Special Topic : Quantum Physics and Devices of Quantum Dots Colloidal quantum-dots surface and device structure engineering for high-performance light-emitting diodes

The application of colloidal quantum dots for light-emitting devices has attracted considerable attention in recent years, due to their unique optical properties such as size-dependent emission wavelength, sharp emission peak and high luminescent quantum yield. Tremendous efforts have been made to explore quantum dots for light-emission applications such as light-emitting diodes (LEDs) and light converters. The performance of quantum-dots-based light-emitting diodes (QD-LEDs) has been increasing rapidly in recent decades as the development of quantum-dots synthesis, surface-ligand engineering and device-architecture optimization. Recently, the external quantum efficiencies of red quantum-dots LEDs have exceeded 20.5% with good stability and narrow emission peak. In this review, we summarize the recent advances in QD-LEDs, focusing on quantum-dot surface engineering and device-architecture optimization.

Although QDs generally possess high PLQY in solution, the realization of high PL QY QD film is much more difficult.Firstly, the ligand loss during film fabrication can bring about the formation of defects.Effective surface-ligand passivation is therefore critical for colloidal QDs to keep their quantum confinement character and high luminescence quantum yield.In addition, in film, carrier transport between REVIEW Shang and Ning 171 QDs may deteriorate carrier recombination, as well as non-radiative Auger recombination.Shell growth is an effective strategy to relax the impact of ligand loss [39,45].In recent years, surface-ligand modification was widely explored to enhance the bonding strength of ligands in order to prevent ligand desorption, improve carrier injection and maintain charge balance [29,46,47].Recently, in-situ growth of ligands was realized by epitaxy growth of perovskite on QD surface in film, which presented a new platform for reducing defects [48].
Apart from the improvement of active materials, great progress has been made in device architecture.In the early stages, QD-LEDs generally utilized QDs embedded in an organic host as the carrier-transporting matrix [49].This strategy endows QDs with high PLQY in film.Nevertheless, carrier injection from matrix to QDs was hampered by the alkyl chain ligands.In recent years, much effort has been devoted to enhancing the carrier transport through the introduction of an inorganic or organic/inorganic hybrid matrix [48,50,51].A sandwiched structure that integrates pure QD films between electrons and hole-transporting layers was explored as well to enhance carrier transport.Meanwhile, much work has been made to optimize carriertransporting layers including tailoring of the band alignment and the use of carrier-blocking layers [29].
In this review, we will firstly introduce QD-LEDs, the concept and figure of merits for readers.We will then discuss QD surface-ligand engineering and the use of core-shell architecture.Device-architecture improvement, such as active-layer and transportinglayer optimization, will subsequently be presented including the analysis of the pros and cons of each device architecture.Finally, a perspective on development trends will be presented.

A BRIEF INTRODUCTION OF QD-LEDs
The typical dimension of QDs lies at around 1-10 nm and they are composed of a few hundred to 10 000 atoms [52].When excitons are confined in three spatial dimensions, discrete electronic states will replace the continuous energy bands of a bulk material, and their electronic band structure becomes strongly size-dependent, as shown in Fig. 1a [53][54][55].QDs therefore exhibit unique properties such as band-gap tunability, sharp emission peak and high PLQY.Figure 1b shows the emission range of classic QDs.A summary of commonly used QDs is shown in Table 1.
A classic device architecture of QD-LEDs is shown in Fig. 1c, which consists of an active emission layer, an electron-transporting layer (ETL) and a hole-transporting layer (HTL).Electrons and holes  are injected from the ETL and HTL, respectively, and recombine at the active layer to drive light emission.By tuning the size and composition of QDs, the emission wavelength can be conveniently tuned.
Figure 1d shows the efficiency development of QD-LEDs in recent decades.Ever since the first demonstration of QD-LEDs in 1994 [82], a considerable improvement in device efficiency and stability have been realized [29,30].Up to now, the demonstrated record peak external quantum efficiency (EQE) and luminance for QD-LEDs have reached 20.5% and 106 000 Cd/m 2 for red [29,83], 14.5% and 218 800 Cd/m 2 for green [30,84] and 12.2% and 7600 Cd/m 2 for blue [85], respectively.Apart from the single-color device, full-color display has also been demonstrated [33,86,87].

Figures of merit for QD-LEDs
Luminance, EQE, luminous efficiency and lifetime are critical figures of merit for judging device perfor-mance.The luminance of a device is defined by the derivative where L v is the luminance (in Cd/m 2 ), d 2 v is the luminous flux (in lm) output from the area d in any direction contained inside the solid angle d , d is an infinitesimal area (in m 2 ) of the source, d is an infinitesimal solid angle (sr) and θ is the angle between the normal n to the surface.
EQE is the generally used parameter for evaluating the performance of LED, which is defined as the ratio of the emitted photons from the device to the injected electrons: where i is the electrical current, e is the electron charge, hν is the photon energy and P 0 is the output power [88,89].

REVIEW Shang and Ning 173
Luminous efficiency is a parameter weighing the conversion efficiency of visible light emission, which is defined as the ratio of luminous flux to power (lm/W) [90].Turn-on voltage is the minimum voltage required to drive the current across the diode and start light-emitting.Full width at half maximum (FWHM) is the width of a spectrum range between points on the y-axis that are half the maximum amplitude.A small FWHM value is required for high color purity.

Surface shell growth
The nanoscale QDs have a high surface-to-volume ratio.A significant fraction of atoms are located on the surface, and some atoms are not fully coordinated, which need to be passivated by surface ligands.The loss of ligands leads to the formation of dangling bonds and defects on the surface, which can be evidenced by low PL QY.Overcoating QDs with higher-band-gap inorganic shell material can confine carriers to the central part of QDs (Fig. 2a) [57,[91][92][93][94][95][96], which can decrease the reactivity of the QD surface with oxygen and water, as well as create the possibility for the formation of a dangling bond [94].
In addition to the energy-level alignment, the match of the crystal structure and the lattice constant between the core and the shell are important to avoid lattice strain at the interface.The first reported prototype core/shell structure is in CdSe/ZnS QDs, in which the ZnS shell significantly improves the PLQY and stability against photo bleaching.When the shell becomes thick, the strain-induced defects increase [97], resulting in the decrease of PLQY.When the shell thickness was more than three monolayers thick, however, the impact of surface states on QDs was relaxed (Fig. 2b.)Apart from ZnS, CdS [56,98] and ZnSe [99] are generally used shell materials to coat CdSe QDs as well.The lattice mismatches are about 4% and 6%, respectively, resulting in much reduced lattice strain at the interface [45].
Lim et al. studied the influence of shell thickness on the performance of corresponding CdSe/Zn1-xCdxS QD-LEDs.Based on the optimization of shell thickness, they realized the highest EQE over 7% and record-high luminance of 105 870 cd/m 2 along with improved device stability.In addition, they revealed that the QD-QD energy transfer, an important source of energy loss, can be inhibited by a thick shell [83].
The Bawendi group studied the effect of shell thickness on emission line widths and photoluminescence blinking.As the shell thickness increases,
Overall, the use of the core-shell structure reduced surface defects, significantly boosted PLQY and improved the stability of the film, which is important for device fabrication.

Surface-ligand engineering
QDs, one kind of nanocrystals (NCs) are typically synthesized in a solution containing surface ligands (e.g.oleic acid, oleylamine) with a headgroup tethered to the QD surface and a hydrocarbon tail directed away [108].The surface ligands act as stoppers to control the QD growth and surfacepassivating ligands prohibit the formation of dangling bonds [53,109].The commonly used headgroups of the ligands are carboxyl, amino, thiol, phosphate group, etc.Recently, halide ligands received great attention for their excellent surfacepassivation capability [11,110,111].
The Owen group systematically summarized the ligand coordinating types on the QD surface, which can be classified into three main categories.The first type is the L-type ligand, i.e. a neutral donor that datively bonds onto the QD surface.Phosphine and amine are the L-type ligands usually used.The bonds between ligands and QDs are typically weak, and they can desorb in the presence of other ligands in the system.Another kind of ligand is the X-type ligands, which include the carboxyl group, thiol group and halides, etc., which usually bond strongly on the QD surface due to their high electron affinity.Cations on QD surfaces can bond with one or two Xtype ligands, depending on their coordinating numbers.Both cations and anions can bond with L-type ligands; however, the bond is generally weak and the ligands are easy to desorb (Fig. 3) [112].
The use of proton solvent in post treatment can cause the desorption of ligands such as carboxyl or amino groups, especially during the solid-state ligand-exchange process.An important strategy to reduce ligand loss is the use of ligands with strong binding characteristics such as thiol ligands.In addition, the long organic alkyl ligands used in synthesis can block carrier injection.Ligand exchange by short ligands is therefore needed to improve carrier transfer and promote the balance transport of electrons and holes.Short alkyl thiol ligands are therefore the most common ligands used in QD surface modification [113][114][115].In 2012, Sun et al. utilized a series of ligands with both thiol and carboxylic acids units to replace the long-chain oleate ligands of PbS QDs.The distance between QDs can be tuned by varying the ligand-chain lengths.Based on the optimization of ligand length to balance carrier injection and effective surface passivation, an order-of-magnitude improvement in device efficiency was achieved (Fig. 4a  and b) [116].
The Peng group reported a solution-processed, high-performance QD-LED based on ligandexchanged CdSe/CdS QDs.They used 1dodecanethiol to replace the original oleate ligands,together with device structural engineering.The device showed the highest EQE up to 20.5%  and a long operational lifetime of more than 100,000 hours at 100 cd/m 2 , which is the best-performing solution-processed red QD-LED so far [29].
Ligands engineering was utilized to improve the efficiency of CdZnS-ZnS QD-based blue lightemitting diodes as well.The replacement of oleic acid ligands on the as-synthesized QDs with 1octanethiol ligands resulted in a double increased electron mobility and greater balanced carrier injection, leading to the highest EQE of 12.2% [46].
Quite recently, branched ligands, namely entropic ligands for QDs, were developed, leading to much improved solubility of QDs by 10 2 −10 6 .High concentration of QDs in solvent is important for large-scale printing techniques, as well as for large-scale application of QD-based optoelectronics (Fig. 4c).Based on branched thiol ligand-coated CdSe-CdScore-shell QDs, a device with better charge transport and higher EQE was achieved (Fig. 4d) [47,117].
Overall, the surface ligands play a crucial role for QD surface passivation, efficient charge transport and carrier balance.Delicate surface-ligand engineering is essential to achieve high efficiency and stable QD-LED devices.

THE OPTIMIZATION OF DEVICE ARCHITECTURE
Apart from the high PLQY of QD active layers, the use of appropriate device architecture to ensure efficient carrier injection and radiative recombination in QDs plays a critical role for high-performance lightemitting diodes.To date, various device architectures, ETLs and HTLs (Fig. 5) have been employed to enhance the device performance.Meanwhile, the interface states between charge-injection layer and active layer have been carefully tailored to reduce non-radiative recombination on the interface.In this section, we will discuss the modification of chargetransport layers and the organic/inorganic matrix for QD-LEDs.We will divide the structures into two parts: planar architecture and QDs in a matrix.For the matrix part, we will classify it into organic matrix, inorganic matrix and hybrid perovskite.

Planar QD-LEDs
At present, QD-LEDs mainly use planar architecture with a flat active layer sandwiched between ETLs and HTLs [27][28][29][30]84,96,118].In this structure, the active layer is typically very thin given the poor conductivity of the QDs.The balanced injection of  electrons and holes is critical for efficient radiative recombination at the active layer, as well as a long device-operation lifetime.The imbalanced carrier injection gives rise to electron charging on QD layers and consequently non-radiative Auger recombination, as well as carrier recombination at the transporting layers [92,119].Energy-level engineering of the transport layer through ion dopant or surface modification was introduced to tailor the band structure of the carrier-transporting layer [120][121][122][123][124][125].
Mg-doped ZnO(Mg:ZnO) nanoparticles (NPs) that possess lower electronic energy levels were employed as the ETL (Fig. 6a).The Mg dopant is beneficial to reduce the carrier-injection barrier and improve the carrier-injection efficiency.Devices with Mg: ZnO NPs ETL show superior performance compared to pure ZnO in terms of luminance and efficiency (Fig. 6b) [126].
Polyethylenimine (PEI) was employed to modify ZnO ETL surfaces, which facilitates the electron injection into CdSe-ZnS QD emissive layers by lowering the work function of the ZnO layer from 3.58 eV to 2.87 eV, resulting in a much reduced interface barrier.Eventually, better charge injection and carrier balance in the emission layer is achieved.As a result, red QD-LEDs with a ZnO/PEI ETL layer exhibit a low turn-on voltage of 2-2.5 V [127].
In addition to the energy-level engineering, extra carrier-blocking layers have been explored in QD-LED devices [29,30,122,[128][129][130][131].The additional carrier-blocking layers help to maintain the charge neutrality of QD emitters and preserve superior emissive properties.Furthermore, an additional buffer layer with a wide band gap can block the reverse transfer of electrons or holes and overshooting of carriers out of the active layer, resulting in better charge balance and suppressed carrier recombination at the active layer.
A deoxyribonucleic acid and cetyltrimetylammonium (DNA-CTMA) complex was employed as an additional hole-transporting and electron-blocking layer (EBL) for CdSe/CdS/ZnS QD-LEDs.With the inclusion of this blocking layer, no electroluminescence (EL) emission from the HTLs/ETLs was observed, indicating that electron injection into HTLs was completely blocked by the DNA-CTMA EBLs [130].Similarly, the incorporation of an insulating polymer buffer layer between ETLs and  Adapted with permission from [133].
QD layers can prevent the overshooting of electrons into active layers (Fig. 7a), giving rise to much better charge balance and a much improved device performance (Fig. 7b).Furthermore, the insertion of this insulation layer can reduce charge recombination at the interface between QD layers and ETLs [29].Based on this principle, PVK was employed as the HTL for blue QD-LEDs (Fig. 7c and d) [30].To date, the highest EQE for red, green and blue QD-LEDs reached up to 20.5%, 14.5% and 10.7%, respectively, and all of them rely on polymer as a buffer layer [29,30].In conclusion, the use of energy-level engineering and a carrier-blocking layer plays an important role for high-performance QD-LED devices.

QDs in organic matrix
In the early stage of QD-LEDs, QDs normally contained a large number of ligands on its surface, partly due to the problem of a lack of strategies for efficient ligand exchange.In order to improve the conductivity of the film, they were normally dispersed in a conductive organic matrix that acted as a carrier-injection medium [49,[132][133][134][135].End group functionalization of conducting organic molecules was explored to enable direct linkage of them on QDs, allowing efficient carrier injection into QDs [49,133,136,137].In addition, organic molecules grafted onto the surface of QDs can protect them from aggregation and enhance solubility in non-polar solvents, facilitating the fabrication of uniform and stable QD films [133].To facilitate efficient carrier injection from polymers to QDs, polymers with large band gaps to form type-I band alignment with QDs are employed [138].Poly(p-methyltriphenylamine) (Poly-TPA) [133], poly(vinyltriphenylamine dimer) (poly-TPD) [138] and their derivatives are common polymer matrices.
CdSe/ZnS QDs and conducting polymer hybrid films by grafting blocking copolymers to them were investigated to enhance carrier transport (Fig. 8a  and b).The hybrid films showed uniform QD distribution, and the device based on it exhibited relatively low turn-on voltage and a three-fold increase in EQE (Fig. 8c) [133].More importantly, the stability of the film was dramatically improved when QDs were mixed with polymers (Fig. 8d).
Side-chain conjugated polymers were used as matrix materials for CdSe/ZnS QD-based QD-LEDs.The devices exhibited the highest EQE of 6.09%.Moreover, the correlation between the highest occupied molecular orbit (HOMO) level of the polymer and device performance was identified.

REVIEW
Figure 9. Flow chart showing the key steps involved in the development of semiconductor matrix-encapsulated nanocrystal arrays (SMENA).These stages include colloidal synthesis of PbS NCs (step 1), growth of the CdS shell (step 2) and spin-coating or dipcoating of NC films, exchange of bulky ligands with thermally degradable molecules (MPA, FA) and crystallographic fusion of NCs; all three were performed via layer-bylayer deposition (steps 3 and 4).In the last step, the pores of the PbS/CdS solid are filled with an additional CdS or ZnS material (step 5).Adapted with permission from [139].
Polymers with low-lying HOMO levels provide a reduced barrier for hole injection, and therefore lead to improved device performance [138,139].
In conclusion, organic matrix could serve as an effective material for pursuing high-performance QD-LEDs.However, the efficiency of the LED with this structure is relatively low, and the thermal stability is another issue to be addressed.

QDs in inorganic matrix
Compared with organic matrix, inorganic matrix possesses better thermal stability and faster carrier- transporting properties.QDs such as PbS and CdSe embedded in a matrix of a wide-band-gap semiconductor material (e.g.CdS, ZnS) were developed.QDs were firstly overcoated with several monolayers of wide-band-gap semiconductor materials, yielding a type-I core-shell structure.The core-shell QDs capped with thermally degradable ligands are subsequently spin-coated onto a substrate, and heated to remove the organic ligands.Finally, the pores generated in the film are filled with an additional wideband-gap semiconductor material through the successive ionic layer adsorption and reaction (SILAR) [51] methods [50,140], as shown in Fig. 9. Inorganic matrix was explored to improve QD surface passivation as well.A solution-phase assembly of PbS QDs into CdS inorganic matrix solids was investigated.As a result, the film exhibits bright IR emission and superior thermal stability [50].Recently, Brovelli et al. developed a dot-in-bulk structure with thick inorganic CdS matrix that shows dual-peak photoluminescence (Fig. 10a and b).Dual-color electroluminescence can be realized as well by controlling the applied bias, which provides a convenient way to develop white light-emitting diodes [141].
In general, the use of inorganic matrix effectively improved the stability of the device.Relatively poor surface passivation and low PL QY of the film form the bottleneck that limited device performance.

QDs in perovskite
To address the problem of interface defects and poor carrier transport, a new matrix passivator with high carrier mobility is desirable.Borrowed from the typical heteroepitaxial growth method that is generally used for semiconductor devices, organohalide perovskite were epitaxialy grown on QDs in situ in the film.The film exhibited remarkable PLQY traceable to their atom-scale surface passivation via perovskite.Meanwhile, the excellent carrier-transporting capability of perovskite allows efficient carrier injection into QDs.For type-I QDs/perovskite composites, electrons and holes in the larger-band-gap perovskites were transferred with 80% efficiency to become excitons in the quantum-dot nanocrystals.It hence provides a new protocol to leverage perovskites' excellent carrier diffusion to produce bright light emission from infrared QDs [48].
By combining the excellent carrier transport of the perovskite matrix and the high radiative efficiency of QDs in film, Gong et al. achieved a record power-conversion efficiency of 4.9% based on the PbS QDs in the MAPbI x Br 3-x film (Fig. 11a and d).Owing to the smaller lattice mismatch, QDin-perovskite composite (Fig. 11b) shows lower defect density at the interface between QD and perovskite, giving rise to enhanced PLQY and carrierinjection efficiency [142].In addition to high device efficiency, the device shows band edge-derived turnon voltage (Fig. 11c).QDs in a perovskite structure therefore provide an opportunity to realize both high PLQY and good carrier-transporting capability at the same time.

CONCLUSION AND PERSPECTIVES
In summary, as benefited from QD surface engineering and device-architecture optimization, the performance of QD-LEDs has been significantly improved, and this shows great promise for future commercialization.Surface engineering, such as the use of core-shell QD structures and thiol ligands, effectively reduced defect concentrations and improved PLQY.Device-architecture engineering including planar or matrix structure and the development of new carrier-transporting and blocking layers prompted radiative carrier recombination.The performance of QD-based LEDs such as light intensity and power-conversion efficiency have exceeded current commercial requirements in recent years, indicating the bright future of QD-LEDs for nextgeneration display applications.
To meet commercial requirements, device stability needs to be much improved.When devices are operating, the non-radiative decay may cause device heating, while the injection of carriers may cause oxidation or reduction of QDs, both of which will cause ligand loss and defect formation, incurring aggravated luminescence efficiency and performance decay.The development of a pure inorganic matrix epitaxially growing on the QD surface is one potential solution.On the other hand, to improve device stability, the defect concentrations need to be reduced further; better solutions for ligand exchange, film fabrication and matrix growth are still highly desirable.Post treatment such as atomic layer deposition could be another potential strategy to improve the stability of the device.
The development of heavy-metal-free QDs like InP or CuInS 2 will be a solution to the toxicity issue.The core-shell QDs based on InP [36,128] and CuInS 2 [7] have recently exhibited over 70% PLQY.Based on these advances, the performance

REVIEW
of heavy-metal-free QD-LEDs has also rapidly improved.However, their broad emission peak limited their color quality.More effort needs to be made to improve PLQY, and reduce the emission peak width.For large-scale application, more attention needs to be paid to the development of scale-up filmfabrication processes.Transfer-printing (Fig. 12a) [33], inkjet-printing (Fig. 12b) [87] and patterning processes [143] could be ideal ways.In addition, large-scale synthesis using a reactor is an effective strategy to reduce the production cost.

Figure 1 .
Figure 1.(a) Electronic states in a bulk semiconductor (left) and a QD made of the same material (right).Continuous bands of a bulk semiconductor with parabolic dispersion of carrier kinetic energies in the valence bands (VBs) and conduction bands (CBs) transform into discrete atomic-like levels in the case of the QDs.(b) Previous heavy-metal-based QDs covering the ultraviolet (CdS), visible (CdSe, CdTe and perovskite) and infrared (PbSe and PbS) regions have heavy metal-free alternatives: ZnSe for ultraviolet; InP, CuInS2 and Si for visible; InAs for infrared.Adapted with permission from [39].(c) Energy-band diagram of a typical QD-LED that outlines the two suspected QD excitation mechanisms: charge injection and energy transfer.(d) Progression of QD-LED performance over time in terms of peak EQE.Red ball, red QD-LEDs; green ball, green QD-LEDs; blue ball: blue QD-LEDs.Inset: four types of device structure.

Figure 2 .
Figure 2. (a) Schematic illustration of shell growth.Route I: layer-by-layer growth of CdSe-CdS core-shell QDs.Route II: The cation-exchange reaction used to convert core-only PbS QDs into core-shell PbS-CdS QDs.(b) Evolution of PL QY of zinc blende (ZB) and wurtzite (WZ) CdSe/CdS core/shell nanocrystals.Adapted with permission from [56].(c) Schematic of the chemical composition and electronic energy level within a graded-shell QD.Adapted with permission from [103].(d) EQE versus current density of QD-LEDs with different types of QDs.The onset of EQE roll-off shifts to higher currents with increasing degree of suppression of Auger recombination.Adapted with permission from [107].

Figure 3 .
Figure 3. (a) Examples of several ligand-exchange reactions are shown.(b)The coordination of different types of ligands in Green's formulation to metal-chalcogenide nanocrystals (such as cadmium selenide) are illustrated.R is analkyl group; Bu is n-butyl.Adapted with permission from[112].

Figure 4 .
Figure 4. (a) Schematic device structure of PbS QD-LEDs.(b) shows the distance between adjacent PbS colloidal QDs.Adapted with permission from [116].(c) Solubility of CdSe QD−ligand complexes with different ligands.(The solubility increased significantly for the entropic ligand-coated CdSe QDs.)(d) External power efficiency of the LEDs with the same CdSe-CdS core-shell QDs and device structure but different ligands.Adapted with permission from [47].

Figure 5 .
Figure 5.The energy-band level of general transport layers.

REVIEWFigure 6 .
Figure 6.(a) Proposed energy-band diagram of solution-processed multilayered QD-LEDs.(b) ZnMgO NP ETL-dependent variations of current efficiency and EQE of CIS/ZnS QD-LEDs as a function of driving voltage.Adapted with permission from [126].(c) Flat-band energy-level diagrams of QD-LEDs illustrate the reduction in the electron-injection barrier between ZnO and QDs due to the presence of the PFN layer.(d) EQE versus current density characteristics of InP@ZnSeS (1.7-nm shell thickness) QLEDs prepared with varied PFN concentrations.Adapted with permission from [128].

2 )Figure 10 .
Figure 10.(a) Left: schematic illustration of CdSe/CdS dot-in-bulk structure and device architecture.Right: Dual-color EL from the dot-in-bulk LEDs.(b) Current density (J; black circles) and luminance (red open circles) versus applied voltage for the QD-LEDs made of CdSe/CdS dot-in-bulk structure.Inset: EQE for emission from the top face of the device as a function of current density.The inset also shows photographs of functioning LEDs biased with 10 V. Adapted with permission from [141].

Figure 11 .
Figure 11.(a) The LED device architecture.(b) Illustration of enhanced electroluminescence efficiency in PbS QDs in MAPbX3 (X = I, Br) perovskite QD-LEDs; the left panel illustrates that radiative recombination dominates when QDs and perovskite are lattice matched, whereas lattice mismatch causes interfacial defects (black dashed line) and non-radiative recombination through traps (right panel).(c) Radiance-voltage characteristics.Inset: Zoom-in region of radiance-voltage plots (0.9-1.3 V) show the device turn-on voltages.(d) Perovskite devices (3.6% volume ratio) matrix versus a pure QD device.Adapted with permission from [142].

Table 1 .
Summary of commonly used QDs.