arX

iv:1

204.

0330

v1 [

phys

ics.

optic

s] 2

Apr

201

2

Optical Yagi-Uda nanoantennas

Ivan S. Maksymov, Isabelle Staude, Andrey E. Miroshnichenko, and Yuri S. Kivshar∗

Nonlinear Physics Centre and Centre for Ultrahigh Bandwidth Devices for Optical Systems (CUDOS),Research School of Physics and Engineering, The Australian National University, Canberra, ACT 0200, Australia

(Dated: April 3, 2012)

Conventional antennas, which are widely employed to transmit radio and TV signals, can beused at optical frequencies as long as they are shrunk to nanometer-size dimensions. Opticalnanoantennas made of metallic or high-permittivity dielectric nanoparticles allow for enhancing andmanipulating light on the scale much smaller than wavelength of light. Based on this ability, opticalnanoantennas offer unique opportunities regarding key applications such as optical communications,photovoltaics, non-classical light emission, and sensing. From a multitude of suggested nanoantennaconcepts the Yagi-Uda nanoantenna, an optical analogue of the well-established radio-frequencyYagi-Uda antenna, stands out by its efficient unidirectional light emission and enhancement.Following a brief introduction to the emerging field of optical nanoantennas, here we review recenttheoretical and experimental activities on optical Yagi-Uda nanoantennas, including their design,fabrication, and applications. We also discuss several extensions of the conventional Yagi-Udaantenna design for broadband and tunable operation, for applications in nanophotonic circuits andphotovoltaic devices.

Keywords: optical nanoantennas, Yagi-Uda antennas, all-dielectric nanoantennas, plasmon-ics, nanoparticles, spectral tuning, Purcell factor

I. INTRODUCTION

Antennas are all around in our modern wireless soci-ety: they are the front-ends in satellites, cell-phones, lap-tops and other devices that establish the communicationby sending and receiving radio waves having frequenciesfrom 300 GHz to as low as 3 kHz. However, accordingto Maxwell’s equations, the same principles of directingand receiving electromagnetic waves should be workingat various scales independently of the wavelength. Thus,one may naturally ask ”Can a TV-antenna send a beamof light?” And the answer is ”Yes, optical antennas can!”The study of optical nanoantennas [1–4] are one of

the most promising areas of activity of the current re-search in nanophotonics due to their ability to bridgethe size and impedance mismatch between nanoemittersand free space radiation [5], as well as manipulate lighton the scale smaller than wavelength of light [6]. Suchdevices possess collective oscillations of conduction elec-trons of metals known as plasmon modes, which increaselight coupling from nanoemitters to the nanoantenna orfrom the nanoantenna to freely propagating light, andvice versa. These intriguing properties implicate greatpotential for the development of novel optical sensors,solar cells, quantum communication systems, and molec-ular spectroscopy techniques, in particular, for the emis-sion enhancement and directionality control over a broadwavelength range. However, on the other hand, plasmonsare also responsible for energy loss in metals at opticalfrequencies, thereby making some of the traditional lowfrequency methods of controlling electromagnetic wavesfutile.

∗ Ivan S. Maksymov mis124@physics.anu.edu.au

In this paper, we provide an overview of ongoing re-search into optical Yagi-Uda nanoantennas – an opti-cal analogue of the well-established radio-frequency (RF)Yagi-Uda antenna, which stands out by its simplicity andexcellent directive properties. By discussing and compar-ing the physics behind electromagnetic wave propaga-tion in RF and optical range we demonstrate the limitsof applicability of classical RF architectures to opticalnanoantennas and, conversely, show what optical Yagi-Uda nanoantenna can do and RF Yagi-Uda antennas can-not. Based on this comparative analysis, we overview themost recent advances in the physics of optical Yagi-Udananoantennas, as well as fabrication and characterizationtechniques. Finally, focusing on novel broadband andwavelength tunable optical Yagi-Uda nanoantennas, wesummarize the potential applications of optical arrayednanoantennas and discuss open issues.

II. HISTORY

Recent advances in nanophotonics confirm that thephrase ”There is nothing new except what is forgotten”[7] is also relevant to novel optical antennas that can re-ceive light waves in the same way as car antennas canpick up radio waves. Indeed, it turns out that most ofthe classical RF antenna design concepts remain applica-ble at the nanoscale [1, 2]. Moreover, the development ofoptical nanoantennas faces similar problems as the pio-neering experiments with RF antennas in the last decadeof the 19th century, including, the major one – the weak-ness of the signals to be received by the antenna.

A simple, efficient, and low-cost solution of this prob-lem for RF antennas was found by a Japanese inventorShintaro Uda, assistant to Prof. Hidetsugu Yagi of To-

2

FIG. 1. (a) Typical RF Yagi-Uda antenna held by one of its inventors Hidetsugu Yagi. The antenna has one reflector andtwo directors. The arrow shows how the length of the antenna elements is related to the incident wavelength λ (reproducedfrom Ref. [8]). (b) Radio Arcala ”Mammoth” 160/80 meter directional Yagi-Uda antenna built in Finland by amateur radiooperators. Each of the four guy wires of the antenna extends 120 meters from the tower representing an area of 170x170meters. The total tower height and the antenna weight are 100 m and 39.6 tons, respectively. Reproduced with permissionfrom www.radioarcala.com.

hoku Imperial University, Sendai, Japan, in 1926 [Fig.1(a)]. He suggested to use multiple single elements toform an array to achieve constructive interference of ra-dio waves in one direction and destructive in the oppo-site. Thanks to this effect the new antenna was capableof capturing weak RF signals [9]. Moreover, by virtue ofelectromagnetic reciprocity, it could also emit RF signalsin a directed manner.

The construction turned out to be so useful that, afterthe translation of Uda’s paper from Japanese to Englishand applications for patents by Yagi (that is why manysources omit the name of Uda), the new antenna design,also known as ”wave channel”, was quickly recognizedin the UK, USA, Germany and Soviet Union and imple-mented both for the transmission and detection of RFwaves on airborne radars during the Second World War.Ironically, the invention received just a little attention inJapan [8, 10].

Nowadays, the conventional Yagi-Uda antenna design,which can be found on almost every roof, not only re-mains the reference for the ham radio [see Fig. 1(b)]and TV antennas [11], but also generated a new thrustof activities in nano-optics [12–15] aimed at achievingbroadband unidirectional light emission and detection atthe nanoscale.

FIG. 2. Schematic of an optical Yagi-Uda antenna consistingof gold nanorods used for the feed, reflector and directorswith the possible dimensions in comparison with the incidentwavelength λ. At the typical telecommunication wavelengthof λ = 1550 nm the area occupied by the nanoantenna is∼ 400x400 nm and its weight is ∼ 70 femtogrammes.

III. OPTICAL YAGI-UDA NANOANTENNAS

The Yagi-Uda antenna is one of the most brilliant ideasof directive antennas, that is antennas radiating greaterpower in only one direction. It is simple to build, and ithas a high efficiency in terms of directivity [11] and lossesthat occur throughout the antenna while it is operatingat a given frequency. Moreover, it can be shrunk to beas small as 1/25 of a human hair with a minimum ofchanges in the construction in order to operate in the

3

infrared and visible regions.A generic Yagi-Uda antenna consists of an actively

driven element called feed surrounded by passive ele-ments, which are not driven [Fig. 1(a)]. These elementsare usually of two types: a resonant reflector and a fewequally spaced directors. The key parameter of the wholestructure is the length of the feeding element, whichshould be at resonance at the desired frequency. Thelengths of the reflector and directors are chosen to be offresonance and designed to add up in phase in the forwarddirection, and to cancel in the reverse direction. For thatpurpose, the director elements are made shorter and thereflector is made longer than the feeding element, whichleads to in– and out– of-phase current distributions.

A. Plasmonic nanoantennas

The choice of the optimal length of the feed for opticalYagi-Uda antennas (see Fig. 2) differs from the standardprocedure adopted in RF antenna design [11]. Metals atRF are highly conductive, and, thus, can be considered asnearly perfect reflectors. The penetration length of thefields (the skin depth) is negligible compared with thedimensions of the antenna elements. However, at opticalfrequencies the skin depth is of the order of 10 nm, and,thus, is comparable with the dimensions of the nanoan-tenna elements. Therefore, the applicability of classicalprinciples of the RF antenna design principle relying onan incident wavelength is reduced, and it must be re-vised by rigorously treating a metal as a strongly coupledplasma. The resulting reduced effective wavelength λeff

seen by optical nanoantennas is related to the incidentwavelength λ by a simple relation [1, 2, 16]:

λeff = a+ b

(

λ

λp

)

, (1)

where a and b are some coefficients with dimensions oflength that depend on geometry and material proper-ties [16], and λp is the plasma wavelength. According tothis wavelength scaling rule, an optical half-wave nanoan-tenna is not λ/2 in length but is of a shorter length λeff/2.The difference between the length of the optical and RFdipolar antennas depends on the antenna geometry andis usually in the range of 2−6 for gold and silver nanopar-ticles [1, 16].The choice of the cross-sectional dimensions of optical

nanoantenna elements and internal edge-to-edge spacingbetween them also differs from the standard RF proce-dure. In the RF regime, the inter-element spacing typi-cally ranges between 0.1λ and 0.315λ and is conditionedby practical rather than physical considerations such asthe simplicity and stability of the antenna construction.As a rule, the diameter of RF antenna elements is smallcompared to the spacing between the elements. At opti-cal wavelengths, however, the antenna element diameter

FIG. 3. Optical Yagi-Uda nanoantenna driven by quantumdots. (a) Scanning electron microscopy image of the fabri-cated 5-element Yagi-Uda nanoantenna. (b) Angular radia-tion diagram for the fabricated Yagi-Uda nanoantenna (blackline) vs. theoretical prediction (red line). Figure is repro-duced from Ref. [12].

becomes a crucial parameter because the inevitable ab-sorption losses in metals can be reduced by making thevolume of the elements larger [17].Another difference between optical and RF antennas

is the excitation of the feeding element. In RF region,Yagi-Uda antennas are locally driven at the feed-gap byelectrically connecting the feed, which is typically a λ/2dipole, to the feed-line. However, most of the opticalnanoantennas investigated so far have been driven fromthe far-field [2]. Such excitation scheme is not optimalfor the Yagi-Uda design because the feed is surroundedby other elements placed at small distances as comparedwith the operating wavelength. Therefore, it would beadvantageous to place the light source in the near-fieldregion of the feed. It makes the optical Yagi-Uda nanoan-tennas perfect candidates for controlling emission of iso-lated quantum nanoemitters, such as, e.g., quantum dots[18] and color centres in nanodiamonds [19].In order to obtain a strong near-field coupling of an

isolated emitter to the fundamental mode of the feed, itis critical to place the emitter at a high electric modedensity point. For the nanoantenna geometry shown inFig. 2 this point corresponds to one of the ends of thefeeding element. A judicious engineering of the feed, suchas, e.g., the introduction of a small (30−50 nm in width)gap in the centre of the feed, can further increase thecoupling strength by exploiting the Purcell effect [20],which ensures preferential emission into a highly-confinedfundamental cavity mode at high rate.The effective wavelength design principle was used for

4

FIG. 4. Phased array of optical Yagi-Uda nanoantennas. (a)Artists view of a 3x3 Yagi-Uda nanoantenna array fed byphase modulating feeding circuits. (b) SEM image of an op-tical Yagi-Uda nanoantenna array. The gold structure is fab-ricated on a glass substrate. The nanorods are embedded inphotopolymer. Scale bar is 500 nm. Figure is reproducedfrom Ref. [14].

the fabrication and experimental demonstration of highlyunidirectional emission of a quantum dot coupled to anoptical Yagi-Uda nanoantenna [12] shown in Fig. 3(a).One of the most important antenna characteristics - theangular emission diagram – was calculated from the ex-perimental Fourier-plane image and agrees well with theresult of numerical simulations [Fig. 3(b)]: the nanoan-tenna emits mainly into the medium which is opticallydenser. This result is in good agreement with the predic-tion that emitters at the surface between two interfacesemit mainly into the medium which is optically denser inthe case that the emission direction is close to the criticalangle [12].

In contrast to earlier experimental study of similarnanoantennas offering directional far-field scattering ofa polarized laser beam [13], the device in Fig. 3(a) en-sures full control of the direction of light emission fromsingle quantum emitters. It represents a decisive step to-ward the development of subwavelength quantum lightsources [21] with directive properties of RF antennas notattainable with optical cavities [18, 20]. The creationof such sources makes it possible to transmit and receivesingle and entangled photons [18] at subwavelength scale,which is an essential feature for quantum cryptography,communication and computing [22].

The near-field excitation design of Yagi-Uda nanoan-tennas also opens up new ways for the implementation ofmore complicated RF antenna concepts at the nanoscale.Recently, arrays of three-dimensional (3D) optical Yagi-Uda nanoantennas (Fig. 4), fabricated using top-downfabrication techniques combined with layer-by-layer pro-cessing, have been suggested [14]. The creation of opti-cal nanoantenna arrays brings the concept of RF phasedantenna arrays to the nanoscale, where each antenna isdriven independently with a defined phase and ampli-tude. It considerably widens the applicability range ofoptical Yagi-Uda nanoantennas, making it possible tosteer the direction of the emitted light beam by address-ing each nanoantenna independently and controlling thephase of the feeding circuit by means of phase modulators[14].It is also worth mentioning the efforts targeted to

reduce absorption losses in Yagi-Uda nanoantennas bymodifying the geometry and the arrangement of thenanoparticles composing the Yagi-Uda architectures [23–27]. These ultracompact nanoantennas look like a Yagi-Uda structure without either reflector or directors. Inter-estingly, a construction that would lead to a poor perfor-mance in RF region not only works well at the nanoscalebut also opens up novel opportunities not possible atlarger scales, such as a very high front-to-back ratio at-tainable with only two coupled nanoparticles separatedby a distance as small as ∼ 1/65 of the operating wave-length. Another extreme case of a Yagi-Uda structure isrepresented by nanoantenna arrays consisting of equallyspaced metal nanoparticles of the same size [28]. Thesearrays possess Bragg modes allowing for long-range andlarge area emission enhancement even if the nanoantennais excited by an emitter with unfavorable dipole momentorientation [29].

B. All-dielectric nanoantennas

Another way for reducing losses and thereby increasingthe antenna efficiency is to utilize high-permittivity low-loss dielectric nanoparticles instead of metallic ones [30].This approach has been already used in RF region [31]where Yagi-Uda dielectric resonator antennas are knownto be broadband in frequency as compared with all-metalconstructions.However, in the optical region, the use of dielectric ma-

terials promises great advantage over highly absorptivemetals. Many studies concentrate on the nanoparticlesmade of silicon – one of the key materials of integratedphotonics. The real part of the permittivity of the siliconis about 16 whereas the imaginary part, indicating dis-sipative losses, is up to two orders of magnitude smallerthan that of silver or gold [32].It has recently been suggested to make optical Yagi-

Uda nanoantennas of all-dielectric elements [33]. All-dielectric Yagi-Uda nanoantennas can be considered asan alternative to their metallic counterparts because

5

FIG. 5. (a) All-dielectric optical Yagi-Uda nanoantenna,consisting of the reflector (sphere 1, right) of the radius of 75nm, and smaller directors (spheres 2-5) of the radii of 70 nm.The dipole source is placed equidistantly between the reflec-tor and the first director at the distance G. The separationbetween the neighboring directors is also G. (b) Directivity ofthe all-dielectric Yagi-Uda nanoantenna as a function of thewavelength for the separation distance G = 70 nm. Insetsshow radiation pattern diagrams at particular wavelengths.

of low losses and, as was suggested earlier [34], sin-gle spherical nanoparticles made of high-permittivity di-electrics may support both electric and magnetic reso-nant modes, which is not attainable with single sphericalmetal nanoparticles. This feature may greatly expandthe applicability of optical nanoantennas for, e.g., detec-tion of magnetic dipole transitions of molecules.

Figure 5(a) shows an example of all-dielectric Yagi-Udananoantenna consisting of four directors and one reflectormade of silicon. The optimal performance of the nanoan-tenna at one operating wavelength can be achieved whenthe director nanoparticles sustain a magnetic resonanceand the reflector nanoparticle sustains an electric reso-nance. Thus, the radii of the directors should be smallerthan the radius of the reflector because the positions ofthe magnetic and electric resonances shift linearly withthe radius of the nanoparticle [33]. Consequently, thenanoantenna investigated in Ref. [33] consists of nanopar-ticles with the radii of 70 nm and 75 nm used for directors

and reflector, respectively.Figure 5(b) plots the directivity spectrum of the all-

dielectric Yagi-Uda nanoantenna excited by a dipolesource placed equidistantly between the reflector and thefirst director surfaces at the distance G = 70 nm. Theinsets in Fig. 5(b) show the radiation patterns at par-ticular wavelengths. One observes a strong maximum atλ ≈ 500 nm. The main lobe of the emission pattern atthis operating wavelength is extremely narrow with thebeam-width about 40o and negligible backscattering.However, the radiation efficiencies of both all-dielectric

and plasmonic Yagi-Uda nanoantennas are nearly thesame for G = 70 nm with the averaged value 70%. Al-though dissipation losses of silicon are much smaller thanthose of silver or gold, the dielectric particles absorb theelectromagnetic energy by the whole spherical volume,while the metallic particles do it at the surface only. Asa result, there is no substantial difference in the over-all performance of these two types of nanoantennas forrelatively large distances between the elements.The difference between all-dielectric and plasmonic

nanoantennas becomes very strong for smaller spacingsbetween the elements: the radiation efficiency of all-dielectric nanoantennas is insensitive to the separationdistance whereas the radiation efficiency of plasmonicnanoantennas drops significantly. This is caused by lo-cal field enhancement between two adjacent metallic sur-faces, which leads to effectively larger absorption of thewhole structure. Thus, the use of all-dielectric nanopar-ticles without the field enhancement allows making morecompact nanoantennas e.g. for on-chip wireless commu-nication. But, in contrast to plasmonic nanoantennas,all-dielectric nanoantennas do not provide large sponta-neous emission enhancement, which is extremely impor-tant for emerging applications such as energy conversion,sensing and quantum information.

IV. BROADBAND NANOANTENNAS

Classical Yagi-Uda antennas are designed to operateat a certain frequency. In RF region, in their basic formYagi-Uda antennas are very narrowband, with typicalbandwidth of ≈ 5%. Using larger diameter antenna ele-ments, the bandwidth can be extended up to ≈ 10%.RF Yagi-Uda antennas can be designed to operate ef-

fectively at multiple frequencies. This can be achieved,for example, by creating electrical breaks along each ele-ment at which point a parallel LC (inductor and capaci-tor) circuit is inserted [11]. This circuit has the effect oftruncating the element at higher frequency band, makingit approximately a half wavelength in length. At lowerfrequencies, the entire element (including the inductanceof the LC circuit) is close to half-wave resonance, imple-menting a different Yagi-Uda antenna. The use of LCcircuits presents a very effective and low-cost solution,which is, however, not without disadvantages: the band-width and electrical efficiency of the antenna are reduced.

6

FIG. 6. (a) Schematic of tapered Yagi-Uda nanoantenna excited by an emitter via one of the possible excitation sites. Theinset shows the definition of the tapering angle α. (b) Scanning electron microscopy images of the fabricated nanoantennaswith different nanorod length variations. (c, d) Close-ups of the highlighted region of one of the nanoantennas [42].

Optical Yagi-Uda nanoantennas demonstrated so far[12–15] are single-frequency and narrowband. However,multifrequency broadband light control is essential forpractical applications and has recently become the sub-ject of intensive research [35, 36]. The aforementionedRF strategies for extending bandwidth of Yagi-Uda an-tennas cannot be applied at optical wavelengths due totwo main reasons: (i) the diameter of optical antenna el-ements is effectively larger that in RF due to fabricationconstraints [1] and (ii) feasible optical analogues of induc-tors and capacitors are not yet available at the nanoscale[37].Another example of a broadband RF antenna from

the Yagi-Uda family is a log-periodic antenna consist-ing of many dipole elements of decreasing length, all ofwhich can play the role of the feeding element [11]. Asthe dipoles are of different lengths, only one of them isresonant at a given frequency. The bandwidth of log-periodic antennas is relatively wide and is usually twotimes greater than that of Yagi-Uda antennas. Conse-quently, the log-periodic design is very attractive for theimplementation in the optical and infrared regions.

A. Tapered nanoantennas

Hereafter, we discuss emerging strategies for the devel-opment of optical Yagi-Uda nanoantennas able to receiveand emit light waves of different lengths with nearly thesame efficiency. As in the preceding sections, we will useand extend ideas known from RF antenna technology.In order to fully explore the potential of the Yagi-Uda

architecture one increases the length of the antenna byincrementing the number of directors. The additionaldirectors help to keep an antenna focused on what is in

front of it only. It increases the antenna efficiency whenthe antenna looks toward the transmitter and ignoressignals coming from other directions.While the use of this method does not present signifi-

cant difficulties at RF frequencies, at optical frequenciesabsorption losses impose severe restrictions on the lengthof the antenna. Luckily, this restriction can be overcomeby slowly tapering the length of the directors. The im-provement of the directivity of RF Yagi-Uda antennas byslowly varying the length of directors was suggested bySengupta [38] in 1959. His idea was abandoned in themodern RF antenna technique because the suggested ta-pering can be applied to long Yagi-Uda antennas only andresults in heavy and unstable constructions. Conversely,at optical frequencies tapering of the plasmonic waveg-uides [39] and metamaterials [40] is a very efficient wayfor nanofocusing of light. The effect of the taper shapeand the optimal taper angle on nanofocusing has beendiscussed actively in the literature (see, e.g., Ref. [41]).Recently, it has been demonstrated that tapering of a

plasmonic Yagi-Uda nanoantenna [see Fig. 6(a)] not onlyserves to focus light and to increase the overall efficiency,but also allows for shrinking down the nanoantenna lon-gitudinal dimension by decreasing the spacing betweenthe elements by a factor of 10. A pronounced improve-ment has been found for a particular configuration with42 nanorods and the tapering angle α = 6.6o (see theinset in Fig. 6(a) for the definition of α) at the principaldesign wavelength 1.515 µm [43–45].Figures 6(b–d) show experimental tapered Yagi-Uda

nanoantennas consisting of 21 nanorods with differentlength variations [42]. The nanoantennas were fabricatedon a glass substrate covered by 5 nm of indium tin ox-ide (ITO) using electron-beam lithography followed byevaporation of 50 nm of gold and a lift-off procedure (see

7

the discussion in the ”Fabrication and Characterization”section). With lateral nanorod diameters of ≈ 45 nmand inter-element spacings of ≈ 35 nm the dimensionsof the experimentally realized structures come very closeto theoretical design parameters [43–45], rendering themsuitable for optical near-field and far-field characteriza-tion.

Each nanorod can be used for the excitation of thetapered Yagi-Uda nanoantenna by quantum emittersmatched spectrally with the nanorod resonant frequencyand placed in the nanorod near-field region [44, 45].Multifrequency broadband operation of the nanoantennaresulting from an overlap of the spontaneous emissionspectra for different emitter positions [Fig. 7(a)] opensup novel opportunities for broadband emission enhance-ment, spectroscopy and sensing. A collective excitationwith narrow emission linewidth emitters converts thenanoantenna into an efficient multifrequency plasmon-enhanced light ”super-emitter” with a discrete spectrum[Fig. 7(b)]. Due to the reciprocity principle, the sametapered nanoantenna can be used both as a transmit-ter and/or as a receiver, thus being useful for creating abroadband wireless communication system.

B. Log-periodic nanoantennas

Tapered Yagi-Uda nanoantennas are conceptually sim-ilar to log-periodic optical nanoantennas inspired by RFlog-periodic architectures [11]. The original idea of op-tical log-periodic antennas probably belongs to Gonza-lez and Boreman [46] who used a lithographic antennato couple infrared radiation into a bolometer with sub-micron dimensions (Fig. 8). The excitation of an op-tical log-periodic antenna with quantum dots has re-cently been demonstrated experimentally [47]. In anal-ogy to RF antennas [11], different log-periodic architec-tures have been considered including arrays of nanorods,nanodipoles and zig-zags.

However, optical log-periodic antennas demonstratedso far just mimic the possible geometries of RF log-periodic antennas such as the length and the spacings ofthe elements. What has not yet been achieved in the op-tical range is the excitation of alternating elements with180o of phase shift from one another. In RF region, it isnormally done by connecting individual elements to al-ternating wires of a balanced transmission line in order toensure that the phasing of the elements is correct. Fur-thermore, as compared with tapered designs, the realiza-tion of optical log-periodic nanoantennas requires higherfabrication precision because both the length and thespacings of the elements shrink toward the nanoantennafront-end.

FIG. 7. (a) Spontaneous emission enhancement spectra of thetapered Yagi-Uda nanoantenna excited with an x -polarizedemitter via different excitation sites (see Fig. 6(a) for thedefinition of the polarization and excitation site numbers).The dashed line, a guide to the eye, connects the maxima ofthe spectra. (b) Spontaneous emission enhancement spectraof the nanoantenna excited simultaneously by 10 narrowbandx-polarized nanoemitters. Figure is taken from Ref. [44].

FIG. 8. Log-periodic optical antenna coupled to a mi-crobolometer. Reproduced from Ref. [46].

8

V. ALL-OPTICAL TUNABILITY

Modern RF antennas are highly optimized in terms oftheir bandwidth and size, and also can be tuned by meansof mechanical reconfiguration of antenna elements or bydriving antennas with electronically switchable circuits.In general, spectral tunability allows a smaller antenna tobehave as a larger antenna or as an array of antennas [11,14], both saving space and improving performance. Forexample, the arms length of today’s ”rabbit ears” digitalTV antennas can be changed in order to tune to thefavorite station and deliver a clear high-definition picture.These old-fashioned TV antennas, which appeared in ourhomes in the 1950’s, are very popular because they arecompact and can substitute several large roof antennas.Consequently, a large and growing body of re-

search investigates tunable optical nanoantennas [37, 48–54]. Many novel control mechanisms try to exploitthe concept of metamaterial-based [55] and non-Fosterimpedance matching circuits [56], where one of thepossible ways for achieving spectral tuning consists inthe use of tunable nanocapacitors and/or nanoinduc-tors [57, 58]. Other approaches rely on vanadium oxidetunable metamaterials [59], mechanically reconfigurablephotonic metamaterials [60] or metamaterials hybridizedwith carbon nanotubes [61]. Large spectral tunabilitycan also be obtained using electrically controlled liquidcrystals [62–64], but a very slow response of liquid crys-tals is not suitable for many application and, in general,a solid-state implementation is more suitable for on-chipintegration of nanoantennas.However, the spectral tuning of optical Yagi-Uda

nanoantennas has so far not been demonstrated becausetheir capability to tune to multiple operating frequen-cies is compromised by their ability to receive and trans-mit light in a preferential direction. Tunable Yagi-Udananoantennas could have technological applications inbuilding broadband optical wireless communication sys-tems, advanced nano-sensor systems, high performancesolar cells as well as wavelength tunable single photonsources and detectors.It has recently been suggested to exploit the free carrier

nonlinearity of semiconductors for a dynamical tuning ofthe operating wavelength of a plasmonic unidirectionalYagi-Uda nanoantenna consisting of plasmonic nanorodsused for the feeding element, reflector and directors [65].The feeding element of the nanoantenna differs from thatin previous designs [12] and consists of a semiconductornano-disk squeezed by two identical nanorods, as shownin Fig. 9(a). The illumination of the feeding elementwith a laser beam modifies the free carrier density inthe semiconductor and changes the conductivity of thenano-disk. This makes it possible to monotonically tunethe operating wavelength of the nanoantenna in a widespectral range as compared with conventionally designedYagi-Uda nanoantennas [12–15].Fig. 9(b) shows the power emitted by the tunable Yagi-

Uda nanoantenna in the maximum emission direction as

FIG. 9. (a) A tunable plasmonic Yagi-Uda nanoantenna con-sisting of plasmonic nanorods. The gray area of the feed corre-sponds to the semiconductor nano-disk used as loading. Thegreen arrow schematically shows the direction of the pumplaser beam. (b) Far-field power spectra of the nanoantennain the maximum emission direction as a function of the wave-length for the free carrier densities in the semiconductor load-ing from 0 cm−3 to 3 ·1021 cm−3 (from bottom curve and up).Red dashed line indicates the operating wavelength of a Yagi-Uda nanoantenna with the same geometry but equipped withan all-metal feeding element.

a function of the wavelength of the pump laser beamfor different free carrier densities N in the semiconduc-tor loading. We see that the increase in the free carrierdensity produces a monotonic tuning of the operatingwavelengths from 1.07 µm [denoted as A in Fig. 9(b)] to0.84 µm (denoted as A’).

Moreover, the power emitted by the nanoantenna canform closed bistability loops at different operating wave-lengths [65]. The formation of closed loops was observedearlier in photonic systems exhibiting nonlinear Fano-Feshbach resonances resulting from the interaction be-tween two Fano resonances located very close to eachother [66]. The closed bistable loops observed may haveappealing applications in realizing ultra-fast all-opticalswitching devices at the nanoscale since the nanoantennaexhibits two different stable states for the same appliedoptical intensity.

9

VI. FABRICATION AND

CHARACTERIZATION

In this section, we overview the techniques used to fab-ricate the Yagi-Uda nanoantennas discussed in the pre-ceding sections. We also survey the experimental tech-niques used to investigate the optical properties of thesenanoantennas in the near– and far–field zones.

A. Fabrication

Optical Yagi-Uda nanoantennas are commonly fabri-cated by standard electron-beam lithography (EBL) fol-lowed by metal evaporation and a liftoff procedure [12–14]. In this way fabrication accuracies below 50 nm canbe obtained. However, this method only allows for therealization of planar antenna geometries. In order to ob-tain stacked structures like the three-element Yagi-Udaantenna arrays shown in Fig. 4(b) [14], a multi-step EBLprocess has to be applied, where several functional lay-ers of metal nanostructures are processed on top of eachother. For the success of this method accurate planariza-tion and alignment procedures are crucial. In Ref. [14],as a first step, gold alignment markers with a thicknessof 250 nm are fabricated on a quartz substrate. Next, afirst planar 30-nm gold nanorod layer is processed usingthe standard fabrication procedure for in-plane nanos-tructures. Afterwards, a 100-nm thick spacer layer con-sisting of the solidifiable photo polymer PC403 (JCR) isspin-coated on top of this layer, followed by a pre- anda hard-bake. In the next step the alignment markers areused for accurately stacking a second nanorod layer ontop of the first one. Finally the planarization, alignment,and in-plane fabrication steps are repeated in order toobtain a third stacked layer.Similarly, a two step EBL process enables the con-

trolled placement of quantum emitters at defined posi-tions with respect to the nanoantenna as demonstrated inRef. [12]. In the first lithography step Yagi-Uda nanoan-tennas are defined on a glass substrate, followed by ther-mal evaporation of a 30-nm gold layer and lift-off. Next,a second lithography step is used to define 70-nm squareareas at the hot spots of the antenna feed elements forthe formation of a self-assembled monolayer of mercap-toundecanoic acid. Core-shell quantum dots can then beimmobilized on these functionalized areas, and in a laststep the unexposed resist is removed. The overall resultof this procedure is displayed in Fig. 3 (b).Other top-down techniques which have been employed

for nanoantenna fabrication include focused ion beam(FIB) milling [67], nanoimprint lithography [68], andnanostencil lithography [69]. While FIB stands out byits versatility to allow for high-precision structuring of al-most any conductive substrate, the latter two techniquesaim at large-area high-throughput fabrication [4].Linear plasmonic nanoantenna arrays, which closely

resemble the Yagi-Uda geometry, have furthermore been

obtained by the controlled assembly of chemically synthe-sized gold nanocubes using a nanomanipulator [70, 71].Similar approaches rely on nanomanipulation of plas-monic particles with an atomic-force microscope tip [72,73]. The growing interest in such bottom-up techniquesis justified by the purity, and well-defined crystallinityof chemically grown nanoparticles, by the wide range ofdifferent nanoparticle shapes and materials/compositesavailable, as well as by the good control of inter-particledistance [4].

B. Characterization

1. Classically designed Yagi-Uda nanoantennas

In analogy to RF antennas, optical nanoantennas canserve both as transmitters and receivers. Because of thereciprocity principle, the emitted radiation pattern of anantenna is equal to its receiving pattern. Experimentally,in RF region it is common to determine the directiveproperties of antennas by measuring the load current ofan antenna receiving radiation from a remote electromag-netic wave source [11]. As this technique cannot be usedin optical region, several near- and far-field spectroscopictechniques have been developed for the characterizationof optical Yagi-Uda nanoantennas [12–15].A straightforward way of measuring the optical prop-

erties of Yagi-Uda nanoantenna arrays is linear-opticalreflectance and transmittance spectroscopy [14]. Whilethe antenna array’s angular radiation pattern cannot beaccessed directly this way, the spectral positions of theplasmonic antenna resonances are identified and evidencefor directional behavior is given by differences betweenreflectance spectra in forward direction, i.e. for light in-cident from the director side, and backward direction, i.e.for light incident from the reflector side.A very elegant method of actually measuring a nanoan-

tenna’s angular radiation pattern is to characterize theantenna in emission mode by back-focal-plane imagingwithin a confocal microscope [12, 74]. To this end, emis-sion from the antenna into the substrate is collected witha high-NA objective, and the intensity distribution on theobjective’s back focal plane or Fourier-plane, which con-tains the directions of emission towards the substrate, isimaged on an electron-multiplying CCD camera.In order to characterize the nanoantenna’s directive

properties in emission mode it is crucial to implement amechanism ensuring that the antenna is only driven at itsfeeding element, while the reflector and director elementsare only excited in a passive manner through coupling tothe feed. In Ref. [13] this is accomplished by tilting thefeeding element by 45o, making it possible to couple toit with light polarized perpendicular to the emitted po-larization of the Yagi-Uda nanoantenna. An alternativeroute is to drive the feed via carefully aligned quantumdots. In Ref. [12] quantum dots are placed exclusivelyin the high-field regions of the feed element. They are

10

FIG. 10. (a) 3D topography of the optical Yagi-Uda nanoantenna. (b-d) Real part of the measured optical signal superimposedonto the topography at different incidents of time: (b) 0o phase, (c) 64o phase, and (d) 128o phase. Figure is reproduced fromRef. [15].

excited far off resonance with respect to other antennaelements, thereby again leading to a selective excitationof the feed element. Both of these techniques furthermoreoffer the advantage to allow for separating the fields emit-ted by the antenna from the exciting fields by means ofadditional polarization or color filters brought in front ofthe detector.

Yagi-Uda nanoantennas in receiving mode have justrecently been investigated experimentally using cross-polarization apertureless near-field optical microscopy[15]. The local out-of-plane electric field components ofthe sample are imaged by raster-scanning it underneaththe tip. Furthermore, the optical amplitude as well asthe optical phase are detected using a homodyne ampli-fication scheme, which allows for visualizing the temporalevolution and the dynamics of the reception process.

Figure 10(a) shows the 3D topography of the investi-gated optical Yagi-Uda nanoantenna fabricated on a glasssubstrate and illuminated by a weakly focused s-polarized(green arrow) laser beam (λ = 1064 nm) under an obliqueincident angle (grey arrow) [15]. Panels (b)-(d) in Fig.10 show the real part of the E -fields measured above thenanoantenna at different snapshots in time denoted as 0o

phase, 64o phase and 128o phase, respectively.

When the phase is of 0o [Fig. 10(b)], the two direc-tors show one positive and one negative field lobe witha field node in between them. The field element shows aweak amplitude, and the reflector shows a dipole point-ing in the opposite direction with respect to the directors.When the phase is of 64o [Fig. 10(c)], all nanoantennaelements show the same colour scheme corresponding to

the same dipole moment orientations. Notice that thefeed element is at its maximum field, brighter that all theother elements at any time. At 128o phase [Fig. 10(d)]the reflector reaches its maximum, but the feed is alreadydeclining and the dipole moments of the directors havealready flipped.

Thus, the directionality of optical Yagi-Uda nanoan-tennas manifests itself by the dependence of the localfield enhancement at the feeding element on the illumi-nation direction. This effect is similar to the increase inthe power measured in experiments with RF antennasreceiving signals from a test antenna [11].

Near-field optical measurement of a rotated (back-ward illuminated) nanoantenna with the same dimen-sions demonstrated that the change of the illuminationdirection leads to a destructive interference suppressingthe resonance of the feeding element. Therefore, the dif-ference between near-field images for forward and back-ward illumination of the same antenna structure is alsoa clear proof of the antenna directionality.

Another experimental technique offering a ∼ 10 nmspatial resolution using a 30 keV focused electron beamas broad-band point dipole source of visible radiation hasbeen recently presented and applied to study the emissionproperties of an optical Yagi-Uda antenna composed ofa linear array of gold nanoparticles [75]. Angle-resolvedspectra for different wavelengths shown in Ref. [75] revealevidence for directional emission of light that dependsstrongly on where the nanoantenna is excited. Theseexperimental results are explained by a simple and intu-itive coupled point dipole model, which includes the effect

11

of the dielectric substrate. In overall, the work estab-lishes angle-resolved cathodoluminescence spectroscopyas a powerful tool to characterize single optical nanoan-tennas.

2. Tapered and log-periodic nanoantennas

Simulations of tapered [43, 45] and log-periodic [47]nanoantennas demonstrated that they emit in the for-ward direction with nearly the same efficiency in al-most all possible excitation scenarios. It simplifies theirexperimental investigations as compared with Yagi-Udananoantennas with a single emitter attached to the feed-ing element only [12]. In Ref. [47], a large number ofquantum dots was randomly distributed around each log-periodic nanoantenna, leading to a uniform quantum dotdistribution with no clustering. These quantum dots ex-cite all nanoantenna elements in an active manner in con-trast to the only actively driven feeding element of theclassically designed Yagi-Uda [12].The optical properties of the log-periodic nanoanten-

nas working in emission mode were investigated by back-focal-plane imaging with a confocal microscope [12, 74].The investigation of the polarization of light emit-ted by log-periodic nanoantennas with different inter-element spacings revealed that samples with large spac-ings present the correct polarization transversal to thenanoantenna axis associated with the forward directionalemission. However, all nanoantennas were also foundto emit light of other polarizations, which differs themfrom classically designed Yagi-Uda nanoantennas excitedby a single nanoemitter and emitting light of only onepolarization [12]. Far-field radiation patterns of the in-vestigated log-periodic nanoantennas were obtained byprocessing confocal microscope images and a reasonableagreement between experiment and theory was found.Interestingly, nanoantennas with small inter-element

spacings showed no preferential polarization. This re-sult was explained by a random distribution of quantumdots. Firstly, some quantum dots may contribute to themeasured intensity without actually being coupled to thenanoantenna elements. It was confirmed by measuringsingle nanorods excited by randomly distributed quan-tum dots. Another reason for not observing directionalemission may lie in the enhancement of dipole mode per-pendicular to the nanoantenna axis or in higher ordermultipole mode contribution. Such a mode triggeringwas demonstrated previously both theoretically and ex-perimentally [76].

VII. APPLICATIONS

In this section we discuss the emerging applicationsof optical Yagi-Uda nanoantennas. We do not pretendto present an exhaustive review, but rather to touch onparticularly promising selected topics.

FIG. 11. Two optical Yagi-Uda nanoantennas can be config-ured as an on-chip optical interconnect.

A. Wireless optical communication

The objective of optical antenna design is equivalentto that of classical antenna design: to optimize the en-ergy transfer between a localized source or receiver andthe free-radiation field [2]. In telecommunication, theantenna concept is combined: an antenna can be usedboth as a receiver and as a transmitter. By utilizingtwo identical RF antennas, the first one operating as atransmitter and the second one as a receiver, one cancreate an RF wireless communication system. The sameconcept can also be used in optics where an on-chip inte-grated optical wireless system, suggested theoretically inRef. [77], might provide an alternative to subwavelengthplasmonic waveguides [6], which suffer from absorptionlosses during light transmission. In analogy to waveg-uides, nanoantennas confine light below the diffractionlimit but, in contrast to waveguides, they transmit itthrough a channel with relatively low losses.

As the transmitter and the receiver of a wireless sys-tem should both have a high directivity, one can envisionan optical interconnect consisting of two (or more) opti-cal Yagi-Uda nanoantennas looking one toward another[78] (Fig. 11). A quantum dot can be placed in the nearfield of the nanoantenna-transmitter so that it drives itsresonant feeding element. The resulting quantum-dot lu-minescence will be strongly polarized and highly directedinto a narrow forward angular cone, and then received bythe nanoantenna-receiver. By placing another nanoemit-ter near the resonant feed of the nanoantenna-receiver,one can communicate energy to, from, and between na-noemitters. Using a tapered Yagi-Uda [43, 45] or a log-periodic [47] nanoantenna one can transmit energy to,from and between many nanoemitters at different fre-quencies simultaneously by precisely placing them nearindividual nanoantenna elements.

Furthermore, tunable Yagi-Uda nanoantennas with asemiconductor loading of the feeding element [65] canperform as active components of optical wireless com-munication systems. They can perform not only as fre-quency selective transmitters or receivers, but also asan all-optical bistable switching device [79] exhibitingemission properties not attainable with actively tunabledipole or bow-tie plasmonic nanoantennas [59, 80, 81].

12

B. Nanoscale spectroscopy

Nanoscale field confinement by optical nanoantennasoffers remarkable opportunities for single molecule de-tection [82–84] and nanospectroscopy [85]. Yagi-Udananoantennas can perform as an optical spectrum an-alyzer at the nanometer scale that is able to distributedifferent frequency contents of the radiation of an opticaldipole source into different directions in space [86]. Thefeasibility of this concept has been demonstrated theo-retically based on optical Yagi-Uda nanoantennas com-posed of plasmonic core-shell nanoparticles. Core-shellplasmonic nanoparticles with the core made of an ordi-nary dielectric and the shell of a plasmonic material, orvice versa, are ideal for this purpose because their res-onant frequencies can be adjusted in a broad range byvarying the ratio of the core radius b to the outer ra-dius a. Specifically, b/a of each particle can be designeddeliberately to ”detune” its resonance so that they mayplay the role of the ”reflectors” or the ”directors” in aconventional RF Yagi-Uda antenna.The beam pattern of a core-shell Yagi-Uda nanoan-

tenna is sensitive to the variation of the operating wave-length for two reasons. Firstly, the relative permittivitiesof the plasmonic materials forming the nanoparticles, andthe resulting electric polarizability of these particles, arewavelength dependent. Secondly, when the sizes of thenanoparticles and their relative positions are decided andthen kept fixed, different operating wavelengths wouldlead to different relative electrical sizes and relative dis-tances, which causes variation in coupling among parti-cles. Therefore, optical nanoantennas composed of plas-monic particles are intrinsically sensitive to wavelengthvariation. Using this unique property, one can envisionan innovative nanoscale ”Yagi-Uda” spectrum analyzerin which the absorption and emission of a fluorescentmolecule can be separated away both spectrally and spa-tially.

C. Photovoltaic devices

Photovoltaics, a promising technology relying on theconversion of sunlight to electricity, may allow for thegeneration of electrical power on a very large scale. Itcould thus make a considerable contribution to solvingthe energy problem faced by our society. Plasmonic exci-tation and light localization can be used advantageouslyin high-efficient photovoltaics [87].Conventionally, photovoltaic absorbers are optically

thick in order to allow better light absorption and pho-tocarrier current collection. However, a large portion ofthe solar spectrum, especially in the 600−1100 nm spec-tral range, is poorly absorbed (Fig. 12). The use of plas-monic nanoantennas of subwavelength dimensions can re-duce the physical thickness of photovoltaic absorber lay-ers while keeping their optical thickness constant. Thesenanoantenna-receivers can trap and couple freely prop-

FIG. 12. AM1.5 solar spectrum, together with a graph thatindicates the solar energy absorbed in a 2-µm-thick crystallineSi film. Figure is reproduced from Ref. [87].

agating sunlight into an absorbing semiconductor thinfilm, increasing the effective absorption cross-section.Broadband tapered Yagi-Uda nanoantennas extend

the list of advantages of plasmonic nanoantennas forphotovolatics because they can be designed to absorbsunlight in a very broad spectral range. In order toachieve this specific goal, the number of nanorods and thenanorod length variation can be chosen [44, 45] such thatthe antenna spectrum covers the AM1.5 solar spectrumincluding the ”problematic” spectral range of 600− 1100nm. Since the reception characteristics of the nanoan-tenna are nearly constant at all operating wavelengths, itis expected that all fractions of the solar spectrum wouldbe absorbed almost equally.

D. Nanosensing

The scientific advances of recent years have recognizedoptical nanoantennas as a technology that could makepossible a range of innovations. One of the importantareas where the nanoantennas have immediate implica-tions are plasmon mediated chemistry [88] and sensing[89–94].Tapered Yagi-Uda nanoantennas can develop signifi-

cantly the idea of the nanoantenna-enhanced gas sensing[94]. In order to achieve this goal, single nanoparticlesmade of gas-sensitive active materials such as platinum(Pt), palladium (Pd) or iridium (Ir) can be preciselyplaced in the near-field region of individual nanorods ofnanoantennas as shown in Fig. 13 [95]. Alternatively, Pt,Pd or Ir can be deposited on the top of the nanoantennaelements. Yet another configuration is a tapered array ofelements made of different metals.A change in the optical properties of the active

nanoparticles will influence the resonances of the individ-ual nanorods, which in turn inevitably changes the far-

13

FIG. 13. (a) Schematic of a broadband plasmonic array Yagi-Uda nanoantenna consisting of metal nanorods of varyinglength. Active nanoparticles are placed in close vicinity ofthe nanorods. (b) Optical sensing principle using the nanoan-tenna interacting with only one active palladium nanoparticleplaced at the excitation site n = 1. A change in the opticalproperties of the nanoparticle on exposure to hydrogen (Pdto PdH) causes a resonance shift (∆λ) in the optical responseof the nanoantenna.

field characteristics of the nanoantenna at different op-erating wavelengths. Thanks to a high sensitivity of theYagi-Uda architecture, the expected spectral shift willbe larger as compared with a several nanometers shiftobserved with single bow-tie nanoantennas [94, 96]. Fur-thermore, as shown in Fig. 7(b), the separation betweenthe peaks in the nanoantenna spectra is large enough toallow for easy reading of data from different nanoparti-cles.

Any change in the far-field response of the nanoan-tenna can be detected using conventional far-field spec-troscopy techniques [2] or with another broadband Yagi-Uda nanoantenna employed as a receiver. The latterconfiguration is attractive for application to horizontalon-chip plasmonic wireless signal-transmission circuits,which can be used for real-time transmission of datamonitored by the nanosensor. Since the data on differ-ent monitored chemical substances can be transmittedsimultaneously at different wavelengths, the suggestednanoantenna-based sensors can be used in the develop-ment of nanoscale optical nose devices that can be tunedboth geometrically and dynamically to monitor differentgases at different operating wavelengths.

E. Quantum communication and information

Efficient sources of single and entangled photons arecrucial for the development of quantum information tech-nology [22]. Single semiconductor quantum dots canemit single photons and degenerate polarization entan-gled photon pairs consisting of photons with the hori-zontal and vertical polarizations [18]. In order to collectthe emitted photons, the quantum dot spontaneous emis-sion can be controlled and funnelled in an optical cavitymode. When a quantum dot is spatially and spectrally

coupled to a cavity mode, its emission rate into the modeis increased by the Purcell factor FP [20] and a fractionFP

FP+1of the quantum dot emission is funnelled into the

mode, which opens up a possibility of a high collectionefficiency of the emitted photons.Yagi-Uda nanoantennas can be used for the emission

of single photons with a high efficiency as compared withoptical cavities because they provide a large and broad-band Purcell factor accompanied by a high directivity[12]. Moreover, using broadband tapered [43, 45] andlog-periodic [47] nanoantennas one can generate singlephotons at different frequencies at the same time. How-ever, the implementation of Yagi-Uda nanoantennas tothe extraction of polarization entangled photon pairs ischallenging because of two different spontaneous emissionenhancement rates for the different photon polarizationsthat eventually erases the entanglement. One of the pos-sible solutions to this problem can be found by fabricat-ing 3D Yagi–Uda nanoantennas one in close proximityto another. This arrangement also known as photonicmolecule [97] can offer identical emission conditions fordifferent photon polarizations.

F. Optical metamaterials

In the previous sections we have demonstrated that ar-rays of plasmonic particles forming Yagi-Uda nanoanten-nas are able to precisely manipulate light in the same wayas arrays of metal bars manipulate radio waves. Nowa-days, a large and growing body of research investigatesnovel artificial materials – metamaterials [98] able to ma-nipulate light in new ways not existing in nature. Meta-materials could make possible a range of optical innova-tions such as, e.g., more powerful microscopes, telecom-munications and computers.One of the approaches for creating metamaterials is to

abruptly change the phase of light, which dramaticallymodifies how light propagates. However, this conceptis not new. Indeed, the idea of controlling the phasewas used by Uda and Yagi in their famous antenna 85years ago. We remind that the reflector of a Yagi-Udaantenna is longer than its resonant length, its impedanceis inductive and thus the current on the reflector lags thevoltage induced on the reflector. The director elementsare shorter than the resonant length, which makes themcapacitive, so that the current leads the voltage. Thiscauses a phase distribution to occur across the elements,simulating the phase progression of a plane wave acrossthe array of elements.However, Uda and Yagi could not know that a long-

held Snell’s law, used to describe how light reflects andrefracts while passing from one material into another,would be modified 85 years after their invention. It hasbeen recently shown in Ref. [99], that the phase of lightand the propagation direction can be changed dramat-ically by using metamaterials based on an array of V-shaped plasmonic nanoparticles. What was pointed out

14

FIG. 14. (a) Schematic view of the metamaterial nanoantenna array. The unit cell of the plasmonic interface (in blue) consistsof eight gold V-shaped nanoparticles of 40-nm arm width and 30-nm thickness, repeated with a periodicity of Λ in the xdirection and λ/8 in the y direction. The phase delay for cross–polarized light increases along the x axis from right to left.(b) Scanning electron microscope images of the unit cells of four antenna arrays with different periodicities fabricated on asingle silicon wafer. (c, d) The false-color maps indicate the experimentally measured relative intensity profiles for the antennaarray labeled (1) (see panel (b), Λ = 1440 nm) with x -polarized excitation. The dashed line shows the theoretical predictionof the peak position using the generalized Snell’s law. (c) Refraction angle versus wavelength for cross-polarized light with 30o

incidence angle. (d) Reflection angle versus wavelength for cross-polarized light with 65o incidence angle. Figure is reproducedfrom Ref. [100].

in Ref. [99] and demonstrated at one wavelength is revo-lutionary. Very recently, this idea has been extended tothe near infrared, which is essential for telecommunica-tions, and more importantly applied to a broad spectralrange from 1 to 1.9 µm [100]. The resulting nanoantennaarray [see Fig. 14(a,b)] offers an unparalleled controlof anomalous reflection and refraction, including nega-tive refraction [Fig. 14(c,d)]. This could lead to a va-riety of applications such as spatial phase modulation,beam shaping, beam steering, and plasmonic lenses, andit could also have an impact on transformation optics andon-chip optics.

VIII. CONCLUSION AND PERSPECTIVES

We have reviewed the recent advances in the studyof unidirectional optical nanoantennas, in particular, theoptical analogue of the famous Yagi-Uda antenna ar-chitecture. Owing to a great potential of this type ofnanoantennas, evidenced by a rapid growth of researchactivity in this direction the presented review is notmeant to be exhaustive, but it gives rather a bird’s-eyeview of optical Yagi-Uda nanoantennas and some of theirapplications.

Suggested over 85 years ago, Yagi-Uda antennas haveplayed a crucial role in the establishment of radar tech-nology, radio communication, television broadcasting,and ham radio. Nowadays, optical analogues of theYagi-Uda architecture open up unique perspectives andgreat potentials to explore new avenues in the fieldsof active nanophotonic circuits, quantum informationtechnology, high-density data storage, optical andbio-sensing, medical imaging systems, photovoltaics andphotodetection.

ACKNOWLEDGMENTS

This work was supported by the Australian ResearchCouncil. The fabrication facilities were supported bythe Australian National Fabrication Facility (ACT node).The authors thank A. Davoyan, C. Simovski, A. Krasnok,P. Belov, M. Decker, D. Neshev, H.H. Tan and C. Ja-gadish for valuable contributions to the work on taperedand all-dielectric nanoantennas. The authors acknowl-edge many useful discussions with their colleagues fromthe Nonlinear Physics Centre and Metamaterial MeetingGroup at the Australian National University.

15

[1] Bharadwaj P, Deutsch B, Novotny L. Optical antennas.Adv Opt Photon 2009, 1, 438–483.

[2] Novotny L, van Hulst NF. Antennas for light. Nat Pho-ton 2011, 5, 83–90.

[3] Giannini V, Fernandez-Domınguez AI, Heck SC, MaierSA. Plasmonic nanoantennas: Fundamentals and theiruse in controlling the radiative properties of nanoemit-ters. Chem Rev 2011, 111, 3888–3912.

[4] Biagoni P, Huang JS, Hecht B. Nanoantennas for vis-ible and infrared radiation. Rep Prog Phys (in press),arXiv:1103.1568v2.

[5] Schuck PJ, Fromm DP, Sundaramurthy A, Kino GS,Moerner WE. Improving the mismatch between lightand nanoscale objects with gold bow-tie nanoantennas.Phys Rev Lett 2005, 94, 017402–4.

[6] Gramotnev DK, Bozhevolnyi SI. Plasmonics beyond thediffraction limit. Nat Photon 2010, 4, 83–91.

[7] This phrase is often credited to Marie-JeanneRose Bertin, the first celebrated fashion designerand dressmaker to Queen Marie Antoinette (athttp://en.wikipedia.org/wiki/Rose_Bertin).

[8] G. Lerosey. Yagi-Uda antenna shines bright. Nat Photon2010, 4, 1–2.

[9] Uda S. High angle radiation of short electric waves. ProsIRE 1927, 15, 377–385.

[10] Sato G. A secret story about the Yagi antenna. IEEEAnt Propagat Mag 1991, 33, 7–18.

[11] Balanis CA. Antenna theory: analysis and design. NewJersey, USA, Wiley Press, 2005.

[12] Curto AG, Volpe G, Taminiau TH, Kreuzer MP,Quidant R, van Hulst NF. Unidirectional emission ofa quantum dot coupled to a nanoantenna. Science 2010,329, 930–933.

[13] Kosako T, Kadoya Y, Hofmann HF. Directional con-trol of light by a nano-optical Yagi-Uda antenna. NatPhotonics 2010, 4, 312–315

[14] Dregely D, Taubert R, Dorfmuller J, Vogelgesang R,Kern K, Giessen H. 3D optical Yagi–Uda nanoantennaarray. Nat Commun 2011, 2, 267–7.

[15] Dorfmuller J, Dregely D, Esslinger M, Khunsin W, Vo-gelgesang R, Kern K, Giessen H. Near-field dynamicsof optical Yagi-Uda nanoantennas. Nano Lett 2011, 11,2819–2824

[16] Novotny L. Effective wavelength scaling for optical an-tennas. Phys Rev Lett 2007, 98, 266802–4.

[17] Hofmann HF, Kozako T, Kadoya Y. Design parametersfor a nano-optical Yagi-Uda antenna. New J Phys 2007,9, 217–12.

[18] Shields AJ. Semiconductor quantum light sources. Na-ture Photon 2007, 1, 215–223.

[19] Castelletto S, Aharonovich I, Gibson BC, Johnson BC,Prawer S. Imaging and quantum-efficiency measure-ment of chromium emitters in diamond. Phys Rev Lett2010, 105, 217403–4.

[20] Vahala KJ. Optical microcavities. Nature 2003, 424,839–846.

[21] Koenderink AF. Plasmon nanoparticle array waveguidesfor single photon and single plasmon sources. Nano Lett2009, 9, 4228–4233

[22] Bouwmeester D, Ekert AK, Zeilinger A. The physics ofquantum information. Berlin, Germany, Springer, 2000.

[23] Pakizeh T, Kall M. Unidirectional ultracompact opticalnanoantennas. Nano Lett 2009, 9, 2343-2349.

[24] Devilez A, Stout B, Bonod N. Compact metallo-dielectric optical antenna for ultra-directional and en-hanced radiative emission. ACS Nano 2010, 4, 3390-3396.

[25] Bonod N, Devilez A, Rolly B, Bidault S, Stout B. Ul-tracompact and unidirectional metallic antennas. PhysRev B 2010, 82, 115429–6.

[26] Stout B, Devilez A, Rolly B, Bonod N. Multipole meth-ods for nano-antennas design: applications to Yagi–Udaconfigurations. J Opt Soc Am B 2011, 28, 1213–1223.

[27] Esteban R, Teperik TV, Greffet JJ. Optical Patch An-tennas for Single Photon Emission Using Surface Plas-mon Resonances. Phys Rev Lett 2010, 104, 026802–4.

[28] Pellegrini G, Mattei G, Mazzoldi P. Tunable, directionaland wavelength selective plasmonic nanoantenna arrays.Nanotechnology 2009,20,065201–5.

[29] Pellegrini G, Mattei G, Mazzoldi P. Nanoantenna arrayfor large–area emission enhancement. J Phys Chem C2011, 115, 24662–24665.

[30] Krasnok AE, Miroshnichenko AE, Belov PA, KivsharYuS. Huygens optical elements and Yagi–Uda nanoan-tennas based on dielectric nanoparticles. JETP Letters2011, 94, 593–598.

[31] Kishk AA. Directive Yagi–Uda dielectric resonator an-tennas. Microwave Opt Techn Lett 2005, 44, 451–453.

[32] Palik ED. Handbook of Optical Constants of Solids.New York, USA, Academic Press, 1985.

[33] Krasnok AE, Miroshnichenko AE, Belov PA, KivsharYuS. All-Dielectric Optical Nanoantennas. submitted

[34] Evlyukhin AB, Reinhardt C, Seidel A, LukyanchukBS, Chichkov BN. Optical response features of Si-nanoparticle arrays. Phys Rev B 2010, 82, 045404–12.

[35] Simovski C, Luukkonen O. Tapered plasmonic waveg-uides with efficient and broadband light transmission.Opt Commun 2012 (in press).

[36] Boriskina SV, Dal Negro L. Multiple–wavelength plas-monic nanoantennas. Opt Lett 2010, 35, 538–440.

[37] Alu A, Engheta N. Tuning the scattering response of op-tical nanoantennas with nanocircuit loads. Nat Photon2008, 2, 307–310.

[38] Sengupta DL. On uniform and linearly tapered longYagi antennas. IRE Trans Antennas and Propagat 1960,8, 11–17.

[39] Stockman MI. Nanofocusing of optical energy in taperedplasmonic waveguides. Phys Rev Lett 2004, 93, 137404–4.

[40] Rockstuhl C, Simovski CR, Tretyakov SA, LedererF. Metamaterial nanotips. Appl Phys Lett 2009, 94,113110–3.

[41] Davoyan AR, Shadrirov IV, Kivshar YuS, GramotnevDK. Optimal tapers for compensating losses in plas-monic waveguides. Phys Status Solidi RRL 2010, 4, 277–279.

[42] Maksymov IS, Staude I, Miroshnichenko AE, DeckerM, Tan HH, Neshev DN, Jagadish C, Kivshar YuS.Arrayed nanoantennas for efficient broadband unidirec-tional emission enhancement. Conference on Lasers andElectro-Optics (CLEO)/USA, under review.

16

[43] Maksymov IS, Davoyan AR, Kivshar YuS. Enhancedemission and light control with tapered plasmonicnanoantennas. Appl Phys Lett 2011, 99, 083304–3.

[44] Maksymov IS, Davoyan AR, Miroshnichenko AE,Simovski C, Belov P, Kivshar YuS. Multifrequency ta-pered plasmonic nanoantennas. Opt Commun 2012,285, 821–824.

[45] Miroshnichenko AE, Maksymov IS, Davoyan AR,Simovski C, Belov P, Kivshar YuS. An arrayed nanoan-tenna for broadband light emission and detection. PhysStatus Solidi RRL 2011, 9, 347-349.

[46] Gonzalez F, Boreman GD. Comparison of dipole,bowtie, spiral and log-periodic IR antennas. InfraredPhys Tech 2005, 46, 418–428.

[47] Pavlov RS. Log-periodic optical antennas. MSc Thesis,Universitat Politecnica de Catalunya, Barcelona, Spain,2005.

[48] Abb M, Albella P, Aizpurua J, Muskens OL. All-opticalcontrol of a single plasmonic nanoantenna – ITO hybrid.Nano Lett 2011, 11, 2457–2463.

[49] Shegai T, Chen S, Milkovic V, Zengin G, Johansson P,Kall M. A bimetallic nanoantenna for directional colourrouting. Nat Commun 2011, 2, 481–6.

[50] Maksymov IS, Miroshnichenko AE. Active control overnanofocusing with nanorod plasmonic antennas. OptExpress 2011, 19, 5888–5894.

[51] Alaverdyan Y, Vamivakas N, Barnes J, LebouteillerC, Hare J, Atature M. Spectral tunability of a plas-monic antenna with a dielectric nanocrystal. Opt Ex-press 2011, 19, 18175–18181.

[52] Large N, Abb M, Aizpurua J, Muskens OL. Photo-conductively loaded plasmonic nanoantenna as build-ing block for ultracompact optical switches. Nano Lett2010, 10, 1741–1746.

[53] Farahani JN, Pohl DW, Eisler HJ, Hecht B. Single quan-tum dot coupled to a scanning optical antenna: A tun-able superemitter. Phys Rev Lett 2005, 95, 017402–4.

[54] Huang F, Baumberg JJ. Actively tuned plasmons onelastometrically driven Au nanoparticle dimers. NanoLett 2010, 10, 1787–1792.

[55] Ziolkowski RW, Erentok A. Metamaterial–based effi-cient electrically small antennas. IEEE Trans AntennasPropag 2008, 54, 2113–2130.

[56] Sussman–Fort SE, Rudish RM. Non-foster impedancematching of electrically-small antennas. IEEE TransAntennas Propag 2009, 57, 2230–2241.

[57] Engheta N, Salandrino A, Alu A. Circuit elements at op-tical frequencies: Nanoinductors, nanocapacitors, andnanoresistors. Phys Rev Lett 2005, 95, 095504–4.

[58] Engheta N. Circuits with light at nanoscales: Opticalnanocircuits inspired by metamaterials. Science 2007,317, 1698–1702.

[59] Seo M, Kyoung J, Park H, Koo S, Kim H-S, Bernien H,Kim BJ, Choe JH, Ahn YH, Kim H-T, Park N, ParkQ-H, Ahn K, Kim D-S. Active terahertz nanoantennasbased on VO2 phase transition. Nano Lett 2010, 10,2064–2068.

[60] Ou JY, Plum E, Jiang L, Zheludev NI. Reconfigurablephotonic metamaterials. Nano Lett 2011, 11, 2142–2144.

[61] Nikolaenko AE, de Angelis F, Boden SA, PapasimakisN, Ashburn P, di Fabrizio E, Zheludev NI. Carbon nan-otubes in a photonic metamaterial. Phys Rev Lett 2010,104, 153902–4.

[62] Kossyrev PA, Yin A, Cloutier SG, Cardimona DA,Huang D, Asling PM, Xu JM. Electric field tuning ofplasmonic response of nanodot array in liquid crystalmatrix. Nano Lett 2005, 5, 1978–1981.

[63] Berthelot J, Bouhelier A, Huang C, Margueritat J,Colas-des-Francs G, Finot E, Weeber J–C, Dereux A,Kostcheev S, Ahrach H. Ibn El, Baudrion A-L, PlainJ, Bachelot R, Royer P, Wiederrecht G P. Tuning of anoptical dimer nanoantenna by electrically controlling itsload impedance. Nano Lett 2009, 9, 3914–3921.

[64] De Angelis C, Locatelli A, Modotto D, Boscolo S, MidrioM, Capobianco AD. Frequency addressing of nano-objects by electrical tuning of optical antennas. J OptSoc Am B 2011, 27, 997–1001.

[65] Maksymov IS, Miroshnicheko AE, Kivshar YuS. Ac-tively tunable optical Yagi-Uda nanoantenna withbistable emission characteristics. arXiv 1202.252v1[physics.optics].

[66] Miroshnichenko AE, Flach S, Kivshar YuS. Rev ModPhysics 2010, 82, 2257–2298.

[67] Huang J-S, Kern J, Geisler P, Weinmann P, Kamp M,Forchel A, Biagioni P, Hecht B. Mode imaging and selec-tion in strongly coupled nanoantennas. Nano Lett 2010,10, 2105–2110.

[68] Boltasseva A. Plasmonic components fabrication viananoimprint. J Opt A: Pure Appl Opt 2009, 11, 114001–11, 2009.

[69] Aksu S, Yanik AA, Adato R, Artar A, Huang M, AltugH. High-Throughput Nanofabrication of Infrared Plas-monic Nanoantenna Arrays for Vibrational Nanospec-troscopy. Nano Lett 2010, 10, 2511–2518.

[70] Chen H-Y, He C-L, Wang C-Y, Lin M-H, Mitsui D,Eguchi M, Teranishi T, Gwo S. Far-Field Optical Imag-ing of a Linear Array of Coupled Gold Nanocubes: Di-rect Visualization of Dark Plasmon Propagating Modes.ACS Nano 2011, 5, 8223–8229.

[71] Yang S-C, Kobori H, He C-L, Lin M-H, Chen H-Y, Li C,Kanehara M, Teranishi T, Gwo S. Plasmon Hybridiza-tion in Individual Gold Nanocrystal Dimers: Direct Ob-servation of Bright and Dark Modes. Nano Lett 2010,10, 632–637.

[72] Merlein J, Kahl M, Zuschlag A, Sell A, HalmA, Boneberg J, Leiderer P, Leitenstorfer A, andBratchitsch R. Nanomechanical control of an optical an-tenna. Nature Photon 2008, 2, 230–233.

[73] Schietinger S, Barth M, Aichele T, and BensonO. Plasmon-enhanced single photon emission from ananoassembled metal-diamond hybrid structure at roomtemperature. Nano Lett 2009, 9, 1694–1698.

[74] Lieb MA, Zavislan JM, Novotny L. Single–molecule ori-entations determined by direct emission pattern imag-ing. J Opt Soc Am B 2004, 21, 1210–1215.

[75] Coenen T, Vesseur EJR, Polman A, Koenderink AF. Di-rectional emission from plasmonic Yagi-Uda antennasprobed by angle-resolved cathodoluminescence spec-troscopy. Nano Lett 2011, 11, 3779-3784.

[76] Chu MW, Myroshnychenko V, Chen CH, Deng JP,Mou CY, Garcıa de Abajo FJ. Probing bright and darksurface-plasmon modes in individual and coupled noblemetal nanoparticles using an electron beam. Nano Lett2009, 9, 399-404.

[77] Alu A, Engheta N. Wireless at the Nanoscale: Opticalinterconnects using matched nanoantennas. Phys RevLett 2010, 104, 213902–4.

17

[78] Novotny L. From near–field optics to optical antennas.Physics Today 2011, 64, 47–52.

[79] Gibbs HM. Optical bistability: Controlling light withlight. Orlando, USA, Academic Press, 1985.

[80] Chen PY, Alu A. Optical nanoantenna arrays loadedwith nonlinear materials, Phys Rev B 2010, 82, 235405–10.

[81] Zhou F, Liu Y, Li Z–Y, Xia Y. Analytical model foroptical bistability in nonlinear metal nano-antennae in-volving Kerr materials. Opt Express 2010, 13, 13337–13344.

[82] Taminiau TH, Stefani FD, Segerink FB, Van Hulst NF.Optical antennas direct single-molecule emission. NatPhoton 2008, 2, 234–237.

[83] De Angelis F, Patrini M, Das G, Maksymov IS, Galli M,Businaro L, Andreani LC, and di Fabrizio E. Nanoscalechemical mapping using three-dimensional adiabaticcompression of surface plasmon polaritons. Nano Lett2008, 8, 2321–2327.

[84] De Angelis F, Das G, Candeloro P, Patrini M, GalliM, Bek A, Lazzarino M, Maksymov IS, Liberale C, An-dreani LC, di Fabrizio E. Nanoscale chemical mappingusing three-dimensional adiabatic compression of sur-face plasmon polaritons. Nat Nanotechnol 2010, 5, 67–72.

[85] Schumacher T, Kratzer K, Molnar D, Hentschel M,Giessen H, Lippitz M. Nanoantenna-enhanced ultrafastnonlinear spectroscopy of a single gold nanoparticle. NatCommun 2011, 2, 333–5.

[86] Li J, Salandrino A, Engheta N. Optical spectrometerat the nanoscale using optical Yagi-Uda nanoantennas.Phys Rev B 2009, 79, 195104–5.

[87] Atwater HA, Polman A. Plasmonics for improved pho-tovoltaic devices, Nat Mater 2010, 9, 205–213.

[88] Chen CJ, Osgood RM. Direct observation of the local–field–enhanced surface photochemical–reactions. PhysRev Lett 1983, 50, 1705–1708.

[89] Zhang WH, Huang LN, Santschi C, Martin OJF. Trap-ping and sensing 10 nm metal nanoparticles using plas-monic dipole antennas. Nano Lett 2010, 10, 1006–1011.

[90] Aimovi SS, Kreuzer MP, Conzalez MU, Quidant R.Plasmon near–field coupling in metal dimers as a steptoward single–molecule sensing. ACS Nano 2009, 3,1231–1237.

[91] Adato R, Yanik AA, Amsden JJ, Kaplan DL, OmenettoFG, Hong MK, Erramilli S, Altug H. Ultra-sensitive vi-brational spectroscopy of protein monolayers with plas-monic nanoantenna arrays. Proc Natl Acad Sci USA2009, 106, 19227–19232.

[92] Hao F, Nordlander P, Sonnefraud Y, Van Dorpe P,Maier SA. Tunability of subradiant dipolar and Fano-type plasmon resonances in metallic ring/disk cavities:Implications for nanoscale optical sensing. ACS Nano2009, 3, 643–652.

[93] Khalavka Y, Becker J, Sonnichsen. Synthesis of rod–shaped gold nanorattles with improved plasmon sensi-tivity and catalytic activity. J Am Chem Soc 2009, 131,1871–1875.

[94] Liu N, Tang ML, Hentschel M, Giessen H, AlivisatosAP. Nanoantenna-enhanced gas sensing in a single tai-lored nanofocus. Nat Mater 2011, 10, 631-636.

[95] Maksymov IS, Miroshnichenko AE, Kivshar YuS. Tun-able Yagi-Uda-type plasmonic nanoantennas: implica-tions for nanoscale optical sensing. Proc SPIE 2012,8351, 835107-7.

[96] Tittl A, Kremers C, Dorfmuller J, Chigrin DN, GiessenH. Spectral shifts in optical nanoantenna-enhanced hy-drogen sensors. Opt Mater Express 2012, 2, 111–118.

[97] Dousse A, Suffczynski J, Beveratos A, Krebs O,Lemaıtre A, Sagnes I, Bloch J, Voisin P, Senellart P.Ultrabright source of entangled photon pairs. Nature2010, 466, 217–220.

[98] Veselago VG. The electrodynamics of substances withsimultaneously negative values of ǫ and µ. Soviet PhysicsUspekhi 1968, 10, 509–514.

[99] Yu N, Genevet P, Kats MA, Aieta F, Tetienne JP, Ca-passo F, Gaburro Z. Light propagation with phase dis-continuities: Generalized laws of reflection and refrac-tion. Science 2011, 334, 333–337.

[100] Ni X, Emani NK, Kildishev AV, Boltasseva A, ShalaevVM. Broadband light bending with plasmonic nanoan-tennas. Science 2012, 335, 427.