Nanophotonics for current and future white light-emitting devices

The development of societies has been historically related to their ability to generate artificial light, from the discovery of fire in prehistoric times to ubiquitous light-emitting diodes (LEDs) of today. It is no wonder the illuminated areas around the world have only increased in recent decades.¹ The advent of blue LEDs in the last few years of the past century enabled the generation of solid-state white light (wLED) in an efficient, durable, reliable, inexpensive, and more environmental-friendly way,² which has opened the door to the widespread use of LED lamps for applications beyond general lighting from horticulture to healthcare or displays.³ In general, solid-state white light is achieved from the combination of LED chips, typically blue or UV, and color conversion layers (CCLs) made of photoluminescent materials, mainly inorganic microparticles doped with rare earth (RE) cations, i.e., phosphors, that provide down-shifting of the LED chip, being the chromaticity of the lamp generally dictated by material choice. Also, phosphor layers usually feature significant light scattering, being mirrors and lenses employed to shape the angular distribution of the emission for applications that demand directional light sources. Although YAG (Y₃Al₅O₁₂) crystals doped with Ce³⁺ (YAG:Ce) represents the current leading solution for color conversion to achieve white light in LED lamps, increasing demands on the properties of the emitting devices drive a continuous search of luminescence materials not only from systematic material synthesis,^4,5 but also from data-driven approaches.^6,7 Indeed, nanophosphors, organic fluorophores, transition metal complexes, biomaterials and semiconductor nanocrystals have also been investigated not only to tailor the chromaticity of LED lamps but also to offer additional properties such as transparency or flexibility, which ultimately provides design freedom.^8,9 However, current technology demands not only efficient luminescent materials but also device architectures that enable an effective extraction of the generated light. In fact, light outcoupling in most emitting devices is still a challenge, as it limits the overall performance of large-area flat emerging devices like organic or quantum dot LEDs.^10–12 From a broader perspective, the ubiquity of LED lamps coupled with the difficulty in achieving a precise control of the directionality of light generated is causing our planet to have never been as illuminated as it is today, especially at night,¹³ which compromises the cycle of bright day and dark night on the basis of which we evolved.¹⁴ Nevertheless, light regulates physiological processes that are essential for our health, impacting our mood and our ability to learn.¹⁵ As a consequence, solid-state lighting technology is urged to address key challenges in the next years in order to push the next generation of emitting devices. The development of novel materials and emitting devices should contribute to devising light sources capable of producing light where it is needed with the required chromaticity, with the aim of improving the efficiency of the lamps and providing them with new functionalities with a reduced environmental impact.

In this context, photonic architectures have already demonstrated a profound impact on the emission properties of light-emitting devices,¹⁶ thanks to their unique ability to enhance the light–matter interaction in the nanoscale.¹⁷ By the same token, they are expected to play a central role in the development of next-generation devices, as we elaborate herein. Specifically, in Sec. II, we discuss the most interesting materials investigated in the context of wLEDs. In particular, we will summarize the main advantages and most promising results attained with organic materials, quantum dots of semiconductors, metal-halide perovskites, or biomaterials, among others. In Sec. III, we show the most relevant achievements using photonic crystals, plasmonic nanostructures, or metasurfaces to tune the properties of emitting devices, while we summarize key opportunities and main challenges ahead in Sec. IV.

The past two decades have witnessed an enormous effort to improve the light quality and luminous efficacy of commercial inorganic white LEDs for lighting as well as for display applications. In parallel, a plethora of new luminescent materials have been considered as active components for this type of devices, which can overcome some of the limitations outlined above and introduce further functionalities such as lightweight, thin and flexible devices, and biocompatible and environmental-friendly approaches. In this section, we highlight the most promising ones, some of which have already been incorporated into commercial devices, and describe their main characteristics from the point of view of material but also device architecture in order to understand how nanophotonics may offer new opportunities to improve their performance or provide them with functionalities demanded by the marketplace. Here, we have focused strictly on those approaches for which devices have been demonstrated.

A vast amount of work has been performed on these devices¹⁸ ever since the first demonstration over two decades ago.¹⁹ The interest in this approach comes from the possibility of profiting from their improved functional characteristics (flexible, thin film lightweight devices) together with high efficiency and low energy consumption for solid-state lighting and display applications. Relying on the generation of light through recombination of singlet and/or triplet excitons in organic compounds, independent of the material employed, a number of device architectures have been employed to generate white light that comprises the use of single emitting layers (where a host–guest implementation is carried out) or a multilayer configuration either stacking different emitting layers or using a tandem structure. With recent approaches to overcome the intrinsic internal quantum efficiency (IQE) limitation of 25% imposed by spin statistics,²⁰ such as the use of phosphorescent materials based on organometallic complexes²¹ or the design of emitters for which thermally activated delayed fluorescence (TADF) is maximized,²² the main issues in this type of devices are still related to limited stability of organic emitters and its thin film architecture, which entails strong losses of the generated light to waveguided and surface-plasmon polariton (SPP) modes that severely hampers the light extraction efficiency (LEE).²³ 

Due to their ease of processing and outstanding emissive properties, such as color tunability and purity as well as high photoluminescence quantum yield (PLQY), quantum dots (QDs) have been thoroughly studied as emitters for display applications over the past two decades showing the potential to outperform OLEDs in terms of the achievable color gamut while sharing some of its benefits such as thin film device configuration or roll-to-roll processability, enabling flexible devices. As a matter of fact, red QDs have been demonstrated to improve backlight color saturation in phosphor-based displays.⁴ While their use as CCLs for displays has been exploited commercially for over 5 years, their use as electroluminescent (EL) material in QLEDs is being actively studied and they are rapidly approaching the performance of other thin film-based approaches such as OLEDs for devices with monochromatic emission.²⁴ In fact, recent simultaneous improvements in luminance and efficiency are bringing these materials closer toward general lighting applications.^25,26 For what concerns QLEDs for white-light emission (WQLEDs), several strategies have been explored over the past few years,²⁷ which are promising but still lack in terms of efficiency with respect to their monochromatic counterparts. These comprise the use of mixed R,G,B QD active layers [see Fig. 1(a)] whose initially low external quantum efficiency (EQE) (10.9%)²⁸ has risen to outstanding values close to 30% upon a precise control of energy transfer between the different QD components, as well as an effective managing of the outcoupling of the emission [Fig. 1(b)].²⁹ A means to circumvent this limitation, albeit complicating the device architecture, is to spatially separate the R, G, and B emitting layers in a stratified structure or using directly a tandem configuration where the latter have shown an outstanding record EQE well above 20%.^30,31 A similar tandem design but combining organic and QD emissive layers has recently been shown to lead to tunable and white light emission with over 20% EQE [Fig. 1(c)].³² 

Lead-halide perovskites (LHPs) have recently entered the list of solution-processed materials with tremendous potential for light-emitting applications,^33,34 after revolutionizing the field of photovoltaics reaching performances pairing those of commercial silicon solar cells in less than a decade.³⁵ Since the first demonstration of a LHP-based LED (PLED) some years ago,³⁶ EQEs have risen to values above 21% for monochromatic green,³⁷ red,³⁸ and NIR emission,³⁹ rivaling other solutions based on wet-processing techniques. Such rapid progress stems both from the excellent optoelectronic properties of these materials⁴⁰ and from lessons learned in device design from other thin film-based light technologies such as OLEDs or QLEDs. This similarity in device architecture also implies similar performance limitations such as the low LEE mentioned above⁴¹ and thus strategies to use nanophotonics to improve LHP-based devices have followed the path made by OLEDs and QLEDs in recent years (see below).⁴² 

While the best performing LHP-based solar cells employ polycrystalline films with large (micrometer) grain sizes, efficient PLEDs demand nanocrystalline films as active layers where spatial excitonic confinement leads to improved EQEs.⁴³ Alternatively, the use of low dimensional 2D and 2D/3D LHP, where mono or multilayers of inorganic lead halides are separated by long organic cations, has also appeared as excellent candidates as efficient light emitters. 2D LHPs, in particular, have emerged as a material presenting WL emission with the potential to be integrated into LEDs.⁴⁴ While PLQYs are still modest (ca. 9%), the possibility of achieving WL without the need of having separated chromatic components makes these materials highly desirable for emitting devices.

Currently, LHPs have been introduced into white light-emitting PLEDs mostly as CCLs incorporating LHP QDs or NPs into transparent matrices that are later coupled to commercial UV or blue LED chips.⁴⁵ Following this architecture, only one all-LHP device has been demonstrated due to the low QY of blue emitting LHP, an issue that extrinsic doping with Mn ions could solve.⁴⁶ This latter approach may benefit from devices presenting recent efficiency records (EQE = 12.3%) for blue PLEDs employing NCs.³⁷ Further, alternatives that incorporate single active LHP WL emitting layers relying on low dimensional perovskites have been presented opening an exciting path for the future avoiding the need for multi-component architectures.^47–49 Finally, one major bottleneck for the commercialization of lead-halide perovskites into which much effort is being placed comprises its instability arising from the interaction with its surroundings.⁵⁰ 

Biological materials are a seamless source of inspiration for technology and the development of wLEDs is no exception.⁵¹ Their relevance comes from a combination of low-cost processing, earth abundance, bio-degradability, and amenability to be implemented into devices through low temperature processes with novel functionalities such as bio-compatibility or thin flexible architectures. Bio-inspired materials have been used to improve or complement different elements of LEDs to the extent of becoming a research field in itself. In fact, some approaches have been proposed to incorporate biological emitters into the active layer of OLEDs which led to proofs of principle but modest performances (the reader is referred to Ref. 51 and references therein). Regarding the role of biological materials as actively emitting components, different biomaterials amenable to be casted as thin films have been used as CCLs. Due to their supramolecular structure, they can host different fluorophores (quantum dots, organic dyes, carbon nanodots, etc.) without affecting their emissive properties or even improving them. By stacking films doped with fluorophores emitting in the R, G, and B layers or using combinations of fluorophores undergoing Förster energy transfer, WL emitting CCLs have been developed using DNA,⁵² rubber encapsulated biomolecules [see Figs. 1(d) and 1(e)],⁵³ bovine serum albumin (BSA) [see Figs. 1(f) and 1(g)],⁵⁴ cellulose,⁵⁵ or silk fibroin⁵⁶ to name but a few. While this approach opens the door to the future incorporation into commercial devices, relevant issues such as the short duration of the emissive properties (chromaticity and intensity) of the fluorophores or thermal instability will have to be tackled.

The search for WL emitting materials is an active research topic nowadays mostly inspired by lighting and display applications. A number of reviews and perspectives have appeared on emerging alternatives, and in this section, we will just provide a brief description of some of the most popular ones.

Nanophosphors are RE-doped inorganic nanoparticles that arise as nanoscale counterparts of standard phosphors. In general, phosphor layers are clusters of micrometer-sized crystals and thus its opacity is unavoidable due to the significant light scattering. Limitations in the physical understanding of the combined multiple light scattering and energy conversion in commonly used phosphors hamper the design and development of phosphor-converted wLEDs with tailored properties. In this context, nanophosphors offer new opportunities to develop transparent emitting thin films with large chemical and thermal stability as shown in Figs. 1(h)–1(j).^57,58

Metal organic frameworks (MOFs) and coordination polymers (CPs) have been thoroughly studied as alternatives to conventional RE-based phosphors for WL generation over the past decade.^59,60 In these systems, a porous crystalline solid is formed by single or clusters of metal ions linked by organic ligands which can further host a vast variety of molecular emitters. The main attraction of these systems from the point of view of WL generation lies in the possibility of designing their emission, which can come from the metal ion, organic ligand, guest species, or processes involving energy transfer among different components.

Several strategies to design the sought for emission have been followed. These comprise the generation of WL from MOFs containing rare earth metals (lanthanides or actinides) either incorporated into the framework or as guest molecules. Also, RE-free approaches have been explored where different molecular emitters such as organic dyes or iridium complexes have been incorporated into the pores of the structure. Beyond their use as phosphors for wLEDs, proof of concept of the use of MOFs and CPs as EL materials have been presented over the past few years with poor (∼1%) quantum efficiencies as a result of the low ligand-metal conductivity. The reader is directed to Refs. 59 and 60 and references therein. Although design flexibility and the possibility of introducing a large variety of components allow for a precise tuning of their chromatic properties and quantum yields close to 100% have been reported, MOFs and CPs suffer from a number of limitations that will need to be tackled in the future. These comprise a difficulty to scale synthesis and, as in the case of hybrid perovksites mentioned above, stability against moisture or heat.

Beyond MOFs and CPs, many hybrid organic-inorganic WL emitters have been developed⁶¹ including different organically templated inorganic compounds (phosphates, sulfates, etc.). Here, the generation of broadband emission relies on different mechanisms depending on the interaction between the organic and inorganic phases. Among the single-phase white light emitters, organic compounds have also attracted interest over the past decade due to the possibility of developing a single emitter with broadband emission which can be casted in the shape of a thin film. To date, a vast range of molecular emitters with broadband emission upon UV/blue excitation have been reported. The mechanism for WL generation ranges mainly from supramolecular aggregation by intermolecular π- π stacking,⁶² excited-state intramolecular proton transfer,⁶³ or aggregation-induced emission.⁶⁴ The main limitation, beyond the instability characteristic of organic compounds, is associated with the need for complicated synthetic routes in many cases.

The use of two-dimensional (2D) materials as transparent electrodes or charge transporting/insulating layers in optoelectronic devices has been widely documented over the past few years. Regarding their use as active materials in LEDs only recently, with the introduction of van der Waals heterostructures, reports have emerged⁶⁵ taking advantage of the possibility of stacking direct-bandgap 2D transition metal dichalcogenides (TMD) without the restriction of lattice matching and tailoring their physical properties by controlling the stacking. This field is still at its infancy and reports for monochromatic LEDs are scarce and present modest efficiencies.⁶⁵ Nevertheless, the demonstration of wLEDs⁶⁶ or broadband emission⁶⁷ with TMD vertical heterostructures is certainly an exciting starting point. While large scale fabrication is surely a shortcoming in the development of this approach, other aspects limiting its performance such as current leakage will need to be tackled. The use of nanophotonics in these structures faces the challenge of selectively acting on the thin (few atomic layers) active area, where plasmonic systems could play a role.

Finally, a non-toxic, chemically stable, and easy to process alternative approach to the above listed luminescent nanomaterials for application in wLEDs is that of carbon dots (CDs). After their proposal in 2004,⁶⁸ this new class of light emitting nanomaterial has been broadly studied and incorporated into several optoelectronic devices, including wLEDs. This has been done both as CCLs,⁶⁹ where performances close to those of commercial devices have been achieved, and as active elements, either emitting or charge injection layers.⁷⁰ In the latter case, reported performances are well below those of other devices with similar architectures. While some fundamental approaches should be followed to improve their efficiency, from preventing aggregation of CDs in the device to improving the CD synthesis, nanophotonics could certainly play a key role in this process. Tailoring the field intensity enhancement within the active layer, following similar approaches used in other thin film devices could improve the CD emission within the host matrix where they are commonly dispersed to reduce aggregation.

Beyond the search for new materials that overcome the above-mentioned issues of commercial wLEDs, the possibility of improving device performance by introducing a structuration of some of its elements has been widely explored over the past few decades. While surface roughening is a well established approach to enhance the outcoupling from LEDs, the advantages of introducing a controlled structuration on the scale of (and below) the wavelength of the emitted light have been actively studied. Such structuration allows for a spatial and spectral redistribution of the density of electromagnetic states through a material on a sub-micrometer scale and this permits a precise control on its emitting properties. The unique advantage of this approach is that it is a transversal one that can be applied to the above plethora of materials that are currently being investigated as an alternative to commercial devices (see Fig. 2).

The efficiency of a light-emitting device is given by the product of three partial efficiencies: (i) excitation efficiency, which represents the fraction of injected electron hole pairs that recombine in the emitting layer for electroluminescent devices or the fraction of the excitation light absorbed by the CCL in color-converting devices; (ii) radiative efficiency, which represents the fraction of injected or optically pumped carriers that recombine radiatively emitting a photon; and (iii) extraction efficiency, which represents the fraction of generated light that escapes from the device. Besides, color rendering index (CRI) and correlated color temperature (CCT) are key parameters to assess the performance of a white light source beyond power conversion efficiency. In particular, high CRI sources are especially required for applications in which faithful reproduction of colors is of utmost importance, such as photography, retail, or museum lighting. CTT also impacts the appearance of the white light source. In general, low CCT (warm) sources are often employed to create relaxing environments, whereas high CCT (cold) sources with higher blue components induce alertness and are normally used at work. There is a growing interest to develop sources in which its color content can be accurately controlled, aiming at mimicking the spectral composition of sunlight throughout the day. This is also significant for the design of outdoor lighting in environmentally sensitive areas since the impact of nighttime illumination on ecological interactions is still mainly unknown.⁷¹ Nanophotonics can introduce extremely large variations in the local density of optical states (LDOS) of a material, which may affect excitation, radiative, and extraction efficiencies. Nevertheless, light-emitting applications demand not only the enhancement of the intrinsic properties of a given material (radiative rate and chromaticity) but also the improvement of the extraction of the generated light from the actual device in a manner that is independent of wavelength and viewing angle and that should not interfere with charge transport across the structure. In this sense, a number of approaches have been proposed over the past 10 years, each carrying its own advantages and limitations. In what follows, we will discuss some of the more promising ones.

Photonic crystals (PhC) rely on the multiple scattering of light by a periodic modulation of the refractive index of a dielectric to alter the way electromagnetic radiation is generated inside or propagates through it. Introducing such structuration in the surface of blue/UV LEDs has been used as a means to improve its outcoupling efficiency^72,73 as well as its directionality^74,75 by coupling the guided modes within the LED to the PhC modes. This improvement route has been further extended to other thin-film technologies suffering from severe guided losses such as OLEDs.⁷⁶ In a recent example, a PhC structure was used to funnel light out of a Br-based PLED as illustrated in Figs. 3(a) and 3(b). The PhC converts guided modes to leaky modes, which favors the outcoupling of the generated light in the nanostructured device, as pointed out by numerical simulations plotted in Fig. 3(c). As a result, the EQE reaches 17.5% (in stark contrast with the 8.2% featured by a reference planar device) due to high LEE (73.6%) in the green.⁷⁷ 

The main issue with this approach is that the discreet nature of the dispersion relation of PhC strongly limits the broadband application of these materials and thus its use for white LEDs has been less exploited. Some reports have appeared on the effect of introducing a periodic patterning on the surface of the phosphor layer of wLEDs containing transparent polycrystalline ceramic conversion plate phosphors as a means to mitigate the drawbacks associated with total internal reflection.^78–82 In these systems, the patterned region (either the top surface of the phosphor or a deposited structured dielectric such as SiO₂, TiO₂, or SiN_X) represents <1% of the total phosphor thickness (on the order of hundreds of micrometers). Here, results on improvements upon surface structuration on light extraction, directionality, and chromatic properties have been reported. While there is a consensus that these changes are due to a combined effect of backreflection of the UV/blue light (enhancing its absorption by the phosphor layer) and extraction of the yellow component by the PhC film, a detailed correlation of the observed changes with the mode structure of the PhC has not been presented and thus a precise design to optimize these effects is lacking.

A route to circumvent the limitations of PhC regarding their discreet Fourier transform is the use of deterministic aperiodic nanostructures.⁸³ Here, one can rely on precise design rules to fabricate 2D arrays of nanostructures possessing a richer Fourier spectrum than periodic systems, which could extend the scattering properties of the photonic architecture to a broader spectral range. This could, in principle, enhance the extraction efficiency of broadband emission generated within a wLED while not introducing artifacts in the directionality associated with diffraction by periodic lattices. This approach was first implemented over a decade ago in the shape of deterministic aperiodic arrays for the case of monochromatic GaN LEDs,⁸⁴ but its applicability certainly becomes more relevant for the case of wLEDs. In particular, demonstration of the application of this type of structure has been presented for devices of different nature with commercial applications. The performance of wOLEDs was improved by introducing a designed “vacuum nanohole array” within the ITO layer [see Fig. 3(d)], which allowed for an impressive EQE of 78% and LE of 161 lm/W in a device containing a dual blue/orange emitting layer to generate WL, as shown in Figs. 3(e) and 3(f).⁸⁵ 

Another recent demonstration of this type of structuration is the use of Vogel spiral arrays defined atop of a standard YAG:Ce CCL deposited on a blue-emitting LED.⁸⁶ It was demonstrated that TiO₂ nanodisks arranged in this particular way [see Figs. 3(g) and 3(h)] lead to both enhanced extraction, 2.7-fold enhancement displayed in Fig. 3(i), and directionality (35% peak forward enhancement) of a device mimicking commercial LEDs. This represents one of the few examples in which photonic design has shown to be effective to improve the performance of standard thick phosphor-converted wLED.

The use of surface plasmon polaritons (SPPs) in nanostructured noble metals has been widely explored in the field of LEDs for more than a decade to improve their performance and overcome their limitations.^16,87 Either if one relies on resonances of isolated plasmonic nanostructures or collective modes resulting from periodic arrays, plasmonic systems can have two effects on the performance of electrically pumped devices. On the one hand, the total electric field redistribution in the close vicinity of metallic nanostructures accompanying the presence of SPP can lead to an enhanced radiative efficiency and QY of the emitter when it is placed close to the nanostructure. On the other hand, SPP arrays can act as efficient scatterers, which provide the means to enhance the extraction efficiency or tailor the directionally of an LED. Devices with thin actively emitting layers are best suited for this approach due to the large spatial confinement associated with SPP. This effect, caused by a modification of the radiative component of the emitter decay rate, is particularly beneficial when the QY of the emitter is low.⁸⁸ Further, it is suited for LEDs with monochromatic emission as a consequence of the narrow-band nature of the SPP and was initially demonstrated for inorganic LEDs,⁸⁹ OLEDs,⁹⁰ and QLEDs.⁹¹ When wLEDs are considered with an emission band much broader than typical SPP modes, alternative approaches are needed. Here, the particular architecture of the device under consideration will determine the strategy to follow. For devices having stratified thin emitters, where different colors are generated at different layers, NP arrays can be judiciously designed and integrated at the appropriate device depth. An example is the introduction of a single film of metallic NPs to enhance simultaneously the EL from the blue emitting layer of an OLED as well as the PL of the red color converting adjacent layer⁹² leading to a 26% improvement in current efficiency. For the case of QLEDs where green and red CCLs are sequentially deposited on top of a blue LED, a proper design of two NP arrays to improve the QY of each type of QD has been demonstrated⁹³ opening the path toward a design concept which could find application in commercial QLEDs present in TV displays. A recent proof of concept,⁹⁴ not implemented in a device yet, made use of a plasmonic array of subwavelength silver pillars on top of which one can deposit a thin film of mixed R-G-B QDs, as illustrated in the sketch of Fig. 4(a). Choosing the appropriate array parameters one can tune its resonance which enhances the PL from each QD component and allows tuning the CIE coordinates of the film emission [see Figs. 4(b) and 4(c)]. In this way, one can circumvent one of the main limitations of thin film-based LEDs, i.e., the definition of pixels with reduced (∼μm²) dimensions, as illustrated in Fig. 4(d). On another recent example, an approach to improve the performance of a wLED with the conventional LED + CCL architecture making use of the potential of SPP to enhance the radiative component of the decay rate of the phosphor was proposed. Here, the SPP from individual Au NP was employed to improve the low conversion efficiency of YAG:Ce³⁺ and thus reduce the amount of phosphor used in a device as well as improving the color rendering index of the LED. This was done by decorating YAG:Ce³⁺ NP particles with Au NP where an enhanced color conversion was attributed to enhanced PLQY of the phosphor NPs as a consequence of the modified LDOS induced by the Au NP.⁹⁵ Other solutions to improve the performance of wLEDs are restricted by the device geometry as is the case of top-emitting OLEDs. Here, a cavity is formed naturally in the device due to the use of metallic electrodes. A proper choice of layer thickness can lead to the formation of optical Tamm modes, characteristic of 1D PhC-metal interfaces, which improves light extraction together with chromatic homogeneization.⁹⁶ 

As for the modification of the emitter QY, enhanced scattering of conventional periodic arrays takes place over a narrow spectral range and thus particular implementations will be demanded by given device geometries. For the case of OLEDs where severe waveguided losses constitute an important drawback, outcoupling strategies represent the main improvement and several approaches have been explored. All of them contain structurations of a different sort that are present across the entire device thickness. One approach comprises the use of buckling patterns that form when depositing aluminum films on top of a PDMS layer. These patterns present a broad periodicity distribution and randomly oriented wave-vectors, which allow for efficiently extracting guided modes over a broad spectral range within the OLED doubling the extraction efficiency.⁹⁷ Another patterning avenue explored the introduction of a dual metallic grating whose resonances, associated with the two periodicities, matched the emission bands of the single emitting layer consisting in a blue emitting fluorescent complex doped with an orange phosphorescent material.⁹⁸ Here, 37% and 48% enhancement in current efficiency and external quantum efficiency, respectively, were reported while maintaining the viewing characteristics. Finally, deterministic aperiodic structures were implemented by means of nanoimprint lithography into a wOLED which, in combination with an external aperiodic array of nanocones imprinted in a resin, could mitigate waveguided and TIR losses doubling the EQE and LE values of a reference planar structure.⁹⁹ For conventional wLEDs comprising a thick phosphor CCL deposited on a blue inorganic LED, it was recently proposed that a plasmonic array could be used to improve the device performance—see pictures displayed in Fig. 4(e). An aluminum nano-cylinder array was grown on a 200 μm YAG:Ce³⁺ layer and a fivefold enhancement in extraction efficiency in the normal direction was reported and associated with scattering by the modes of the array [see Fig. 4(f)].¹⁰⁰ Besides, the plasmonic array enhances the overall efficiency of the CCL by up to a factor of 2.3. As a result, with the aid of plasmonics, the 200 μm plate behaves roughly as if it would be 350 μm thick, as shown in Fig. 4(g). For LEDs possessing a thin emitting layer, a different approach has been explored over the past few years comprising the use of arrays of metal NPs.¹⁰¹ Here, the limited spatial reach of localized SPP associated with individual metal NPs is compensated by the appearance of collective modes originating in the scattering by the NP array and, in the presence of an optical WG, waveguide plasmon polaritons (WPPs) which extend throughout the volume of the WG. Here, the possibility of enhancing the emission and improving the directionality, as illustrated in Figs. 4(h) and 4(i), of a red phosphor consisting of a 700 nm polymer film doped with organic dyes was demonstrated.¹⁰² This latter approach could be used for other devices showing a similar architecture such as LEDs comprising emerging materials as CCLs (see Sec. II).

Finally, it is worth mentioning that the use of plasmonic structures to improve LED performance can involve, as mentioned above, the incorporation of metallic structures in the region comprised between electrodes. This can compromise charge transport within the LED and thus, beyond a detailed optical design of the device, an electrical study should be carried out in parallel, something which is seldom tackled due to the difficulty entailed by simulating the role in charge transport of subwavelength metallic structures.

Optical metasurfaces (OMs), in the shape of subwavelength patterned layers, have emerged as a powerful tool for light manipulation over the past decade.¹⁰³ Relying on the scattering properties of subwavelength structures of different nature, they can manipulate the polarization, phase, or dispersion of light and have been incorporated into conventional monochromatic LEDs as proofs of principle to improve their performance. Beyond their novel functionalities, the possibility to achieve them with elements of such reduced dimensions has gathered the interest of part of the nanophotonics community. OMs have been used as backreflectors to enhance polarization selectivity,¹⁰⁴ combined with a vertical cavity, as a means to improve directionality as well as controlling the wavefront of the LED [see Figs. 5(a)–5(c)],¹⁰⁵ or to embed quantum well structures able to emit unidirectionally at devised angles [see Figs. 5(d)–5(f)]¹⁰⁶ or focused beams.¹⁰⁷ OMs can be fabricated from metallic or all-dielectric components, avoiding the ever-present limitation of plasmonic systems associated with ohmic losses. Further, and contrary to other nanophotonic approaches exhibiting a narrow spectral response, OMs can be tailored to display broadband behavior and have been integrated into wOLEDs where a combination of a grating and an OM has been demonstrated as a powerful means to enhance light extraction from the device while improving its polarization properties.¹⁰⁸ While fabrication approaches based on commonly used electron beam lithography or focused ion beam offer outstanding control on the precise translation of the metasurface design onto a given substrate, they clearly pose a great challenge toward their industrial commercialization. In this sense, the introduction of alternative fabrication approaches such as nanoimprint lithography or photolithography could overcome this limitation. For the particular case of photolithography-based techniques, the demonstration of large scale (using up to 12 in. wafers) fabrication on silicon or glass substrates brings the use of this nanophotonic approach a step closer to commercialization.¹⁰⁹ 

The field of nanophotonics offers a range of phenomena regarding the light–matter interaction with an exciting prospect for circumventing some of the limitations of current state-of-the-art wLEDs for display and lighting. Due to the wide variety of device geometries, where light generation can take place within the active layer between the electrodes or at a CCL above, a good-for-all recipe cannot be defined a priori. Thus, depending on the device architecture, some guidelines can be extracted that could pave the road to devices with improved light extraction as well as tailored chromatic and directional properties.

• For devices possessing a thick (100s μm) CCL, approaches based on local field enhancement are not well suited. Instead, nanophotonic arrays that can improve outcoupling through multiple scattering could be a better solution. In particular, those relying on aperiodic arrays with an associated broad Fourier space that can, upon a proper design, improve extraction and directionality over a broad spectral range.

• For devices relying on thin active emitting films contained between electrodes, several approaches can be followed. Plasmonic arrays can be introduced in close vicinity of the emitters to simultaneously enhance its QY and outcoupling of light. Further, an external structuration can be used with appropriate designs to improve extraction and directionality.

While the appropriate combination of photonic nanostructures and optoelectronic devices has been demonstrated to lead to an enhanced performance, the feasibility of mass production may pose a strong limitation on the short-term implementation of nanophotonic lighting. More so in an industrial field where many device designs are robust and durable, and altering the fabrication process can introduce elevated costs. Nevertheless, the recent developments in large-scale fabrication of nanostructures mentioned above can pave the way to their incorporation into real-life applications in the coming years. As mentioned throughout the paper, these improvements will not only benefit commercial wLEDs available in the market but likely next generation devices relying on thin film active layers (such as QLEDs or PLEDs). Beyond propelling the development of more efficient lamps or displays, nanophotonics offers new functionalities such as precise polarization or wavefront control, as demanded for visible light communication or optogenetic systems, which could be added to next-generation LEDs. It also opens the door to the design of light sources in which the spectral content of the emission can be precisely tailored. This might be relevant for horticulture or healthcare but also for the environment as it could mitigate the impact that artificial lighting has on the ecology. In short, nanophotonics represent the most promising control tool to generate light with specific characteristics just where it is required, which is key to attain smart light sources that are sustainable and environmentally responsible beyond being efficient.

Financial support was provided by the European Research Council (ERC) under the European Union's Horizon 2020 Research and Innovation Programme (NANOPHOM, Grant Agreement No. 715832).

Data sharing is not applicable to this article as no new data were created or analyzed in this study.