Complex Photonic Structures for Light Harvesting.

www.advopticalmat.de

REVIEW

www.MaterialsViews.com

Complex Photonic Structures for Light Harvesting Matteo Burresi, Filippo Pratesi, Francesco Riboli, and Diederik Sybolt Wiersma* sources covering less than 0.1% of the global energy needs. A worldwide adaptation to solar energy sources could be favored by improving solar cell efficiency and hence lowering the material needs and costs. Photonic techniques can play an important role by increasing the capabilities of a material to trap light, thus facilitating absorption. Ideally, one wants to maximize the coupling of light into a structure while minimizing the amount of light that escapes. In an absorption-free medium, this requirement is in contrast with reciprocity in Maxwell’s equations, but it is allowed in a material with nonnegligible absorption. In the ideal situation, light should easily enter, but then remain as long as possible inside so that efficient absorption and conversion from light to electricity can take place. In the field of solar cell applications, a planar geometry is usually employed, where the system is much wider than it is thick. This geometry is dictated by the need of intercepting as many photons as possible while keeping the building material to a minimum. The reason for the latter is not only cost-effectiveness. In general, photogenerated charge carriers drift once they are in proximity of the electrostatic field of the p–n junction; however, when they are far from it they diffuse for a particular length—the so-called minority carrier diffusion length— after which they recombine with holes, dissipating their kinetic energy as heat. Hence, the ultimate goal is to make the material as thin as possible while still maintaining good absorption characteristics: indeed when the film is too thin, part of the light is transmitted and, thus, lost. The role of photonic structures is then to make the material optically thick, by coupling light into the so-called guided modes of the thin film, to ‘store’ it for long enough to allow a relevant absorption to occur. Ideally, the coupling mechanism should work efficiently for any direction of incident light. This holds well if there is a high density of optical modes in the plane of the structure with, at the same time, a certain amount of coupling between these modes and the free space outside. In past years, several nanophotonic strategies have been proposed, and nowadays the field is mature enough for the fabrication of prototypes. This is a delicate matter, since the realization of photonic nanostructures demands a significant modification of the cell surfaces and crystalline bulk structures, which is generally detrimental for the transport properties of electrons. In essence, improved light absorption would be worthless if it is not followed by a net increase of current generation. Hence, in this review, we will focus on the latest results in the literature with particular interest in those studies concerning the realization of photonic-nanostructured solar cell prototypes. The manuscript is structured into two main sections:

Over the last few years, micro- and nanophotonics have roused a strong interest in the scientific community for their promising impact on the development of novel kinds of solar cells. Certain thin- and ultrathin-film solar cells are made of innovative, often cheap, materials which suffer from a low energy conversion efficiency. Light-trapping mechanisms based on nanophotonics principles are particularly suited to enhance the absorption of electromagnetic waves in these thin media without changing the material composition. In this review, the latest results achieved in this field are reported, with particular attention to the realization of prototypes, spanning from deterministic to disordered photonic architectures, and from dielectric to metallic nanostructures.

1. Introduction Sunlight has by far the highest theoretical potential of the Earth’s renewable energy sources. The solar constant (the solar flux intercepted by the Earth) is 1.37 kW m−2. Taking into account the cross-sectional area of the Earth intercepting this flux, losses by atmospheric scattering or absorption, and other relevant factors, the available solar power is 89 300 TW. This theoretical potential represents more energy striking the Earth’s surface in 2.5 h than the worldwide energy consumption in the year 2013 from all sources combined.[1] This theoretical potential could be used to generate 15 TW of carbon-neutral power (the target value for year 2050, which is believed will stabilize atmospheric CO2 values) from 10% efficient solar-conversion systems covering only 0.17% of the Earth's surface area.[2] For a country like the USA, the potential solar energy from the surface covered by public roads equals roughly the total energy need of the country. Currently, this potential remains largely unused, with photovoltaic energy Dr. M. Burresi, F. Pratesi, Prof. D. S. Wiersma European Laboratory for Non-linear Spectroscopy (LENS) Università di Firenze via Nello Carrara 1, 50019 Sesto Fiorentino, (FI), Italy E-mail: [email protected] Dr. M. Burresi Istituto Nazionale di Ottica (CNR-INO) Largo Fermi 6, 50125 Firenze, (FI), Italy Dr. F. Riboli Dipartimento di Fisica Università di Trento Via Sommarive 14, 38123 Povo, (TN), Italy Prof. D. S. Wiersma Department of Physics Università di Firenze via Nello Carrara 1, 50019 Sesto Fiorentino, (FI), Italy This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

DOI: 10.1002/adom.201400514

722

wileyonlinelibrary.com

© 2015 The Authors. Published by WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Adv. Optical Mater. 2015, 3, 722–743

www.advopticalmat.de www.MaterialsViews.com

2. Deterministic Photonic Structures 2.1. Deterministic Dielectric Photonic Structures The refractive index contrast between air and the dielectric material of the cell is often pronounced, leading to a relevant reflection of light at the cell interface, which reduces light absorption. This effect becomes more important for large incidence angles, as predicted by the Fresnel equations.[3] In the past decades, various approaches have been exploited to fabricate broadband all-angle antireflective structures. Simpler ones, e.g., textures on bulk silicon devices,[4] are based on a ray optics approach. Others obtain constructive interference inside the absorbing film, exploiting single or multiple index-matched layers with thicknesses on the order of a quarter of a wavelength.[5] Coatings based on graded refractive index front layers have also been developed. Such structures are composed by successive layers with an optical thickness increasing according to a polynomial law.[6] In this respect, 1D or 2D patterns have also been successfully implemented on the front surfaces of solar cells.[7,8] Light with an energy close to the electronic band edge of the active material is scarcely absorbed, especially in the case of thin-film devices whose absorption lengths are longer than the thickness of the cell. Generally, in this case, absorption enhancement is achieved by developing strategies to increase the optical path inside the material, often achieved by exploiting total internal reflection.[9,10] Theoretical limits for the possible optical path enhancement of bulk architectures have long been established to be 4n2 for unpolarized light in the ray optics approximation.[11] Such a limit has been extended in the past few years to light-trapping approaches, where the wave nature of light is dominant, e.g., in thin-film materials.[12,13] Given the demand for efficient and cost-effective energy-harvesting devices, thin-film devices have been considered, offering the potential to reduce the use of expensive materials and to fully exploit advanced photonic light-trapping approaches. Such methods not only involve structures placed on top or behind the dielectric layer(s), but they can themselves be realized within the active materials. In contrast to previous architectures based on geometric optics, these structures fully exploit the wave nature of light, thus storing light inside the material for long time. This translates into a pronounced absorption enhancement.[14] Furthermore, the local density of photonic states can be tailored to accomplish a pronounced energy harvesting.[15]

2.1.1. Structures Based on Ray Optics The market of solar cells is currently dominated by bulk silicon devices. For these systems, texturing of sizes much larger


REVIEW

i) deterministic photonic structures (a morphology that can be designed by a specific mathematical rule), and ii) the emerging disordered photonic structures (a morphology that can be designed by statistical principles). These two sections will consider both dielectric and metallic photonic architectures. At the end of the manuscript a Glossary Section has been included to help the reader with the technical terminology.

Matteo Burresi is a Research Fellow of the Italian Institute of Optics (CNR) and Project Leader at the European Laboratory for Non-Linear Spectroscopy (LENS). He obtained his PhD at the University of Twente, supervised by Prof. L. (Kobus) Kuipers at the FOM Institute AMOLF. Engineered disordered systems are his current research topic, spanning from fractal disordered (Lévy glasses) to correlated disordered systems (amorphous photonic materials). Filippo Pratesi obtained his PhD at the European Laboratory for Non-Linear Spectroscopy (LENS). He joined the group of Optics of Complex Systems supervised by Prof. D.S. Wiersma during his previous formation, which took place at both the University of Florence and Copenhagen University (Erasmus project). His research interests focus on the transport of waves in heterogeneous disordered systems and their application for energy harvesting. Diederik Sybolt Wiersma is professor in physics at the University of Florence and in charge of the micro- and nanophotonics research area at the European Laboratory for Non-Linear Spectroscopy (LENS). His research interests lie in the fundamental physics and applications of photonic materials, with an emphasis on disorder-related structures. In addition, he has recently started a new research program on elastomer photonics and photonic nanorobotics. than the incident wavelengths have been employed, trying to enhance the absorption in the infrared part of the spectrum and to reduce the amount of expensive materials. Different shapes and aspect ratios were considered for texturing. A recently proposed concept considers inverted pyramids, at the facet of which light experiences multiple reflections (see Figure 1a). The probability of light being absorbed is thus enhanced as compared to a simple flat surface, where light is immediately back-reflected into the environment.[16,17] With this method, a power conversion efficiency as high as 24% has been demonstrated and scaled to mass production.[18]



723


REVIEW

grating and λ is the incident wavelength. The incident angle is θ1, θ2 is the diffracted angle, and m is the diffraction order. While for small periods (d λ/n2, light can be diffracted into the film with an angle that satisfies total internal reflection,[3] thus trapping light in guided modes. The shape of the grooves comprising the structure is related instead to the intensity delivered to the various diffractive orders. For this reason, different shapes have been considered, such as rectangular or triangular grooves in mono-periodic gratings,[34,35] and direct[36] or inverted pyramids,[29,37] cones,[28] and holes[38,39] in planar ones. When the real shape of the scattering objects is taken into account and the periodicity of the gratings becomes comparable to the wavelength, advanced numerical modeling such as rigorous coupled wave analysis (RCWA) becomes necessary.[40–42] Figure 1. Scanning electron micropscope (SEM) images. a) Inverse pyramids on the front Given the relative simplicity of grating surface of a bulk c-Si solar cell. Reproduced with permission.[16] Copyright 2013, Elsevier. theory[31] and the high-level expertise [ 19 ] b) A cylindrical microlens quartz array. Reproduced with permission. Copyright 2013, Elsereached by the fabrication methodologies vier. c,d) Honeycomb microtextured substrate fabricated above a thin-film a-Si solar cell and nowadays, several studies on solar cell pro[ 20 ] cross-cut. Reproduced with permission. Copyright 2013, Elsevier. totypes have been published. For example, through nanoimprint techniques, Chen et al.[33] fabricated a grating structure of inverted pyramids in The development of structures performing wide-angle acceptance and designed according to ray optics concepts[21] is a conjugated polymer layer, namely, poly(3,4- ethylenedioxyth iophene):poly(styrenesulfonate) (PEDOT:PSS), on a c-Si wafer still under consideration due to the recent advancements of fabsubstrate (Figure 2c). Such a polymer material is transparent rication techniques.[9,22,23] Microlenses, for example (Figure 1b), and is able to form a radial Schottky junction with n-type silfocus the light incident from large angles inside the active icon.[32] These devices are interesting due to the cheap matelayer.[24] Texturing placed above the active layer is advantageous because it avoids possible damage of the film during fabricarials employed, the fabrication technique, and the particular tion.[25] In addition, microlenses have been suggested as a valid junction topology which improves the internal quantum efficiency for all the frequency ranges (300–1100 nm) solution to reduce the shading of front metallic contacts.[19] (Figure 2d–f). All-back-contact architectures have also been Micrometer-sized structures have also been used recently considered in combination with front-side light-trapping as a backbone for novel conformal p–n junction topologies structures.[43] Jeong et al.[28] successfully implemented nano(Figure 1c,d), where, e.g., a thin-film active layer is deposited above the scattering structure.[17,20,26] cones of 450 nm and height 400 nm on a 10 µm-thick c-Si active layer (Figure 3). The external quantum efficiency, that is, the ratio of extracted charge carriers to the number of photons in the incident beam, was found to be above 80% 2.1.2. Grating Structures in the range 400–800 nm. A power conversion efficiency of 13% was obtained, overcoming the 10% result of the nonpatThe idea of exploiting the wave nature of light to enhance the terned cell. efficiency of solar cells using diffractive grating structures was suggested long ago.[27] Devices with active layers made of high refractive-index materials, e.g., crystalline silicon, have been proven to strongly benefit from the presence of periodic modu2.1.3. Thin-Film Solar Cells lations on their front[28] or back surfaces.[29,30] Here, we focus Thin-film materials could open the way to low-cost solar cells. on the main physical mechanism for light scattering from an The reduced thickness of the active material has a beneficial ideal transmissive 1D dielectric grating, as described by the difeffect on photogenerated charge collection, reducing the disfraction equation:[31] tance between contacts. Unfortunately, it concurrently affects mλ the absorptive capability of the system, since most of the light n1 sin θ 1 − n2 sin θ 2 = (1) with energy at the edge of the electric bandgap of the material d will only be minimally absorbed. For this reason, light-trapping structures become crucial in thin-film devices. However, from where n1 and n2 are the indices of refraction of the air and the a design viewpoint, when using gratings with periodicity (L) dielectric material of the layer, respectively, d is the period of the

724





REVIEW

one channel with several resonances, leading to an absorption enhancement exceeding the classical limit (Figure 4d). For ultra-thin films, that is, when the thickness is half the incident wavelength or smaller, it can be shown that, to fully exploit the nanophotonics capabilities, the designed structures must yield deep-subwavelength field confinement and/or low group velocity.[12] The fabrication of 1D or 2D gratings has been implemented in solar cells using different active materials.[38,46–48] Glass nanograting structures on a-Si:H, for example, have been recently investigated (see Figure 5).[49] By introducing a lighttrapping structure, the authors succeeded in enhancing the short-circuit current density and the power conversion efficiency of the device up to 50%, while leaving the other electrical parameters such as the open-circuit voltage Voc and the Fill Factor (FF)[10] primarily unaffected. The effect of light-trapping structures on polymeric solar cells is currently under investigation.[8,50,51] Cheng et al.,[7] for example, used capillary force lithography to imprint a regular grating on the transparent conductive oxide (TCO) layer, or directly in the active layer, of an inverted organic solar cell (Figure 6a). Even if more Figure 2. a) SEM image of an hexagonal structure of air holes in a silicon layer and b) cross-cut efficient solar cells have been fabricated with of the structure showing the topology of the p–n junction. Reproduced with permission.[32] Copy- polymeric materials,[52] the introduction of right 2013, Elsevier. c) SEM image of the nanopyramid structures coated with the PEOT:PSS layer, nanostructures into the active layer results in d) external quantum efficiency, e) J–V characteristics, and f) calculated internal quantum effia 27% increase of the power conversion efficiency (IQE) of the devices with and without the nanopyramid structures.[33] Reproduced with ciency (Figure 6b). This shows the suitability [ 33 ] permission. Copyright 2013, AIP Publishing LLC. of capillary force imprint techniques for the fabrication of photonic structures in these devices, preventing the deterioration effects that typically affect comparable with the incident wavelength (λ) on top of films standard imprinting methodologies. with a thickness close to λ, grating theory is no longer accurate. Under such conditions, interference effects take place, and a A further relevant characteristic of light-trapping structures theoretical description taking into account wave optical effects in thin-film solar cells is their nontrivial angular response.[53,54] is necessary. A statistical temporal coupled-mode theory[44] has Investigations of this aspect are crucial in view of the fact that the operation of these devices occurs at an oblique incidence been successfully used to calculate an upper limit for absorpfor most of the time. From a photonic point of view, an increase tion enhancement in these structures.[12] It has been shown of the incidence angle corresponds to a shift of the wave vecthat deterministic nanophotonics structures are expected to tors of the grating in the reciprocal space. An angular characsignificantly exceed the standard absorption enhancement terization of the external quantum efficiency has been recently limit, 4n2—also known as the Yablonovitch limit[45]—which performed for amorphous silicon solar cells, where monoperihas been derived for bulk low-absorbing structures with roughodic sine-shaped grating structures were fabricated on the front ened surfaces, giving rise to Lambertian scattering (see Secsurface.[46] They estimated for wavelengths in the range 590– tion 3.1 below). In this model, the absorption of the system is proportional to the number of resonances M (Figure 4a–c) 750 nm and unpolarized light a short-circuit current density corresponding to frequencies that are able to couple to guided enhancements of 20% larger than what is expected with Lammodes in the thin film, and inversely proportional to the bertian roughening. The effect of the modal dispersion on the number of channels N available in free space. Let us consider external quantum efficiency was obtained experimentally, and the case of 2D gratings above a bulk absorbing film. In this compared with numerical calculations (Figure 7). condition, only light with a wave vector k = m2π/L with m = Since the optical behavior of microgratings and photonic 0, ±1, ±2,… can be used to excite the modes in the film and, structures on thin films has been studied quite extensively concurrently, only modes with k ≤ k0 ≡ w/c, where w is the frein the past years through numerical calculations,[55–58] their quency and c the speed of light—the so-called light line—can full exploitation in real devices is currently under investigabe excited from outside (Figure 4b). When L ⯝ λ, there is only tion.[39,59–62] Depauw et al.[38] fabricated a periodic pattern of




725


REVIEW


Figure 3. a) Optical image of the back (top, left) and front (top, right) of a 10 mm-thick Si solar cell. SEM images of cross-sectional view of the device (bottom, left) and cross-sectional view of the nanocones (bottom, right). b) Schematic illustration of the device. c) External quantum efficiency data of the device. d) J–V characteristics of the two devices in (c). Reproduced with permission.[28] Copyright 2013, Macmillan Publishers Ltd.

holes on the front surface of a 1 µm-thick crystalline silicon solar cell. The TCO layer on the front of the cell has the same morphology of the photonic crystal etched in the silicon layer (Figure 6c,d). A topological, optical, and electrical characterization was performed on the samples to identify the effect of this structure on charge collection. Significant enhancements of the short-circuit current density up to approximately 20% have been measured for angles up to 60°. Unfortunately, the patterning induces a significant reduction of the Voc, and quantum efficiency measurements suggest an increase in surface recombination, which in turn decreases the overall efficiency of the device.

2.1.4. Aperiodic Photonic Structures Up to this point, we have considered solar cell architectures with periodic light-trapping patterns. However, the absorption enhancement obtained with this approach usually takes place in relatively small frequency bandwidths. Interests has recently been devoted to other types of deterministic structures achieved, for example, by modifying the structural properties of the system in such a way as to reach an intermediate optimal response.[63,64] One of the more straightforward strategies is the introduction of controlled defects in photonic crystals,[65] which results in a so-called superlattice. This method leads to the creation of additional modes that propagate in the structure, and thus yield richer resonance spectra.[63,66] The local density

726


of states of photons can be manipulated as well as the in-coupling with the absorbing structure (Figure 8). In the literature, a few works can be found exploring the possibility to exploit nonperiodic deterministic structures—that is, without translational invariance—to improve the optical response of thin-film solar cells. These systems are usually generated by some substitution rule of two elementary objects, e.g., two values of the dielectric constant or two different thicknesses (1D), or radii (2D). A subset of these, quasi-crystals, can be generated through projections of periodic structures from higher dimensions[67] and present long-range ordered features.[68] The interest in these structures for light management lays in the control of their structural order at different length scales and the exploitation of noncrystallographic symmetries. Their high-order rotational symmetries, compared to periodic ones, is a crucial characteristic to have both enhanced scattering angles and broader wavelength response. Indeed, the order of local rotational symmetry is closely related to the number of nearest neighbors of each scatterer and, more generally, to how evenly the points in the pattern spread out in space.[67] These properties are studied from the Fourier transform of the positions of the scattering elements in real space, also known as the structure factor S(k).[63] With respect to the in-coupling mechanism for periodic structures, quasi-crystal structures have a denser spectrum in reciprocal space, resulting in a larger probability of satisfying the matching condition for in-coupling. As a consequence, quasicrystalline structures can enhance the absorption as much as periodic ones, but with a more isotropic angular response.[69–71]




REVIEW

Numerical calculations have been performed on quasicrystal patterns of various symmetries, and their light-trapping performance has been compared with other strategies involving periodic[71,72] and disordered structures.[67,73,74] Quasi-crystalline structures have also been practically realized with nanoimprint lithography (NIL), or focused ion-beam (FIB) milling (Figure 9),[67,74–77] and have been implemented as architectures for solar cells.[78,79]

2.2. Deterministic Plasmonic Structures

Figure 4. a) 1D grating structure with mirror symmetry. Air slot width is 40% of the period. b) Absorption spectra obtained by numerical simulations for s and p polarized light. Light is incident from the normal direction. c) Channels in 2D k-space. Blue and gray dots represent channels when the incident light comes in from normal and off-normal directions, respectively. The radius of the circle is k0 . d) Limit of absorption enhancement in 2D grating structures. Reproduced with permission.[54] Copyright 2010, OSA.

In metals, nanophotonics is subject to different optical phenomena, since conducting electrons are free to move when an electric field is applied. As a result, a slowly ( k0 and, thus, SPPs are trapped at the metal interface, with its intensity exponentially decaying from the interface. In the limit of lossless metals (perfect metals), kx is purely imaginary and, thus, no propagating SPPs can be sustained. This corresponds to the case of very low-frequency electromagnetic fields, for which the metal is nearly perfect. The most striking property of SPPs is the possibility to concentrate light beyond the diffraction limit.[81] Surface plasmon polariton modes, sustained on nanowaveguides with diameters much smaller than the plasmon wavelength, have been repeatedly reported.[82,83] As it has been argued,[84] this can be seen as a concentration of electromechanical, rather than electromagnetic, energy. The energy of the system is thus the sum of the potential and mechanical kinetic energies of the free electrons, as proposed in the Drude–Sommerfeld model.[80] This field concentration is not due to interference effects of multiply scattered waves, as in the case of dielectric nanophotonic structures (see Sections 2.1 and 3.1), and thus it is not bound to the Abbe diffraction limit. Strong light confinement leads to larger field enhancement, which can be exploited to locally increase

728


the light absorption of materials in the vicinity of metals, with evident advantages for photovoltaic applications. Such a field enhancement, however, comes with a cost. As the field is more concentrated, it penetrates further into the metal, which is a lossy material. In order to use SPPs for photovoltaics, a careful trade-off between field concentration and metallic absorption has to be found by properly choosing the materials employed and tailoring the metal surfaces.[85] Surface tailoring at the nanoscale is necessary not only to concentrate SPPs, but also to excite them. Being waves bound to the surface of a metal, SPPs exhibit a momentum mismatch with free propagating light (as discussed above), which is generally circumvented by nanostructuring the metal surface, for example, with gratings. 2.2.1. Gratings and SPPs Since the realization of the first solar cell prototypes, their back contact has been made of metal. Although there is nowadays a large variety of proposed architectures for solar cells, involving both organic or solid state solutions, most of them still use




REVIEW

Section 2.1). This enhanced absorption comes at a cost, that is, the frequency dependency of the light-trapping mechanism as governed by grating theory (see Section 2.1). Thus, the grating period is usually designed to capture red or near-infrared radiation, for which the absorption length of the active layer is longer. For shorter wavelengths, in thin films made, for instance, from µc-Si, a-Si, CdTe, and other solid state materials, the absorption length is much shorter than its thickness, thus, no light-trapping mechanism is required.[86,87] For ultrathin films, generally used in so-called third-generation solar cells, the thickness of the active layer is sub-micrometric, sometimes less than 200 nm. Trapping light by photonic means (see Section 2.1) in such a thin active layer leads to several difficulties. Firstly, the presence of the back contact metal gives rise to a cut-off frequency for the guided modes of the film, preventing light trapping at relatively long wavelengths. Secondly, a significant percentage of the total intensity of the guided light is not confined to the active region, especially for the low-refractive-index organic solar cells, thus decreasing the absorption efficiency. Under such conditions, the presence of SPPs can significantly increase solar cell performance. SPP modes do not suffer from a frequency cut-off. In addition, the intensity of an SPP is strongly enhanced in the vicinity of the metal, i.e., very close to the active layer. These enhanced plasmon-trapping effects come at the cost of losses due to the metal absorption, which dissipates energy as heat. With the proper choice of materials and their geometry it is possible to find the best tradeoff between reduced plasmon losses and pronounced field confinement. Plasmonic architectures are predicted to significantly increase the performance of ultrathin solar cells, both solid state and organic,[88] and several prototypes have been already realized and Figure 7. a) Current density of an a-Si solar cell with a 1D light-trapping grating (left and right), [89–92] investigated. Ferry and co-workers[93] integrated over the wavelength range 590–750 nm. Circles, squares, and triangles denote p-, s-, and unpolarized results, respectively. The solid lines illustrate a cosine angle-dependence developed a plasmonic solar cell based on a starting from the unpolarized value at 0°. b) Experimental angular maps of external quantum 2D grating which shows a 26% (see Figure 10) efficiency of the cell with periodicity P = 550 nm. The left and right panels correspond to s and p enhancement in short-circuit current density. polarizations. Superimposed symbols represent the calculated dispersion of waveguide modes. This was obtained by using active layers of less Reproduced with permission.[46] Copyright 2014, AIP Publishing LLC. than 200 nm, for which optical guided modes are not supported, exploiting the coupling to SPPs at the metal interface. Furthermore, the device was fabrimetal back contacts. As the demand for thinner active layers cated by nanoimprint lithography, which is a relatively cheap, increased, it became natural to try to introduce some kind of large-area fabrication method, showing that plasmonic nanocorrugation in this metal film to scatter light parallel to the thin structures are promising also for industrial implementation. active layer. The initial solutions were to realize either a disordered texturing of the metal film, achieving a Lambertian scattering of light, or periodic patterns, i.e., gratings, which scatter 2.2.2. Complex SPP Architectures light at specific angles. Some of these angles are beyond the critical angle for total internal reflection and the diffracted light More sophisticated solutions have been investigated in the last is trapped inside the active film, thus ensuring absorption (see years regarding the evolution of gratings. Modified gratings




729


REVIEW


Figure 8. a) Two different superlattice geometries and their spectral absorption b) as a function of superlattice air hole diameter. For reference, the vertical dotted white lines indicate when the superlattice air hole diameter is the same as the background lattice air hole diameter (i.e., no superlattice). c) Absorbed current as a function of superlattice air hole diameter for four different configurations obtained by integrating the spectral absorption and weighting by the AM 1.5 solar spectral current. Reproduced with permission.[15] Copyright 2012, the Amercan Chemical Society.

have been proposed and are expected to improve absorption in thin-film materials due to the combination of grating properties and the plasmonic field enhancement capabilities.[94] Gratings with inclusions or indentations have been also considered to significantly rise the absorption of, for instance, organic materials.[95,96] The general idea is to combine the coupling properties of a metallic grating with the field enhancement occurring when two metal interfaces are in very close proximity (see Section 3.2). In such a configuration, not only is light trapped in SPP modes, but its intensity is also strongly localized in the active material between the metals and thus efficiently

730


Figure 9. a) Moiré nanolithography as described in.[75] (Reproduced with permission.[75] Copyright 2012, the Amercan Chemical Society). Fabrication scheme of moiré patterns from a hexagonal mask exposed twice at an angular offset α. SEM images and (insets) Fourier transforms of the first-order Bragg peaks for b) a hexagonal array (α = 0°), c) a superperiodic array (α = 10°), d) a superperiodic array (α = 20°), and e) a 12-fold array (α = 30°). f) Absorption spectra for the optimized periodic and quasi-periodic structures. The AM 1.5 solar spectrum (magenta-dotted) and the absorption spectrum of the solar cell with a silver mirror as a back-reflector (black-dotted) are also shown. g) Absorption enhancement versus incident angle for optimized periodic structures. Reproduced with permission.[71] Copyright 2013, AIP Publishing LLC.




REVIEW

firstly adjusted a standard grating design in order to achieve a 33% relative absorption enhancement (see Figure 11). They then introduced an indentation in the grating which gives rise to an absorption resonance in the NIR, leading to a 41% relative absorption enhancement. Surface plasmon polaritons can be launched also by approaches other than gratings.[97] Ferry et al.[98] proposed to enhance the absorption of a thin-film material, fabricating an array of grooves in a metallic back contact (Figure 12). Each groove efficiently couples the solar radiation to SPPs due to their large scattering cross-section. These nanoindentations are arranged in a periodic manner, in order to homogeneously cover the film. The distance between grooves was chosen depending on the SPP absorption length and their orientation was designed so to have an efficient scattering for different polarizations. The advantage of such a configFigure 10. a) Schematics of the patterned solar cell. b) Snapshot of an imprinted, patterned, uration is that the incident angular dependsolar cell chip. Each colored square is a different area where the pattern has different particle ency of the relative absorption enhancediameter and pitch. c) SEM of Ag-coated patterns (290 nm diameter particles with 500 nm ment (reported to be 150%) is extremely pitch). d) Electrical measurements for a-Si:H with intrinsic layer thickness 160 nm. Reproduced minute. with permission.[93] Copyright 2010, OSA. A different kind of periodic plasmonic nanostructure which has been investigated for harvesting the solar radiation impinging on ultra-thin film absorbed. By engineering the hybridization of the localized SPP materials is an array of nanoparticles. These are arranged and the SPP modes in nanovoids, Lee and co-workers[95] introwith a period which is shorter than the wavelength they are duced a tunable resonance into the absorption spectra. They

Figure 11. a) Schematics of a solar cell with a nanovoid indented grating. b,c) Absorption enhancements for different geometries of the nanovoids. d–g) Magnetic field amplitude distribution at points E, F, G, and H, respectively, from (b). Reproduced with permission.[95] Copyright 2010, OSA.




731


REVIEW


3. Disordered Photonic Structures 3.1. Disordered Dielectric Photonic Structures One of the first optical mechanisms considered to increment the absorbance performance of active materials for photovoltaics was the random scattering of light, generally achieved by disordered structures. According to the scattering strength of the disordered system, different regimes of light transport are available. For (quasi-)single scatFigure 12. a) Schematics of a solar cell with subwavelength grooves on a back Ag con- tering, the direction of light is randomized, tact. b) Absorption enhancements as a function of angle of incidence for 700−1100 nm generally with a Lambertian angular distriincident light. Reproduced with permission.[98] Copyright 2008, the Amercan Chemical bution, which can be used to increase the Society. light path in the absorbing material. For the case of multiple scattering, light diffusion takes place[107] when the scattering strength is weak. When designed to capture.[99,100] These metallic nanoparticles are placed in such close proximity in order to support propagating the scattering strength is so strong enough that interference modes across the particle chains, thus trapping light.[101] Lin effects become predominant, the light undergoes subdiffusion, et al. employed energy transfer (ET) between nanoparticles, which is a form of optical waveguiding, to achieve and efficient light absorption.[102] Surface plasmon ET occurs when the localized plasmon resonances of individual metallic particles (see Section 3.2.1) couple to each other (Figure 13c).[103] Lin et al. found that, compared to a conventional waveguide, surface plasmon ET provides larger scattering angles, finally yielding, after optimization of the geometry with a genetic algorithm, a relative absorption enhancement of 46.95% for a silicon film of 300 nm. Very recently, a novel plasmonic architecture, initially envisioned for plasmon management,[104] was proposed for harvesting solar radiation. These plasmon cavities[105] are based on the coupling of plasmon resonances achieved by tailoring metallic films, with optical Fabry–Perot like cavities. Chou and Ding[106] investigated such a structure with the purpose of realizing a solar cell prototype (Figure 14). They fabricated a centimeter-scale device where a hole array in a gold layer was placed on top of an organic ultrathin film (the active layer). The back contact was an Al film, which acted also as reflector for the Fabry–Perot cavity of the plasmon cavity. In contrast with ref. [104], where absorption coming from the spacer between the two metal films is negligible, for this device, the plasmonic properties of the structure had to be engineered to compensate for the absorption occurring in the active layer. The author realized a plasmonic– Figure 13. a) Absorption spectrum of an array of nanoparticles designed according to ET organic solar cell with a short-circuit current principles. b) Angular distribution the scattered light (Poynting vector). c) Electric field distridensity increase of 41% and a power conver- bution for ET between nanoparticles at wavelength 642 nm. Reproduced with permission.[102] sion efficiency enhancement of 52%. Copyright 2012, OSA.

732





REVIEW Figure 14. a) Schematic of the plasmonic cavity with a subwavelength hole-array solar cell (PlaCSH). b) SEM of the Au nanostructured film (175 nm diameter and 200 nm pitch hole array). c) Snapshot of 4 inch nanoimprint mold used. d) EQE spectrum of PlaCSH solar cell (SC), ITO-SC, and the absorption length in P3HT:PCBM (85 nm thick on glass). Measured EQE spectrum of PlaCSH and solar cell without nanostructuring of the ITO film; measured absorption spectrum the bare active layer. Reproduced with permission.[106] Copyright 2013, OSA.

eventually leading to a complete halt of light transport, the socalled Anderson localization of light.[108] A (sub-)diffusive process has the clear advantage of slowing down light transport to increase the path length of light inside a material and, thus, the probability for light to be absorbed. The idea to exploit the interference effects arising from multiple scattering of light to increase its absorption dates back to the eighties.[109] In this pioneering work, S. John demonstrated that the effect of a small imaginary part of the dielectric constant on the propagation of waves in a disordered medium close to the Anderson transition induces a divergence of the absorption coefficient. This effect is attributed to the extremely slow diffusive propagation of energy in the disordered medium. The transport of light in this class of system is driven by modes characterized by resonances of finite spectral width and spatial extent, generated by the interference between multiple scattered light waves.[108] Indeed, every scattering event randomizes the direction of propagation and the phase of the electromagnetic wave, yet giving the possibility to define photonic modes with a characteristic spatial profile and lifetime. In diffusive systems, the scattering mean free paths—the distance covered by light between two scattering events[108]—is much longer than its wavelength λ, and the modes spectrally overlap and are spatially delocalized over the sample. When the scattering mean free path, ls, becomes comparable to the wavelength, the modes are spectrally well separated and spatially isolated, and the propagation of light through the system is inhibited.[108] Figure 15a and b schematically depict the spatial distribution of a diffusive mode and a localized mode, respectively. For the sake of clarity, the modes are represented as two-dimensional.


However, it has to be pointed out that the Anderson transition occurs only in 3D, whereas in 2D the modes are always localized,[108] albeit their extension can vary significantly.[110] In such a subdiffusive system, light performs many oscillations before escaping the sample, thus increasing the probability that it will be absorbed.

3.1.1. Early Designs: Lambertian Scattering To date, the most practical solutions make use of disordered photonic materials, employing either disordered gratings or rough surfaces. The integration of these surfaces just above or

Figure 15. Intensity profile of a delocalized mode (left), and a localized mode (right). In both cases the modes are composed by collection of speckles. For delocalized modes the envelop of the speckle intensity is almost constant as a function of the position and it depends only on the bound ary of the samples. In the localized regime the envelope of the speckle pattern intensity decreases exponentially in all directions.



733


REVIEW


Figure 16. a) The ability of an antireflection coating (ARC) to maximize optical absorption is limited because the optical path length is not altered. b,c) Incoherent or geometric light-trapping scheme, where random scattering is used to increase the optical path length inside the photoactive region. d) Coherent light-trapping approach where interference, diffraction, and optical electric-field concentration are exploited to enhance the likelihood that an incident photon is absorbed in the photoactive region.

below the active material allows spreading of the incident light, increasing the optical path of each light ray and, therefore, the probability of it being absorbed.[111–114] Other advantages come from the decrease in the reflection of the incident light, thanks to better refractive index matching between the absorber and the external environment.[115–117] Moreover, this approach does not directly modify the active material and does not influence its electronic properties, as would happen if the scattering centers were embedded in the active layer. The upper limit for the absorption enhancement that can be achieved using a diffusive mechanism has been calculated by E. Yablonovitch.[45] Within the ray optics approximation, this limit sets the maximum absorption enhancement of a rough surface (in doublepass configuration) with respect to flat (single pass) geometry. As mentioned in Section 2.1, this limit is 4n2, where n is the refractive index of the active layer. For instance, by roughening a slab of silicon and using at the bottom of it a back-reflector, it is theoretically possible to enhance the absorption by a factor of 50, in the visible range.[118] Although this limit does not take into account interference effects, it is often chosen as a reference value for absorption enhancement in photovoltaic materials. A conceptually simple implementation of photonic disorder for solar cells can be achieved, for example, by smearing the surface of the active material with a disordered arrangement of scatterers, e.g., small dielectric spheres, or more simply by gently roughening its surface. Figure 16 shows schematic representations of solar cells. The simpler architecture is a flat surface (Figure 16a), where a possible antireflection coating is deposited on top. The optical path length is not altered, and absorption is limited by the single or double pass of the photon through the active layer. Only by increasing its thickness it is possible to enhance the absorption efficiency. Figure 16b and c show two possible implementations of how to take advantage of random scattering to enhance the absorption efficiency: In the first case, the presence of disordered dielectric spheres on the top of the active layer is able to spread the light through the active material,[119] while in the second case, similar results are achieved by roughening its surface.[120–126] Note that the aim here is to scatter enough light so as to have a significant increase of absorption, avoiding, though, the induction of a multiple light scattering process that would reflect most of the impinging light. In most cases, this haze is obtained through

734


texturing of the silicon surfaces or TCO layers.[127–131] A novel realization of this concept was recently proposed by Fang et al.[132] They developed a new kind of cellulose paper with ultrahigh optical transparency (≈90%) and ultrahigh haze (≈60%) to be employed as light scattering substrate for photovoltaic devices. When this cellulose system is placed above a silicon wafer, a relative absorption enhancement of 10−18% is measured in the range 400−1000 nm. This material was also able to enhance the power conversion efficiency of an organic solar cell by 10%, by simply being added on its surface.

3.1.2. Coherent Light Trapping in Thin Films through Photonic Disordered Modes Nowadays, it is well accepted that the most appropriate photonic architectures to be employed in ultrathin film solar cells are designed in planar waveguide configurations. To fully exploit the potential of disordered photonic systems and to benefit from coherent light-trapping mechanisms, the scattering elements have to be embedded into the active material. Figure 16d shows the basic principle of the coherent light-trapping mechanism based on modal coupling. In this case, the wave nature of light is completely exploited, being the thickness of the active layer and the typical pitch of the photonic material comparable with the incident wavelength (see Section 2.1). The amount of coupled and reflected light, the acceptance angles, and the polarization response can be engineered by playing with the size, depth, and positional correlation length of the scattering centers.[133] All these considerations hold for both deterministic and disordered 2D photonic structures, but investigations of the latter for harvesting solar radiation are scarce. 2D architectures show several advantages with respect to 3D arrangements of scattering centers. The most important is that the probability to trap light and create long-living modes is much higher than in three dimensions.[134] Other advantages, common also to periodic structures, are related to the intrinsic 2D nature of the system that allows direct access to quasi-guided modes from the third dimension[133] and to local engineering of the material.[135] Planar architectures are also compatible with standard growing processes, and they can be integrated with electronic components. Coherent light-trapping mechanisms were initially developed in 2D deterministic photonic structures and nowadays




REVIEW Figure 17. a) SEM image of a 2D disordered photonic material on a suspended membrane. b) Schematic representation of the multiple scattering of guided waves. The balance between the amount of guided and radiative number of resonances determine the coupling efficiency from the outside and the lifetime of localized modes. This is calculated in (c) and (d) for a random and a correlated disordered hole pattern, where the dashed line correspond to the light cone. Reproduced with permission.[133] Copyright 2012, Macmillan Publishers Ltd.

are often investigated (see Section 2.1). Nevertheless, such an ability generally comes at the cost of a limited set of operating conditions, i.e., a small frequency bandwidth, a narrow acceptance angle range, and a single polarization operation. All these limitations, which can be detrimental for photovoltaic applications, can be overcome by considering 2D disordered modes. As mentioned above, these modes are localized, by virtue of the dimensionality of the system, and intrinsically lossy. Indeed, even considering an absorption-free material, the out-of-plane scattering induced by the in-plane multiple scattering would be non-negligible. Although the in-plane transport is inhibited, the lifetime of disordered modes decreases significantly with respect to deterministic structures, thus incrementing the vertical light-coupling efficiency. These characteristics are also followed by an increase of the acceptance angles and a flattening of the spectral response. The first work to investigate how the nature of these optical modes can be exploited for photovoltaic applications was carried on by Vynck et al.[133] They studied a disordered photonic slab whose architecture is similar to that of Figure 17, i.e., a dielectric membrane perforated with cylindrical air holes acting as scattering centers.[136] Through a detailed analysis of the modes present in the system, they calculated the Fourier resonances M (see Section 2.1) for a large spectral bandwidth, as shown in Figure 17 in normalized units. For low frequencies around t/λ ⯝ 0.6 (t being the film thickness), the resonance distribution exhibits a well-defined peak, corresponding to a propagating mode with an effective refractive index. In this situation, only a few M resonances are present above the light line (see Section 2.1). Increasing the normalized


frequency, a progressive broadening of the peak, associated with a stronger light scattering by the holes, is observed, and the M-resonances broadly spread over the reciprocal space, leading to an increase of the wavevector density above the light line. As shown in Section 2.1, the amount of wavevectors above the light line are directly related to the coupling efficiency of light inside the dielectric membrane. The almost uniform distribution of the resonances above the light line, in frequency and reciprocal space, indicates that the light-coupling efficiency depends only weakly on the wavelength and angle of incidence. This lead to an increase of the absorption in thin films, which is broadband, omnidirectional, and polarization-insensitive, as experimentally proven by Burresi et al.[137] Vynck et al.[133] also proposed a novel kind of architecture for harvesting solar radiation: By introducing spatial correlations in the disordered distribution of scatterers, they obtained a disordered (nondeterministic) photonic structure with short-range correlations,[110] as shown in the inset of Figure 17d. The effect of these characteristic structural properties can be clearly observed in the photonic dispersions of Figure 17c,d. It is easy to see that the amount of resonances lying above the light line in the presence of spatial correlations depends on the normalized frequency, t/λ. Indeed, by playing with structural parameters it is possible to engineer and redistribute the amount of leaky resonances, thus enhancing or depleting the probability to couple light inside the dielectric film. The calculated relative absorption enhancement—with respect to the bare slab—of nanopatterned thin a-Si films integrated in the 600−800 nm wavelength range of the solar spectrum is 230%



735


REVIEW


for a completely random pattern of scattering centers, and 260% for the correlated case. These values are to be compared with the same quantity in the completely ordered counterpart, that is, 230%. These initial results have driven recent research efforts around disordered nanophotonic materials to enhance absorption efficiency. Several investigations, both theoretical[138–141] and experimental,[142] have looked at more realistic geometries for thinfilms solar cells. All these investigations showed that disordered photonic nanostructures yield comparable if not superior absorption efficiency with respect to their deterministic counterparts. This is extremely promising for future applications. Given the statistical nature of the optical properties of disordered photonic nanostructures, they are intrinsically more robust towards the structural imperfections arising during the fabrication processes. The displacement of scatterers or the variation in their geometry with respect to the intended pattern would not change the average properties, leaving unaltered the overall performance of the device. This robustness suggests that less accurate, and thus less expensive, fabrication methodologies could be developed for a cost-effective realization of photonic-based solar cells. In this respect, very recently, Paetzold et al.[143] performed an experimental investigation of the role of disorder in photonic light-trapping structures etched in the front TCO layer of a-Si:H solar cells (Figure 18a). The impact of Gaussian stochastic displacement (Figure 18b) of initially periodic arrangements of nanostructures was studied. While at small wavelengths, light is absorbed in the first tenth of a nanometer from the surface, at long wavelengths, strong coupling to specific leaky waveguide modes occurs (Figure 18c). Increasing the displacements of the scatterers leads to broader resonances in the medium, which improves the external quantum efficiency in the bandwidth around 700 nm. For this reason, relative enhancements up to 60% of the short-circuit current Jsc, compared to those for periodic structures, are observed (Figure 18d). In these peculiar disordered structures, a trade-off between the mode lifetime τ (defining the coupling efficiency from the vertical direction) and the absorption time τa (the average time after which light is absorbed by the dielectric medium) is crucial to properly engineer the structure. When τ is too large, light will not couple to the disordered modes. When τ is too short with respect to τa, light coupled to the modes will escape before being absorbed. How these two lifetimes relate to the inplane transport properties is currently under investigation.

3.2. Disordered Plasmonic Structures Metals do not only sustain surface (see Section 2.2) but also volume charge density oscillations. This is the case for small metal particles, that is, particles whose size is comparable to the penetration depth of the electromagnetic waves. At optical frequencies, such a penetration depth is around 10 nm and, thus, for particles of a few hundreds of nanometers (smaller than the excitation wavelength), the optical response of the nanoparticle is dielectric-like. Given the very high complex-valued dielectric constant of metals, a series of spectral resonances in the scattering and absorption cross-section arises: so-called localized

736


surface plasmon resonances (LSPRs). As a result, these wavelength-dependent cross-sections can be several times the cross-sectional area of the particle. The extinction, scattering, and absorption efficiency, defined as the ratio between the corresponding cross-sections and the cross-sectional area, can be exactly calculated for spherical nanoparticles of radius R by Mie theory as[144] Qe =

2 ∑ (2l + 1) ℜ (al + bl ) x2

Qs =

2 ∑ (2l + 1) (al2 ) + (bl2 ) x2

{

(4)

}

Qa = Qs + Qe

(5)

(6)

bl =

n2ψ l ( y )ψ l ′ ( x ) − n1ψ l ′ ( y )ψ l ( x ) n1ψ l ( y )ηl ′ ( x ) − n1ψ l ′ ( y )ηl ′ ( x )

(7)

al =

n1ψ l ( y )ψ l ′ ( x ) − n2ψ l ′ ( y )ψ l ( x ) n2ψ l ( y )ηl ′ ( x ) − n2ψ l ′ ( y )ηl ′ ( x )

(8)

Here, ψl and ηl are the Riccati–Bessel functions of the first and third kind, x = n2kR and y = n1kR are normalized parameters, n1 and n2 are the complex-valued refractive indexes of the sphere and environment, respectively, and k is the wavevector of light. Depending on the relation between λ and R, i.e., on the normalized parameters x and y, several resonances appear in Qs and Qa, which are associated with different multipole moments of the particle (dipole, quadrupole, octupole moments, both electric and magnetic). Figure 19 shows the scattering and absorption coefficients for silver and aluminium nanoparticles in a polymeric environment (refractive index 1.5) for different radii. The dielectric constants as a function of frequency of these materials have been taken from the literature.[145] Clearly, by properly choosing the material and the dimension of the nanoparticle, it is possible to tune the resonances so as to cover the entire visible spectrum, a property extremely useful for solar applications,[146] as we will now show. One of the most striking properties of LSPRs is the capability to concentrate the optical field close to the particle surface.[147] For the case of SPPs (Section 2.2), this is due to the accumulation of free electrons which oscillate following the direction of the driving electric field. As a result, since the field is higher where the charges accumulate the most, the location of the field enhancement is determined by the light polarization and the shape of the particle. In case of an isotropic particle, only the polarization will determine the field distribution (Figure 20a). In the case of particular geometries, the field tends to accumulate where the radius of curvature is minimum, similar to the lightning-rod effect at low frequencies (Figure 20b). At optical frequencies, however, the presence of LSPRs increments the field enhancement much more with respect to the lightningrod effect. In this regard, metal nanoparticles can be considered optical antennas,[149,150] which improve the light–matter




REVIEW Figure 18. a) SEM image of cross sections of a periodically structured a-Si:H solar cell. b) Surface textures with different amounts of disorder measured by atomic force microscopy. c) Absorbance of the solar cell with a periodic and random light trapping structure. d) Enhancement of the Jsc of a-Si:H solar cells prepared on the nanostructured superstrates with average distances of neighboring nanostructures of 500 nm (squares) and 800 nm (spheres). Reproduced with permission.[143] Copyright 2014, AIP Publishing LLC.

interaction through near-field coupling. Systems composed of nanoparticles of various geometries have been investigated in order to enhance more effectively the local field. This phenomenon was used in a variety of applications, ranging from surface-enhanced Raman scattering[151,152] to sensing,[153] to nonlinear optics[154,155] and photovoltaics. As a final remark, it should be noted that the realization of plasmonic nanoparticles with different shapes and types of aggregates are often obtained with self-assembly fabrication techniques.[156] In general, photonic architectures designed to improve the absorption of ultrathin solar cells that can be fabricated by self-assembly techniques are particularly welcomed by industry, given their low-cost implementation. Self assembly, if not specifically controlled, naturally leads to the realization of disordered structures.

3.2.1. Plasmonic Nanoparticles and Photovoltaics

Given their large scattering efficiency, plasmon nanoparticles can be used to efficiently couple light into the absorbing thin films.[157–159] One of the first attempts to use metal nanoparticles on top of an a-Si solar cell was carried out by Derkacs et al.,[157] who found an 8% increase in short-circuit current density. In general, this strategy turns out to be particularly suited to relatively thick films with a high refractive index.[158,160,161] When a nanoparticle is placed at the interface between two dielectric materials, most of the scattered light couples to the high-refractive-index medium.[158] Figure 20c shows the fraction of scattered light in an absorbing film as a function of the shape of the particles at the film interface. By tuning the size and material, the resonant scattering cross-section of the nanoparticles can be employed to increase the absorption of a portion of the solar spectrum. This can lead to the realization of thin films which, for instance, absorb in the UV but are semitrasparent in the visible and NIR regions.[162] The use of nanoparticles to increase light harvesting must deal with their ever-present absorption, which can become extremely pronounced for particles of a few tens of nanometers (see Figure 19). Very recently, scientists have been trying to exploit this absorption by collecting the optically generated hot carriers of the metal itself.[163–165] These hot carriers are injected into the conduction bands Figure 19. a,b) Spectral scattering and absorption efficiency for silver and aluminium nanopar- of materials in close proximity with the particles. Shao et al. deposited gold nanoparticles with different radii. The dashed and continuous line correspond respectively to absorpticles on a TiO2 nanotube array (Figure 21) tion and scattering efficiency.




737


REVIEW


Figure 20. a,b) Calculations of the electric-field distribution near to a circular disk and a triangular disk of gold. The exciting electric field is oriented along the x direction. Reproduced with permission.[148] Copyright 2010, JMOe. c) Percentage of light scattered into a silicon substrate by Ag particles with various shapes laying on top of a silicon substrate. The cylinder has a 100 nm diameter and is 50 nm high; the hemisphere and sphere are both 100 nm in diameter; a 150 nm diameter sphere is also shown. Reproduced with permission.[158] Copyright 2008, AIP Publishing LLC.

and carried out photocurrent experiments.[163] They found an enhancement of current (145%) when illuminating the sample with respect to the current measured in darkness, which is attributed to the hot carriers injected into the TiO2. An interesting alternative approach for harvesting solar radiation employing plasmonic nanoparticles was recently implemented by Fahim et al.[166] The authors succeeded in preparing a layer of random gold nanoparticles embedded in SiNx to use as an antireflection coating on top of a c-Si solar cell. The work presented a study of the photocurrent enhancement as a function of the distance between the plasmonic nanoparticles and the c-Si substrate (see Figure 22). At 100 nm, the photocurrent enhancement was minute and mostly due to the strong scattering of the gold nanoparticles, which increased the light path. As the distance between c-Si interface and the plasmonic nanoparticles was enlarged, the photocurrent enhancement increased significantly, up to an overall 6% augmentation for a distance of 10 nm. This effect was attributed to the large field enhancement close to the nanoparticle surface, which concentrates the light intensity and thus increases the optical absorption.

functionalizations[168,169] have been envisioned, mostly with the intent to increase the concentration of the field or the scattering efficiency while keeping the plasmon absorption as low as possible. One of the first ideas to improve the plasmon–matter interaction, not only for photovoltaics but also for sensing and other applications, is the use of dimers of nanoparticles.[147] The plasmon resonance of two or three metal nanoparticles in very close proximity can couple.[170,171] Such a coupling can be tailored to achieve increased scattering efficiency and field enhancement in the nanogap between particles. Developing a cheap fabrication methodology to realize dimers or trimers with the proper nanogap width, which can be as low as 10 nm, is extremely challenging. Recently, Pastorelly et al.[170] fabricated an organic solar cell in which plasmonic oligomers, obtained by self assembly with a roll-to-roll compatible process, were integrated. The optoelectronic characterization of this device obtained by a scalable approach shows an increase in short-circuit current density of about 12%. As these nanoparticles can be considered optical nanoantennas, the attention of scientists has focused on their optical response as a function of their shape, such as rods,[149,171] stars,[172] prisms,[173,174] and triangles,[147] in order to spectrally tune and manipulate the absorption and field enhancement. 3.2.2. Complex Local Surface Plasmon Architectures Tailoring the shape of metallic nanoparticles allows control over their spectral response, which can be tuned to increase The optical response of nanoparticles has been thorthe weak absorption of active materials in a certain spectral oughly investigated, and new shapes,[147] designs,[167] and region, generally in the NIR. One of the most interesting solutions is the use of plasmon nanoshells.[175] These are made of dielectric particles coated with a very thin layer of metal, usually gold or silver. Such particles have plasmon resonances very sensitive to the relative size of their core and shell layers. This sensitivity arises from the so-called hybridization between the plasmons of the inner and outer metallic interfaces.[175] The resonance frequencies of nanoshells can be extended from the visible to the NIR domain.[176] Very recently, Paz-Soldan et al. realized a quantum dot solar cell in which Figure 21. a) SEM of TiO2 nanotubes film. The inset shows a bottom-view picture. b) TiO2 nanotubes film with gold nanoparticles after preparation. c) Measured on/off current ratio a plasmon nanoshell (Figure 23) was inte[177] This architecture yields a scatas a function of wavelength of the excitation laser with and without the nanoparticle coating. grated. [ 163 ] tering cross-section much higher than that Reproduced with permission. Copyright 2014, AIP Publishing LLC.

738





REVIEW

and are suited for relatively thick solar cells. This class of structure allows one to decrease the amount of reflected light by texturing the front surface of the film, avoiding possible damage to the active material and thus preventing the creation of recombination centers for the free carrier. The introduction of thin-film solar cells allows a reduction of the amount of raw materials, hence reducing costs and also minimizing the path length that the phoFigure 22. Schematic representation of a crystalline silicon solar cell where gold nanoparticles togenerated carriers have to cover to reach the contacts. Ideally, the thickness should are embedded within a silicon nitride layer. b) Current density enhancement as a function of be reduced even further in order to also capspacing layer thickness obtained with gold nanoparticles 60 nm in diameter. Reproduced with permission.[166] Copyright 2012, AIP Publishing LLC. ture so-called hot electrons, whose diffusion length is extremely short. However, such a minute thickness prevents the absorption of optical energies of standard particles, whereas the absorption cross-section is close to the edge of the electronic bandgap of the active matecomparable with the latter. Exploiting the large field enhancerial. In this range of energy, the absorption length of light is ment at the metal interface, the active layer of quantum dots much longer than the thickness of the film and, consequently, increases light absorption in the NIR domain, yielding an the radiation is not absorbed. Ray optics approaches are not overall 35% enhancement in photocurrent. sufficient to significantly improve the absorption of such thin films, and nanophotonic methodologies which fully exploit the wave nature of light have to be employed. Light-trapping 4. Conclusions architectures, spanning from deterministic to disordered, are designed to couple and store light inside the active layer. In this review, we have discussed the state-of-the-art of microAlternatively, the extraordinary field enhancement of surface and nanophotonics for photovoltaic applications. We focused plasmon polaritons in close proximity of the active layer can be our attention on the optical performances of the specific exploited to increase the light absorption. However, to induce a designs of photonic and plasmonic structures that can be significant absorption enhancement, the active material should used to maximize the absorption of the incident light, made be directly engineered with nanopatterning. with deterministic or disordered architectures. Although it is Generally, this leads to a decrease in the electronic perforwell known that these structures can enhance significantly the mance by creating electronic surface defects between the bulk absorption of dielectric films, even for extremely thin films, and the surfaces of the scattering elements, thus increasing only recently has the impact of photonic architectures on the unwanted recombination mechanisms. We show in this review actual current production been tested. Nowadays, the first prothat important progress in this respect has been made by the totypes based on ray optics mechanisms, light trapping, modal scientific community which will pave the way for future induscoupling approaches, and even plasmonic principles have been trial applications. fabricated. In all these cases it is possible to take advantage of The main contribution to the Schockley–Queisser limit for deterministic and disordered photonic/plasmonic designs. the efficiency of a single-junction solar cell comes from the Dielectric structures engineered on length scales longer than thermalization of hot carriers and the absorption bandwidth the incident optical wavelengths work in the ray optical regime of the cell active material. The most efficient solutions that allow one to overcome these problems are multi-junction and quantumdot-based solar cells, which exhibit efficiencies beyond 40%. Nevertheless, these architectures are very expensive in terms of production processes and the cost of raw materials, and thus are not suitable on an industrial scale. All these issues cannot be overcome by simply increasing the light absorption, which accounts for a small percentage of the enhancement in terms of energy conversion. Hence, the nanophotonic photovolFigure 23. a) Schematic of a quantum dot heterojunction device with embedded gold taics community aims to drastically modify nanoshells. b) Top-view SEM showing the nanoshells after preparation (scale bar 1 µm). c) TEM the standard architecture of solar cells. showing a cross-sectional image of single gold nanoshells embedded in a quantum dot film (scale bar 100 nm). d) Measured J–V characteristics under irradiation with a solar simulator. Light-trapping schemes, together with plasThe current density enhancement is about 13%. Reproduced with permission.[177] Copyright monic effects, have to be integrated in order to gain the best from each approach. The 2013, the Amercan Chemical Society.




739


REVIEW


Schockley–Queisser limit depends on the radiation intensity and, thus, by increasing its concentration with nanophotonic methodologies, would raise the efficiency limit. Extremely thin films engineered by nanophotonics principles can be used to realize solar cells in which the distance covered by the carriers to reach the contacts is reduced. For example, radial p–n junction Si nanowires have been successfully realized, in which the photogenerated carriers only need to cross the nanowire transversal size in order to be collected.[178] In this respect, inspiration can come from organic photovoltaic devices and dye-sensitized solar cells, where the n-type and the p-type materials are morphologically mixed, making the interaction between electrons and holes less probable. Another possible scenario is offered by quantum dot solar cells that can be realized at room temperature from colloidal solutions.[179,180] The advantage of semiconductor nanocrystals is to have an electronic bandgap that depends on the dot size, increasing the bandgap by more than 1 eV compared with the bulk value. This allows us to properly engineer the width of the electronic bandgap, making possible the use of unconventional materials as active materials for solar cells. The road that researchers have to follow towards novel solar cells made compatible with industrial-scale requirements is still long, and only by merging different solutions will it be possible to get closer to the Schockley–Queisser limit and beyond that, by taking advantage of innovative trapping schemes and electronic bandgap engineering.

5. Glossary Absorption Enhancement: This is defined as AE = Aph/A, where Aph is the absorption of a nanophotonic solar cell and A the absorption of bare solar cell. Anderson Localization of Light: This phenomenon, occurring when light experiences an extremely strong multiple scattering, leads to a complete halt of light transport in disordered systems. Diffraction Limit: Theoretical maximum resolution of an optical microscope which, according to Abbe’s calculations, is d = λ/NA, where λ is the wavelength of light and NA is the numerical aperture of the optical system employed. Diffusion: A phenomenon occurring when light propagates through a system and undergoes multiple scattering events. Generally it is described by the diffusion equation. External Quantum Efficiency: The ratio between the number of photogenerated electrons and the number of photons incident on the solar cell. Fill Factor: The ratio between the product Jsc × Voc and the actual maximum obtainable power PMAX. Guided Modes: A waveguide mode is a spatial distribution of optical energy that is confined within the waveguide by total internal reflection. Hot Electrons: Electrons with energy significantly above the electronic bandgap, generally generated with the absorption of blue light. Internal Quantum Efficiency: The ratio between the number of absorbed photons and the number of photogenerated electrons. Lambertian Scattering: Lambert’s cosine law says that the intensity of the diffusively reflected light from a rough surface is proportional to the cosine of the angle between the normal axis to the surface and the direction of observation. Open-Circuit Voltage: Defined as the voltage Voc at the cell terminals when the current is zero. p–n junction: A boundary interface between two semiconductors with different doping types.

740


Power Conversion Efficiency: Efficiency of a solar cell defined as η = ( ff ⋅ J sc ⋅ Voc ) / Pinc, where ff is the fill factor, Jsc the short-circuit current density, Voc the open-circuit voltage and Pinc the incident power density. Relative Absorption Enhancement: Defined as AE = Aph/A−1. Scattering Mean Free Path: Average distance between scattering events in a system where light undergoes multiple scattering events. Schockley–Queisser Limit: Maximum theoretical power conversion efficiency (33.7%) for a single p–n junction solar cell. Schottky Junction: A boundary interface between a semiconductor and a metal. Short-Circuit Current Density: Defined as the density of the current Jsc when the voltage at the cell terminals is zero. Structure Factor: A mathematical tool commonly used in crystallography that describes how a material scatters incident radiation. Subdiffusion: A phenomenon that can occur when light propagates through a system and undergoes multiple scattering events due to the presence of strong scattering. The spreading of light is slower than in the diffusive regime, and it cannot be described by the standard diffusion equation. Total Internal Reflection: When light is incident upon a dielectric surface above a certain critical angle it undergoes complete reflection. Transparent Conductive Oxide: A semitransparent conductive layer that works as an ohmic contact, most commonly in the form of indium tin oxide (ITO).

Acknowledgements The research leading to these results has received funding from European Research Council under the European Union’s Seventh Framework Programme (FP7/2007- 2013) / ERC grant agreement n? [291349] on photonic micro robotics and the Network of Excellence on Nanophotonics for Energy Efficiency. We gratefully thank the entire Optics of Complex Systems group at LENS for feedback and discussions. Received: November 2, 2014 Revised: January 20, 2015 Published online: March 25, 2015

[1] Key world energy statistics. Technical report, International Energy Agency, Paris, 2014, http://www.iea.org, accessed: March, 2015. [2] J. Tsao, N. Lewis, G. Crabtree, Solar FAQs, US Department of Energy, USA 2006. [3] M. Born, E. Wolf, Principles of Optics: Electromagnetic Theory of Propagation, Interference and Diffraction of Light, Cambridge University Press, UK 1999. [4] D. Smith, P. Cousins, A. Masad, A. Waldhauer, S. Westerberg, M. Johnson, X. Tu, T. Dennis, G. Harley, G. Solomon, S. Rim, M. Shepherd, S. Harrington, M. Defensor, A. Leygo, P. Tomada, J. Wu, T. Pass, L. Ann, L. Smith, N. Bergstrom, C. Nicdao, P. Tipones, D. Vicente, at the 38th IEEE Photovoltaic Specialists Conference (PVSC), Austin, Texas, June 2012, pp.1594–1597. [5] D. Hocine, M. Belkaid, M. Pasquinelli, L. Escoubas, J. Simon, G. Rivire, A. Moussi, Mater. Sci. Semicon. Proc. 2013, 16, 113. [6] Y. Liu, O. J. Guy, J. Patel, H. Ashraf, N. Knight, Microelectron. Eng. 2013, 110, 418. [7] Y.-S. Cheng, C. Gau, Sol. Energ. Mat. Sol. C 2014, 120 B, 566. [8] S. Chattopadhyay, Y. Huang, Y. Jen, A. Ganguly, K. Chen, L. Chen, Ser. Mat. Sci. Engn. R 2010, 69, 1. [9] C. Trompoukis, I. Abdo, R. Cariou, I. Cosme, W. Chen, O. Deparis, A. Dmitriev, E. Drouard, M. Foldyna, E. G. Caurel, I. Gordon, B. Heidari, A. Herman, L. Lalouat, K.-D. Lee, J. Liu, K. Lodewijks, F. Mandorlo, I. Massiot, A. Mayer, V. Mijkovic, J. Muller, R. Orobtchouk, G. Poulain, P. Prod’Homme, P. R. i. Cabarrocas,




[11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23]

[24] [25]

[26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38]

[39] [40] [41]


REVIEW

[10]

C. Seassal, J. Poortmans, R. Mertens, O. E. Daif, V. Depauw, Phys. Status Solidi A 2014, 212, 140. A. Luque, S. Hegedus, Handbook of Photovoltaic Science and Engineering, John Wiley & Sons, Chichester, UK 2011. E. Yablonovitch, J. Opt. Soc. Am. 1982, 72, 7899. Z. Yu, A. Raman, S. Fan, Proc. Natl. Acad. Sci USA 2010, 107, 17491. Z. Yu, A. Raman, S. Fan, Phys. Rev. Lett. 2012, 109, 173901. K. Sakoda, Optical Properties of Photonic Crystals, Vol. 80, Springer, Berlin 2005. D. M. Callahan, J. N. Munday, H. A. Atwater, Nano Lett. 2012, 12, 1214. Y. Choi, D. Kim, E. C. Do, D. Kim, J. Mun, J. W. Lee, Y. Lee, Y. G. Kim, Sol. Energy 2014, 100, 94. S. M. Iftiquar, J. Jang, H. Park, C. Shin, J. Park, J. Jung, S. Kim, J. Yi, Phys. Status Solidi A 2014, 211, 4924. A. Blakers, N. Zin, K. R. McIntosh, K. Fong, at the Asia Pacific Conference Energy Procedia, Singapore, October 2012. M. Nam, K. Kim, J. Lee, K.-K. Lee, S. S. Yang, Sol. Energy 2013, 91, 374. S. J. Baik, Y. Lee, K. S. Lim, K.-T. Kim, Renewable Energy 2014, 6498. R. Winston, Opt. Photonics News 1995, 6, 1133. S.-H. Hong, H. Kim, H. Kim, J. Kim, Mater. Lett. 2013, 112, 62. M. de Jong, J. Rath, R. Schropp, P. Sonneveld, G. Swinkels, H. Holterman, J. Baggerman, C. van Rijn, E. Hamers, J. Non-Cryst. Solids 2012, 358, 2308. Y. Chen, M. Elshobaki, Z. Ye, J.-M. Park, M. A. Noack, K.-M. Ho, S. Chaudhary, Phys. Chem. Chem. Phys. 2013, 15, 4297. O. Douhret, A. Swinnen, S. Bertho, I. Haeldermans, J. D’Haen, M. D’Olieslaeger, D. Vanderzande, J. V. Manca, Prog. Photovoltaics Res. Appl. 2007, 15, 8713. E. Lee, Y. Kim, M. Gwon, D.-W. Kim, S.-H. Baek, J. H. Kim, Sol. Energ. Mat. Sol. C 2012, 103, 93. P. Sheng, A. N. Bloch, R. S. Stepleman, Appl. Phys. Lett. 1983, 43, 6579. S. Jeong, M. D. McGehee, Y. Cui, Nat. Commun. 2013, 4, 2950. G. Li, H. Li, J. Y. L. Ho, M. Wong, H. S. Kwok, Nano Lett. 2014, 14, 2563. M. Wellenzohn, R. Hainberger, J. Photonics Energy 2013, 3, 034595. C. Palmer, E. Loewen, Diffraction Grating Handbook, 6th Ed., Newport Corporation, Rochester, New York, USA 2005. F. Zhang, D. Liu, Y. Zhang, H. Wei, T. Song, B. Sun, ACS Appl. Mater. Interfaces 2013, 5, 4678. J.-Y. Chen, M.-H. Yu, S.-F. Chang, K. W. Sun, Appl. Phys. Lett. 2013, 103, 133901. A. A. Abouelsaood, S. A. El-Naggar, M. Y. Ghannam, Prog. Photovoltaics Res. Appl. 2002, 10, 513. F. Llopis, I. Tobias, Prog. Photovoltaics Res. Appl. 2005, 13, 27. D. Liang, Y. Kang, Y. Huo, Y. Chen, Y. Cui, J. S. Harris, Nano Lett. 2013, 13, 4850. A. Mavrokefalos, S. E. Han, S. Yerci, M. S. Branham, G. Chen, Nano Lett. 2012, 12, 2792. V. Depauw, X. Meng, O. El Daif, G. Gomard, L. Lalouat, E. Drouard, C. Trompoukis, A. Fave, C. Seassal, I. Gordon, IEEE J. Photovoltaics 2014, 4, 215. C.-H. Du, T.-Y. Wang, C.-H. Chen, J. A. Yeh, Thin Solid Films 2014, 557, 372. T. K. Chong, J. Wilson, S. Mokkapati, K. R. Catchpole, J. Opt. 2012, 14, 024012. K. Huang, Q. Wang, X. Yan, M. Yu, X. Shen, L. Chen, J. Chen, Opt. Commun. 2014, 320, 169.

[42] L. Wang, Z.-H. Lu, X.-F. Lin, Q.-D. Chen, B.-B. Xu, H.-B. Sun, J. Lightwave Technol. 2013, 31, 276. [43] N. Zin, A. Blakers, E. Franklin, T. Kho, K. McIntosh, J. Wong, T. Mueller, A. G. Aberle, Y. Yang, X. Zhang, Z. Feng, Q. Huang, Energy Procedia 2012, 25, 1. [44] H. Haus, Waves and Fields in Optoelectronics, Prentice-Hall Series in Solid State Physical Electronics, Prentice Hall, Inc., Englewood Cliffs, New Jersey, USA 1984. [45] E. Yablonovitch, J. Opt. Soc. Am. 1982, 72, 899. [46] F.-J. Haug, K. Söderström, A. Naqavi, J. Li, C. Ballif, Appl. Phys. Lett. 2013, 103, 131108. [47] H. Stiebig, N. Senoussaoui, C. Zahren, C. Haase, J. Meller, Prog. Photovoltaics Res. Appl. 2006, 14, 13. [48] Y. Yang, K. Mielczarek, M. Aryal, A. Zakhidov, W. Hu, ACS Nano 2012, 6, 2877. [49] T.-G. Chen, P. Yu, Y.-L. Tsai, C.-H. Shen, J.-M. Shieh, M.-A. Tsai, H.-C. Kuo, Opt. Express 2012, 20, A412. [50] X. Yu, X. Yu, J. Zhang, Z. Hu, G. Zhao, Y. Zhao, Sol. Energ. Mat. Sol. C 2014, 121, 28. [51] J.-H. Choi, H.-J. Choi, J.-H. Shin, H.-P. Kim, J. Jang, H. Lee, Org. Electron. 2013, 14, 3180. [52] Y. Liang, Z. Xu, J. Xia, S.-T. Tsai, Y. Wu, G. Li, C. Ray, L. Yu, Adv. Mater. 2010, 22, E135. [53] K. Söderström, F.-J. Haug, J. Escarré, O. Cubero, C. Ballif, Appl. Phys. Lett. 2010, 96, 213508. [54] Z. Yu, A. Raman, S. Fan, Opt. Express 2010, 18, A366. [55] A. Bozzola, M. Liscidini, L. C. Andreani, Opt. Express 2012, 20, A224. [56] J. L. Donnelly, B. C. P. Sturmberg, K. B. Dossou, L. C. Botten, A. A. Asatryan, C. G. Poulton, R. C. McPhedran, C. M. de Sterke, Opt. Express 2014, 22, A1343. [57] X. Ziang, W. Wei, Q. Laixiang, X. Wanjin, G. G. Qin, Opt. Express 2013, 21, 18043. [58] X. Fang, C. Zhao, H. Bao, J. Quant. Spectrosc. Radiat. Transfer 2014, 133, 579. [59] I. Y. Bu, T. Hsueh, Sol. Energy 2012, 86, 1454. [60] A. Mihi, F. J. Beck, T. Lasanta, A. K. Rath, G. Konstantatos, Adv. Mater. 2014, 26, 443. [61] J. Bai, C. C. Yang, M. Athanasiou, T. Wang, Appl. Phys. Lett. 2014, 104, 051129. [62] G. Li, J. Y. L. Ho, H. Li, H.-S. Kwok, Appl. Phys. Lett. 2014, 104, 231113. [63] E. R. Martins, J. Li, Y. Liu, J. Zhou, T. F. Krauss, Phys. Rev. B 2012, 86, 041404. [64] C. Battaglia, C.-M. Hsu, K. Söderström, J. Escarr, F.-J. Haug, M. Charriire, M. Boccard, M. Despeisse, D. T. L. Alexander, M. Cantoni, Y. Cui, C. Ballif, ACS Nano 2012, 6, 2790. [65] L. Peres, V. Vigneras, S. Fasquel, Sol. Energ. Mat. Sol. C 2013, 117, 239. [66] D. M. Callahan, K. A. W. Horowitz, H. A. Atwater, Opt. Express 2013, 21, 30315. [67] P. W. Flanigan, A. E. Ostfeld, N. G. Serrino, Z. Ye, D. Pacifici, Opt. Express 2013, 21, 2757. [68] L. Dal Negro, S. Boriskina, Laser Photonics Rev. 2012, 6, 178. [69] C. Bauer, H. Giessen, Opt. Express 2013, 21, S3, A363. [70] P.-C. Tseng, M.-H. Hsu, M.-A. Tsai, C.-W. Chu, H.-C. Kuo, P. Yu, Org. Electron. 2011, 12, 886. [71] A. Micco, A. Ricciardi, M. Pisco, V. La Ferrara, L. V. Mercaldo, P. Delli Veneri, A. Cutolo, A. Cusano, J. Appl. Phys. 2013, 114, 063103. [72] J. Trevino, C. Forestiere, G. D. Martino, S. Yerci, F. Priolo, L. D. Negro, Opt. Express 2012, 20, S3, A418. [73] P. Postigo, J. Llorens, at the 15th International Conference on Transparent Optical Networks (ICTON), Cartagena, Spain, June 2013.



741


REVIEW

www.MaterialsViews.com [74] V. Ferry, M. Verschuuren, C. van Lare, R. E. I. Schropp, A. Polman, H. A. Atwaterat the 37th IEEE Photovoltaic Specialists Conference (PVSC), Seattle, USA, June, 2011. [75] S. M. Lubin, W. Zhou, A. J. Hryn, M. D. Huntington, T. W. Odom, Nano Lett. 2012, 12, 4948. [76] A. Singh, C. Dickinson, K. M. Ryan, ACS Nano 2012, 6, 3339. [77] S. Wang, B. D. Weil, Y. Li, K. X. Wang, E. Garnett, S. Fan, Y. Cui, Nano Lett. 2013, 13, 4393. [78] A. Crescitelli, A. Ricciardi, M. Consales, E. Esposito, C. Granata, V. Galdi, A. Cutolo, A. Cusano, Adv. Funct. Mater. 2012, 22, 4389. [79] A. Ricciardi, M. Pisco, A. Cutolo, A. Cusano, L. O’Faolain, T. F. Krauss, G. Castaldi, V. Galdi, Phys. Rev. B 2011, 84, 085135. [80] N. W. Ashcroft, N. D. Mermin, Solid State Physics, 1st Ed., Holt, Rinehart and Winston, USA 1976. [81] D. K. Gramotnev, S. I. Bozhevolnyi, Nat. Photon. 2010, 4, 83. [82] E. Verhagen, M. Spasenovi, A. Polman, L. K. Kuipers, Phys. Rev. Lett. 2009, 102, 203904. [83] J. Tian, Z. Ma, Q. Li, Y. Song, Z. Liu, Q. Yang, C. Zha, J. Akerman, L. Tong, M. Qiu, Appl. Phys. Lett. 2010, 97, 231121. [84] M. I. Stockman, Opt. Express 2011, 19, 22029. [85] H. A. Atwater, A. Polman, Nat. Mater. 2010, 9, 205. [86] C. Eisele, C. E. Nebel, M. Stutzmann, J. Appl. Phys. 2001, 89, 7722. [87] C. Haase, H. Stiebig, Appl. Phys. Lett. 2007, 91, 061116. [88] C. Rockstuhl, S. Fahr, F. Lederer, J. Appl. Phys. 2008, 104, 123102. [89] X. Li, W. C. H. Choy, X. Ren, J. Xin, P. Lin, D. C. W. Leung, Appl. Phys. Lett. 2013, 102, 153304. [90] Y. Jin, J. Feng, X.-L. Zhang, M. Xu, Y.-G. Bi, Q.-D. Chen, H.-Y. Wang, H.-B. Sun, Appl. Phys. Lett. 2012, 101, 163303. [91] M. Theuring, P. H. Wang, M. Vehse, V. Steenhoff, K. von Maydell, C. Agert, A. G. Brolo, J. Phys. Chem. Lett. 2014, 5, 3302. [92] K. Hara, N. Hathaithip, A. Baba, K. Shinbo, K. Kato, F. Kaneko, at the Proc. Int. Conf. on Electrical Insulating Materials (ISEIM), Niigata City, Japan, June 2014. [93] V. E. Ferry, M. A. Verschuuren, H. B. T. Li, E. Verhagen, R. J. Walters, R. E. I. Schropp, H. A. Atwater, A. Polman, Opt. Express 2010, 18, S2, A237. [94] P.-Y. Chen, H.-H. Hsiao, C.-I. Ho, C.-C. Ho, W.-L. Lee, H.-C. Chang, S.-C. Lee, J.-Z. Chen, I.-C. Cheng, Opt. Express 2014, 22, S4, A1128. [95] S. Lee, S. In, D. R. Mason, N. Park, Opt. Express 2013, 21, 4055. [96] K. Q. Le, A. Abass, B. Maes, P. Bienstman, A. Alù, Opt. Express 2011, 20, S1, A39. [97] L. Novotny, B. Hecht, Principles of Nano-optics, Cambridge University Press, UK 2006. [98] V. E. Ferry, L. A. Sweatlock, D. Pacifici, H. A. Atwater, Nano Lett. 2008, 8, 4391. [99] I. Kim, D. S. Jeong, T. S. Lee, W. S. Lee, K.-S. Lee, J. Appl. Phys. 2012, 111, 103121. [100] C. Simovski, D. Morits, P. Voroshilov, M. Guzhva, P. Belov, Y. Kivshar, Opt. Express 2013, 21, S3, A714. [101] W. E. I. Sha, W. C. H. Choy, Y. G. Liu, W. C. Chew, Appl. Phys. Lett. 2011, 99, 113304. [102] A. Lin, S.-M. Fu, Y.-K. Chung, S.-y. Lai, C.-W. Tseng, Opt. Express 2013, 21, S1, A131. [103] J. R. Krenn, Nat. Mater. 2003, 2, 210. [104] D. Chanda, K. Shigeta, T. Truong, E. Lui, A. Mihi, M. Schulmerich, P. V. Braun, R. Bhargava, J. A. Rogers, Nat. Commun. 2011, 2, 479. [105] N. C. Lindquist, W. A. Luhman, S.-H. Oh, R. J. Holmes, Appl. Phys. Lett. 2008, 93, 123308. [106] S. Y. Chou, W. Ding, Opt. Express 2013, 21, S1, A60. [107] E. Akkermans, G. Montambaux, Mesoscopic Physics of Electrons and Photons, 1st Ed., Cambridge University Press, UK 2007. [108] P. Sheng, Introduction to Wave Scattering, Localization and Mesoscopic Phenomena, Springer, Berlin/Heidelberg 2006. [109] S. John, Phys. Rev. B 1985, 31, 304.

742


[110] G. M. Conley, M. Burresi, F. Pratesi, K. Vynck, D. S. Wiersma, Phys. Rev. Lett. 2014, 112, 143901. [111] C. Rockstuhl, S. Fahr, K. Bittkau, T. Beckers, R. Carius, F.-J. Haug, T. Söderström, C. Ballif, F. Lederer, Opt. Express 2010, 18, S3, A335. [112] X. Sheng, S. G. Johnson, J. Michel, L. C. Kimerling, Opt. Express 2011, 19, S3, A841. [113] P. Kowalczewski, M. Liscidini, L. C. Andreani, Opt. Lett. 2012, 37, 4868. [114] S. Fahr, T. Kirchartz, C. Rockstuhl, F. Lederer, Opt. Express 2011, 19, S4, A865. [115] S. Koynov, M. S. Brandt, M. Stutzmann, Appl. Phys. Lett. 2006, 88, 203107. [116] S. Koynov, M. S. Brandt, M. Stutzmann, J. Appl. Phys. 2011, 110, 043537. [117] R. Smaali, E. Centeno, A. Moreau, Appl. Phys. Lett. 2013, 103, 113905. [118] C. S. Schuster, A. Bozzola, L. C. Andreani, T. F. Krauss, Opt. Express 2014, 22, S2, A542. [119] Y.-J. Hung, S.-S. Hsu, Y.-T. Wang, C.-H. Chang, L.-Y. Chen, L.-Y. Su, J.-J. Huang, Nanotechnology 2011, 22, 485202. [120] A. Bessonov, Y. Cho, S.-J. Jung, E.-A. Park, E.-S. Hwang, J.-W. Lee, M. Shin, S. Lee, Sol. Energ. Mat. Sol. C 2011, 95, 2886. [121] B. Janthong, Y. Moriya, A. Hongsingthong, P. Sichanugrist, M. Konagai, Sol. Energ. Mat. Sol. C 2013, 119, 209. [122] J.-S. Cho, S. Baek, S.-H. Park, J. H. Park, J. Yoo, K. H. Yoon, Sol. Energ. Mat. Sol. C 2012, 102, 50. [123] U. W. Paetzold, W. Zhang, M. Prömpers, J. Kirchhoff, T. Merdzhanova, S. Michard, R. Carius, A. Gordijn, M. Meier, Ser. Mat. Sci. Engn. B 2013, 178, 617. [124] B. B. Van Aken, J. Löffler, M. C. Heijna, W. J. Soppe, J. Non-Cryst. Solids 2012, 358, 2268. [125] Q. Wang, T. Soderstrom, K. Omaki, A. Lennon, S. Varlamov, Energy Procedia 2012, 15, 220. [126] X. Yu, X. Yu, J. Zhang, Z. Hu, G. Zhao, Y. Zhao, Sol. Energ. Mat. Sol. C 2014, 121, 28. [127] B. Dou, R. Jia, H. Li, C. Chen, Y. Sun, Y. Zhang, W. Ding, Y. Meng, X. Liu, T. Ye, Sol. Energy 2013, 91, 145. [128] S. Yang, Y. Wang, L. Bai, B. Liu, J. Fan, X. Yang, H. Zhao, C. Wei, Q. Huang, X. Chen, G. Wang, Y. Zhao, X. Zhang, RSC Adv. 2013, 3, 208. [129] A. Bozzola, M. Liscidini, L. C. Andreani, Prog. Photovoltaics Res. Appl. 2014, 22, 1237. [130] P. Kowalczewski, M. Liscidini, L. C. Andreani, Opt. Express 2013, 21, S5, A808. [131] P. M. Sberna, G. G. Scapellato, N. Piluso, S. Boninelli, M. Miritello, I. Crupi, E. Bruno, V. Privitera, F. Simone, S. Mirabella, Appl. Phys. Lett. 2013, 103, 221902. [132] Z. Fang, H. Zhu, Y. Yuan, D. Ha, S. Zhu, C. Preston, Q. Chen, Y. Li, X. Han, S. Lee, G. Chen, T. Li, J. Munday, J. Huang, L. Hu, Nano Lett. 2014, 14, 765. [133] K. Vynck, M. Burresi, F. Riboli, D. S. Wiersma, Nat. Mater. 2012, 11, 1017. [134] E. Abrahams, P. W. Anderson, D. C. Licciardello, T. V. Ramakrishnan, Phys. Rev. Lett. 1979, 42, 673. [135] F. Riboli, N. Caselli, S. Vignolini, F. Intonti, K. Vynck, P. Barthelemy, A. Gerardino, L. Balet, L. H. Li, A. Fiore, M. Gurioli, D. S. Wiersma, Nat. Mater. 2014, 13, 720. [136] F. Riboli, P. Barthelemy, S. Vignolini, F. Intonti, A. De Rossi, S. Combrie, D. S. Wiersma, Opt. Lett. 2011, 36, 127. [137] M. Burresi, F. Pratesi, K. Vynck, M. Prasciolu, M. Tormen, D. S. Wiersma, Opt. Express 2013, 21, A268. [138] A. Oskooi, P. A. Favuzzi, Y. Tanaka, H. Shigeta, Y. Kawakami, S. Noda, Appl. Phys. Lett. 2012, 100, 181110. [139] R. Peretti, G. Gomard, L. Lalouat, C. Seassal, E. Drouard, Phys. Rev. A 2013, 88, 053835.





REVIEW

[140] F. Pratesi, M. Burresi, F. Riboli, K. Vynck, D. S. Wiersma, Opt. Express 2013, 21, A460. [141] A. Oskooi, Y. Tanaka, S. Noda, Appl. Phys. Lett. 2014, 104, 091121. [142] M. Burresi, F. Pratesi, K. Vynck, M. Prasciolu, M. Tormen, D. S. Wiersma, Opt. Express 2013, 21, S2, A268. [143] U. W. Paetzold, M. Smeets, M. Meier, K. Bittkau, T. Merdzhanova, V. Smirnov, D. Michaelis, C. Waechter, R. Carius, U. Rau, Appl. Phys. Lett. 2014, 104, 13. [144] H. C. v. d. Hulst, Light Scattering by Small Particles, Dover Publications, New York 1981. [145] E. D. Palik, Handbook of Optical Constants of Solids, Academic Press, London, UK 1991. [146] A. Paris, A. Vaccari, A. C. Lesina, E. Serra, L. Calliari, Plasmonics 2012, 7, 525. [147] E. Hao, G. C. Schatz, J. Chem. Phys. 2003, 120, 357. [148] K. Q. d. Costa, V. Dmitriev, J. Microwaves Optoelectron. Electromagnet. Appl. 2010, 9, 123130. [149] P. Mehlschlegel, H.-J. Eisler, O. J. F. Martin, B. Hecht, D. W. Pohl, Science 2005, 308, 57281607. [150] M. Agio, A. Ali, Optical Antennas, Cambridge University Press, UK 2013. [151] K. Kneipp, Y. Wang, H. Kneipp, L. T. Perelman, I. Itzkan, R. R. Dasari, M. S. Feld, Phys. Rev. Lett. 1997, 78, 1667. [152] S. Nie, S. R. Emory, Science 1997, 275, 53031102. [153] A. Dmitriev, C. Hagglund, S. Chen, H. Fredriksson, T. Pakizeh, M. Kall, D. S. Sutherland, Nano Lett. 2008, 8, 3893. [154] A. Bouhelier, M. Beversluis, A. Hartschuh, L. Novotny, Phys. Rev. Lett. 2003, 90, 013903. [155] M. Lippitz, M. A. van Dijk, M. Orrit, Nano Lett. 2005, 5, 799. [156] M. Sharma, P. R. Pudasaini, F. Ruiz-Zepeda, E. Vinogradova, A. A. Ayon, ACS Appl. Mater. Interfaces 2014, 6, 15472. [157] D. Derkacs, S. H. Lim, P. Matheu, W. Mar, E. T. Yu, Appl. Phys. Lett. 2006, 89, 093103. [158] K. R. Catchpole, A. Polman, Appl. Phys. Lett. 2008, 93, 191113. [159] S. Pillai, K. R. Catchpole, T. Trupke, M. A. Green, J. Appl. Phys. 2007, 101, 093105. [160] K. R. Catchpole, A. Polman, Opt. Express 2008, 16, 21793. [161] Y. Wang, B. Shao, Z. Zhang, L. Zhuge, X. Wu, R. Zhang, Opt. Commun. 2014, 316, 37.

[162] R. Mupparapu, K. Vynck, I. Malfanti, S. Vignolini, M. Burresi, P. Scudo, R. Fusco, D. S. Wiersma, Opt. Lett. 2012, 37, 368. [163] Z. Shao, X. Wang, S. Ren, Z. Tian, S. Fan, S. Sun, S. Liu, Q. Wang, J. Appl. Phys. 2013, 114, 063510. [164] F. B. Atar, E. Battal, L. E. Aygun, B. Daglar, M. Bayindir, A. K. Okyay, Opt. Express 2013, 21, 7196. [165] C. Clavero, Nat. Photon. 2014, 8, 95. [166] N. F. Fahim, Z. Ouyang, B. Jia, Y. Zhang, Z. Shi, M. Gu, Appl. Phys. Lett. 2012, 101, 261102. [167] P. J. Schuck, D. P. Fromm, A. Sundaramurthy, G. S. Kino, W. E. Moerner, Phys. Rev. Lett. 2005, 94, 017402. [168] Q. Xu, F. Liu, W. Meng, Y. Huang, Opt. Express 2012, 20, A898. [169] M. Zhang, T. H. Ngo, N. I. Rabiah, T. P. Otanicar, P. E. Phelan, R. Swaminathan, L. L. Dai, Langmuir 2014, 30, 75. [170] F. Pastorelli, S. Bidault, J. Martorell, N. Bonod, Adv. Opt. Mater. 2014, 2, 171. [171] D. M. O’Carroll, C. E. Hofmann, H. A. Atwater, Adv. Mater. 2010, 22, 1223. [172] D. Kozanoglu, D. H. Apaydin, A. Cirpan, E. N. Esenturk, Org. Electron. 2013, 14, 1720. [173] W. Jiang, M. Salvador, S. T. Dunham, Appl. Phys. Lett. 2013, 103, 183303. [174] J. E. Millstone, S. J. Hurst, G. S. Mtraux, J. I. Cutler, C. A. Mirkin, Small 2009, 5, 646. [175] S. J. Oldenburg, R. D. Averitt, S. L. Westcott, N. J. Halas, Chem. Phys. Lett. 1998, 288, 243. [176] J. R. Cole, N. J. Halas, Appl. Phys. Lett. 2006, 89, 153120. [177] D. Paz-Soldan, A. Lee, S. M. Thon, M. M. Adachi, H. Dong, P. Maraghechi, M. Yuan, A. J. Labelle, S. Hoogland, K. Liu, E. Kumacheva, E. H. Sargent, Nano Lett. 2013, 13, 1502. [178] M. D. Kelzenberg, S. W. Boettcher, J. A. Petykiewicz, D. B. Turner-Evans, M. C. Putnam, E. L. Warren, J. M. Spurgeon, R. M. Briggs, N. S. Lewis, H. A. Atwater, Nat. Mater. 2010, 9, 239. [179] J. Zhang, J. Gao, E. M. Miller, J. M. Luther, M. C. Beard, ACS Nano 2013, 8, 614. [180] C.-H. M. Chuang, P. R. Brown, V. Bulovi, M. G. Bawendi, Nat. Mater. 2014, 13, 796.



743

Replication of Leaf Surface Structures for Light Harvesting.

Light-harvesting pigment-protein complex deficiency in Hosta (Liliaceae).

Light-harvesting chlorophyll pigments enable mammalian mitochondria to capture photonic energy and produce ATP.

Aperiodic TiO2 nanotube photonic crystal: full-visible-spectrum solar light harvesting in photovoltaic devices.

Enhanced method for determining the optical response of highly complex biological photonic structures.

Aluminum Nanoarrays for Plasmon-Enhanced Light Harvesting.

Natural strategies for photosynthetic light harvesting.

Panchromatic absorbers for solar light-harvesting.

Membrane crystals of plant light-harvesting complex II disassemble reversibly in light.

b protein complex.

Evidence for a light-harvesting chlorophyll a-protein complex in a chlorophyll b-less barley mutant.

Mussel-inspired plasmonic nanohybrids for light harvesting.

PHOTOSYSTEM II SUBUNIT R is required for efficient binding of LIGHT-HARVESTING COMPLEX STRESS-RELATED PROTEIN3 to photosystem II-light-harvesting supercomplexes in Chlamydomonas reinhardtii.

Parasitic loss suppression in photonic and plasmonic photovoltaic light trapping structures.

Biogenesis of light harvesting proteins.

Nature-replicated nano-in-micro structures for triboelectric energy harvesting.

Enhanced light harvesting in dye-sensitized solar cells coupled with titania nanotube photonic crystals: a theoretical study.

Electromagnetic energy transport in finite photonic structures.

Assessing modulation of stromal and thylakoid light-harvesting complex-II phosphatase activities with phosphopeptide substrates.

Dephosphorylation of the thylakoid membrane light-harvesting complex-II by a stromal protein phosphatase.

Crystallization of the B800-820 light-harvesting complex from Rhodopseudomonas acidophila strain 7750.

Static and Dynamic Disorder in Bacterial Light-Harvesting Complex LH2: A 2DES Simulation Study.

A light-harvesting polyoxometalate-polypyridine hybrid induces electron transfer as its Re(I) complex.

b-protein complex in three different media.