Recent Progress in Emerging Near-Infrared Emitting Materials for Light-Emitting Diode Applications

Yingqi Zheng; Xiaozhang Zhu

doi:10.1055/s-0040-1716488

Organic Materials, Table of Contents

CC BY-NC-ND 4.0 · Organic Materials 2020; 02(04): 253-281
DOI: 10.1055/s-0040-1716488

Review

Recent Progress in Emerging Near-Infrared Emitting Materials for Light-Emitting Diode Applications

Yingqi Zheng

^aBeijing National Laboratory for Molecular Sciences, CAS Key Laboratory of Organic Solids, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190, China

^bSchool of Chemical Sciences, University of Chinese Academy of Sciences, Beijing 100049, China

,

Xiaozhang Zhu

^aBeijing National Laboratory for Molecular Sciences, CAS Key Laboratory of Organic Solids, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190, China

^bSchool of Chemical Sciences, University of Chinese Academy of Sciences, Beijing 100049, China

› Author Affiliations

Abstract

Full Text

PDF Download All articles of this category

Key words

near-infrared emitting materials - materials design - organic light-emitting diodes - perovskite light-emitting diodes - quantum dot emitting diodes

Introduction

ISO 20473 defines infrared (IR) light as an electromagnetic radiation with a wavelength between 780 and 1 mm, which can be divided into three parts: near-IR (NIR; 780–3000 nm), mid-IR (MIR; 3–50 μm), and far-IR (FIR; 50–1000 μm).[1] Almost half the energy reaching the earth's surface from the sun is in the form of NIR radiation.[2] Thus, materials used in photovoltaics must absorb NIR radiation effectively. NIR light has been widely applied in biomedicine, night vision devices, and telecommunications.[3] [4] [5] For example, NIR light is advantageous for biomedical and biosensing applications because of the weak absorption and autofluorescence of biological tissues in the NIR region. Moreover, 700–1000 nm are included in the “semitransparent window” for biological tissues, so it is possible for NIR light-emitting diodes (LEDs) to be integrated into photodynamic therapeutic and bioimaging applications due to the enhanced blood circulation or photoactivation of some drugs by NIR light.[6] NIR devices can be integrated into through-space optical communication applications, which present the characteristics of anti-interference, security, high-speed, and invisibility as a result of the penetration and independent transmission of IR light.[4] [5] In addition, NIR LEDs are used in biometrics such as identification of finger veins and irises[7] and in night-vision readable displays.[1] Thus, development of effective NIR-emitting materials is important for many areas of technology.

There has been much recent progress in the field of LEDs with enhanced performance, and various commercial applications have been developed. Nowadays, many LEDs are based on a structure in which the emitting layer is sandwiched between the cathode and the anode for electron and hole injection. For a multilayer structure, the device may contain electron- or hole-transport layers. Driven by the applied voltage, injected electrons and holes can drift and diffuse in the device, ending up with bounded electron–hole pairs, namely excitons, for radiative and nonradiative recombination [Auger recombination or Shockley–Read–Hall (SRH) recombination; [Figure 1]]. There are several key parameters, including turn-on voltage (V _on), driven voltage, external quantum efficiency (EQE), photoluminescence (PL) quantum yield (PLQY), luminance (cd m⁻²), and full width at half maximum (FWHM), which serve as figures of merit for performance evaluation of the LEDs. V _on is the voltage at which the device's luminance achieves 1 cd m⁻² and the driven voltage is the voltage at which the device is working properly. V _on indicates whether the electrodes contact well with the active layers in the devices and the driven voltage determines the power efficiency (defined as the ratio of the luminance to the driven power, which is the product of the driven voltage and current) and the device lifetime. Furthermore, their luminance and FWHM are important parameters indicating the brightness to the human eye and the color purity of the LEDs, respectively. Different from luminance, radiance (W sr⁻¹ m⁻²) is defined as the flux density of radiant energy per solid angle and per unit of projected area of radiating surface and is measured by a machine, not by human eyes. In general, the EQE, defined as the ratio of the number of photons emitted from the device to the number of charge carriers injected into it, is the key parameter for LEDs. Because light can be absorbed, trapped, or reflected by the multiple layers in the device, EQE is the product of light outcoupling efficiency (η _out), namely light extraction efficiency, and the internal quantum efficiency (IQE), defined by the ratio of photons emitted from the emitting layer to the charge injected into the device. Thus, the parameter can be calculated by [Eq. (1)]:

where γ is the recombination efficiency of injected electrons and holes; in other words, the balance between injected electrons and hole population and under ideal conditions, γ = 100%. The term η _r can be defined as the radiative exciton ratio in the emitting layer and η _PL is the PLQY of the emitting layer. For traditional organic fluorescent materials, η _r(max) is 25% because of the spin selection rule in the exciton formation process, which states that traditional organic fluorescent materials can only harvest singlet excitons for emission in electroluminescence (EL).[8] However, organic phosphorescent, thermally activated delayed fluorescence (TADF), and emitting materials based on a “hot-exciton mechanism” can make full use of triplet excitons, so their theoretical η _r(max) is 100%. The value of η _r is 1 for perovskite materials because they are direct-bandgap semiconductors. All these important metrics are used in LED evaluation, and the development of new materials for high-performance LEDs is very important.

Figure 1 Schematic representation of organic, perovskite, and QD materials and the typical structure of OLEDs, QLEDs, or PeLEDs.

Generally speaking, NIR-emitting materials can be roughly classified into inorganic (III–V semiconductors and nanocrystals), organic (small molecules, polymers, organometallic complexes), and hybrid organic–inorganic [quantum dots (QDs), perovskites] materials. In view of the high temperature and vacuum-based processing of the traditional inorganic semiconductors such as the III–V semiconductors, organic materials with easy chemical modification and solution processability, QD materials with high luminescence efficiency and size-dependent emission, and perovskite materials with excellent carrier-transport properties have emerged as suitable candidates for NIR-emitting materials. For example, organic materials with nontoxic elements exhibit good biocompatibility, so they can be used in biomedical in vivo applications,[6] and their solution processability enables large-area fabrication of organic LEDs (OLEDs) when compared with inorganic ones. QD LEDs (QLEDs) are also promising applications for large-area flat-panel displays because of the solution processability, high color purity, and luminescence efficiency of QDs. Considering the high carrier mobility, the low-to-moderate ionization energy, the tunable emission depending on the component ratios, and the two- or three-dimensional (2D or 3D) structure of perovskite materials, it is reasonable to assume a luminous future for the application of perovskite LEDs (PeLEDs).

In this review, we focus on three kinds of emerging materials, i.e., organic materials, QDs, and perovskites as emitting layer materials for NIR emission with wavelengths ranging from 670 to 1400 nm. In the following sections, representative published works in the NIR LED region are reviewed, the challenges that have been and need to be addressed are discussed, and an outlook for the developments in these three kinds of materials is presented.

Organic NIR Fluorescent Materials

Organic materials are widely applied emitting materials because of their tailorable molecular structures, low cost, flexibility, light weight, and tunable emission. To obtain organic materials that emit NIR light, the basic principle is to lower the optical bandgap that is determined by the HOMO and LUMO levels according to molecular orbital theory. The bond-length alternation (BLA) is the average difference in bond length between adjacent single and double C–C bonds in a conjugated molecule, which indicates the degree of electron delocalization. The introduction of different electron-rich (electron donor, D) and electron-deficient (electron acceptor, A) units wisely to build a push–pull system is an established approach that is widely used for tuning the molecular energy levels and bandgaps in organic materials due to the hybridization of the energetic levels of the donor and acceptor units and the changed BLA. Generally, in a conjugated system the bandgap narrows when the BLA is reduced by judicious modifications of molecular structures. The introduction and stabilization of quinoidal resonance in aromatic conjugated molecules is also an effective way to reduce the bandgap, which can decrease BLA, and increase the HOMO and lower the LUMO levels.[1] [9] [10] [11]

Organic conjugated molecules with narrow bandgaps usually have the disadvantage of a relatively strong vibrational coupling between the high vibrational energy levels in the ground state and the low vibrational levels in the first excited state, so the nonradiative transition rate increases exponentially with the decreasing bandgap, which is also called the “energy-gap law.” Moreover, due to their extended π-conjugated lengths, most NIR-conjugated molecules favor a planar structure, which tends to be affected by the aggregation-induced quenching effect. For example, porphyrin is a planar molecule with a four-pyrrole ring and exhibits strong absorption in the red-to-NIR region due to its extended conjugation, but it also presents strong aggregation-induced quenching, which is considered disadvantageous. Intriguingly, in 2001 Tang et al. proposed that molecules with a propeller-shaped structure illuminate brightly in the condensed state, although in solution the emission is very weak; this effect was named “aggregation-induced emission (AIE).”[12]

In EL, the ratio of singlet excitons to triplet excitons is 1:3 in traditional organic emitting materials due to the spin selection rule.[13] Thus, the IQE is limited to 25% in fluorescent OLEDs based on traditional materials because fluorescence originates from the radiative recombination of singlet excitons. To circumvent the drawback, some molecules with new luminous mechanisms were developed, such as TADF, triplet fusion (TF), and doublet emission, which may break the spin selection rule. For OLEDs based on phosphorescent emitters, the radiative exciton ratio can reach up to 100% in EL. Thus, phosphorescent organometallic materials are alternative choices to avoid the disadvantage and will be discussed in section “NIR Organometallic Phosphorescent Materials.” The traditional organic NIR fluorescent materials that obey the spin selection rule and the unconventional organic NIR fluorescent materials that violate the spin selection rule are discussed in the following sections ([Table 1]).

Table 1
Summary of typical organic NIR-emitting materials and their corresponding EQE values and EL maximums
Emitter	Host	EL peak (nm)	Doping ratio (wt%)	EQE (%)	Ref.
S1	Alq₃	752	2	1.12	[68]
S2	Alq₃	748	2	1.23	[68]
S3	Alq₃	823	2	0.27	[68]
S4		690		3.13	[69]
S5	PIDT-2TPD	840		1.15	[33]
S6	Alq₃	749 802	2	1.15	[52]
S6	Alq₃	749 802	2	0.43	[52]
S7		864		0.20	[52]
S8		706		0.89	[52]
S9		749		0.29	[52]
S10	MEH-PPV/PBD	725		0.5	[72]
S11		675		2.58	[71]
S12	F8BT	720	0.5	1.1	[64]
S13		1050		0.16	[51]
S14		1050		0.33	[51]
S15		1080		0.73	[51]
S19	SY-PPV	730		0.65	[28]
S20		760		1.9	[47]
P1		740		0.2	[16]
P2		708 712		0.15	[43]
P2		708 712		0.05	[43]
P3		723		0.30	[43]
P5	P8	895	1	0.091	[44]
P7	P8	939	3	0.06	[44]
P12		880		0.27	[55]
P14		824		0.3	[49]
P18		650		0.7	[61]
P19		670		1	[39]
P23		900		0.15	[35]
P26		970		0.05	[54]
T1	TPBi	668		0.9	[91]
T2	TPBi	728	100	3.9	[87]
T2	TPBi	644	15	14.5	[87]
T3		728		0.064	[86]
T4		715		0.254	[86]
T5	Zn(BTZ)₂	710 (10 wt%)	10	7.8	[93]
T5	Zn(BTZ)₂	728 (20 wt%)	20	5.1	[93]
	ADO-TPA	735		2.7	[93]
	TPBi	777	0	2.19	[92]
	TPBi	693	10	10.19	[92]
T6	CBP	716	4	8.53 (±0.27)	[95]
	CBP	721	6	9.69 (±0.11)
	CBP	730	7	8.09 (±0.08)
T7	DMAC-PN	730	4	2.65	[99]
T8	CBP	712 (60 wt% doped)	60	0.7	[94]
T9	Ga₂(saph)₂q₂	700	2	2.1	[82]
T10		700		1.54	[70]
TTM-3NCz	CBP	710 (3 wt% doped)	3	27	[84]
TTM-3NCz	CBP	710 (3 wt% doped)	3	27	[84]
Po1	MEH-PPV	730		0.56	[18]
Po2	PVK:PBD	720	5	0.2	[19]
Po3	PVK:PBD	820	2	0.10	[19]
P6BP	F8BT	883		0.1	[20]

Traditional Organic NIR Fluorescent Materials Obeying Selection Rule

In general, traditional organic NIR fluorescent dyes are characterized by being metal-free, environmentally friendly, and cost-effective, and also by their flexibility and biocompatibility. Thus, they offer great potential in applications such as bioimaging, night vision, telecommunications, and sensors.[2] [3] [4] [5] [6] [7] [14] Based on traditional organic NIR fluorescent materials, the maximum IQE of these devices is limited to 25% according to the 1:3 ratio of singlet to triplet excitons[13]; despite this, the fluorescence lifetimes are of the order of a few nanoseconds, which is ideal for high-transmission-rate applications such as “light fidelity” (Li-Fi), which is a kind of optical communication technique.[4] [5] In this section, we discuss different classes of widely studied NIR fluorescent emitters including porphyrin-based materials, D–A-type small-molecule fluorescent materials, and D–A-type polymeric materials.

Porphyrin-Based NIR Fluorescent Materials

The first fluorescent OLEDs utilizing conjugated polymers as an active layer were reported by Burroughes et al. in 1990,[15] but only in 1995 were NIR fluorescent OLEDs first reported by Baigent et al.[16] These latter used a cyano-substituted thienylene phenylenevinylene copolymer (P1 in [Figure 5]) as an emitter to fabricate LEDs that show a broad EL spectrum peaked at 740 nm with a relatively high EQE of 0.2% at that time.[16] After this report, it was found that porphyrin derivatives can be integrated as NIR emitters in fluorescent and phosphorescent OLEDs due to their intense absorption and emission in the red-to-NIR region.[17] [18] [19] [20] [21] [22] [23] [24] [25] [26] Porphyrins, consisting of four pyrrole rings interconnected by ethylene groups and (usually) a central metal ion kept in the center by the correlation bonds with the four N atoms, are highly stable to light, heat, and O₂, and possess excellent fluorescence in solution due to their robust and planar structure.[11] Moreover, the photophysical properties of porphyrin can be modified by changing substituents or central metal ions.

Figure 2 Chemical structures of porphyrin-based NIR fluorescent materials and their hosts.

Grafting the emitting dyes to polymers could be a feasible approach to solving phase separation and aggregation-induced quenching issues in OLEDs, which may cause poor fluorescence efficiencies in solid-state blending thin films.[17] [27] Polymers with the porphyrin units (Po1 in [Figure 2]) attached to poly(2-methoxy-5-ethylhexyloxy-phenylene vinylene) (MEH-PPV; [Figure 2]) at different weight loadings (5% and 8%) were synthesized and fabricated as active layers in LEDs by Iqbal et al. in 1999.[27] The fluorescence of two MEH-PPV porphyrin copolymers was proven to be dominated by the porphyrin units in the PL and EL spectra due to the fast energy transfer from the polymer to the porphyrins. However, the aggregation-induced quenching effect of the porphyrin units arising from the strong intermolecular interaction between chains, which decreases the quantum efficiency of the copolymer, cannot be ruled out completely. Similarly, by tuning the concentration of tetraphenylporphyrin units grafted to MEH-PPV, in 2000, Morgado et al. fabricated some OLEDs based on the same molecule exhibiting emission peaks at 660 and 730 nm with an EL efficiency of 0.56%. Unfortunately, the tetraphenylporphyrin units grafted to MEH-PPV also suffered severely from the aggregation-induced quenching effect and the limited spectral overlap between the emission band of MEH-PPV and the absorption band of porphyrin units, which reduced the fluorescence efficiency dramatically.[18]

Figure 3 (a) Chemical structures of c-P6T, P6, and P6BP. (b) Absorption (full lines) and PL (dashed lines) spectra of pure F8BT and blend films. Insets show the absorption and emission of the hexamer component of the spectrum. (c) Excitation–emission spectrum of F8BT:c-P6T blends. Reprinted with permission from Ref. 20. Copyright 2011 American Chemical Society.

Another polymer-based LED using an ethyne-bridged porphyrin oligomer to construct promising emitters was fabricated by Ostrowski et al. in 2003.[19] By varying the conjugated length or doping concentration of oligo[(porphinato)zinc(II)] (Po2 and Po3 in [Figure 2]) in two polymer host materials in the active layer, namely poly(N-vinylcarbazole (PVK): 2-(4-biphenyl)-5-(4-tert-butylphenyl)-1,3,4-oxadiazole (PBD) (60:40 wt%) and MEH-PPV ([Figure 2]), they successfully modulated the EL wavelength over a large spectral domain spanning from the low-energy visible region to the high-energy NIR region. The two typical OLEDs exhibit emission peaks at 720 and 820 nm with an EQE of 0.20% and 0.10%, respectively. With a similar strategy, Fenwick et al. designed OLEDs incorporating meso-butadiyne-linked linear and cyclic Zn porphyrin hexamers named P6 and c-P6T ([Figure 3]) that had redshifted emission (λ _PL = 873 and 920 nm, respectively) compared to single porphyrin rings by extending their π-conjugation.[20] In this work, the authors used a new approach to redshift the EL emission by introducing a cyclic porphyrin hexamer as an emitter in the 9,9′-dioctylfluorene-alt-benzothiadiazole (F8BT in [Figure 2]) host and demonstrated that the curved π-surface could suppress aggregation. The nanoring c-P6T in the F8BT matrix shed a significantly bathochromic shift with the EL peaking at 960 nm and the linear P6 in F8BT matrix exhibited an EL peak at 883 nm. Although both devices showed relatively low EQEs of 0.024% for c-P6T and 0.009% for P6, the addition of 4-benzyl pyridine to the linear porphyrin hexamer (P6BP in [Figure 3]) provides additional steric hindrance against aggregation and a weaker intermolecular interaction to provide an order of magnitude increase in EQE up to 0.10% at 883 nm in EL.[20]

Figure 4 Chemical structures of typical electron-rich and electron-deficient moieties in NIR polymers and small molecules.

D–A Type Polymeric NIR Fluorescent Materials

Among the different organic fluorescent materials investigated so far, the D–A copolymers displayed excellent transport properties and tunable energy levels, and are thus promising candidates for NIR-emitting materials.[28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] By appropriately covalent coupling electron-rich D units to electron-deficient A units to synthesize D–A type molecules, we can narrow the bandgaps of materials to extend the emission in the NIR region, because the emission can be ascribed to the intramolecular charge-transfer state at lower energy compared to localized excited states.[1] [9] [10] [11] [29] [30] [31] [32] The commonly used electron-rich units are thiophene (T), 3,4-ethylenedioxythiophene (EDOT), triphenylamine (TPA), fluorene, indacenodithiophene (IDT), indacenodithieno-thiophene (IDTT), etc., and the commonly used electron-deficient units are benzo[c,1,2,5]thiadiazole (BT), [1,2,5]thiadiazolo[3,4-g]quinoxaline (TQ), benzo[1,2-c:4,5-c]bis[1,2,5]thiadiazole (BBT), [1,2,3]triazolo [4′,5′:4,5]benzo[1,2-c,1,2,5]thiadiazole (BTT), diketopyrrolo[3,4-c]pyrrole (DPP), thieno[3,4-c]pyrrole-4,6-dione (TPD), boron-dipyrromethene (BODIPY), etc. ([Figure 4]).[34] [35] [36] [37] [38] [40] They all show excellent photophysical and charge carrier-transport properties in organic semiconductor devices such as OLED, organic photovoltaics (OPVs), and organic field-effect transistors. In general, the electron-rich units can raise the materials' HOMO level and the electron-deficient units can lower the LUMO level according to the molecular orbital theory.[41]

Figure 5 Chemical structures of NIR copolymers and their hosts.

Substitution of S by heavier atoms such as Se or Te in the molecular backbone is an alternative way for redshifting the emission wavelength into the NIR region.[40] [42] [43] [44] [45] [46] [47] However, the introduction of heavy atoms is usually detrimental to fluorescence emission, as it may cause fluorescent quenching due to enhancement of the spin-orbital coupling effect according to the heavy-atom effect. In 2015, Yang at el. reported on a series of conjugated D–A polymers incorporating poly(9,9-dioctylfluorene) (PFO) as the electron-rich unit and either 4,7-diselenophen-2-yl-2,1,3-benzothiadiazole (SeBT) or 4,7-diselenophen-2-yl-2,1,3-benzoselenadiazole (SeBSe) as the electron-deficient units (P2 and P3 in [Figure 5]). For the blend of thin films with SeBT or SeBSe loading as low as 1%, the PL emission consisted exclusively of single peaks derived from SeBT or SeBSe units in the copolymers, whereas the emission derived from the fluorene segment was completely quenched. In solution, the emission from the fluorene segment did not disappear instantly with a low content of SeBT or SeBSe units in copolymers and could only be quenched at a certain concentration or content of acceptor units. To shed light on the phenomenon, the authors modulated the copolymer solution concentration and content of the narrow-bandgap units to determine the relative role of the intra- and interchain energy transfer. With the increase in the ratios of narrow-bandgap units, the threshold concentration (the concentration for entirely energy transfer) of copolymers decreased by orders of magnitude. As a result, the efficient energy transfer from the fluorene segment to a SeBT or SeBSe unit proved to be dominated by an intramolecular trapping mechanism. However, the threshold concentrations of PFO–SeBSe in solution for each copolymer were about 5–10 times larger than that of the corresponding PFO–SeBT, which proved that more large-size Se atoms in PFO–SeBSe made the interchain interaction weaker than that in PFO–SeBT. Intriguingly, the copolymer optical bandgap drops dramatically from 2.92 to 1.88 or 1.78 eV, respectively, when adding a small amount of SeBT or SeBSe units (5%) to the main chain. The PL emission peaks of PFO–SeBSe were redshifted significantly to the NIR range from 734 for PFO–SeBSe1 (1%) to 790 nm for PFO–SeBSe15 (15%) when compared to the PFO–SeBT counterparts, the emission peaks of which ranged from 671 nm for PFO–SeBT1 (1%) to 713 for PFO–SeBT50 (50%). However, the PLQYs decreased from 22% for PFO–SeBSe1 to only 1% for PFO–SeBSe. Polymer LEDs (PLEDs) using those emitters delivered a low quantum efficiency of 0.3% when the EL emission peaked at 723 nm, but the efficiency further decreased to 0.02% with increasing SeBSe loadings.[43]

Heteroannulation derivatives based on BBT and TQ have been developed as narrow-bandgap guests, the energy levels of which can be well matched with many wide-bandgap polymers or small molecules.[33] [48] [49] [50] [51] [52] [53] [54] In 2015, Tregnago et al. fabricated a series of PLEDs incorporating a bisthienyl(benzotriazolothiadiazole) unit or a bisthienyl(benzotriazoloselenadiazole) unit as the emitter (P4–P7 in [Figure 5]) with phthalimide-thiophene (P8 in [Figure 5]) as the host polymer. The EQE of one device with 1% loading of the bisthienyl(benzotriazoloselenadiazole) unit reached up to 0.018% with the EL emission peaking at 990 nm.[44] Similarly, Steckler et al. reported a single-layer NIR PLED with an EQE of 0.27% at 885 nm based on phthalimide-thiophene copolymerized with bisthienyl(thiadiazoloquinoxaline) or bisthienyl (benzotriazolothiadiazole) as the NIR-emitting segments (P9–P13 in [Figure 5]).[55] Therefore, it can be concluded that the heavier atoms extend the emission wavelength to the NIR region and quench the fluorescence concomitantly.[43] [55] In 2007, Sun et al. synthesized a series of conjugated copolymers composed of PFO and 6,7-dimethyl-4,9-di(4-hexylthien-2-yl)-[1,2,5]thiadiazolo[3,4-g]quinoxalines (P14–P17 in [Figure 5]) by Pd-catalyzed Suzuki coupling reactions. PLEDs based on these copolymers showed NIR emission between 800 and 900 nm and the highest EQE reached up to 0.30% at 824 nm from PFO-DDTQ1-based devices.[49] In 2015, Marder et al. investigated heteroannulated acceptors based on benzothiazole and authored a comprehensive review which offered interesting insights into the electronic properties and structures of these materials.[36]

Recently, DPP derivatives have drawn much attention because of their wide applications in laser printers, information storage systems, erasable optical memory devices, and emitting layers.[40] [52] [53] [54] [56] [57] [58] [59] [60] In 2002, Beyerlein et al. reported on a new hairy rod-type conjugated polymer consisting of N-alkylated diketopyrrolopyrrole (DPP) diphenylene and alkoxy-substituted phenylene units in the backbone (P18 in [Figure 5]) featuring a characteristic bright-red PL peaking at ca. 640 nm. The devices based on this material exhibited an EL emission peaking at 650 nm and a maximum EQE of 0.7% with a V _on of 5 V, although they suffered from a severe efficiency roll-off at high voltages. Their EQE dramatically dropped from 0.7% to 0.4% with a change in the applied voltage from 5 to 20 V.[61] After a few years, F8BT ([Figure 2]) was proved to behave as the best host polymer in view of previous NIR OLED reports and the efficiency roll-off could be effectively solved by copolymerizing DPP-based D–A–D units with F8BT.[20] [39] [62] [63] [64] [65] In 2013, Fenwick et al. reported two efficient deep-red emitters, DPP1 and DPP3 (1% and 3% correspond to the initial monomer ratios), consisting of a small number of DPP-based moieties copolymerized with F8BT (P19 in [Figure 5]), which reached up to a 1% EL quantum efficiency at 670 nm and about 67% solid-state PL quantum efficiency. The efficiency roll-off effect, which arises from excess unbalanced carriers passing through the devices and then recombining at the anode or cathode, could be eliminated by replacing the hole-transport layer poly(3,4-ethylenedioxythiophene) (PEDOT:PSS) with an electron-blocking poly(9,9′-dioctylfluorene-alt-N-(4-butylphenyl)-diphenylamine) (TFB) interlayer.[39]

Another intriguing NIR-emitting D–A alternating copolymer, poly[N-9′-heptadecanyl-2,7-carbazole-alt-5,5-(4′,7′-di-2-thienyl-2′,1′,3′-benzothiadiazole) (PCDTBT in [Figure 5]) as the polymer main chain incorporating thiophene-benzothiadiazole-thiophene (TBT) as electron-deficient units and carbazole (Cz) as electron-rich units was reported by Lombeck et al. in 2016. They introduced a side-chain-controlled strategy to modulate the copolymers' solubility and photophysical properties by the addition of a hexyl side chain to TBT units. With increasing amounts of hex-TBT units from 0% to 100% in the copolymer, the backbone twisted, and the absorption spectrum shifted to the blue region. Both the PL and EL quantum efficiencies increased with increasing hex-TBT content (which was of considerable interest). It was shown that the EL quantum efficiency of the fully hexylated PCDTBT reached up to 1.1% with the EL peaking at 680 nm, and with luminance up to 2500 cd m⁻², it was enhanced by 2 orders of magnitude compared to standard PCDTBT.[66] To extend the emission of materials further into the NIR region, Zampetti et al. reported a new series of copolymers based on the triazolobenzothiadiazole (BTT) derivatives 6-(2-butyloctyl)-4,8-di(thiophen-2-yl)-[1,2,3]triazolo[4′,5′:4,5]benzo[1,2-c]-[1,2,5]thiadiazole (TBTTT) as emitters included in the D–A copolymer (P20–P25 in [Figure 5]), with alternating bithiophene and thienopyrroledione (TPD) moieties (P2TTPD in [Figure 5]) as the host matrix. Intriguingly, the PLEDs fabricated with the pure host polymer had a maximum EQE of 0.49% at 690 nm with a V _on as low as 2.4 V. NIR emission peaking at about 900 nm with EQE up to 0.15% was obtained by P23-based PLEDs with a V _on of only 5 V. In this work it was demonstrated that such a low V _on could be attributed to the efficient intrachain energy transfer to the TBTTT moieties.[35]

As discussed above, copolymerizing NIR moieties with a wide-bandgap polymer host main chain is a feasible way of avoiding the aggregation-induced quenching effect and reduces phase segregation to some extent. However, copolymerization of emitters and host main chain seems to be a better approach to NIR emission and effective luminescence than blending, but there still are some drawbacks. The disadvantages are as follows: (i) chromophore aggregation and close packing due to either interchain or intrachain interaction still take place in copolymers. It is hard for us to determine and control the distribution of different moieties in each polymer chain for the copolymer and the exact unit amount for each chain; (ii) by copolymerization, the frontier orbital energy level may change for each moiety; thereby, the formed heterojunction may not end up being a type-I heterojunction, which is commonly preferred to energy transfer over charge transfer; (iii) the copolymerization of D and A units is more costly than the blend-based method.[67]

D–A-Type Small-Molecule NIR Fluorescent Materials

The same D/A building blocks mentioned above can be combined to synthesize D–A–D or A–D–A small molecules as emitters in active layers.[33] [51] [52] [64] [68] [69] [70] [71] In 2009, Qian et al. reported three kinds of D–A–D chromophores based on TQ as an electron-deficient unit and TPA as electron-rich units (S1–S3 in [Figure 6]). By changing the electron-rich substituted units on the TQ core, they successfully fabricated OLEDs with tunable EL peaks ranging from 768 to 870 nm. The best OLED based on 2 wt% S2 doped in Alq₃ ([Figure 6]) showed an exclusive NIR emission at 748 nm with an EQE of 1.23% over a wide range of current density, and a maximum radiance of 2880 mW Sr⁻¹ m⁻² at 15 V.[68] Ledwon et al. reported a new organic material with the D–π–A–π–D structure utilizing Cz as the electron donor and benzothiadiazole as the electron acceptor (S4 in [Figure 6]). A strong solvatochromic effect in its EL spectrum and the electron paramagnetic resonance demonstrated its intramolecular charge-transfer property and the delocalization of radical cations and anions over the molecular backbone. OLEDs based on S4 exhibited an EL peak at 690 nm with a V _on of 4 V, a maximum EQE of 3.13% that was a relatively high value among other Cz-based and benzothiadiazole-based materials.[69]

Figure 6 Chemical structures of NIR small molecules and their hosts.

Blending of an appropriate host and emitter to produce a thin film as the emitting layer is a practical method of high-efficiency OLED fabrication. Thus, the selection of matched host and emitter becomes crucial in the OLED fabrication process. In 2018, Minotto et al. reported on a new electroluminescent blend film incorporating S5 as the emitter and PIDT-2TPD ([Figure 6]) as the host matrix. OLEDs based on the blend films afford excellent spectral and transport properties such as the EL peaking at 840 nm, a V _on of 1.7 V, and an EQE up to 1.15%. The improved device performance was attributed to the good spectral overlap and charge transfer in the D–A–D structure of S5 and the high PLQYs of both PIDT-2TPD and S5. The host matrix, PIDT-2TPD, contains an IDT as the electron-rich unit and two TPDs as the electron-deficient units in each monomer. The presence of two TPD units was beneficial: a good electron-transport property of PIDT-2TPD, good spectral overlaps between their photoemission and the absorption of S5 were achieved (better than F8BT), and the PLQY of the thin film (PIDT-2TPD doped with 0.5 wt% S5) attained 18%. As a consequence of the excellent spectral overlap between the absorption of S5 and the emission of PIDT-2TPD and their high PLQYs, the maximum EQE of 1.15% is the highest among the reported undoped devices, peaking at around 840 nm. To the best of our knowledge, the device performance is the best ever reported for the EL wavelength above 800 nm with a purely organic and solution-processed active layer, not leveraging a triplet-assisted emission.[33] In 2012, a family of D–A–D type NIR fluorophores containing nonplanar conjugated tetraphenylethene (TPE) moieties was developed with electron-deficient TQ or BBT as A units (S6–S9 in [Figure 6]). The twisted TPE units that restrict the intramolecular rotation and reduce the π–π stacking give rise to a good AIE enhancement (AIEE) property. The PL spectra of these fluorophores range from 600 to 1100 nm with their optical bandgaps ranging from 1.85 to 1.50 eV. Undoped OLEDs based on these fluorophores exhibited an EL emission peaking from 706 to 864 nm with the EQEs spanning from 0.89% to 0.20%. This work demonstrated that compounds with an AIEE property are suitable for fabricating efficient undoped NIR OLEDs.[52] Another series of multi-heterocycle D–A–D telechelic-conjugated oligomers, based on electron-rich thiophene, phenylene, and EDOT units as donor units along with electron-deficient benzothiadiazole as acceptor units (S10 in [Figure 6]), were designed and investigated by Ellinger et al. in 2011. These authors observed that S10 has a strong tendency to aggregate regardless of the host matrix. By blending the oligomer S10 in a MEH-PPV/PBD matrix, the EQE of the OLEDs achieved ∼0.5% with an EL peak at 725 nm.[72]

Impressive D–A type chromophores may guide scientists to find new luminous mechanisms and design a strategy for organic materials. In 2015, Han et al. reported a new D–A–D type NIR fluorescent compound, 2,3-bis(4′-(diphenylamino)-[1,1′-biphenyl]-4-yl)fumaronitrile (S11 in [Figure 6]), that also exhibited strong NIR fluorescence with a corresponding thin-film quantum efficiency of 33% due to its “hybridized local and charge transfer” (HLCT) and AIE property. Undoped and metal-free OLEDs based on the S12 feature a high maximum EQE of 2.58% and the EL emission peaked at 675 nm with a V _on as low as 3.6 V.[71] In 2017, a new A–D–A oligomer S12 ([Figure 6]) incorporating two α,β-unfunctionalized 4,4-difluoro-4-bora-3a,4a-diaza-s-indacene (BODIPY) as A units and two ethylene-bridged bithiophenes was synthesized and characterized by Zampetti et al.[64] Because of the delocalization of the BODIPY low-lying LUMO over the oligothienyl moieties (demonstrated by density functional theory [DFT]), the PL emission extends to the NIR region with a PL efficiency of 20% in the solid state. According to the DFT calculations, such an intramolecular energy structure should favor a “hole funneling” effect towards the central bithienyl moieties in the presence of a substantially homogeneous distribution of electrons on a relatively low-lying LUMO, which could be effectively populated from a host polymer such as F8BT in the guest–host systems. In this regard, the authors blended F8BT as the host matrix with S12 as the emitter to form a straddling heterojunction to localize the exciton in S13 for the following radiative decay. PLEDs based on S12/F8BT blends exhibited an EL that peaked at 720 nm with a maximum EQE of 1.1%. The two examples prove that the A–D–A molecules are promising candidates for NIR light-emitting materials.[64]

Although much progress has been made, there are still numerous challenges in regard to NIR emission above 1000 nm. A representative work was reported by Qian et al., who synthesized a series of simple NIR D–A–D molecules consisting of benzo(1,2-c:4,5-c)bis((1,2,5)thiadiazole) (BBTD) as the acceptor and four different donor moieties (S13–S16 in [Figure 6]). The OLEDs based on these materials exhibit EL emission peaks ranging from 1050 to 1220 nm, with a maximum EQE up to 0.28% at a current density of 10 mA cm⁻².[51]

Squaraine dyes have drawn much attention because of their sharp and intense absorption in the long-wavelength region, and the good photostability from their rigid structure.[28] [73] In 2012, Würthner and coworkers reported bright NIR fluorophores based on squaraine with an unexpected halogen effect that violates our conventional perception. In their work, the introduction of halogen atoms induced an increase in the molar absorption coefficient (ε) and a redshift that increased in the sequence of H, Cl, Br, I-substituted squaraine derivatives (S17 and S18 in [Figure 6]). The spectral emission region of three kinds of halogen-substituted squaraine derivatives could be extended to above 900 nm. They attributed the halogen effect to an electron density transfer from the substituents to the chromophore core and a strengthened conjugation along the π system. With increasing polarizability of halogen substitutes, the halogen effect became more pronounced. This work proved that squaraine dyes have a promising prospect for NIR luminescence.[73] In 2013, Stender reported a solution-processible NIR OLED consisting of an organic host–guest system, incorporating SY-PPV ([Figure 6]) as the host and a bromoindolenine squaraine dye S19 ([Figure 6]) as the guest, featuring a highly efficient energy transfer from the visible spectral range to the NIR (about 750 nm). The OLED with a doping concentration of 0.05 wt% achieved a maximum EQE of 0.65% with EL emission peaking at 730 nm.[28]

Unconventional Organic NIR Fluorescent Materials Breaking the Selection Rule

Much progress has been achieved in recent years toward development of materials for OLEDs due to the tunability of organic materials by molecular tailoring. In typical OLEDs, holes and electrons injected from the anode and cathode respectively combine in the emitting layer to form singlet and triplet excitons at a ratio of 1:3 according to the spin statistics.[13] [74] Therefore, the IQE of OLEDs based on conventional organic materials is limited to 25%. TADF, which was also called “E-type delayed fluorescence (DF),” was first rationalized by Perrin in 1929.[75] Furthermore, the so-called E-type DF was also reported in eosin and benzil by Parker and Hatchard, respectively.[76] [77] The TADF mechanism was recently reinvigorated by Adachi, who fabricated OLEDs that exploit nonradiative triplet excitons and thus the overall EL efficiency is enhanced. The energy gap of organic materials between singlet and triplet excited states can be lowered. Typically, TADF organic molecules are characterized by a highly efficient spin upconversion from nonradiative triplet states to radiative singlet states with a high radiative decay rate exceeding 10⁶ decays per second. Therefore, the molecules can harvest both singlet and triplet excitons for EL and even feature relatively high IQEs exceeding 90%.[78] [79] [80] [81] To date, this mechanism has been studied extensively, and in this section we will focus on TADF NIR fluorescent materials. In addition, triplet–triplet annihilation (TTA), by which two triplet excitons fuse into one singlet exciton, is an alternative way of fully utilizing triplet excitons for fluorescence emission.[80] [82] Fluorescence emission from radicals' doublet excitons can be another way to make full use of excitons in EL.[83] [84]

TADF NIR-Emitting Materials

To narrow the energy gap of singlet and triplet excited states in organic molecules, many researchers combined strong D and A moieties to obtain TADF molecules for the formation of intramolecular charge transfer (ICT) in excited states. The general TADF molecular design strategy is to spatially separate D and A moieties bridged by aromatic units or to twist the D and A moieties in the molecular backbone to localize LUMO in A moieties and HOMO in D moieties to limit the overlap between HOMO and LUMO in ICT excited states.[80] [81] [85] [86] [87] [88] A poor overlap between LUMO and HOMO leads to a small singlet–triplet energy gap, as well as to a lower oscillator strength, which may decrease the radiative rates according to Fermi's golden rule.[78] Due to the limitation of the intrinsic energy-gap law and the aggregation-induced quenching effect, the NIR OLEDs based on TADF emitters have to date seldom exhibited EQE values above 5%. Nevertheless, EQE values close to 30% were achieved for the red TADF OLEDs[89] [90] and the first deep-red TADF OLED with an EQE value of nearly 10% and an EL that peaked at 668 nm was first reported by Wang et al. with a novel NIR TADF V-shaped D–π–A–π–D molecule T1 ([Figure 8]) with a small singlet–triplet splitting (ΔE _ST) of 0.13 eV originating from its ICT feature and rigid structure. The optimized OLEDs, based on blends of 1,3,5-tris(N-phenyl-benzimidazol-2-yl)benzene (TPBi in [Figure 8]) as the host matrix and T1 as the dopant, exhibited a maximum EQE value of 9.8% with peak EL emission at 668 nm. Intriguingly, they demonstrated that T1 is also an AIE-active material with increased fluorescence quantum efficiencies in the solid state as compared to solution due to its twisted structure. The undoped devices achieved a maximum EQE of 2.1% because of its high PLQY in the solid state. These deep-red materials laid a good foundation for the further study of NIR materials.[91] Two years later, Wang et al. reported NIR OLEDs based on the TADF emitter T2 ([Figure 8]) doped in TPBi and exhibiting a high EQE of 14.5% at 644 nm and 3.9% at 728. Through X-ray diffraction (XRD) analysis, they demonstrated an edge-to-edge aggregation in the crystal of T2; it shows a strong deep-red or NIR emission in a neat thin film (λ _max = 733 nm, PLQY = 21%).This work demonstrated that a rational design of a supramolecular structure of a molecule is a feasible way to synthesize high-efficiency NIR organic materials.[87]

Figure 7 (a) Doublet emission after photo- or electrical excitation. The right-hand illustration shows the electron spin vector directions for doublets. (b) Chemical structures of TTM, TTM-3NCz, and TTM-3PCz. (c) EL (solid lines) and PL (dotted lines) spectra for TTM-3NCz (red) and TTM-3PCz (black) LEDs. Reprinted with permission from Ref. 84. Copyright 2018 Springer Nature.

In 2017, two dendrimer-type NIR TADF emitters, 2,6-bis(4-(bis(4-((6-(9H-carbazol-9-yl)hexyl)oxy)phenyl)amino)phenyl)anthracene-9,10-dione (T3 in [Figure 8]) and 2,6-bis(4-(bis(4-((6-(9′H-[9,3′:6′,9″-terbenzo[b]indol]-9′-yl)hexyl)oxy)phenyl)amino)phenyl)anthracene-9,10-dione (T4 in [Figure 8]), were reported by Sun et al. and the undoped OLEDs based on T4 exhibited a maximum EQE of 0.254% with EL peaked at 715 nm. The dendrimers exhibit a self-host feature and the peripheral Cz/tricarbazole dendrons encapsulate the core to suppress aggregation.[86] Furthermore, in the same year, Yuan et al. reported a wedge-shaped D–π–A–π–D TADF emitter 3,4-bis(4-(diphenylamino)phenyl)acenaphtho[1,2-b]pyrazine-8,9-dicarbonitrile (T5 in [Figure 8]) with a small singlet–triplet splitting (ΔE _ST) of 0.14 eV, synthesized through combining an acenaphtho[1,2-b]pyrazine-8,9-dicarbonitrile acceptor core with two TPA donor units. The undoped NIR devices featured a maximum EQE of 2.19% with the EL emission peaking at 777 nm, whereas the doped devices based on blends of a TPBi host matrix doped with 10 wt% of T5 were characterized by an extremely high EQE of 10.19% and an EL peak at 693 nm. The high quantum efficiencies were attributed to a moderate orbital overlap between the HOMO and the LUMO in T5.[92] The same group later reported a series of T5-based OLEDs by varying the host materials such as ADO-TPA ([Figure 8]) for adjusting the neighboring surroundings. It turned out that an OM compound bis[2-(2-benzothiazolyl-N3)phenolato-O]zinc (Zn(BTZ)₂ in [Figure 8]) as the host material matched best with T5 and the doped devices achieved rather high EQE values of 12.3% and 5.1% by varying the doping concentration from 2 wt% to 20 wt%, and EL peaks ranging from 668 to 728 nm. The relatively high fluorescence efficiencies of the doped films can be ascribed to the sufficient spectral overlap between the emission spectrum of Zn(BTZ)₂ and the absorption spectrum of T5. The redshifted emission peaks can be ascribed to the solid-state solvation effect of Zn(BTZ)₂ because of the largest dipole moment among all the used hosts.[93]

In contrast to the reported twisted TADF molecules, curcuminoid derivatives with no twisted structure exhibited exceptional photophysical properties in OLED applications. In 2018, Kim et al. reported on two D–A–D-type boron difluoride curcuminoid derivatives consisting of one or two triphenylamine donor groups and one acetylacetonate boron difluoride acceptor unit as effective TADF emitters (T6 and T8 in [Figure 8]).[94] [95] For T6, unlike the general strategy to obtain TADF molecules with the spatially separated HOMO and LUMO, the large spatial overlap between the hole and electron wave functions of the boron difluoride curcuminoid derivative T6 gives rise to a high molar extinction coefficient and oscillator strength, which positively correlate with the high radiative decay rate, rendering it a good emitting material for OLEDs. Interestingly, an efficient TADF activity was observed in its thin film. Moreover, its absorption and emission bands can be tuned in a broad range due to the large ground- and excited-state dipole moments. Quantum-chemical calculations demonstrated that the vibration and spin–orbit coupling assisted TADF mechanism involved in T6 can be attributed to an enhanced nonadiabatic coupling effect between the low-lying excited states. In conclusion, the NIR OLEDs based on blends of the 4,4′-di(9H-carbazol-9-yl)-1,1′-biphenyl (CBP in [Figure 8]) host and the curcuminoid derivative T6 exhibited a maximum EQE of nearly 10% at 721 nm and a maximum radiance value of 3 × 10⁶ mW sr⁻¹ m⁻². By modulating the polarity of the active medium, the EL peaks can be tuned from 700 to 780 nm. This work provides new insights into the strategy of designing fascinating TADF materials and represents an important step in the field of NIR OLED.[95]

It is difficult to achieve a small ΔE _ST and a high radiative decay rate in a compound simultaneously by a common strategy of spatially separating the HOMO and LUMO, and thus TADF NIR emitters with PLQY are quite rare. In view of the abundant conventional fluorescent dopant (CFD) mentioned earlier, it is feasible to use TADF materials as hosts or sensitizers and CFDs as dopants to fabricate emitting layers in OLEDs.[96] [97] [98] [99] [100] In this way, holes and electrons recombine in the host or sensitizer to form singlet and triplet excitons at a 1:3 ratio and then triplet excitons absorb the heat in the environment to form singlet excitons via a reversed intersystem crossing (RISC) due to the relatively small ΔE _ST. Radiative excitons are transferred to a fluorescent or phosphorescent dopant and then radiative decay for emitting photons through Förster resonance energy transfer (FRET, a nonradiative mechanism of energy transfer from a fluorophore in the excited state to a fluorophore in the ground state via dipole–dipole coupling) in PL or a charge carrier trapping mechanism in EL. In 2014, Adachi et al. used TADF molecules as assistant dopants that efficiently transfer all the electrically singlet and triplet excitons resonantly to the fluorescent emitters with different emission peaks to fabricate a series of OLEDs for blue, green, yellow, and red light emission. The performance, with EQEs as high as 13.4–18% for different colors, was good and demonstrated that in this way we can extend the emission of OLEDs to the NIR region.[97] In 2019, Duan et al. fabricated a series of OLEDs to manipulate the interplay of host and TADF sensitizer to maximize the device performance. Adopting the TADF sensitization strategy, Qiao et al. fabricated OLEDs based on the TADF host DMAC-PN doped with T7 ([Figure 8]), whose molecular structure consists of 1,2-bis(9,9-dimethyl-9,10-dihydroacridine)-4,5-dicyanobenzene (DMAC-PN in [Figure 8]) and 4,9-bis(4-(diphenylamino)phenyl)-naphtho[2,3-c,1,2,5]selenadiazole (T7), exhibiting a maximum EQE of 2.65% with an EL peak at 730 nm and a small roll-off. In this work, the authors found that dopants with lateral bulky substituents could suppress the Dexter energy transfer (a short-range nonradiative process via collisional or exchange energy transfer) so that more singlet excitons could decay radiatively in the dopant. Therefore, the small roll-off can be ascribed to a relatively low concentration of triplet excitons, which may be quenched by TTA in the emitting layer.[100] Meanwhile, Adachi et al. demonstrated that NIR electrophosphorescence up to 1100 nm can be achieved by using the TADF molecule 2-phenoxazine-4,6-diphenyl-1,3,5-triazine (PXZ-TRZ) as the host matrix and as a triplet sensitizer to provide better balance between holes and electrons and a higher radiative exciton ratio in NIR-electrophosphorescence OLEDs as a result of the balanced bipolar charge-transport properties of PXZ-TRZ ([Figure 8]). The balanced carrier-transport properties were attributed to the spatially separated HOMO and LUMO distributions in PXZ-TRZ. Although optimized OLEDs based on blends of PXZ-TRZ as the TADF host and copper phthalocyanine (CuPc) or phthalocyanine (PtPc) as the phosphorescent emitter featured rather low EQE values of less than 0.1%, limited by the relatively low phosphorescence efficiency of CuPc and PtPc, Adachi et al. believe that the issue will be addressed by using high-efficiency NIR emitters and a TADF-sensitized electrophosphorescence system.[97]

TF (TTA) NIR-Emitting Materials

Triplets fusion, also called TTA, is an alternative strategy to effectively exploit triplet excitons, and OLEDs based on this mechanism may achieve an η _r(max) value of 62.5% because theoretically, 3/4 triplet excitons and 1/4 singlet excitons are generated by electricity and then two triplet excitons combine to form one singlet exciton.[80] In 2016, Qiao et al. reported a high-performance NIR OLED based on TF, reaching a maximum EQE of 2.1% at 700 nm and maintaining a high EQE of about 2% in a wide range of current densities. The device was fabricated with a NIR emitter, 4,9-bis(4-(2,2-diphenylvinyl)phenyl)-naphtho[2,3-c,1,2,5]selenadiazole (T9 in [Figure 8]), with a large LUMO/HOMO overlap and ambipolar character acting as a dopant in a bipolar host Ga₂(saph)₂q₂ ([Figure 8]). In this system, recombination of electrons and holes as well as TF occur in the host material and the singlet excitons are transferred to T9 via a FRET mechanism for radiative decay.[82] Qiao et al. offer us a new device design strategy and molecular structure for NIR emission.

NIR-Emitting Materials with Doublet Emission

In 2015, Li et al. proposed a new open-shell molecule as an emitter in OLEDs, in which only one electron is populated in the highest singly occupied molecular orbital (SOMO) and can be excited to the lowest singly unoccupied molecular orbital (SUMO). Theoretically, the transition between SOMO and SUMO is spin-allowed, so the upper limit of IQE in the OLEDs is 100%. Photons are generated by D₁–D₀ transitions in these neutral radicals; in this way, the poor exciton utilization efficiency of OLEDs can be solved.[83] Later in 2018, Li et al. made a breakthrough in the NIR OLED field by fabricating high-performance OLEDs emitting at 710 nm (TTM-3NCz; [Figure 7]) and 703 nm with maximum EQEs of 27% and 17%, respectively.[84] This work demonstrated that open-shell doublet donor-radical molecules are promising candidate materials for high-efficiency OLED fabrication.

Figure 8 Chemical structures of unconventional NIR molecules and their hosts.

NIR-Emitting Materials with “Hot-Exciton Mechanism”

Sometimes, molecules with a certain structure may leverage excitons in higher excited states, such as the third-excited-triplet excitons (T₃) for light emission in EL; this is also called the “hot-exciton” mechanism. In 2014, Yao et al. reported on the butterfly-shaped A–D–A type chromophore T10 ([Figure 8]), which exhibited a strong NIR fluorescence that peaked at 700 nm in film and the corresponding quantum efficiency reached 16%. Remarkably, the NIR OLED based on T10 features a maximum EQE of 1.54% and a high radiative exciton ratio of 48% with quite a low efficiency roll-off. At a high current density of 300 mA cm⁻², the EQE of the S11-based device remained as high as 1.17%. These authors established that a HLCT state exists for T10 in a low-polarity solvent or in the thin-film state and the high radiative single exciton ratio can be ascribed to the facile T₃ to S₁ RISC process in the HLCT state demonstrated by the similar natural transition orbitals of S₁ and T₃ states. Given the large dipole moment of the CT state and the certain degree of orbital overlap of the LE state, S11 displayed strong NIR emission and AIE behavior in the aggregated state. Despite D–A–D fluorophores having been widely studied, A–D–A-type molecules have been scarcely used because of the difficult monofunctionalization of the units in terms of synthesis.[70]

NIR Organometallic Phosphorescent Materials

With regard to phosphorescent OLEDs, organometallic phosphorescent materials have several advantages over traditional fluorescent materials. First, because phosphorescence originates from triplet excitons' radiative decay, phosphorescent OLEDs can harvest both singlet and triplet excitons for light emission, so the ideal IQE may attain 100%. Second, phosphorescence shows a longer decay lifetime from microsecond to millisecond. Third, the heavy metal ions in organometallic phosphorescent materials can facilitate the spin–orbit coupling effect, so the intersystem-crossing rates will increase and the phosphorescence radiative transitions from excited triplet to ground states will dominate in exciton deactivation pathways. The emissive mechanisms of metal complexes is complicated and the possible luminous mechanisms of metal complex materials consist of triplet metal-to-ligand charge transfer (³MLCT), intraligand charge transfer (³ILCT), ligand-to-ligand charge transfer (³LLCT), ligand-to-metal charge transfer (³LMCT), metal–metal-to-ligand charge transfer (³MMLCT), and metal-to-ligand–ligand charge transfer (³MLLCT). The impact factors of excited states include the metal centers, chemical structures, and triplet-state energy levels of the ligands, local environment, and intermolecular interaction, so the photophysical properties of metal complexes such as absorption, emission wavelength, phosphorescence lifetime, and quantum efficiency can be modulated by varying the ligand structures and metal ions.[101] In the following section, we summarize the phosphorescent NIR materials and classify them by central metals under appropriate optimization ([Table 2]).

Table 2
Summary of chemical structures of NIR phosphorescent organometallic materials and their corresponding EQE and EL peaks
Emitter	Host	EL peak (nm)	Doping ratio (wt%)	EQE (%)	Ref.
M1	PVK	1540	80	–	[104]
M2	F8BT	890		[62]
M3(Ln = Nd)		890, 1070		[110]
M3(Ln = Er)		1540		[110]
M3(Ln = Yb)		985		[110]
M4(Ln = Yb)	PPP-OR1	977		0.001	[111]
M4(Ln = Er)	MEH-PPV	1560		[111]
M5	Alq₃	772		8.5	[25]
M5	Alq₃	772	6	6.3	[22]
M6	PVK:PBD	896	4	3.8	[24]
M6	PVK:PBD	898	2	0.75	[21]
M6	Alq₃	900	4	3.8	[21]
M7	Alq₃	773	4	9.2	[21]
M8	PVK:PBD	774	2	3.0	[21]
M8	PVK:PBD	1005	4	0.20	[21]
M9	Alq₃	848	4	2.8	[23]
M10	Alq₃	846	4	1.5	[23]
M11		760		0.03	[126]
M12		705		10.5	[127]
M13		700		14.5	[128]
M13		715		10.5	[127]
M14		720		8.5	[127]
M15		800		1.2	[125]
M16	PVK	720	5	0.1	[129]
M18	M17	620	0.3	9.2	[131]
M19	PVK:PBD	788	20	0.5	[124]
M20	PVK:PBD	791	20	0.5	[124]
M22	DIC-TRZ	760		4.5	[132]
M23	PVK:OXD7	709	5	1.28	[121]
M24	PVK:OXD7	714	5	3.07	[121]
M25	PVK:OXD7	714	5	2.44	[121]
M26	PVK:PBD	690	20	0.67	[123]
M27	PVK:PBD	753	20	0.3	[123]
M28	Ga₂(saph)₂q₂	780	20	2.2	[120]
M29	Alq₃	718		2.7	[133]
M30	Alq₃	814	6	1.5	[133]
M31:M32	BP4mPy	770		0.067	[134]
[Pt(fprpz)₂]		740		24	[118]

Lanthanide-Based Materials

In general, organic dyes usually possess broad emission spectra, which is disadvantageous for the color purity of devices. On the contrary, lanthanide metal ions exhibit extremely sharp emission bands because their 4f orbitals are effectively shielded by the external forces of overlying 5s² and 5p[6] orbitals. For this reason, the emission and absorption bands (f–f transition) of metal ions become extremely sharp when electronic transitions occur from one J state of an fⁿ configuration to another J state of the same configuration, and the metal ions are generally excited by intramolecular energy transfer from the triplet states of a π-conjugated ligand in lanthanide organometallic complexes. Therefore, the color purity of lanthanide-based OLEDs can be very high.[102] Complexes incorporating lanthanide materials like Er(III), Yb(III), and Nd(III) have attracted considerable attention due to their NIR emission.[103]

However, the luminescence efficiencies of lanthanide complexes also suffer from the “energy-gap law” and the low absorption coefficient of parity-forbidden f–f transitions.[104] To enhance the luminescence efficiencies of lanthanide complexes, several approaches have been employed, such as constituting a host–guest system in the emitting layer, changing the ITO or silicon-based substrates, and copolymerizing the organic ligands with a polymer main chain.[62] [104] [105] [106] [107] In 1999, the first phosphorescent OLEDs incorporating tris(8-hydroquinoline)erbium(III) (ErQ) as the neat emitting layer with an EL peak at 1.54 μm were fabricated by Curry et al.,[108] [109] and were based on ITO or a silicon substrate. The room-temperature EL at 1.54 μm was ascribed to the intra-atomic transitions between the ⁴I_13/2 to ⁴I_15/2 levels in the Er³⁺ ion but at that time the authors were unable to establish the exact luminescence efficiencies for the OLEDs. Furthermore, by blending the lanthanide complex into a polymer matrix as the emitting layer, NIR OLEDs with higher efficiency were fabricated. The most common host polymers are MEH-PPV, PVK, and F8BT, whose emission peaks are 550, 450, and 550 nm, respectively.[59] [105] [106] [110]

In 2000, Sun et al. reported OLEDs using tris(acetylacetonato)(1,10-phenanthroline)erbium (M1 in [Figure 10]) doped into PVK with a doping ratio of 80 wt% as the emitting layer, exhibiting a NIR EL emission peaking at 1.54 μm.[104] In 2001, Slooff et al. reported an 890-nm luminescence from neodymium-doped PLEDs based on blends of F8BT and a lissamine-functionalized terphenyl-based neodymium complex (M2 in [Figure 10]) as the emitting layer. In this work, lissamine acted as a sensitizer to transfer charges to Nd³⁺ followed by direct formation of triplets on the metal ions.[62] In the same year, Kawamura et al. reported a series of OLEDs based on a tris(dibenzoylmethanato)(monobathophenanthroline)lanthanide(III) complex (lanthanide = Nd, Er, Eu, and Yb, M3 in [Figure 10]), displaying the EL emission bands ranging from 900 to 1600 nm at room temperature with EQEs below 0.03%.[110] Harrison and coworkers fabricated NIR PLEDs using blends of MEH-PPV or bis-alkoxy-substituted poly(p-phenylene) (PPP-OR11 in [Figure 10]) as host materials and Yb-TPP-based or Er-TPP-based complexes as emitters (M4 in [Figure 10], TPP = 5,10,15,20-tetraphenylporphyrin). The Yb-based devices utilizing MEH-PPV or PPP-OR11 as hosts exhibited an EL peak at 977 nm and Er-based devices utilizing MEH-PPV as the host exhibited an EL peak at 1560 nm. The visible part of EL was strongly quenched in all devices and in the PPP-OR11-based devices, the effect become significant due not only to a good spectral overlap between the Q-absorption bands of TPPs and MEH-PPV fluorescence but also to the better spectral overlap between the fluorescence of PPP-OR11 and the Soret absorption band of TPPs, which facilitate the FRET process. Interestingly, in the EL process, TPPs acted as sensitizers to transfer triplet excitons to the lanthanide ions followed by the phosphorescence radiative recombination in the metal ions. The Yb-based device using PPP-OR1 as the host showed a maximum EQE of 0.001% at 7 V, which is five times higher than that of a Yb-based device using MEH-PPV as the host.[111]

Figure 9 (a) Chemical structures of Pt(II) complexes 1–4. (b) Absorption spectra of 1–3 in THF (left, solid lines) and corresponding absorption spectra (right, unified symbols) and PL spectra (right, filled symbols) of 1–3 in solid film normalized at the peak wavelength are also shown. Reprinted with permission from Ref. 118. Copyright 2016 Springer Nature.

Transition Metal-Based Materials

Transition metal complexes with strong phosphorescence emission have attracted intensive attention in recent years, thanks to their intriguing photophysical properties, especially for Ir(III) and Pt(II) complexes.[108] [109] [110] [112] [113] [114] [115] [116] [117] The square-planar platinum complexes have a rigid molecular configuration, which could suppress nonradiative decay. However, the concomitant aggregation-induced quenching effect and relatively long phosphorescence lifetime may also lead to severe efficiency roll-off at a high current density in devices. Remarkably, the extremely high EQE record of 24% at 740 nm among NIR OLEDs was set by homoleptic Pt(II) complexes.[118] Compared to the square-planar Pt(II) complexes, octahedral Ir(III) complexes have relatively short phosphorescence lifetimes, so they exhibit less efficiency roll-off in devices.[111] [118] [119] [120] [121] [122] [123] [124] The maximum EQEs of 3.4% at 702 nm,[121] 3.1% at 714 nm,[122] and 2.2% in the 750–800 nm range have been achieved by Ir(III) complex-based devices.[111]

Pt-Based Phosphorescent Materials

Since the first phosphorescent OLED based on Pt-porphyrin complex (PtOEP) doped into Alq₃ was reported by Baldo et al.,[113] Pt-porphyrin complexes have drawn extensive interest in recent years due to the significant absorption and emission in the red-to-NIR region.[21] [22] [23] [24] [25] [119] [125] As described before, the lone pairs on nitrogen atoms can coordinate to various metals and the porphyrin ring can be functionalized by a variety of groups; thus, rich and fascinating photophysical properties can be found in porphyrin-based materials. In the previous section, we have discussed porphyrin-based fluorescent materials, so in this subsection we will focus on Pt-porphyrin complexes for phosphorescence emission.

In 2007, Sun et al. studied the photophysical properties of several Pt-porphyrin derivatives in depth and fabricated devices using Alq₃ or CBP as the host matrix doped with Pt(II)-tetraphenyltetrabenzoporphyrin (M5 in [Figure 10]) as a phosphor, and dimethyl-4,7-diphenyl-1,10-phenanthroline (BCP) as an electron-blocking layer, which exhibited a maximum EQE of 8.5% with the EL peaking at 772 nm and good color purity.[25] Performances of the device exceeded those reported previously in works based on M5 doped in Alq₃ (EQE was 6.3% at 765 nm).[22] In a different approach to that of others, these authors fabricated the OLEDs with an emitting layer of CBP doped with 4 wt% M5 and 2 wt% PQIr (Ir(III) bis(2-phenylquinolyl-N, C²′)acetylacetonate) simultaneously, to suppress TTA and the aggregation of M5. Interestingly, they discovered that the field-induced charge pair dissociation mechanism dominates in the efficiency roll-off process rather than the TTA mechanism, from which they inferred that co-doping induced efficient cascade energy transfer from CBP through PQIr to M5. Therefore, the accumulation of high-density triplets in the emitting layer can be avoided. The co-doped OLED showed a maximum EQE of 8.5% with a less pronounced efficiency roll-off.[25]

To extend the NIR emission above 800 nm, Sommer et al. used a blend of Pt(tptnp) as the dopant (tptnp: tetraphenyltetranaphtho[2,3]porphyrin, M6 in [Figure 10]) and PVK/PBD (6:4) as the host as an emitting layer to fabricate PLEDs with a strategy of extending conjugation in the porphyrin ligand. The devices based on the spin-coated emitting layer showed an EL peak at 896 nm, and a maximum EQE of 0.4%. Moreover, the optimized devices incorporating blends of M6 and CBP as the emitting layer, which were prepared by vapor deposition, featured a maximum EQE of 2.8% at 896 nm.[24] In 2011, with the similar strategy, these authors synthesized a family of π-extended Pt(II) porphyrins to fabricate a series of solution-processed PLEDs and vapor-deposited multilayer OLEDs with EL peaks ranging from 771 to 1005 nm. The optimized PLED and OLED based on π-extended Pt porphyrins (M7 and M8 in [Figure 10]) exhibited EQEs of 9.2% at 773 nm and 3.0% at 774 nm, respectively at that time.[21] In 2016, Huang et al. reported vacuum-evaporated NIR OLEDs based on two Pt(II) azatetrabenzoporphyrin complexes (M9 and M10 in [Figure 10]), which featured the maximum EQE of 2.8% at 848 nm and 1.5% at 846 nm, respectively.[23]

In the previous section we mentioned that copolymerizing the porphyrin rings with the polymer host can be an alternative method of achieving fluorescence emission.[17] [18] [27] Regarding phosphorescent materials, in 2015, Freeman et al. copolymerized a dimesityl diphenyl porphyrin platinum (M11 in [Figure 10]) into a PFO host main chain connected by phenyl groups. By varying the loading of M11 (0.5%, 1%, 2%, and 5%), the PLEDs incorporating these copolymers as a neat emitting layer showed EL peaks at 665, 736, and 818 nm with a shoulder peak at 642 nm and EQEs ranging from 0.17% to 0.48%.[26] In 2016, the same group reported a PLED incorporating M11 copolymerized with poly(diphenylanthracene) as an emitting layer. The device with 10% porphyrin loading showed an EL peak at 760 nm and an EQE up to 0.03%, revealing a remarkable redshift compared to its PL peak at 666 nm. They attributed the redshift phenomenon to the increased aggregation of the novel polymer backbone, which induced either polymer–porphyrin or porphyrin–porphyrin aggregation.[126]

In addition to Pt porphyrins, the other Pt complexes may exhibit NIR emission by appropriate molecular design. In 2007, Cocchi and coworkers successfully fabricated OLEDs based on a series of terdentate cyclometalated phosphorescent Pt(II) complexes (M12–M14 in [Figure 10]) as emitters.[127] Thanks to the high excimer emission quantum yields and the confined recombination zone in the complexes, the OLEDs emitting at 705 and 720 nm displayed excellent performances with EQEs up to 10.5%.[127] In 2008, the same group reported an extremely efficient NIR OLED with an EL peak at ∼700 nm and a very high EQE of 14.5%. The authors used the same series of terdentate cyclometalated efficient phosphorescent Pt(II) complexes (M12–M14) as pristine emitting layers to fabricate high-performance excimer-based NIR OLEDs. They demonstrated that inserting a PbO₂ buffer layer between the Ca cathode and the electron-transporting layer can improves the device performance owing to the changed band bending between the electron-transport layer and the cathode.[128] In 2013, Nisic and coworkers synthesized two kinds of Pt complexes, and the OLED based on the latter one (M31) showed a maximum EQE of 1.2% at 800 nm.[125] Moreover, several high-efficiency NIR OLEDs that emit light at 740 nm with a maximum EQE of 24 ± 1% and a radiance of (3.6 ± 0.2) × 10⁵ mW sr⁻¹ m⁻² were reported by Tuong Ly et al. in 2017. It was demonstrated by them through XRD and theoretical calculations that the high performances can be ascribed to the high PLQY (about 81%) and a superior edge-on orientation of the Pt(II) complex [Pt(fprpz)₂] in solid states.[118] In this work, the authors used three kinds of 2-pyrazinylpyrazolate Pt(II) phosphor complexes: [Pt(fprpz)₂], [Pt(fprpz)(fppz)], and [Pt(fprpz)(tbfppz)] ([Figure 9]), which were characterized by rather high PLQYs of 81% at 740 nm, 55% at 703 nm, and 82% at 673 nm, respectively, in the thin-film state, to fabricate OLEDs. To gain in-depth insights into the transition process, the authors employed wide-angle X-ray scattering, angle-dependent luminescence, and time-dependent DFT to determine the self-assembly and MMLCT properties at the molecular level. They demonstrated that in the solid state, the aggregation property of Pt(II) complexes is through π-interaction in the d_z2 direction, and the MMLCT property along the Pt–Pt linkage seems to be an exciton-like model. Thanks to the short packing distance, the interaction among 5d_z2 (HOMO) in Pt(II) and π* orbitals (LUMO) in the ligands is strong. As a result, the excitons featuring a long diffusion length can be prevented from exciton-vibrational coupling and the nonradiative deactivation process will be suppressed.[118]

Figure 10 Chemical structures of organometallic NIR phosphorescent materials and their hosts.

Ir-Based Materials

The OLED featuring exclusive NIR emission was reported by Williams et al. in 2006,[129] which exhibited a maximum EQE of nearly 0.25% at 720 nm, including blends of iridium(III) bis(1-pyrenyl-isoquinolinato-N,C)acetylacetonate (M16 in [Figure 10]) and polymer hosts PVK and PBD. Furthermore, the optimized devices with an extra hole-blocking layer of BCP and an electron-injecting layer Alq₃ showed a higher performance but less NIR purity, which can be attributed to emission from exciplex formation among BCP, PVK, and PBD.[129] Wong and coworkers synthesized a family of borylated Ir(III) cyclometalated complexes with oligothiophene-based ligands to fabricate OLEDs. The EL spectra's relative intensities in the NIR region are significantly higher than those in PL spectra. Although the performances of devices needed to be improved, it is believed that by modulating the ligands, the Ir(III) complexes with an electron-withdrawing borylated substituent would show greater potential in NIR emission.[130]

In 2007, Tsuzuki and coworkers fabricated OLEDs using a host/guest system of M17 and M18 ([Figure 10]) [where M17 is bis(2-phenylpyridinato-N,C ²′)iridium(acetylacetonate) and M18 is tris(1-phenylisoquinolinolato-C², N)iridium(III)] because of the relatively larger bandgap, higher T₁ state energy, and good hole-transport property of M17. The green phosphorescence of M17 was effectively quenched by the 1 wt% M18, suggesting that the triplet energy of M17 was efficiently transferred to M18. The maximum EQE of the devices was 9.2% with an EL peak at 620 nm at 0.3 wt% M18, and the driving voltage was 4.4 V at a luminance of 100 cd m⁻².[131] Furthermore, the extension of π-conjugated systems and decoration of the ligands have been demonstrated as effective ways to extend the emission into the NIR region and maintain a rigid molecular configuration concomitantly although it may bring difficult problems in synthesis and purification, which can be attributed to the increased steric hindrance and decreased solubility of the materials.[62] In 2015, Xin et al. designed two NIR-emitting cationic iridium(III) complexes M19 and M20 ([Figure 10]) and the OLEDs based on them displayed a constant EQE of around 0.5% over a large current density range of 1–100 mA cm⁻².[124] With the same synthetic strategy, in 2017, Xue et al. reported two novel NIR-emitting homoleptic facial Ir(III) complexes based on extended π-conjugated benzo[g]phthalazine ligands, tris[1,4-di(thiophen-2-yl)benzo[g]phthalazine]iridium(III) (M21 in [Figure 10]) and tris[1-(2,4-bis(trifluoromethyl)phenyl)-4-(thiophen-2-yl)-benzo[g]phthalazine] iridium(III) (M22 in [Figure 10]), whose PLQYs in solution are 5.2% at 824 nm and 17.3% at 765 nm, respectively. The authors fabricated NIR OLEDs using 2,4-diphenyl-6-bis(12-phenylindolo[2,3-a]-carbazole-11-yl)-1,3,5-triazine (DIC-TRZ in [Figure 10]) as the host matrix doped with two Ir(III) complexes, featuring an ultrapure EL emission at 760 nm and a maximum EQE of 4.5%. EQEs of these devices remain as high as 3.5% at 100 mA cm⁻². This work may provide us with new insights into the fabrication of high-efficiency and small roll-off NIR OLEDs.[132] Similarly, by introducing the electron-rich thiophene groups in ligands, Kesarkar et al. reported three cyclometalated Ir(III) complexes doped in PVK and OXD7 ([Figure 10]) with different β-diketonate ancillary ligands with increasing conjugation, which exhibited efficient NIR emission, namely M23, M24, and M25 ([Figure 10]). The device based on M24 showed a remarkable EQE of 3.07% with an EL peak at 714 nm.[121] In 2012, two NIR cationic iridium (III) complexes M26 and M27 were synthesized and characterized by Rao and coworkers. The use of sp²-hybridized N opposite the chelating N atom in the ligand of M27 caused the emission to redshift 60–80 nm as compared to M26.[123] In 2013, these authors changed part of the ligands in M27 and synthesized a new cationic complex M28 with a strong NIR emission. The device based on M28 as an emitter and the bipolar host Ga₂(saph)₂q₂ exhibited a maximum EQE of 2.2% and a constant EQE of around 2% over a wide range of current densities from 3 to 100 mA cm⁻².[120] These works indicate that Ir-based materials hold great promise for the usage in NIR OLEDs in terms of small efficiency roll-off.

Other Transition-Metal Materials

In 2009, Lee at el. developed a series of charge-neutral Os(II) isoquinolyl triazolate complexes (M29 and M30 in [Figure 10]) with trans- and cis-configurations to fabricate a series of devices based on Alq₃ , TPBi, and TAZ ([Figure 10]) as hosts. It is noteworthy that by using Alq₃ as the host, a device with EL peaking at 814 nm and an EQE of 1.5% is achieved, and a device with EL peaking at 718 nm and an EQE of 2.7% is achieved, indicating that the Os(II) complexes with lower intrinsic excited states, short radiative lifetime, and nonionic natures are promising in NIR OLEDs.[133] In 2014, by employing a highly conjugated isoquinolinyl azolate chelate as ligands to decrease the excited-state energy level, Liao and coworkers reported a NIR OLED based on Os(bpftz)₂(PPh₂Me)₂ and [Os(bpy)(dttz)(PPhMe₂)₂] (M31 and M32 in [Figure 10]) doped in BP4mPy ([Figure 10]) emitting at 770 nm with an EQE of 0.067%.[134]

NIR QD Materials

QD nanocrystals (QDs) display readily tunable luminescence throughout the whole NIR region due to their large Bohr exciton radius (20 nm for PbS and 46 for PbSe). Unlike typical organic materials, QDs are not subject to strong vibrational coupling between the ground and excited states, which causes broadened spectra and increased nonradiative decay rates. Moreover, QDs also feature a higher luminance efficiency, higher photostability, and narrower spectral emissions than organic materials. The PL and EL emission wavelengths can be tuned by varying the QD size according to the quantum confine effect, so QDs outperform organic materials at wavelengths beyond 1000 nm and are promising candidates for NIR light-emitting materials.[135] [136] [137] [138] [139] [140] [141] [142] [143] [144] [145] [146] However, because the surface states on QDs can quench excitons, their surface structures should be carefully processed. In addition, the interdot carrier transport can lead to unfavorable transport-assisted trapping, which results in self-quenching in QDs. Passivation with a wide-bandgap inorganic or organic shell, capping with insulating organic ligands, and incorporation into a host matrix are the established methods applied to improve device performance. These methods, however, may bring some problems such as low current density, charge injection imbalance, and exciton ionization induced by large bias voltages. In this section, typical NIR QLEDs with emitting layers that were treated by different passivation methods will be discussed.[136] [142] There is difficulty in the synthesis of monodispersed QDs, which might increase the cost of QLEDs.[147]

The earliest NIR QLED was reported by Banin et al. in 2008, who utilized core–shell InAs–ZnSe QDs embedded in MEH-PPV to fabricate QLEDs with EQE values of about 0.5% and EL peaks varying from 1 to 1.3 μm. In this work, they demonstrated that the optimized core–shell nanocrystal structure shields the excitons localized in the core from the host deactivation path and maintains the energy transfer from the host through a Förster or Dexter mechanism.[135] In the same year, a highly efficient hybrid organic–inorganic nanocrystal NIR QLED based on PbS nanocrystals that emit at 1.2 μm with a maximum EQE of 1.15% was reported by Bourdakos et al.,[142] outperforming the previously reported nanocrystal-based device.[135] The excitons are directly created on the oleic acid-capped PbS nanocrystals, thus removing the requirement of efficient energy transfer from organic molecules to nanocrystals. A solution-processed hybrid polymer-nanocrystal NIR LED based on PbSe nanocrystals embedded in a conjugated polymer MEH-PPV that emits at 1280 nm with an EQE of 0.83% was reported by Choudhury et al. in 2009. In this work, the PbSe nanocrystals acted as carrier traps where the injected electrons and holes recombine to form excitons for radiative recombination. They demonstrated that increased carrier trapping leads to more efficient exciton formation on the emitting sites of PbSe nanocrystals, resulting in a higher quantum efficiency.[144]

In addition to IV–VI semiconductor nanocrystals such as PbS, PbSe, indirect-bandgap-semiconductor Si nanocrystals can be used in NIR QLEDs. In 2011, Cheng et al. obtained highly efficient Si nanocrystal LEDs with EQEs of 8.6% and 6.7% at emission wavelengths of 853 and 777 nm, respectively. To our knowledge, this is the best performance among nanocrystal-based LEDs emitting at similar wavelengths. The high device performance can be ascribed to the high intrinsic PLQYs of 45% and 43% at the peak wavelengths of 853 and 777 nm for Si nanocrystals of 5- and 3-nm diameter, respectively, that were chemically passivated with ligands of 1-dodecene in a uniformly dispersed solution. This work demonstrated that nanocrystals of indirect-gap Si semiconductors have exceptional photophysical properties in the NIR region.[137]

By using different organic ligands for surface passivation, the interdot distance of QDs can be tuned and their performance may be improved. In 2012, Hu et al. fabricated QLEDs that showed a maximum EQE of 0.73% and an EL peak at 1412 nm by using ZnO nanocrystals as the electron-transporting layers and a ligand replacement technique to activate PbSe QDs. After treatment with 1,2-ethanedithiol (EDT), the absorption peaks and PL spectrum of the PbSe QD film exhibited significant redshifts due to QD film densification.[146] Moreover, by tuning the distance between adjacent PbS QDs and varying their size, Sun et al. fabricated a series of thin-film QLEDs with EL emission peaks ranging from 884 to 1618 nm, and the best one showed a peak EQE of 2.0% at 1054 nm with a maximum radiance of 6.4 W sr⁻¹ m⁻². In this work, the distance between adjacent PbS dots is modulated by varying the lengths of the mercapto alkyl carboxylic acids from 3-mercaptopropionic acid, 6-mercaptohexanoic acid, and 8-mercaptooctanoic acid (MOA) to 11-mercaptoundecanoic acid, which acts as bifunctional linker molecules. It was proven that a small interdot distance in the QD layer facilitates efficient charge transport and charge injection as well as leads to undesired exciton dissociation. By contrast, a larger interdot distance promotes exciton radiative recombination, but charge injection becomes difficult and the screening of the external field is weakened, which may cause exciton ionization. Considering all the factors, the authors found that the optimum charge injection balance and radiative exciton recombination were achieved by using MOA as linker molecules.[138] Similarly, in 2012, Ma et al. reported a new way of producing high-quality PbS nanocrystalline films acting as both electron-transporting and emitting layers in NIR LEDs by the benzenedithiol (BDT) cross-linking method, which exhibits a 20% quantum yield in solution. The BDT-treated PbS nanocrystal films have lower doping concentration, lower conductivity, and less free-carrier density than the conventional EDT-treated nanocrystal films. Therefore, the undesirable nonradiative Auger recombination, which is proportional to the cube of the carrier density, can be suppressed in the BDT-treated films. Such devices exhibit a maximum EQE of 0.72% at 1050 nm.[140]

Furthermore, properly processed QDs can be used as electron-transporting or hole-transporting layers. Yang et al. reported on NIR LEDs based on QD materials as the hole-transporting, electron-transporting, and light-emitting layers, which are the so-called “all-quantum-dot IR LEDs” ([Figure 11]). By tuning the particle sizes, the QLEDs emit from 1220 to 1622 nm and the best one, emitting at 1350 nm, achieved a maximum EQE of 1.58% with a low V _on of 1.2 V. Because of the low mobility of colloidal QDs (CQDs), the injected carriers accumulated in the active layer mostly for radiative or Auger recombination rather than the SRH recombination. The simplified material requirements and solution-processed methods provide us with a new way to fabricate cost-effective NIR LEDs.[141] In 2015, Yan et al. used a blue GaN LED as the excitation source and PbSe QDs as the NIR luminescent material to fabricate NIR LEDs with a maximum EQE of 2.52% at 950 nm. Moreover, EQE values of 1.83% and 0.67% were also achieved for 1550 and 1960 nm QLEDs by changing the active-layer thickness.[139]

Figure 11 (a) Schematic illustration of the device architecture. Arrows represent the layers consisting of CQDs with different sizes and ligand passivation. (b) QLED energy diagram. The solid line in the active layer represents the CQDs emitting at 1622 nm and the dashed line represents the CQDs emitting at 1220 nm. (c) CQD absorption spectra acting as charge-transporting layers. (labs = 1178, 1320, 1427, and 1572 nm). Reprinted with permission from Ref. 141. Copyright 2015 American Chemical Society.

As described above, the use of core–shell QDs can serve as a useful way to inhibit surface oxidation and nonradiative pathways in QLEDs.[135] In 2015, Supran et al. reported that LEDs based on PbS–CdS core–shell QDs showed a peak EQE of 4.3% with a V _on as low as 0.6 V and EL peaks ranging from 1163 to 1341 nm; the performance of which significantly surpassed those of devices based on PbS core-only QDs ([Figure 12]). It was shown that the CdS shell synthesized by cation exchange significantly mitigates the in situ nonradiative pathways in PbS QDs, such as multicarrier Auger recombination, exciton energy transfer to conductive charge-transporting layers (CTLs), and exciton dissociation at CTL/QD interfaces; thus, the device performances can be enhanced significantly.[142]

Figure 12 (a) Schematic illustration of the cation-exchange reaction used to convert core-only PbS QDs into core–shell PbS–-CdS QDs. (b) Device architecture (left) and cross-sectional SEM image (right) of the QLED based on these QDs. (c) The QLED's flat-band energy level diagram. The electron affinity of PbS is tuned from approximately 3.8 to 3.9 eV by the reduction in core size (from 4.0 to 3.6 nm) accompanying cation exchange. Reprinted with permission from Ref. 142. Copyright 2015 Wiley-VCH.

Like OLEDs or PLEDs, by embedding QDs into a suitable matrix, we can obtain high-performance QLEDs due to the decreased interdot quenching. In 2016, Bansal et al. reported on highly luminescent CdS QDs, which featured a comparatively high PLQY of 69% in solution and 34% in a neat thin film. They were embedded in an organic semiconductor matrix to fabricate a NIR solution-processed hybrid LED. The QLED based on a blend of QDs with a CBP matrix, with a volume ratio of 0.05:0.95, shows a maximum EQE of 0.62% at a peak emission wavelength of 760 nm. For the CBP host QLEDs, the hole injection barrier is as high as 0.45 eV; thus, the EQE is much lower than might be expected for the high PLQY of CdS QDs, and the devices need to be optimized further.[136] According to previous reports, CQD films that have high luminescence efficiency and high carrier mobility simultaneously cannot be obtained, because high-mobility materials also feature a high exciton dissociation rate.[142] Nevertheless, Gong et al. addressed this issue with a strategy of embedding nanocrystals in a hybrid perovskite matrix with high carrier mobility. With the benefit of the excellent carrier-transport property and the suitable perovskite matrix consisting of a certain I/Br ratio, the transport-assisted trapping losses such as SRH recombination and the self-quenching on the CQD films can be inhibited, so the device emitting at 1391 nm achieves a maximum EQE of 5.2% and a record high power conversion efficiency (defined as the ratio of the energy of a photon emitted to the energy of an electron injected) of 4.9% so far with emission beyond 900 nm.[145]

Considering the size-dependent emission of QDs and their large exciton Bohr radius of commonly used Pb chalcogenide, it is vital to find an effective way of synthesizing monodisperse QDs for QLED fabrication. As described above, researchers have applied suitable core–shell structures or organic ligands to passivate the surface defects and suppress interdot quenching. Through the modulation of organic ligands on the surface of QDs, they have successfully tuned the interdot spacing and synthesized high-efficiency and high-uniformity QD films for NIR emission ([Table 3]). Therefore, it is reasonable to assume a bright future for QDs used in NIR QLED.

Table 3
Summary of NIR QD and perovskite materials and their corresponding EQE and EL peaks
Materials	EL Peak (nm)	EQE (%)	Ref.
MAPbI₃ QDs	748	10.4	[161]
CdS QDs[a]	760	0.62	[136]
PEA₂(MA)_n−1Pb_nI_{3n + 1} QWs	760	8.8	[163]
NMA₂(FA)_n−1Pb_n(I/Br)_{3n + 1} QWs	763	11.7	[151]
Si QDs (3 nm)	777	6.7	[137]
Bulk FAPbI₃	803	20.7	[150]
Si QDs (5 nm)	853	8.6	[137]
FPMAI-MAPb_0.6Sn_0.4I₂Br	854	1.4	[147]
FPMAI-MAPb_0.6Sn_0.4I_2.2Br_0.8	871	1.8	[147]
FPMAI-MAPb_0.8Sn_0.2I₃	877	0.7	[147]
PbS QDs	884	0.43	[138]
	1054	2.0
	1232	0.45
	1432	0.035
	1618	0.035
FPMAI-MAPb_0.6Sn_0.4I_2.4Br_0.6	885	1.7	[147]
FPMAI-MAPb_0.6Sn_0.4I_2.6Br_0.4	900	2.6	[147]
FPMAI-MASnI₃	901	0.9	[147]
FPMAI-MAPb_0.6Sn_0.4I_2.8Br_0.2	909	3.3	[147]
FPMAI-MAPb_0.6Sn_0.4I₃	917	5	[147]
FPMAI-MAPb_0.4Sn_0.6I₃	928	2.7	[147]
FPMAI-MAPb_0.2Sn_0.8I₃	943	0.6	[147]
MASnI₃	945	0.72	[156]
CsSnI₃	950	3.8	[155]
PbSe	950	2.52	[139]
	1550	1.83
	1950	0.67
PbS (BDT-treated)	1050	0.72	[140]
PbS CQDs	1220	0.79	[141]
	1350	1.58
	1450	0.37
	1622	0.14
PbS-CdS QDs	1230	4.3	[142]
PbS	1250	1.15	[143]
PbSe[b]	1280	0.83	[144]
PbS QDs	1391	4.7	[145]
PbSe (EDT-treated)	1412	0.73	[146]

^a Emitters doped in CBP host material.

^b Emitters doped in MEH-PPV (6 wt%) host material.

NIR Perovskite Materials

Considering the “energy-gap law” for organic materials and the difficulty in synthesizing monodispersed QDs, it is very important to investigate alternative NIR-emitting materials. Organometal halide perovskites with a high PLQY, excellent carrier mobility, tunable bandgap, long exciton diffusion length, solution processability, and good color purity are promising light-emitting materials for LED fabrication. In recent years, organometal halide perovskites have been found to possess excellent EL properties ranging from the visible to the NIR range. However, nonradiative recombination caused by defects and leakage current on the surface limits their EL performance. These issues have been addressed in numerous studies.[148] [149] [150] [151]

Typical organometal halide perovskites can be described with the general formula of L₂(SMX)_n-1MX₄, where M is a divalent metal cation, X is a halide, L and S are organic cations with long and short chains, respectively, and n is the number of semiconducting MX₄ monolayer sheets within the two organic insulating layers. When n = ∞, the structure corresponds to the 3D perovskite SMX₃; when n = 0, the structure corresponds to the 2D perovskite L₂MX₄; and n values in between are quasi-2D perovskites ([Figure 14c]). Hybrid 3D perovskites feature high PLQYs, high carrier mobility, and small exciton binding energy; thus, the 3D perovskite materials can achieve high PLQY at high-excitation photon fluences where radiative bimolecular recombination dominates. The 2D perovskites generally exhibit good film-formation properties, a larger exciton-binding energy than 3D perovskites, and low PLQYs at room temperature due to the fast exciton-quenching rates. Nevertheless, the quasi-2D layered perovskites (also called a “Ruddlesden–Popper structure”) with a self-organized quantum well (QW) structure show exceptional optical properties when used in PeLEDs.[151]

Figure 13 Device structure of PeLED. Inset: schematic representation of a nanometer-sized grain with n-butylammonium cations passivating their surfaces. Reprinted with permission from Ref. 161. Copyright 2016 Springer Nature.

3D Perovskite Materials

In 2015, Gil-Escrig et al. reported on PeLEDs based on CH₃NH₃PbI₃ exhibiting a peak EQE of 0.04% at an emission peak of 765 nm. Given the very low PLQY of the active layer, this work implied that the prospect for improvement in EL performance was bright for NIR PeLEDs.[152] Later in 2015, Kumawat et al. studied the structural, electrical, and optical properties of four 3D perovskite semiconductors with the formula CH₃NH₃PbX₃ (where X is a halide ion) with different bandgaps ranging from the visible to the NIR region by changing the halide ion, substrate wetting conditions, and crystallite size. A maximum EQE of 0.28% was achieved for a NIR device with an EL peak of approximately 760 nm.[153] In the same year, Wang et al. introduced the multifunctional interfacial layer of polyethyleneimine (PEI), which is conducive to forming high-quality perovskite thin films with good crystallinity features, long PL lifetimes, and good surface coverage for the solution-processed PeLEDs. They concluded that an organic polyamine can be used to improve the interface between the oxide electron-transporting layer and the perovskite emissive layer to produce a low-defect and high-quality heterojunction to improve device performance. Consequently, these authors successfully fabricated solution-processed PeLEDs with an EL peak of 768 nm and a maximum EQE of 3.5% (average EQE is 2.2%).[154] Unlike OLEDs or QLEDs, the EQE of the NIR PeLEDs increases with the increase in current density because higher excitation densities facilitate radiative bimolecular recombination in PeLEDs instead of the exciton–exciton and exciton–charge quenching that occurred in OLEDs and QLEDs.

To further effectively extract the light trapped in LEDs, namely to improve the light-outcoupling efficiency, many methods such as using low-index grids and buckling patterns have been investigated. However, these methods are always associated with complicated fabrication processes with the emission spectra and directionality distorted.[154] [155] [156] [157] [158] [159] Intriguingly, Cao et al. have exploited perovskites prepared by introducing amino-acid additives in the precursor solutions to fabricate efficient and bright NIR PeLEDs that can make full use of the light trapped in the devices because the spontaneously formed submicrometer-scale structures significantly increase the outcoupling efficiency. Because of the amino-acid additives, the perovskite surface defects are passivated and nonradiative recombination can be reduced; thus, the maximum EQE of 20.7% at a current density of 18 mA cm⁻² was achieved for the PeLED emitting at 803 nm.[150] In 2018, Giuri et al. obtained high-efficiency NIR formamidinium lead iodide (FAPbI₃) PeLEDs by introducing starch to manipulate the nanostructure of perovskite films, which exhibits a high radiance up to 206.7 W sr⁻¹ m⁻² at the very high current density of 1000 mA cm⁻², EQEs of approximately 5%, and EL wavelengths ranging from 751 to 801 nm. The addition of starch helped the formation of an α–δ mixed phase that can improve the radiative bimolecular recombination and circumvent losses at the interface between the emissive layer and the hole/electron-transporting layer; the improved substrate coverage of the perovskite film can inhibit current leakage and modulate the dielectric landscape, so that the light-outcoupling efficiency can be increased.[160]

Many reports have demonstrated that the photophysical properties of perovskite materials can be modulated by changing the halide composition. In 2017, Xiao et al. proposed that long-chain ammonium halides can act as surfactants that terminate the grain growth of 3D perovskite in the perovskite solutions during the film-forming process, producing small-grain-sized, ultrasmooth, pinhole-free, and compact perovskite films. After the addition of long-chain ammonium halides, the 3D nanometer-sized perovskite grains reduce their dimensions and start a transition from 3D to quasi-2D-like layered perovskite structures. By varying the ratio of BAX:MAPbX₃ (BAX = n-butylammonium halides; X = Br and I; MA = methylammonium) in perovskite precursor solutions and incorporating them in LEDs, they realized that the EQEs of 20:100 Br-PeLEDs emitting at 513 nm and 20:100 I-PeLEDs emitting at 748 nm reached 9.3% and 10.4%, respectively ([Figure 13]).[161] In 2019, Qiu et al. reported mixed Pb–Sn halide perovskite-based NIR LEDs with tunable emission peaks from 850 to 950 nm by varying the Pb:Sn ratio or introducing a bromide. With the addition of 4-fluorobenzylammonium iodide (FPMAI) additives in the perovskite precursor, they obtained ultrasmooth nanocrystalline perovskite films and fabricated a PeLED with a maximum EQE of 5.0%, a low V _on of 1.65 V, and an emission peak at 917 nm by using Pb–Sn mixed perovskite with a Pb:Sn ratio of 3:2.[147]

Figure 14 (a) Unit cell structures of PEA₂(CH₃NH₃)_n−1Pb_nI_{3n +1} perovskites with different (n) values from 2D (n = 1) to 3D (n = ∞). (b) EL spectra of PeLEDs based on PEA₂(CH₃NH₃)_n−1Pb_nI_{3n + 1} with different <n> values.[163] (c) Schematic representation of the structures of (NMA)₂(FAPbI₃)_n−1PbI₄. (d) Schematic cascade energy transfer in MQWs. Excitation energy is transferred from small-n QWs to large-n QWs followed by the emission in the large-n QWs.[151] Reprinted with permission from Ref. 163. Copyright 2016 Springer Nature. Reprinted with permission from Ref. 151. Copyright 2016 Springer Nature.

Considering the use of toxic elements such as Cd and Pb in QLEDs and PeLEDs, which raises environmental and health issues, Hong et al. reported on a lead-free NIR PeLED based on a low-temperature solution-processed CsSnI₃ perovskite. They used two methods, namely, the one-pot solution synthesis and the toluene dripping method, to prepare the CsSnI₃ films and discovered that the CsSnI₃ films prepared by the latter method showed compact micrometer-sized CsSnI₃ grains with few pinholes and cracks at the grain boundary. The NIR LEDs based on such high-quality CsSnI₃ films achieved a maximum EQE of 3.8% at 4.5 V, a maximum radiance of 40 W sr⁻¹ m⁻², and an EL peak at 950 nm. The characterization showed that the CsSnI₃ film made by toluene dripping exhibits a faster PL decay, lower nonradiative recombination rates, and lower trap density of states than the CsSnI₃ film prepared by one-pot solution synthesis, which explains the high performance of the former.[155] Additionally, Tan et al. fabricated methylammonium tin iodide (MASnI₃)-based PeLEDs with a peak EQE of 0.72% and an EL peak at 945 nm and achieved spectrum tunability by varying the Br content in these perovskite materials.[156] In 2017, Zhao et al. reported a nanocrystal film preparation strategy of using 4-fluorophenylmethylammonium iodide (FPMAI) as additives in situ to confine the crystal growth of the perovskite and achieved a high-quality perovskite film. Because of the good perovskite film, NIR LEDs based on CH₃NH₃PbI₃ emitting at 749 nm exhibited a maximum EQE of 7.9%, 40 times higher than the one without additives. It can be concluded that the additives passivate the defects on the perovskite surface, thereby enhancing the LED performance.[162]

Quasi-2D Layered Perovskite Materials

In 2016, Yuan et al. investigated a series of perovskite mixed material based on (C₈H₉NH₃)₂(CH₃NH₃)_n−1Pb_nI_3n+1 consisting of different quantum-size-tuned grains that transfer photogenerated excitons to the lowest-bandgap emitter in the mixture. In this work, they increased the number of PbI₆ monolayers <n> to make the average dimensionally tuned perovskite in the solid state for exploring how <n> affects the luminescence property of quasi-2D and 3D perovskite materials ([Figure 14a, b]). In consequence, they found that the EQE and the radiance for n = 5 reach up to 8.8% for 200-nm films and 80 W sr⁻¹ m⁻² respectively, with an EL peak at about 750 nm. It was demonstrated that the fast PL decay at shorter wavelengths in these materials can be ascribed to the intraband carrier funneling process, and the relatively lower PLQY can be attributed to the dominance of trap-assisted nonradiative recombination. They suggested that there is an increased local exciton density in low-bandgap and large-n grains from the energy funnel process, so the high PLQY under low excitation can be achieved effectively.[163] Similarly, Wang et al. presented solution-processed PeLEDs based on self-organized multiple QWs (MQWs) with good film morphology. To make good MQW perovskite films, they dissolved a precursor solution of 1-naphthylmethylamine iodide (NMAI), formamidinium iodide (FAI), and PbI₂ with a molar ratio of 2:1:2 in N,N-dimethylformamide (DMF) to deposit perovskite films (named NFPI₇). The authors then used NMAI, formamidinium bromide (FABr), and PbI₂ with a molar ratio of 2:1:2 in DMF to deposit new perovskite MQW films NFPI₆B with better crystallinity than NFPI₇ because of the changed halide composition in the precursor solution. A peak EQE of 11.7% at 2.6 V with a current density of 38 mA cm⁻² and an EL emission peak at 763 nm was achieved in the LEDs based on NFPI₆B ([Figure 14c, d]). In the MQW perovskite films, exciton decay is slower than the energy transfer. Most of the photogenerated excitons can be transferred from the small-n QWs to the large-n QWs, in which n is the number of MX₄ layers, so the local exciton density is high in the large-n QWs, which results in radiative recombination.[151]

Researchers have exploited 3D, quasi-2D, and 2D hybrid perovskite materials to make full use of the advantages of these materials, such as high carrier mobility, high PLQY, and tunable spectra ([Table 3]). Although long-term instability remains an issue and health or environmental concerns still exist, efforts are being made to find appropriate materials and optimized devices for lowering the driving voltages and circumventing the health and environmental risks.

Conclusions and Outlook

In conclusion, many difficulties were encountered when discovering and designing high-efficiency NIR-emitting materials. For organic materials, the performance is seldom satisfactory in the long-wavelength region, according to the “energy-gap law” and the aggregation-induced quenching effect. Extending the emission range further into the NIR region in OLEDs requires careful management of both the aggregation situation and the nonradiative deactivation pathways of the excited states in active layers.[9] Researchers carefully design the molecular structure of organic emitters and tune their energy levels to extend their emission to the NIR region through introducing different donor or acceptor moieties. By rigidifying the molecular backbone and by judicious molecule design, we can circumvent some negative factors to the greatest possible extent. The so-called “energy-gap law” is only applicable to organic molecules with similar structures, and we may thus find some exceptions and then summarize them to form a feasible molecule design strategy.[9] Doping and copolymerizing emitters into hosts are established methods for the fabrication of emitting layers with the aim of suppressing the aggregation-induced quenching effect. In these LEDs, the emitters accept excitons from hosts, followed by the radiative recombination of excitons in the emitting center. To efficiently leverage all the excitons generated in OLEDs, researchers have synthesized materials with TADF, TTA, a doublet radical mechanism, and organometallic materials. The organometallic materials perform best in all kinds of organic materials from the perspective of the EQEs of LEDs with emission peaks near 700 nm.[118] As mentioned in the Introduction, the potential applications of NIR LEDs in general span from security to biomedicine (e.g., blood oximetry, photodynamic therapy instruments, night-vision display). In addition to some common NIR applications, low-toxicity organic materials have a definite advantage over inorganic counterparts in the territory of in vivo applications such as bioelectronics, though the competition by inorganic NIR emitters is strong in terms of efficiency, radiance, and durability. Besides, the relatively low-cost organic materials have great potential in large-area fabrication due to their solution-processability and flexibility.

As for QD materials featuring high PLQY and brightness, small FWHM, and tunable emission, they still suffer from severe self-quenching problems in solid-state films because of the enhanced transport-assisted trapping caused by the coupling between QDs.[146] Another issue is that the surfaces of QDs should be carefully processed because active surface states may quench radiative excitons. The spectral tunability originates from the quantum size effect of QDs, by which the bandgaps of QDs increase as their sizes decrease. Thus, QDs perform well in the wavelength range above 1000 nm when compared with organic materials and perovskite materials. Moreover, the spectrally narrow emission of QDs compared with organic and inorganic emitters makes QDs excellent luminescent sources. Thus, NIR QLEDs have the potential in military applications, such as night-vision-readable displays[1] and iris identification systems,[7] due to their narrow emission. Similarly, considering the longer penetration depths and lower damage to living tissues of NIR light, the NIR QLEDs with emission from 800 to 1700 nm have a competitive advantage in the field of biomedical imaging and optical diagnosis.[6]

Perovskite materials feature high color purity and quantum efficiencies, good color tunability, and high flexibility, but their exciton binding energy is low, so they exhibit severe exciton dissociation when used as emitting materials at room temperature. In addition, it is hard to fabricate high-stability perovskite devices because the PeLEDs are usually driven at a high voltage. Moreover, Most QD materials and perovskite materials contain Pb and Cr, elements that are detrimental to human health and the environment, and this limits their future application in daily life and in vivo applications. By the careful choice of suitable additives in the precursor solutions and changing the content ratios of perovskite materials, researchers have successfully made high-quality thin films for PeLED fabrication and tuned the emission peaks of the devices further to the NIR region. Furthermore, the self-organized QW structure from the quasi-2D perovskites helps in understanding the structure–property relationships of perovskite materials and alternative candidates for NIR emission. The perovskite materials show best performances in terms of EQE when compared with QDs and organic materials. The high EQEs and radiances of PeLEDs also make perovskite materials very promising for night-vision display.

We hope that this review of NIR emitting materials for LED fabrication will stimulate further investigation in this area, at both industrial and academic levels. And further basic studies in the luminescence mechanism will promote the understanding of material design. All in all, the existing problems will be solved at some point, and undoubtedly, a bright future can be assumed for commercial applications of these NIR LEDs.

References

References
1 Qi J, Qiao W, Wang ZY. Chem. Rec. 2016; 16: 1531
2 Pansare V, Hejazi S, Faenza W, Prud'homme RK. Chem. Mater. 2012; 24: 812
3 Smith AM, Mancini MC, Nie S. Nat. Nanotechnol. 2009; 4: 710
4 Le ST, Kanesan T, Bausi F, Haigh PA, Rajbhandari S, Ghassemlooy Z, Papakonstantinou I, Popoola WO, Burton A, Le Minh H, Cacialli F, Ellis AD. Opt. Lett. 2014; 39: 3876
5 Haigh PA, Bausi F, Ghassemlooy Z, Papakonstantinou I, Le Minh H, Fléchon C, Cacialli F. Opt. Express 2014; 22: 2830
6 Maggini L, Cabrera I, Ruiz-Carretero A, Prasetyanto EA, Robinet E, De Cola L. Nanoscale 2016; 8: 7240
7 Kono M, Ueki H, Umemura S. Appl. Opt. 2002; 41: 7429
8 Segal M, Baldo MA, Holmes RJ, Forrest SR, Soos ZG. Phys. Rev. B: Condens. Matter 2003; 68: 075211
9 Zampetti A, Minotto A, Cacialli F. Adv. Funct. Mater. 2019; 29: 1807623
10 Gao Q, Wang ZY. Chem. Asian J. 2010; 5: 1006
11 Dou L, Liu Y, Hong Z, Li G, Yang Y. Chem. Rev. 2015; 115: 12633
12 Luo J, Xie Z, Lam JW. Y, Cheng L, Tang BZ, Chen H, Qiu C, Kwok HS, Zhan X, Liu Y, Zhu D. Chem. Commun. 2001; 18: 1740
13 Baldo MA, O'Brien DF, Thompson ME, Forrest SR. Phys. Rev. B: Condens. Matter 1999; 60: 14422
14 Veldhuis SA, Boix PP, Yantara N, Li M, Sum TC, Mathews N, Mhaisalkar SG. Adv. Mater. 2016; 28: 6804
15 Burroughes JH, Bradley DD. C, Brown AR, Marks RN, Mackay K, Friend RH, Burns PL, Holmes AB. Nature 1990; 347: 539
16 Baigent DR, Hamer PJ, Friend RH, Moratti SC, Holmes AB. Synth. Met. 1995; 71: 2175
17 Iqbal R, Moratti SC, Holmes AB, Yahioglu G, Milgrom LR, Cacialli F, Morgado J, Friend RH. J. Mater. Sci. - Mater. Electron. 2000; 11: 97
18 Morgado J, Cacialli F, Friend RH, Iqbal R, Yahioglu G, Milgrom LR, Moratti SC, Holmes AB. Chem. Phys. Lett. 2000; 325: 552
19 Ostrowski JC, Susumu K, Robinson MR, Therien MJ, Bazan GC. Adv. Mater. 2003; 15: 1296
20 Fenwick O, Sprafke JK, Binas J, Kondratuk DV, Di Stasio F, Anderson HL, Cacialli F. Nano Lett. 2011; 11: 2451
21 Graham JP, Mininni PD, Pouquet A, Shelton AH, Schanze KS, Xue J. Phys. Rev. E: Stat. Nonlinear Soft Matter Phys. 2011; 84: 016314
22 Borek C, Hanson K, Djurovich PI, Thompson ME, Aznavour K, Bau R, Sun Y, Forrest SR, Brooks J, Michalski L, Brown J. Angew. Chem. 2007; 119: 1127
23 Huang L, Park CD, Fleetham T, Li J. Appl. Phys. Lett. 2016; 109: 233302
24 Sommer JR, Farley RT, Graham KR, Yang Y, Reynolds JR, Xue J, Schanze KS. ACS Appl. Mater. Interfaces 2009; 1: 274
25 Sun Y, Borek C, Hanson K, Djurovich PI, Thompson ME, Brooks J, Brown JJ, Forrest SR. Appl. Phys. Lett. 2007; 90: 213503
26 Freeman DM. E, Tregnago G, Rodriguez SA, Fallon KJ, Cacialli F, Bronstein H. J. Org. Semicond. 2015; 3: 1
27 Iqbal R, Yahioglu G, Milgrom L, Moratti SC, Holmes AB, Cacialli F, Morgado J, Friend RH. Synth. Met. 1999; 102: 1024
28 Stender B, Völker SF, Lambert C, Pflaum J. Adv. Mater. 2013; 25: 2943
29 Havinga EE, ten Hoeve W, Wynberg H. Polym. Bull. 1992; 29: 119
30 Brocks G, Tol A. J. Phys. Chem. 1996; 100: 1838
31 Brocks G, Tol A. Synth. Met. 1996; 76: 213
32 Ellinger S, Graham KR, Shi P, Farley RT, Steckler TT, Brookins RN, Taranekar P, Mei J, Padilha LA, Ensley TR, Hu H, Webster S, Hagan DJ, Van Stryland EW, Schanze KS, Reynolds JR. Chem. Mater. 2011; 23: 3805
33 Minotto A, Murto P, Genene Z, Zampetti A, Carnicella G, Mammo W, Andersson MR, Wang E, Cacialli F. Adv. Mater. 2018; 30: 1706584
34 Hendriks KH, Li W, Wienk MM, Janssen RA. J. Am. Chem. Soc. 2014; 136: 12130
35 Murto P, Minotto A, Zampetti A, Xu X, Andersson MR, Cacialli F, Wang E. Adv. Opt. Mater. 2016; 4: 2068
36 Parker TC, Patel DG, Moudgil K, Barlow S, Risko C, Brédas J.-L, Reynolds JR, Marder SR. Mater. Horiz. 2015; 2: 22
37 Kronemeijer AJ, Gili E, Shahid M, Rivnay J, Salleo A, Heeney M, Sirringhaus H. Adv. Mater. 2012; 24: 1558
38 Yuen JD, Kumar R, Zakhidov D, Seifter J, Lim B, Heeger AJ, Wudl F. Adv. Mater. 2011; 23: 3780
39 Fenwick O, Fusco S, Baig TN, Di Stasio F, Steckler TT, Henriksson P, Fléchon C, Andersson MR, Cacialli F. APL Mater. 2013; 1: 032108
40 Shahid M, Ashraf RS, Huang Z, Kronemeijer AJ, McCarthy-Ward T, McCulloch I, Durrant JR, Sirringhaus H, Heeney M. Chem. Sci. 2012; 3: 181
41 Pandey L, Risko C, Norton JE, Brédas J.-L. Macromolecules 2012; 45: 6405
42 Yang R, Tian R, Hou Q, Yang W, Cao Y. Macromolecules 2003; 36: 7453
43 Yang R, Tian R, Yan J, Zhang Y, Yang J, Hou Q, Yang W, Zhang C, Cao Y. Macromolecules 2005; 38: 244
44 Tregnago G, Steckler TT, Fenwick O, Andersson MR, Cacialli F. J. Mater. Chem. C 2015; 3: 2792
45 Kronemeijer AJ, Gili E, Shahid M, Rivnay J, Salleo A, Heeney M, Sirringhaus H. Adv. Mater. 2012; 24: 1558
46 Yang J, Jiang C, Zhang Y, Yang R, Yang W, Hou Q, Cao Y. Macromolecules 2004; 37: 1211
47 Sharbati MT, Panahi F, Shourvarzi A, Khademi S, Emami F. Optik 2013; 124: 52
48 Sun M, Jiang X, Liu W, Zhu T, Huang F, Cao Y. Synth. Met. 2012; 162: 1406
49 Sun M, Jiang X, Wang L, He C, Du B, Yang R, Cao Y. J. Polym. Sci., Part A: Polym. Chem. 2008; 46: 3007
50 Qian G, Dai B, Luo M, Yu D, Zhan J, Zhang Z, Ma D, Wang ZY. Chem. Mater. 2008; 20: 6208
51 Qian G, Zhong Z, Luo M, Yu D, Zhang Z, Wang ZY, Ma D. Adv. Mater. 2009; 21: 111
52 Du X, Qi J, Zhang Z, Ma D, Wang ZY. Chem. Mater. 2012; 24: 2178
53 Yang Y, Farley RT, Steckler TT, Eom S.-H, Reynolds JR, Schanze KS, Xue J. Appl. Phys. Lett. 2008; 93: 163305
54 Chen M, Perzon E, Andersson MR, Marcinkevicius S, Jönsson SK. M, Fahlman M, Berggren M. Appl. Phys. Lett. 2004; 84: 3570
55 Steckler TT, Lee MJ, Chen Z, Fenwick O, Andersson MR, Cacialli F, Sirringhaus H. J. Mater. Chem. C 2014; 2: 5133
56 Nielsen CB, Turbiez M, McCulloch I. Adv. Mater. 2013; 25: 1859
57 Zambounis JS, Hao Z, Iqbal A. Nature 1997; 388: 131
58 Bronstein H, Chen Z, Ashraf RS, Zhang W, Du J, Durrant JR, Shakya Tuladhar P, Song K, Watkins SE, Geerts Y, Wienk MM, Janssen RA. J, Anthopoulos T, Sirringhaus H, Heeney M. J. Am. Chem. Soc. 2011; 133: 3272
59 Stas S, Balandier J.-Y, Lemaur V, Fenwick O, Tregnago G, Quist F, Cacialli F, Cornil J, Geerts YH. Dyes Pigm. 2013; 97: 198
60 Zhang K, Tieke B. Macromolecules 2008; 41: 7287
61 Beyerlein T, Tieke B, Forero-Lenger S, Brütting W. Synth. Met. 2002; 130: 115
62 Slooff LH, Polman A, Cacialli F, Friend RH, Hebbink GA, van Veggel FC. J. M, Reinhoudt DN. Appl. Phys. Lett. 2001; 78: 2122
63 Li P, Fenwick O, Yilmaz S, Breusov D, Caruana DJ, Allard S, Scherf U, Cacialli F. Chem. Commun. 2011; 47: 8820
64 Zampetti A, Minotto A, Squeo BM, Gregoriou VG, Allard S, Scherf U, Chochos CL, Cacialli F. Sci. Rep. 2017; 7: 1611
65 Sassi M, Buccheri N, Rooney M, Botta C, Bruni F, Giovanella U, Brovelli S, Beverina L. Sci. Rep. 2016; 6: 34096
66 Lombeck F, Di D, Yang L, Meraldi L, Athanasopoulos S, Credgington D, Sommer M, Friend RH. Macromolecules 2016; 49: 9382
67 Cacialli F, Stoneham M. J. Phys.: Condens. Matter 2002; 14: V9
68 Qian G, Zhong Z, Luo M, Yu D, Zhang Z, Ma D, Wang ZY. J. Phys. Chem. C 2009; 113: 1589
69 Ledwon P, Zassowski P, Jarosz T, Lapkowski M, Wagner P, Cherpak V, Stakhira P. J. Mater. Chem. C 2016; 4: 2219
70 Yao L, Zhang S, Wang R, Li W, Shen F, Yang B, Ma Y. Angew. Chem. Int. Ed. 2014; 53: 2119
71 Han X, Bai Q, Yao L, Liu H, Gao Y, Li J, Liu L, Liu Y, Li X, Lu P, Yang B. Adv. Funct. Mater. 2015; 25: 7521
72 Ellinger S, Graham KR, Shi P, Farley RT, Steckler TT, Brookins RN, Taranekar P, Mei J, Padilha LA, Ensley TR. Chem. Mater. 2011; 23: 3805
73 Mayerhoffer U, Fimmel B, Würthner F. Angew. Chem. Int. Ed. 2012; 51: 164
74 Pope M, Kallmann HP, Magnante P. J. Chem. Phys. 1963; 38: 2042
75 Delorme R, Perrin F. J. Phys. Radium 1929; 10: 177
76 Parker CA, Hatchard CG. Trans. Faraday Soc. 1961; 57: 189423
77 Parker CA, Joyce TA. Chem. Commun. 1968; 1421
78 Uoyama H, Goushi K, Shizu K, Nomura H, Adachi C. Nature 2012; 492: 234
79 Adachi C, Baldo MA, Thompson ME, Forrest SR. J. Appl. Phys. 2001; 90: 5048
80 Dias FB, Bourdakos KN, Jankus V, Moss KC, Kamtekar KT, Bhalla V, Santos J, Bryce MR, Monkman AP. Adv. Mater. 2013; 25: 3707
81 Freeman DM. E, Musser AJ, Frost JM, Stern HL, Forster AK, Fallon KJ, Rapidis AG, Cacialli F, McCulloch I, Clarke TM, Friend RH, Bronstein H. J. Am. Chem. Soc. 2017; 139: 11073
82 Xue J, Li C, Xin L, Duan L, Qiao J. Chem. Sci. 2016; 7: 2888
83 Peng Q, Obolda A, Zhang M, Li F. Angew. Chem. Int. Ed. 2015; 54: 7091
84 Ai X, Evans EW, Dong S, Gillett AJ, Guo H, Chen Y, Hele TJ. H, Friend RH, Li F. Nature 2018; 563: 536
85 Lee SY, Yasuda T, Komiyama H, Lee J, Adachi C. Adv. Mater. 2016; 28: 4019
86 Sun K, Chu D, Cui Y, Tian W, Sun Y, Jiang W. Org. Electron. 2017; 48: 389
87 Li C, Duan R, Liang B, Han G, Wang S, Ye K, Liu Y, Yi Y, Wang Y. Angew. Chem. Int. Ed. 2017; 56: 11525
88 Hu Y, Yuan Y, Shi Y.-L, Lin J.-D, Jiang Z.-Q, Liao L.-S. J. Mater. Chem. C 2018; 6: 1407
89 Li J, Nakagawa T, MacDonald J, Zhang Q, Nomura H, Miyazaki H, Adachi C. Adv. Mater. 2013; 25: 3319
90 Zeng W, Lai HY, Lee WK, Jiao M, Shiu YJ, Zhong C, Gong S, Zhou T, Xie G, Sarma M, Wong KT, Wu CC, Yang C. Adv. Mater. 2018; 30: 1704961
91 Wang S, Yan X, Cheng Z, Zhang H, Liu Y, Wang Y. Angew. Chem. Int. Ed. 2015; 54: 13068
92 Yuan Y, Hu Y, Zhang Y.-X, Lin J.-D, Wang Y.-K, Jiang Z.-Q, Liao L.-S, Lee S.-T. Adv. Funct. Mater. 2017; 27: 1700986
93 Hu Y, Yuan Y, Shi Y.-L, Li D, Jiang Z.-Q, Liao L.-S. Adv. Funct. Mater. 2018; 28: 1802597
94 D'Aléo A, Sazzad MH, Kim DH, Choi EY, Wu JW, Canard G, Fages F, Ribierre JC, Adachi C. Chem. Commun. 2017; 53: 7003
95 Kim D.-H, D'Aléo A, Chen X.-K, Sandanayaka AD. S, Yao D, Zhao L, Komino T, Zaborova E, Canard G, Tsuchiya Y, Choi E, Wu JW, Fages F, Brédas J.-L, Ribierre J.-C, Adachi C. Nat. Photonics 2018; 12: 98
96 Nakanotani H, Higuchi T, Furukawa T, Masui K, Morimoto K, Numata M, Tanaka H, Sagara Y, Yasuda T, Adachi C. Nat. Commun. 2014; 5: 4016
97 Nagata R, Nakanotani H, Adachi C. Adv. Mater. 2017; 29: 1604265
98 Zhang D, Duan L, Li C, Li Y, Li H, Zhang D, Qiu Y. Adv. Mater. 2014; 26: 5050
99 Xue J, Liang Q, Zhang Y, Zhang R, Duan L, Qiao J. Adv. Funct. Mater. 2017; 27: 1703283
100 Song X, Zhang D, Lu Y, Yin C, Duan L. Adv. Mater. 2019; 31: 1901923
101 Xiang H, Cheng J, Ma X, Zhou X, Chruma JJ. Chem. Soc. Rev. 2013; 42: 6128
102 Kido J, Ikeda W, Kimura M, Nagai K. Jpn. J. Appl. Phys. 1996; 35: L394
103 Wei H, Yu G, Zhao Z, Liu Z, Bian Z, Huang C. Dalton Trans. 2013; 42: 8951
104 Sun RG, Wang YZ, Zheng QB, Zhang HJ, Epstein AJ. J. Appl. Phys. 2000; 87: 7589
105 Harrison BS, Foley TJ, Bouguettaya M, Boncella JM, Reynolds JR, Schanze KS, Shim J, Holloway PH, Padmanaban G, Ramakrishnan S. Appl. Phys. Lett. 2001; 79: 3770
106 Curry RJ, Gillin WP, Knights AP, Gwilliam R. Appl. Phys. Lett. 2000; 77: 2271
107 Khreis OM, Curry RJ, Somerton M, Gillin WP. J. Appl. Phys. 2000; 88: 777
108 Curry RJ, Gillin WP. Appl. Phys. Lett. 1999; 75: 1380
109 Gillin WP, Curry RJ. Appl. Phys. Lett. 1999; 74: 798
110 Kawamura Y, Wada Y, Yanagida S. Jpn. J. Appl. Phys. 2001; 40: 350
111 Cao X, Miao J, Zhu M, Zhong C, Yang C, Wu H, Qin J, Cao Y. Chem. Mater. 2015; 27: 96
112 Baldo MA, O'Brien DF, You Y, Shoustikov A, Sibley S, Thompson ME, Forrest SR. Nature 1998; 395: 151
113 Baldo MA, Lamansky S, Burrows PE, Thompson ME, Forrest SR. Appl. Phys. Lett. 1999; 75: 4
114 Cleave V, Yahioglu G, Barny PL, Friend RH, Tessler N. Adv. Mater. 1999; 11: 285
115 Adachi C, Baldo MA, Forrest SR, Thompson ME. Appl. Phys. Lett. 2000; 77: 904
116 Lamansky S, Djurovich P, Murphy D, Abdel-Razzaq F, Lee HE, Adachi C, Burrows PE, Forrest SR, Thompson ME. J. Am. Chem. Soc. 2001; 123: 4304
117 Harrison BS, Foley TJ, Knefely AS, Mwaura JK, Cunningham GB, Kang T.-S, Bouguettaya M, Boncella JM, Reynolds JR, Schanze KS. Chem. Mater. 2004; 16: 2938
118 Tuong Ly K, Chen-Cheng R.-W, Lin H.-W, Shiau Y.-J, Liu S.-H, Chou P.-T, Tsao C.-S, Huang Y.-C, Chi Y. Nat. Photonics 2017; 11: 63
119 Qiao J, Duan L, Tang L, He L, Wang L, Qiu Y. J. Mater. Chem. 2009; 19: 6573
120 Tao R, Qiao J, Zhang G, Duan L, Chen C, Wang L, Qiu Y. J. Mater. Chem. C 2013; 1: 6446
121 Kesarkar S, Mroz W, Penconi M, Pasini M, Destri S, Cazzaniga M, Ceresoli D, Mussini PR, Baldoli C, Giovanella U, Bossi A. Angew. Chem. Int. Ed. 2016; 55: 2714
122 Rayabarapu DK, Paulose BM. J. S, Duan JP, Cheng CH. Adv. Mater. 2005; 17: 349
123 Tao R, Qiao J, Zhang G, Duan L, Wang L, Qiu Y. J. Phys. Chem. C 2012; 116: 11658
124 Xin L, Xue J, Lei G, Qiao J. RSC Advances 2015; 5: 42354
125 Nisic F, Colombo A, Dragonetti C, Roberto D, Valore A, Malicka JM, Cocchi M, Freeman GR, Williams JA. G. J. Mater. Chem. C 2014; 2: 1791
126 Freeman DM. E, Minotto A, Duffy W, Fallon KJ, McCulloch I, Cacialli F, Bronstein H. Polym. Chem. 2016; 7: 722
127 Cocchi M, Virgili D, Fattori V, Williams JA. G, Kalinowski J. Appl. Phys. Lett. 2007; 90: 023506
128 Cocchi M, Kalinowski J, Virgili D, Williams JA. G. Appl. Phys. Lett. 2008; 92: 113302
129 Williams EL, Li J, Jabbour GE. Appl. Phys. Lett. 2006; 89: 083506
130 Ho C.-L, Yao B, Zhang B, Wong K.-L, Wong W.-Y, Xie Z, Wang L, Lin Z. Organomet. Chem. 2013; 730: 144
131 Tsuzuki T, Tokito S. Adv. Mater. 2007; 19: 276
132 Xue J, Xin L, Hou J, Duan L, Wang R, Wei Y, Qiao J. Chem. Mater. 2017; 29: 4775
133 Lee T.-C, Hung J.-Y, Chi Y, Cheng Y.-M, Lee G.-H, Chou P.-T, Chen C.-C, Chang C.-H, Wu C.-C. Adv. Funct. Mater. 2009; 19: 2639
134 Liao JL, Chi Y, Liu SH, Lee GH, Chou PT, Huang HX, Su YD, Chang CH, Lin JS, Tseng MR. Inorg. Chem. 2014; 53: 9366
135 Tessler N, Medvedev V, Kazes M, Kan S, Banin U. Science 2002; 295: 1506
136 Bansal AK, Antolini F, Zhang S, Stroea L, Ortolani L, Lanzi M, Serra E, Allard S, Scherf U, Samuel ID. W. J. Phys. Chem. C 2016; 120: 1871
137 Cheng KY, Anthony R, Kortshagen UR, Holmes RJ. Nano Lett. 2011; 11: 1952
138 Sun L, Choi JJ, Stachnik D, Bartnik AC, Hyun BR, Malliaras GG, Hanrath T, Wise FW. Nat. Nanotechnol. 2012; 7: 369
139 Yan L, Shen X, Zhang Y, Zhang T, Zhang X, Feng Y, Yin J, Zhao J, Yu W. RSC Advances 2015; 5: 54109
140 Ma X, Xu F, Benavides J, Cloutier SG. Org. Electron. 2012; 13: 525
141 Yang Z, Voznyy O, Liu M, Yuan M, Ip AH, Ahmed OS, Levina L, Kinge S, Hoogland S, Sargent EH. ACS Nano 2015; 9: 12327
142 Supran GJ, Song KW, Hwang GW, Correa RE, Scherer J, Dauler EA, Shirasaki Y, Bawendi MG, Bulović V. Adv. Mater. 2015; 27: 1437
143 Bourdakos KN, Dissanayake DM. N. M, Lutz T, Silva SR. P, Curry RJ. Appl. Phys. Lett. 2008; 92: 153311
144 Choudhury KR, Song DW, So F. Org. Electron. 2010; 11: 23
145 Gong X, Yang Z, Walters G, Comin R, Ning Z, Beauregard E, Adinolfi V, Voznyy O, Sargent EH. Nat. Photonics 2016; 10: 253
146 Hu W, Henderson R, Zhang Y, You G, Wei L, Bai Y, Wang J, Xu J. Nanotechnology 2012; 23: 375202
147 Qiu W, Xiao Z, Roh K, Noel NK, Shapiro A, Heremans P, Rand BP. Adv. Mater. 2019; 31: 1806105
148 Chiba T, Hayashi Y, Ebe H, Hoshi K, Sato J, Sato S, Pu Y.-J, Ohisa S, Kido J. Nat. Photonics 2018; 12: 681
149 Lin K, Xing J, Quan LN, de Arquer FP. G, Gong X, Lu J, Xie L, Zhao W, Zhang D, Yan C, Li W, Liu X, Lu Y, Kirman J, Sargent EH, Xiong Q, Wei Z. Nature 2018; 562: 245
150 Cao Y, Wang N, Tian H, Guo J, Wei Y, Chen H, Miao Y, Zou W, Pan K, He Y, Cao H, Ke Y, Xu M, Wang Y, Yang M, Du K, Fu Z, Kong D, Dai D, Jin Y, Li G, Li H, Peng Q, Wang J, Huang W. Nature 2018; 562: 249
151 Wang N, Cheng L, Ge R, Zhang S, Miao Y, Zou W, Yi C, Sun Y, Cao Y, Yang R, Wei Y, Guo Q, Ke Y, Yu M, Jin Y, Liu Y, Ding Q, Di D, Yang L, Xing G, Tian H, Jin C, Gao F, Friend RH, Wang J, Huang W. Nat. Photonics 2016; 10: 699
152 Gil-Escrig L, Longo G, Pertegás A, Roldán-Carmona C, Soriano A, Sessolo M, Bolink HJ. Chem. Commun. 2015; 51: 569
153 Kumawat NK, Dey A, Narasimhan KL, Kabra D. ACS Photonics 2015; 2: 349
154 Wang J, Wang N, Jin Y, Si J, Tan ZK, Du H, Cheng L, Dai X, Bai S, He H, Ye Z, Lai ML, Friend RH, Huang W. Adv. Mater. 2015; 27: 2311
155 Hong WL, Huang YC, Chang CY, Zhang ZC, Tsai HR, Chang NY, Chao YC. Adv. Mater. 2016; 28: 8029
156 Lai ML, Tay TY, Sadhanala A, Dutton SE, Li G, Friend RH, Tan ZK. J. Phys. Chem. Lett. 2016; 7: 2653
157 Matterson BJ, Lupton JM, Safonov AF, Salt MG, Barnes WL, Samuel ID. W. Adv. Mater. 2001; 13: 123
158 Ziebarth JM, Saafir AK, Fan S, McGehee MD. Adv. Funct. Mater. 2004; 14: 451
159 Sun Y, Forrest SR. Nat. Photonics 2008; 2: 483
160 Giuri A, Yuan Z, Miao Y, Wang J, Gao F, Sestu N, Saba M, Bongiovanni G, Colella S, Esposito Corcione C, Gigli G, Listorti A, Rizzo A. Sci. Rep. 2018; 8: 15496
161 Xiao Z, Kerner RA, Zhao L, Tran NL, Lee KM, Koh T.-W, Scholes GD, Rand BP. Nat. Photonics 2017; 11: 108
162 Zhao L, Yeh YW, Tran NL, Wu F, Xiao Z, Kerner RA, Lin YL, Scholes GD, Yao N, Rand BP. ACS Nano 2017; 11: 3957
163 Yuan M, Quan LN, Comin R, Walters G, Sabatini R, Voznyy O, Hoogland S, Zhao Y, Beauregard EM, Kanjanaboos P, Lu Z, Kim DH, Sargent EH. Nat. Nanotechnol. 2016; 11: 872

Figures