depending on which symmetry is ultimately required

and it enables multiple inexpensive high-resolution production

FEBRUARY 2010, Vol. 132 / 011014-5

E license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

tof

vrs

FerqlsteTa

0

Downools to be made from each template generation, allowing theriginal master template to be preserved. A schematic of the toolormation and polymer replication process is given in Fig. 12a.

7.3 Fabrication Steps. Several alternate pathways were de-eloped and investigated in order to convert the polymer templateeplicas into a complete nantenna structure. One such pathway is

ig. 10 a SEM top view of Si wafer after anisotropic oxidetch. Narrow ridges along walls and star shape in square areesidual e-beam resist layer which are removed after subse-uent oxide strip. b SEM cross section of patterned oxide

ayer of Fig. 11a prior to DRIE step. Due to the unusual de-ign of the structure, the initial etch width was slightly largerhan the design for these experiments but higher accuracy isxpected in the future as the etch process is further refined. cilted view SEM image of patterned oxide layer of Figs. 11and 11b.hown schematically in Fig. 12b. Here, the master template

11014-6 / Vol. 132, FEBRUARY 2010

loaded 26 Sep 2012 to 219.223.192.230. Redistribution subject to ASMdepth is designed to correspond to them 1/4 wave separation dis-tance between the ground-plane and the antenna loop layer andthis separation is reproduced in the depth of the replicated poly-mer structure. SEM images of polymer replicas made from theproduction tools in turn made from the silicon master templateare shown in Figs. 13a and 13b. Vacuum deposition of themetal layer is used to form the antenna layer bottom of deeptrench and partial ground-plane top surface, and the trenchesare then filled with the same polymer that comprises the substrate.Due to its IR transparency, polyethylene was chosen as the sub-strate and a polyethylene dispersion was used to fill the trenches.A continuous ground-plane is then deposited, and because thecomplete structure at this point is less than 12 m thick, an op-tional backside film is laminated to the ground-plane to providemechanical support and for ease of handling the structure.

Fig. 11 Cross section of master template after DRIE step,showing oxide mask layer over bulk silicon

Fig. 12 a Schematic of pattern replication process. b Sche-

matic of antenna and ground plane process.

Transactions of the ASME

E license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

lfcmpcfi

ndept1at

rtse

Ftv

J

DownThe fidelity between the nantenna master template and the rep-icas is good, as seen in Figs. 10, 11, and 13. In addition, iterativefine tuning of the master template etch process can be used tourther improve the match between the design and template. Toompensate for physical characteristics of different types of poly-eric substrates e.g., coefficient of thermal expansion, etc., a

recompensation step in the template design and/or etch processan also be used, since the transfer function between design andnal replica has been found to be very consistent.

7.4 Flexible Substrate NEC Prototype Devices. The tech-iques described in the previous sections have been used to pro-uce prototype NEC structures in thin 12 m flexible poly-thylene substrates. Using a semi-automated process, we haveroduced a number of 4 in. 100 mm square coupons Fig. 14ahat were tiled together to form sheets of NEC structures Fig.4b. The nantenna prototypes, manufactured with semi-utomated methods, are in the process of being further experimen-ally tested.

Initial thermal characterization has been performed using high-esolution thermal cameras Fig. 15a, providing proof of selec-ive thermal energy collection. Figure 15b shows a circular NECtructure at the center of a metal plate. When excited by thermal

ig. 13 a and b SEM image of polymer replicas made fromooling not shown from wafer master template top, normaliew; bottom, tilted viewnergy, the metal reference plate and the NEC structure have dis-

ournal of Solar Energy Engineering

loaded 26 Sep 2012 to 219.223.192.230. Redistribution subject to ASMtinct emissivity contrasts Fig. 15c, based on the modeled de-sign. This behavior will be further optimized to increase the col-lection efficiency of thermal radiation in the 10 m solarspectrum.

Because multiple pathways have been developed to produce thecomplete nantenna structure, future work will include a study ofthe tradeoffs inherent in these process variants. Tooling and pro-cesses are currently being optimized using semi-automated labo-ratory equipment. These results will be used to assess the impactof process variables on costs, yields and throughput and will serveas the basis for designing a multistep R2R pilot process for thecontinuous fabrication of these unique optical devices.

8 ApplicationsEconomical large-scale fabrication of NEC devices would sup-

port applications such as covering the roofs of buildings andsupplementing the power grid. NEC devices can also be optimizedfor collection of discrete bands of electromagnetic energy. Avail-able flux that reaches the earths surface in the 0.80.9 m rangeis about 800 W /m2 at its zenith. While visible light flux is de-pendent on cloud cover and humidity, incident light in this rangeis a constant during daylight hours. Double-sided panels couldabsorb a broad spectrum of energy from the sun during the day,while the other side might be designed to take in the narrowfrequency of energy produced from the earths radiated heat orpotentially residual heat from electronic devices.

This technology may also support several emerging applica-tions, including passive thermal management products, such asbuilding insulation, window coatings, and heat dissipation in elec-

Fig. 14 a Prototype FSS structure on flexible substrate andb 300 mm600 mm nantenna sheet stitched together from18 couponstronic consumer products, such as computers. The nantennas are

FEBRUARY 2010, Vol. 132 / 011014-7

E license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

brtsw

Fpp

0

Downroadband collectors of energy with a tailorable spectral emissionesponse. This in effect generates a frequency selective distribu-ion of energy that potentially can collect unwanted energy re-idual or incident heat and redistribute it at other innocuous

ig. 15 a Experimental setup for thermal characterization ofrototypes. b and c Optical top and graphical bottom ex-erimental results from thermal characterization.avelengths.

11014-8 / Vol. 132, FEBRUARY 2010

loaded 26 Sep 2012 to 219.223.192.230. Redistribution subject to ASMOther applications for this technology are very diverse. It isconceivable that nantenna collectors, combined with appropriaterectifying elements, could be integrated into the skin of con-sumer electronic devices to continuously charge their batteries.Due to the ability to integrate the nanostructures into polymermaterials, it is possible that, for example, NECs could even bedirectly fabricated into polymer fabrics.

9 Conclusions and Future WorkBoth modeling and experimental measurements have demon-

strated that the nantennas described in this work can absorb closeto 90% of the available in-band energy. Optimization techniques,such as increasing the radial field size, could potentially increasethis efficiency to even higher percentages. More extensive re-search needs to be performed on energy conversion methods toderive overall system electricity generation efficiency. The circuitscan be made from any number of different conducting metals andthe nantennas can be formed on thin, flexible materials like poly-ethylene.

Further laboratory evaluations of the flexible substrate NECprototypes are planned. Manufacturing methods will continue tobe refined to support roll-to-roll manufacturing of the nanostruc-tures. Future work will focus on designing the nantenna structurefor operation in other wavelengths. By further shaping the spectralemission of the NEC, it may be possible to concurrently collectenergy in the visible, near-infrared and midinfrared regions.

This research is at an intermediate stage and may take years tobring to fruition and into the market. The advances made by ourresearch team have shown that some of the early barriers to thisalternative PV concept have been crossed and this concept has thepotential to be a disruptive and enabling technology. We encour-age the scientific community to consider this technology alongwith others when contemplating efforts and resources for solarenergy.

AcknowledgmentThe authors would like to thank Glenn Boreman and staff at the

CREOL, University of Central Florida, for their work in prototyp-ing of FSS structures and radiometric measurements, and MEMSand Nanotechnology Exchange Reston, VA for large-area tem-plate patterning of silicon wafers and related SEM images. Wealso recognize the support of Ben Munk and Ron Marhefka ofOhio State University for their support in electromagnetic model-ing. The authors would like to thank Judy Partin for her contribu-tions in thermal characterization and analysis. The authors extenda special thank you to Jay Rsykamp at NAVSEA for his supportand much needed direction. This work was supported by the U.S.Department of Energy, Office of Nuclear Energy, under DOEIdaho Operations Office Contract DE-AC07-05ID14517.

References1 Consolmagno, G. J., and Schaefer, M. W., 1944, Worlds Apart: A Textbook in

Planetary Sciences, Prentice-Hall, Englewood Cliffs, NJ.2 http://www.mtt.org/awards/WCB%27s%20distinguished%20career.htm3 http://www.kurasc.kyoto-u.ac.jp/plasma-group/sps/history2-e.html4 Wilke, I., Oppliger, Y., Herrmann, W., and Kneubuhl, F. K., 1944, Nanometer

Thin-Film Ni-NiO-Ni Diodes for 30 THz Radiation, Appl. Phys., A58, pp.329341.

5 Krishnan, S., Bhansali, S., Buckle, K., and Stefanakos, E., 2006,Fabricationand Characterization of Thin-Film Metal-Insulator-Metal Diode for Use inRectenna as Infrared Detector, Mater. Res. Soc. Symp. Proc., 935, pp. 4047.

6 Alda, J., Rico-Garca, J., Lpez-Alonso, J., and Boreman, G., 2005, OpticalAntennas for Nano-Photonic Applications, Nanotechnology, 16, pp. S230S234.

7 Hoertz, P. G., Staniszewski, A., Marton, A., Higgins, G. T., Incarvito, C. D.,Rheingold, A. L., and Meyer, G. J., 2006, Toward Exceeding the ShockleyQueisser Limit: Photoinduced Interfacial Charge Transfer Processes that StoreEnergy in Excess of the Equilibrated Excited State, J. Am. Chem. Soc., 128,pp. 82348245.

8 Munk, B. A., 2000, Frequency Selective Surfaces: Theory and Design, Wiley,New York, NY, pp. 223.

9 Gonzalez, F. J., Ilic, B., Alda, J., and Boreman, G. D., 2005, Antenna-

Coupled Infrared Detectors for Imaging Applications, IEEE J. Sel. Top.

Transactions of the ASME

E license or copyright; see http://www.asme.org/terms/Terms_Use.cfm

J

DownQuantum Electron., 111, pp. 117120.10 Henderson, L. W., 1993, Introduction to PMM, Version 4.0, Ohio State

University, EletroScience Laboratory, Columbus, OH, Technical Report No.725 347-1, Contract SC-SP18-91-0001.

11 Monacelli, B., Pryor, J., Munk, B., Kotter, D., and Boreman, G., 2005, Infra-red Frequency Selective Surfaces Based On Circuit-Analog Square Loop De-sign, IEEE Trans. Antennas Propag., 532, pp. 745752.

12 Monacelli, B., Pryor, J., Munk, B., Kotter, D., and Boreman, G., 2005, Infra-red Frequency Selective Surfaces: Square Loop Versus Square-Slot ElementComparison, Paper No. AP0508-0657.

13 Land, E. H., 1977, An Introduction to Polavision, Photograph. Sci. Eng.,215, pp. 225236.

14 Slafer, W. D., Walworth, V. K., Holland, A. B., and Cowan, J. J., 1987, In-vestigation of Arrayed Silver Halide Grains, J. Imaging Sci., 31, pp. 117125.

15 Cowan, J. J., and Slafer, W. D., 1986, Holographic Embossing at Polaroid:ournal of Solar Energy Engineering

loaded 26 Sep 2012 to 219.223.192.230. Redistribution subject to ASMThe Polaform Process, Progress in Holographic Applications-1985, SPIE Pro-ceedings Vol. 600, pp. 4956.

16 Slafer, W. D., Kime, M., Monen, M., Horton, W., and Wan, L., 1992, Con-tinuous Web Manufacturing of Thin Coversheet Optical Media, SPIE Pro-ceedings, Optical Data Storage, Vol. 1663, pp. 324335.

17 Praino, R. F., and Slafer, W. D., 2006, Techniques for Patterning ConductiveLayers, Invited Paper Presented at the 2006 AIMCAL Fall Technical Confer-ence, Reno, NV, Oct. 2225.

18 Slafer, W. D., and Praino, R. F., 2008, Progress in Roll-to-Roll Patterning ofTransparent & Metallic Conductors, Proceedings of USDC/FlexTech AllianceFlexible Electronics and Displays Conference, 9.3, Phoenix, AZ, Jan. 2124.

19 Isailovic, J., 1985, Videodisc and Optical Memory Systems, Prentice-Hall,Englewood Cliffs, NJ.

20 http://www.ecma-international.org/publications/standards/Ecma-272.htm21 Ansoft Corporation, 2005, Ansoft High Frequency Structure Simulator, Users

Guide, Vol. 10.FEBRUARY 2010, Vol. 132 / 011014-9

E license or copyright; see http://www.asme.org/terms/Terms_Use.cfm