Droop-Free Colloidal Quantum Dot Light-Emitting ... - ACS Publications

Sep 10, 2018 - Department of Chemical Engineering & Department of Energy System Research, Ajou University, Suwon 16499, Republic of Korea. ⊥...
0 downloads 0 Views 1MB Size
Subscriber access provided by University of Sunderland

Communication

Droop-free colloidal quantum dot light emitting diodes Jaehoon Lim, Young-Shin Park, Kaifeng Wu, Hyeong Jin Yun, and Victor I. Klimov Nano Lett., Just Accepted Manuscript • DOI: 10.1021/acs.nanolett.8b03457 • Publication Date (Web): 10 Sep 2018 Downloaded from http://pubs.acs.org on September 10, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Droop-Free Colloidal Quantum Dot Light Emitting Diodes Jaehoon Lim,†, ‡ Young-Shin Park,†,⊥ Kaifeng Wu,† Hyeong Jin Yun,† and Victor I. Klimov*,† †

Chemistry Division, Los Alamos National Laboratory, Los Alamos, New Mexico 87545, USA



Department of Chemical Engineering, Ajou University, Suwon 16499, Republic of Korea



Center for High Technology Materials, University of New Mexico, Albuquerque, New Mexico

87131, USA *Address correspondence to: [email protected] ABSTRACT: Colloidal semiconductor quantum dots (QDs) are a highly promising materials platform for implementing solution processable light emitting diodes (LEDs). They combine high photostability of traditional inorganic semiconductors with chemical flexibility of molecular systems, which makes them well suited for large-area applications such as television screens, solid-state lighting, and outdoor signage. Additional beneficial features include size-controlled emission wavelengths, narrow bandwidths, and nearly perfect emission efficiencies. State-of-theart QD-LEDs exhibit high internal quantum efficiencies approaching unity. However, these peak values are observed only at low current densities (J), and correspondingly low brightnesses, while at higher J, the efficiency usually exhibits a quick roll-off. This efficiency droop limits achievable brightness levels and decreases device longevity due to excessive heat generation. Here we demonstrate QD-LEDs operating with high internal efficiencies (up to 70%) virtually droop-free up to unprecedented brightness of >100,000 cd m-2 (at ~500 mA cm-2). This exceptional performance is derived from specially engineered QDs that feature a compositionally graded interlayer and a final barrier layer. This QD design allows for improved balance between electron and hole injections combined with considerably suppressed Auger recombination, which helps mitigate efficiency losses due to charge imbalance at high currents. These results indicate a significant potential of newly developed QDs as enablers of future ultra-bright, highly efficient devices for both indoor and outdoor applications. KEYWORDS: Quantum dot, light emitting diode, LED, efficiency droop, charged exciton, trion, Auger recombination

1 ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Light emitting diodes (LEDs) are ubiquitous in everyday life and can be found in numerous devices from home lighting and displays to vehicle headlamps and projectors. Most of the commercially available LEDs are based on III-V semiconductors fabricated via vacuum-based epitaxial techniques.1 An emerging class of LED materials is highly emissive organic molecules. Organic LEDs or OLEDs can be processed via low-cost, high-throughput solution-based methods, which makes them well-suited for large-area applications such as television screens and solid-state lighting.2 A more recent addition to the family of solution-processable LEDs is devices based on colloidal quantum dots (QDs).3 These materials combine properties of traditional inorganic semiconductors such as high photostability with chemical flexibility of organic molecules. In addition, they feature size-controlled emission wavelengths, excellent color purity resulting from narrow emission bandwidths3, and near-unity emission quantum yields (QYs).4-7 More than two decades of QD-LED research have resulted in the development of highly efficient devices operating over a wide range of visible8-15 and near-infrared wavelengths.16-19 One still existing challenge in this area is a so-called efficiency droop, that is, a progressive (but reversible) deterioration of LED performance with increasing current density (J) or luminance (L). The droop effect limits practically achievable brightness levels and leads to excessive heat generation, which increases power consumption and reduces device longevity. As a result, QDLEDs show the best performance at low brightnesses4 that satisfy requirements for mobile displays and television screens (< 1,000 cd m-2), as well as indoor lightings (< 5,000 cd m-2),8-14, 20, 21

however, are still insufficient for applications such as daylight displays, digital signage,

traffic lights, and window-integrated see-through screens. Therefore, resolving the droop

2 ACS Paragon Plus Environment

Page 2 of 30

Page 3 of 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

problem is important for both achieving industrial standards in the stability of QD-based devices as well as extending their utility to applications requiring ultra-high brightnesses. Original explanations of droop in QD-LEDs invoked EL quenching due to exciton dissociation by applied electric field.22,

23

While this effect might indeed contribute to the

observed decrease in the device efficiency, a growing amount of evidence suggests that the prevailing reason for efficiency droop is nonradiative Auger decay wherein the exciton recombination energy is rapidly transferred to the extra carriers instead of being released as a photon.24 This process can be triggered, for example, by imbalance in electron and hole injection rates which could lead to formation of charged excitonic species.3, 20, 21 In the present study, we combine device- and QD-level controls of charge injection with a goal to realize LEDs wherein charge balance in the QD emitting layer is maintained over a wide range of current densities and hence device brightnesses. For this purpose, we re-evaluate the origin of droop by analysing the performance of LEDs made on three types of heterostructured QDs that have the same-size CdSe core but distinct structures of the thick outer shell designed so as to allow for controlling charge injection and the rate of Auger recombination. Using these samples, we analyze effects of improving change balance and Auger decay suppression (as well as their combination) on LED performance. Guided by these studies, we identify the most effective strategy for mitigating the droop problem, which allows us to demonstrate red-emitting LEDs with a virtually droop-free performance up to 100,000 cd m-2 (efficiency reduction 70%. The emission is centred at 2.03 eV and displays a fairly narrow linewidth of 61 meV. The PL lifetime is ~12 ns (Figure 2c, blue trace), which is close to the ON-state emission lifetime observed in single-dot measurements (Figure S1). This allows us to assign it to a radiative single-exciton lifetime ( τ X 0 = τ r , X 0 ). It further suggests that the non-ideal PL QY is due to fast de-excitation not resolved in our measurements (200 ps resolution), which can also be 6 ACS Paragon Plus Environment

Page 6 of 30

Page 7 of 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

interpreted in terms of the existence of a subset of non-emissive QDs (~30% of the total number). To evaluate the degree of Auger decay suppression due to the cg-layer, we conduct measurements of negative-trion lifetimes. To prepare negatively charged QDs, we employ photochemical reduction with lithium triethylborohydride as a hole scavenger34 (see Methods). Using this method, we inject ~0.5 electron per dot on average, and then, applying low-intensity pulsed excitation to generate neutral excitons and negative trions. The latter species manifest as a fast initial component in PL time transients followed by a slower neutral-exciton decay (Figure 2c, red trace). After isolating the X– component (green symbols in Figure 2c) we find the trion lifetime τ X − = 4.3 ns. Based on statistical scaling24, the X– radiative lifetime τ X − = τ r , X 0 / 2 = 6.1 ns, which yields the Auger time constant τ A, X − = τ X −τ r , X − /(τ r , X − − τ X − ) = 14.6 ns. In standard core-only CdSe QDs with the same confinement energy as in cg/B-QDs, τ A, X − is ~300 ps.3 This value is more than 20 times shorter than that in the cg/B-QD sample, indicating a very considerable suppression of Auger decay achievable through compositional grading. This suppression translates into a strong enhancement in the X– QY (Q1). Defining a relative PL QY of negative trions as q1 = Q1 / Q0 = 2τ X − / τ X 0 , we find that for the cg/B-QD sample, q X − = 70.5% versus only ~3% for standard CdSe dots (estimated assuming τ r , X 0 = 20 ns35). Next, we investigate the effect of the B-layer on carrier injection. For this purpose, we compare cg/B-QDs to the reference sample of the same structure but without the final barrier layer (referred to as cg-QDs; compare Figure 3a and 3b). We incorporate these two types of the QDs into “single-carrier” electron- or hole-only non-emitting devices in the form of thin QD layers. In the electron-only devices, the dots are sandwiched between ZnO layers that serve as 7 ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

electron-transporting/hole-blocking layers. In the hole-only devices, the QDs are placed between the two hole-transporting/electron-blocking layers based on organic molecules: poly(9,9dioctylfluorene-alt-N-(4-sec-butylphenyl)-diphenylamine) (abbreviated as TFB) spin-coated on top of the ITO/PEDOT:PSS and (tris(4-carbazoyl-9-ylphenyl)amine) (abbreviated as TCTA) deposited on top of the Al/MoOx. The comparison between the cg/B- and cg-QDs indicates that the barrier layer significantly reduces the current density (by a factor of 15 to 20 at the average electric field F = 0.1 – 0.5 MV cm-1) in electron-only devices (Figure 3c). However, its effect on the current density in hole-only devices is considerably weaker (Figure 3d). Specifically, for F = 0.1 MV cm-1, the attenuation factor is only ~2. Thus, incorporation of the B-layer allows one to selectively suppress electron injection, without considerably affecting hole injection, which is a particularly useful capability in the case of the inverted LED architecture, which tends to oversupply electrons. The difference in the effect of the B-layer on electron and hole injections can be attributed to compositional inhomogeneities within the ZnSeyS1-y shell generated, for example, by fluctuations in the reaction environments and/or the facet-dependent QD reactivity. Specifically, the existence of ZnSe-rich domains can selectively facilitate hole injection into the emitting core as the VB edge of ZnSe is energetically close to that of CdSe. On the other hand, the ZnSe CB edge is considerably higher than that of CdSe and thus would still have a blocking effect on the electron. As a final step, we evaluate the performance of cg/B-QDs in p-i-n inverted LEDs (Figure 1a), benchmarking it against devices based on cg-QDs and core/shell (C/S) CdSe/Cd0.7Zn0.3Se QDs with ungraded shell. All three QD samples have the same core radius (r

8 ACS Paragon Plus Environment

Page 8 of 30

Page 9 of 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

= 2 nm) and the overall shell thickness (d + h ≈ 7 nm) (see Figure S2 for structural details). They are characterized by similar PL peak energies (1.94–2.03 eV) and single-exciton PL QYs (Q0 = 55–72%). However, due to the cg-layer, the first two samples have a considerably higher X– QYs than the third sample (~50% versus 18%; see Table 1). Another distinction is the presence of the barrier layer in the cg/B-QDs, which selectively inhibits electron injection; this effect is absent in the cg- and C/S-QD samples (see Figure 3c,d and Figure S2). To fabricate LEDs, we use TCTA as an HTL and a mixture of hybridized ZnO with insulating poly(vinylpyrollidone) (PVP) as an ETL. We found that, not only the asymmetric electron and hole injection barriers, but the presence of a large contact resistance at a TCTA−MoOx/Al interface also deteriorates the charge balance in our devices. By controlling relative fractions of ZnO and PVP we can tune the ETL conductivity, which provides an additional means for balancing electron and hole injections (Supporting Note 1 and Figures S3 and S4). As illustrated in Figure 3e,f, all fabricated devices display typical diode characteristics and turn-on voltages closed to the QD band gap (~2.0 V). In all LEDs, electroluminescence (EL) is exclusively due to QDs without measurable parasitic emission from charge-transport layers (inset of Figure 3f), indicating the absence of any appreciable leakage current. The presence of the barrier layer reduces the current density in the cg/B-QD-LEDs compared to that in devices based on two other QD samples (Figure 3e). However, this occurs without a significant decrease in brightness. In fact, at higher biases (V > 5 V), the luminance of the cg/B-QD-LED is comparable to that of the LED based on the cg-QDs and is considerably higher than that of the C/S-QD-LED (Figure 3f). These observations suggest that the cg/B-QD-based LED has a greater L-to-J ratio, that is, a higher EQE than two other devices, at least at higher biases.

9 ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

To compare the performance of different LEDs in more quantitative terms we analyse the dependence of EQE on luminance (Figure 4a) and current density (Figure 4b). The LEDs based on cg- and C/S QDs exhibit a pronounced maximum in the ηext-versus-L (or J) dependence, which defines their peak EQEs (9.3% and 3.7%, respectively). Both devices exhibit a characteristic droop behaviour past the EQE maximum. Usually, the effect of droop is quantified in terms of the luminance or current density for which the EQE drops by 50% from the peak value (L50 and J50, respectively).2,29 In our best devices, however, droop is extremely weak (see below), and J50 is pushed to the region where fabricated LEDs may suffer from irreversible degradation. Therefore, here we use L90 and J90 (values corresponding to the 90% of the peak EQE) as droop characteristics2. Based on the measurements of the C/S-QD-LEDs, J90 = 287 mA cm-2 and L90 = 7,690 cd m-2. In the case of the cg-QD sample, J90 increases to 458 mA cm-2 and L90 to 78,357 cd m-2 (Table 1). This improvement is a consequence of suppression of Auger recombination leading to increased emissivity of charged species. The importance of such species in LEDs based on C/S- and cg-QDs is indicated by the quantitative analysis of the observed EQE as elaborated below. According to spectroscopic studies (Figure S5 and Table 1), PL QYs of negative trions in C/S- and cg-QDs are 18% and 43%, respectively. If X– were the LED emissive species, the resulting EQEs would be, respectively, 3.6% and 8.6%, as defined by ηoutQ1. Both of these values are remarkably close to the measured peak EQEs (3.7 and 9.3%, respectively; Figure 4c), suggesting that even under most optimal pumping conditions when EQE is maximized, the dots are charged with one extra electron on average. The observed EQE drop past the peak point is likely due to the increasing degree of charging, which leads to the increased Auger-decay rate. These observations are consistent with a previous assessment that the inverted LED architecture 10 ACS Paragon Plus Environment

Page 10 of 30

Page 11 of 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

with a ZnO-ETL tends to oversupply electrons, which increases the abundance of negatively charged excitons.20 The introduction of the ZnSeyS1-y barrier layer into the QD design is expected to improve charge balance by partially inhibiting electron injection (see Figure 3a). The measurements of devices made of the cg/B-QDs indeed indicate that the observed EQEs are consistent with values expected for neutral excitons. Based on Q0 of 72% and ηout of ~20%, the expected ideal EQE is ~14%. This is very close to the measured peak EQE of 13.5% (Figure 4c), suggesting that the primary emitting species in this case are indeed neutral excitons. Importantly, the charge balance is maintained across a wide range of biases, which allows for virtually droop-free operation up to very high brightness of 100,000 cd m-2. The droop onset corresponds to L90 of ~200,000 cd m-2 (J90 = 861 mA cm-2), which is a record high value realized with either QD-LEDs or OLEDs (Figure 5). The brightness can be further pushed to ~320,000 cd m-2 by increasing J to 2,062 mA cm-2. This strong performance is a direct consequence of a special design of the cg/B-QDs, which allows for simultaneous control of both change balance and Auger recombination. To quantify the effect of charge imbalance on EQE, we express it as ηext = ηout(p0Q0 + p1Q1 + p2Q2 + …), where Qi is the PL quantum yield of an exciton coexisting with i extra electrons, and pi is the corresponding charging probability. The values of pi are normalized so as p0 + p1 + p2 + … = 1. Based on Figure 4b,c, most of the measured EQEs reside within the ranges expected for EL due to neutral (p0 = 1, top of the bars in Figure 4c), singly-charged (p1 = 1, intermediate lines within the bars), or doubly-charged (p2 = 1, bottom of the bars) excitons, or a mixture of these states. This suggests that even at the highest biases, the maximum value of i does not exceed 2, and hence, we can limit our consideration to p0, p1, and p2. In this case, the average per-dot number of excess electrons (〈Ne〉) can be found from 〈Ne〉 =p1 + 2p2. 11 ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Our earlier analysis indicates that prior to the onset of droop, the primary emissive species in LEDs based on cg/B-QDs are neutral excitons. Thus, near the peak-EQE point, p0 ≈ 1, p1 ≈ 0, and p2 ≈ 0. After the droop onset, the increased applied bias leads to a progressive increase in the occurrence of singly charged dots, i.e., increased p1. The fraction of doubly charged QDs, however, still likely remains low (i.e., p2 ≈ 0) even at high current densities of ~1 A cm-2 because injection of the second electron into the charged dot is inhibited by strong Coulombic repulsion. This assessment is supported, in particular, by the absence of EL from doubly-charged excitons (inset of Figure 3f) which would have been observed at ~150 meV above the band-edge. This situation, is likely maintained until all dots in the EML are charged with a single electron, and only after that, the sub-ensemble of doubly charged dots starts to build up via electron injection into singly charged dots, leading to a further drop in the EQE (Figure 4b, red triangles). These considerations suggest that charging occurs in a stepwise manner, so as for any given bias only two of the charging probabilities (pi and pi+1) are nonzero, and hence, they can be connected by pi + pi+1= 1. Using this stepwise charging arguments, we can back out the degree of charging from the measured ηext-versus-J dependence (Supporting Note 2 and Figure S6) and plot these dependences for three different QDs in Figure 4d as 〈Ne〉-versus-J. As was indicated by our earlier analysis, in cg/B-QD-LED, prior to the onset of droop, the dots are primarily charge neutral (〈Ne〉 = 0.22), while at the droop onset (defined by L90), 〈Ne〉 increases to 0.55, indicating that EQE reduction is due to a buildup of the population of negative trions leading to increased carrier losses due to Auger decay. A distinct situation is realized in LEDs based on cg- and C/S-QDs and. In these devices, near the peak-EQE point, 〈Ne〉 is 0.88 for cg-QDs and 0.99 for C/S-QDs, suggesting the 12 ACS Paragon Plus Environment

Page 12 of 30

Page 13 of 30 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

dominant contribution of negative trions to the observed EL. At the L90 point, 〈Ne〉 is 1.01 in the case of the cg-QDs and 1.17 for the C/S-QDs. Here the main reason for EQE reduction is enhancement of Auger recombination due to increasing occurrence of doubly charged excitons that replace singly charged species. The QD charging probabilities at the L90 point (pi,90) can be directly related to the PL QYs of QD states with various degrees of charging. For example, if the peak EQE is solely realized with neutral dots (the cg/B-QD case), p1,90 can be found from 0.9Q0 = (1 – p1.90)Q0 + p1,90Q1, which yields p1,90 = 0.1/(1 – q1). This relationship suggests that the larger the negative trion QY is, the higher degree of charging can be sustained by a device without appreciable EQE deterioration. This emphasizes again the importance of control of Auger recombination for realizing droop-free devices. For example, if we consider standard CdSe QDs, we obtain that p1,90 is only ~0.10. On the other hand, p1,90 increases to 0.34 for the cg/B-QDs.

This indicates a

considerably higher tolerance of devices based on “Auger-decay-engineered” QDs to imbalanced charge injection. The above analysis indicates that by endowing a QD with structural features for simultaneous control of both carrier-injection and Auger decay is a highly effective approach for mitigating the efficiency-droop problem, as in particular evident from the comparison of droop behaviors of cg/B-QD-LEDs and LEDs utilizing engineered QDs of previous generations. An example of such a comparison is given in Figure 5a, wherein we plot peak-normalized EQE-vs-J characteristics of the cg/B-QD LED (triangles) and LEDs of ref 20 based on “nongraded” core/shell CdSe/CdS QDs with an additional higher-band-gap Cd0.5Zn0.5S shell (C/S/S QDs; pentagons) and partially graded CdSe/CdSe0.5eS0.5/CdSe core/alloy/shell QDs (C/A/S; diamonds). As was pointed out in ref 20, introduction of the Cd0.5Zn0.5S barrier allowed for 13 ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

improving charge balance by impeding electron injection. This produced a boost in the peak EQE to 7.5% versus 1.8% in the case of C/A/S QDs. However, the distortion of charge balance at higher biases resulted in a quick EQE drop (to ~65% of the peak value at 500 mA cm-2) due to fast (not controlled) Auger recombination. In C/A/S QDs, on the other hand, Auger decay was considerably suppressed. As a result, the droop effect was less pronounced, and at 500 mA cm-2 the EQE was around 80% of its peak value. Our newly developed cg/B-QDs combine both wellcontrolled charge balance and strong Auger-decay suppression, which results in a remarkable improvement in both the peak EQE and the droop characteristics over the C/S/S and C/A/S QDs. Specifically, the peak efficiency is boosted to 13.5%, and importantly, this value is maintained (