Interfacial structural analysis with X-ray reflectivity for elucidation of alkyl-chain effects on hole injection in quantum-dot light-emitting diodes

Abstract

Stable quantum-dot light-emitting diodes (QD-LEDs) require sufficient hole injection, which requires understanding and controlling the hole transporting (HT) polymer/QD interface structure. We successfully illustrated and quantified the interface structure by applying X-ray reflectivity (XRR) analysis and clarified the dominant factors of hole injection. We developed HT polymers with various alkyl chains to investigate the relationship between the interface structure and hole injection. We revealed that longer alkyl chains interact more strongly with QDs, causing QD sinking (= interfacial mixing) and the formation of an interface favorable for hole injection. Thus, interfacial structure analysis with XRR is useful for designing HT polymers with enhanced hole injection.

---

Introduction

Quantum-dot light-emitting diodes (QD-LEDs) are attracting much attention for next-generation displays, due to their excellence in colour gamut, brightness, and efficiency.^1–12 Luminance stability is one of the most important problems to solve when it comes to their practicable implementation in displays. There have been reports on the working mechanism of QD-LEDs.^13–20 In state-of-the-art QD-LEDs, in which a QD emissive layer is sandwiched between a hole transporting (HT) polymer and inorganic electron transporting layer (ETL), the ETL can provide effective electron injection, so that a lack of hole injection leads to unbalanced electrons and holes; thus, low device characteristics.^13–17 Therefore, sufficient hole injection is required for stable QD-LEDs. In other words, developing HT polymers having enhanced hole injection is necessary.

The energy-level offset at the HT polymer/QD interface and hole mobility are important parameters for controlling hole injection. However, even if the energy offset and hole mobilities are the same, hole injection is not necessarily the same.¹⁷ This is thought to be due to differences in the interface state. Therefore, controlling hole injection requires understanding the structure of the HT polymer/QD interface and clarifying the relationship between the interface structure and hole injection. It should be noted that QD-LEDs are fabricated using a wet process. The HT polymer/QD interface is probably roughened, as the underlying HT polymer is somewhat eroded during the QD coating process,^21,22 differing from the interface structure of vapour-deposited organic LEDs.

Transmission electron microscopy (TEM) is a powerful tool for directly observing an interface structure, however, quantitative analysis is difficult. Therefore, we focused on X-ray reflectivity (XRR), considering the density profile obtained from XRR measurement is quite useful to analyse the HT polymer/QD interface structure due to the following reasons.^23–25 A physical parameter of density can be obtained, which enables us to easily assign the HT polymer and QD. This is because the density difference is large between an organic HT polymer and inorganic QD. In other words, the area with low density can be assigned as the HT polymer, area with high density as the QD, and area where density changes as the HT polymer/QD interface. The interfacial state (for example, interfacial mixing) can be investigated from how density changes at the interface region. The QD arrangement can also be investigated on the basis of QD density. Density profiles can be thus quantitatively compared among different samples.

There are various factors controlling the interface state, such as surface energy, thermal properties, and solubility in QD solvents. Considering that QD particles have long chain alkyl ligands for stability,²⁶ the interaction between a ligand and HT polymer should be one of the most important factors controlling the interfacial state.^27,28 Therefore, we focused on an alkyl substituent as an interacting unit with a QD ligand, assuming that a long alkyl chain has strong interactions with QD particles, to show enhanced hole injection.

We designed and synthesized HT polymers to investigate the relationship among alkyl-chain length and interaction, hole injection, and device lifetime. We also revealed the relationship between interaction and hole injection from the viewpoint of the interface structure.

Experimental

Materials

Red-emitting QDs were synthesized according to the literature.⁹ The QDs are composed of an indium phosphide (InP) core of 3.3 nm, zinc selenide (ZnSe) shell of 4.5 nm, and zinc sulfide (ZnS) shell of 0.2 nm. The surfaces of the QDs are covered with long alkyl oleic acid (OA) ligands. Zinc magnesium oxide (ZnMgO) nanoparticles were prepared according to the literature.²⁹ Poly(3,4-ethylenedioxythiophene) and polystyrene sulfonate (PEDOT:PSS) dispersed in aqueous solution (Baytron CleviousTM AI4083) was purchased from H. C. Starck GmbH. The 9-([1,1′-Biphenyl]-3-yl)-9′-([1,1′-biphenyl]-4-yl)-9H,9′H-3,3′-bicarbazole was purchased from TCI and used after sublimation. The 1,4,5,8,9,11-hexa-azatriphenylene hexacarbonitrile (HAT-CN) was purchased from LUMTEC. Aluminum pellets were purchased from iTASCO.

Simulation

All quantum chemical calculations were conducted using the Gaussian 16 program.³⁰ The B3LYP functional³¹ with the 6-31G(d,p) basis set³² adding the D3 version of Grimme's dispersion with Becke–Johnson damping³³ was used as the density functional theory (DFT) method incorporating intermolecular interactions. Interaction energies (IEs) between the QD and HT polymers having various alkyl chains were simulated according to the following protocol. (1) The model molecule of the QD surface modified with OA anions (QD/OA) was constructed with a single layer of ZnS wurtzite³⁴ (Zn₂₃S₂₂) and four OA anions dispersed on it by contacting the carboxyl part with the ZnS surface (total charge: 0, spin multiplicity: singlet). The geometries of OAs were then optimized using the DFT method under the condition of the frozen coordinates of the ZnS (Fig. S1a, ESI†) (obtained total energy: E(QD/OA)). This QD/OA model structure with a planar QD surface may have somewhat different morphology from typical QDs with spherical surfaces; however, systematic application of this model to the target HT polymers (shown below) is expected to give fair correspondence between the HT polymers and their IEs. (2) The single repeat unit of our designed HT polymers (with phenyl and hydrogen termination on each end) with alkyl chains of C(CH₃)₃ (tertiary butyl), n-C₄H₉, C₆H₁₃, C₈H₁₇, C₁₀H₂₁, C₁₂H₂₅, and C₁₄H₂₉ (HT-4′, HT-4, HT-6, HT-8, HT-10, HT-12, and HT-14, respectively) were geometry optimized using the DFT method (obtained total energy: E(HT polymer)). (3) The QD/OA + HT-14 input structure was constructed as the n-C₁₄H₂₉ chain of HT-14 placed between the four OAs of the QD/OA. The geometries of OAs and HT-14 were then optimized using the DFT method under the condition of the frozen coordinates of the ZnS (Fig. S1b, ESI†) (obtained total energy: E(QD/OA + HT polymer)). (4) For the HT polymers other than HT-14, the QD/OA + HT input structure was constructed from the optimized structure of the QD/OA + HT-14 by contracting the C₁₄H₂₉ chain from the end. The geometries of OAs and the alkyl chain of the HT polymer were then optimized using the DFT method under the condition of the frozen coordinates of the ZnS and aryl part of the HT polymer (obtained total energy: E(QD/OA + HT polymer)). The frozen coordinates of the aryl part of the HT polymer are based on the assumption of the existence of dense OA on the QD surface, which prohibits the sinking of the aryl part of the HT polymer between the QAs. (5) The IEs for QD/OA and each HT polymer were IE = E(QD/OA) + E(HT polymer) − E(QD/OA + HT polymer). A large IE means stronger interaction between the QD/OA and HT polymers.

Synthesis and characterization

The monomers 3-hexyl-1,5-dibromobenzene, 3-decyl-1,5-dibromobenzene, and 3-dodecyl-1,5-dibromobenzene were purchased from Tokyo-Kasei Co. Ltd and used without any purification. The details of the synthesis are described in the ESI.† Finally, five HT polymers were synthesized: HT-4′, 6, 8, 12 and 14, with the alkyl substitution of C(CH₃)₃, C₆H₁₃, C₈H₁₇, C₁₂H₂₅ and C₁₄H₂₉, respectively. The highest occupied molecular orbital (HOMO) energy was measured with photoelectron yield spectroscopy (Riken Keiki, AC-3). The glass transition temperature (T_g) was measured with thermogravimetry-differential thermal analysis (TA Instruments, Discovery DSC25). The average molecular weight was determined from gel-permeation chromatography by using polystyrene standards. The insolubility of the HT polymer film was confirmed by the fact that the thickness did not change after octane (QD solvent) immersion. Thickness was determined with ellipsometry. Insolubility was further confirmed by the fact that the absorption spectra of HT-14 before and after octane immersion were the same. HT-14 is considered to have the highest solubility due to the lowest molecular weight and longest alkyl chain among the five HT polymers. The absorption spectra were measured with an UV-vis spectrophotometer (SHIMADZU, UV-1800).

Fabrication of devices and model films

The QD-LEDs, consisting of PEDOT:PSS (35 nm)/HT polymer (25 nm)/QD (20 nm)/ZnMgO (60 nm)/Al (100 nm), were fabricated on 150-nm-thick indium tin oxide (ITO)-patterned glass substrates. The substrates were cleaned with acetone and isopropyl alcohol (IPA) in an ultrasonic bath then dried using an IPA dryer. Oxygen plasma treatment was undertaken for 20 min using an ultraviolet ozone cleaner (Jelight, UVO144AX-220). A PEDOT:PSS dispersion was spin-coated onto the ITO-pre-patterned substrate with multiple-steps of 500 (5 s) and 3000 rpm (50 s), followed by a pre-baking at 110 °C for 10 min in air to remove any residual moisture and hard baking of 150 °C for 30 min in a N₂-filled glovebox (<0.1 ppm H₂O, <0.1 ppm O₂). About 1.0 wt% of HT polymers in o-xylene was spin-coated at 2000 rpm and baked at 150 °C for 30 min in the glovebox. The QD solution in octane (15 mg mL⁻¹) was spin-coated at 3000 rpm, followed by baking at 120 °C for 30 min. The ZnMgO solution (70 mg mL⁻¹ in ethanol) was spin-coated at 4000 rpm and thermally treated at 140 °C for 30 min in a N₂ atmosphere. Aluminium was deposited using a thermal evaporator at a rate of 0.1 nm s⁻¹ under a vacuum pressure less than 5 × 10⁻⁷ Torr as the cathode, and the device was encapsulated. The treatment temperature of each layer was determined from the thermal stability of the lower-layer material and QD-LED performance. The hole only device (HOD) structure is the same as the QD-LEDs, except that ZnMgO was replaced with an electron-blocking layer.‡ All solution layers were fabricated in accordance with the same method described above. The electron-blocking layer and Al cathode were fabricated by thermal evaporation, and the device was encapsulated. HODs without QDs were also fabricated. The model samples for XRR analysis, which were single layers of HT polymer, QD, and ZnMgO, and a stacked structure of HT polymer/QD/ZnMgO, were fabricated on quartz substrate in accordance with the same method as QD-LEDs. The treatment temperature of ZnMgO was varied at 120, 140, and 180 °C. A model sample for TEM analysis, i.e., Si/HT polymer/QD/ZnMgO, was also prepared in the same manner as for QD-LEDs.

Measurements and analysis

The current–voltage–luminance characteristics of QD-LEDs were measured using a system incorporating a spectroradiometer (Konica Minolta, CS-2000A) and source-meter unit (Keithley instruments, SMU2635B). The external quantum efficiency was calculated on the assumption that light would have Lambartian distribution. The lifetime results of the QD-LEDs were measured at the initial luminance of 4500 cd m⁻² using a commercial multi-channel lifetime test system with an embedded photodiode located in a temperature- and humidity-controlled chamber. The current density–voltage characteristics of the HODs were measured with a source meter (Keithley instruments, 2400). XRR analysis was conducted on an X-ray diffractometer (Rigaku, SmartLab) using a CuKα1 source (λ = 0.154 nm) operated at 45 kV and 200 mA. The 2 theta was scanned from 0° to 4.0° with a 0.004° step. The experimental data were analysed with the Rigaku SmartLab Studio software. TEM images were obtained on an ARM200F (JEOL) operated at 200 kV.

Results and discussion

Design and properties of hole-transporting polymers

We designed side-chain-type HT polymers, as shown in Fig. 1, to introduce an alkyl substituent at the main-chain part and HT moiety at the side-chain part since the electronic properties should not be affected by the difference in the alkyl structure. The arylamine moiety was selected as the HT unit, considering its HT performance, energy level, and device performance in organic LEDs.³⁵ Long normal alkyl chains were examined as an interaction substituent since red QDs composed of InP/ZnSe/ZnS have long OA on their surfaces,¹⁰ as shown in Fig. 2. A normal alkyl substituent was systematically examined from C₄H₉ to C₁₄H₂₉, limiting the maximum chain length to 14 due to concerns about solubility in the QD solvent (octane). A tertiary butyl (C(CH₃)₃) substituent was also examined as a different type of alkyl chain.

Fig. 1 Molecular structure and concept of our designed side-chain-type polymers.

Fig. 2 Structure of red-emitting InP/ZnSe/ZnS QD.

The interaction strength between the HT polymers and a QD was estimated using the DFT method as follows. IEs were simulated as the difference in the total energies: IE = E(QD/OA) + E(HT polymer) − E(QD/OA + HT polymer). Here, E(QD/OA) is the total energy of the QD surface model (a single layer of ZnS wurtzite³³) with four OA anions (Fig. S1a, ESI†), E(HT polymer) is the total energy of the single repeat unit of the HT polymer, and E(QD/OA + HT polymer) is the total energy of the model structure, where the alkyl chain of the HT polymer is placed between the four OAs of the QD/OA (Fig. S1b, ESI†). The estimated IEs were found to increase in accordance with the alkyl length, as shown in Table 1, except for C(CH₃)₃. The IE for C(CH₃)₃ was similar to that for C₆H₁₃. Accordingly, we selected C(CH₃)₃, C₆H₁₃ and C₈H₁₇ as the representatives with small IE, and C₁₂H₂₅ and C₁₄H₂₉ as those with large IE. Thus, five HT polymers were synthesized; HT-4′, 6, 8, 12, and 14, having C(CH₃)₃, C₆H₁₃, C₈H₁₇, C₁₂H₂₅, and C₁₄H₂₉, respectively.

Table 1 Estimated interaction energies between the HT polymer and QDs

Alkyl chain	C(CH3)3	C4H9	C6H13	C8H17	C10H21	C12H25	C14H29
Interaction energy (eV)	0.425	0.102	0.463	0.640	0.837	1.076	1.482

---

The measured properties of the developed HT polymers are listed in Table 2. The HOMO energies of all polymers were similar, about −5.50 to −5.55 eV (Fig. S4, ESI†). This indicates that the difference in energy level offset does not need to be considered. The T_g decreased in accordance with the alkyl-chain length except for HT-4′ (Fig. S5, ESI†). The T_g of HT-4′ was similar to that of HT-12. In other words, HT-4′ has IE similar to HT-6 but different T_g, and T_g similar to HT-12 and 14 but different IE. It was confirmed that all the HT polymer films have sufficient insolubility for octane, the solvent for QD stacking (Fig. S6, ESI†). Hole mobilities were investigated using HODs without QDs. Fig. S7 (ESI†) illustrates that the HODs showed almost the same current density–voltage characteristics, which means little difference in their hole mobilities. Current densities at 8 V are listed in Table 2 as representatives of HT performance. Accordingly, the developed HT polymers have almost the same HOMO energy and hole mobility and sufficient insolubility, differing in only T_g and IEs. Therefore, we can isolate the thermal property and interaction strength and study their impact on hole injection.

Table 2 Properties of the developed HT polymers

		HT-4′	HT-6	HT-8	HT-12	HT-14
Current density at 8 V for HODs without QDs and consisting of QDs are representatives for HT performance of HT polymers and hole-injection performance from HT polymers to QDs, respectively. Parameters in bold are those showing differences among HT polymers.
Design	Alkyl chain	C(CH3)3	C6H13	C8H17	C12H25	C14H29
	Interaction energy (eV)	0.425	0.463	0.640	1.076	1.482
Physical properties	HOMO (eV)	−5.50	−5.55	−5.55	−5.55	−5.55
	T g (°C)	146	216	197	152	141
	Insolubility for octane	√	√	√	√	√
Hole transporting performance	HOD (without QDs) (mA cm^−2)	1475	1378	1383	1394	1302
	HOD (with QDs) (mA cm^−2)	0.51	0.41	0.45	1.39	1.50

---

Hole-injection performance

HODs consisting of QDs were fabricated to investigate the relationship between IEs and hole injection. The HT-12 and 14 polymers, with large IE, indicate higher current density than HT-4′, 6 and 8, with small IE, as shown in Fig. 3. Comparing HT-4′ and HT-6, having similar IE but different T_g, the current densities were similar. Comparing HT-4′ and HT-12, 14, having similar T_g but different IE, the current density was higher when IE was large. In other words, IE is the dominant factor for hole injection and T_g is not. For detailed investigation, the temperature dependence was examined for HT-12 and HT-4′, as shown in Fig. 4a and b, respectively. QD baking temperatures were varied at 120, 140, 160, and 180 °C, where 120 and 140 °C are below T_g and 160 and 180 °C are above T_g. The HODs consisting of HT-12 with large IE showed temperature dependence, while the HODs consisting of HT-4′ with small IE showed no temperature dependence. This means that IE is the dominant factor for hole injection and that there is an additional temperature effect when IE is large. In the following subsections, we discuss the mechanism of why hole injection is enhanced when IE is large and why temperature dependence exists when IE is large from the viewpoint of the interface structure.

Fig. 3 Current density–voltage characteristics of HODs.

Fig. 4 Temperature dependence of current density–voltage characteristics of HODs consisting of (a) HT-12 and (b) HT-4′.

Interface-structure analysis with X-ray reflectivity

We applied XRR to the interface-structure analysis. We analysed the HT polymer/QD/ZnMgO structure model film, which is the main part around the emissive layer in QD-LEDs. Zinc magnesium oxide is an electron-transporting material, and the thickness and process temperature of each layer are the same as those for QD-LEDs. The results for a single film of each layer and the results for HT polymer/QD/ZnMgO stacked film are as follows.

The structure of the QD single film was first analysed. A QD-film thickness of 20 nm corresponds to 2 steps of QD particles since the QD size is about 10 nm. The experimental reflection pattern could not be reproduced with a structure model consisting of a single uniform layer, as shown in Fig. 5a. This means that the QD layer is not a uniform structure, i.e., a density distribution in accordance with the QD steps exists in the film-thickness direction. Thus, the structure model was modified to consist of three layers, two steps of QDs and the inter step, to reproduce the experimental pattern, as shown in Fig. 5b. From the analysis, the density profile was obtained as shown in Fig. 5c, where the first and second QD steps were about 3.2–3.3 g cm⁻³ and the inter step was about 3.0 g cm⁻³. The two steps of QD particles were observed with TEM, as shown in Fig. 5d. This good agreement indicates that the stacking state of QDs could be analysed with XRR. Regarding the single film of HT-12 and ZnMgO, the reflection pattern could be analysed with a structure model consisting of a single uniform layer (Fig. 5e and f, respectively). The densities of HT-12 and ZnMgO were about 1.1 and 2.9 g cm⁻³, respectively. The ZnMgO layer can be regarded as a uniform layer, probably because the thickness of 60 nm is sufficient compared with particle size (about 3 nm).

Fig. 5 XRR patterns of QD single film analysed with the (a) 1-layer model and (b) 3-layer model. Inset is the analysis model. (c) Density profile of the QD single film obtained with the 3-layer model. (d) TEM image of the HT-12/QD/ZnMgO stacked film (focused on QD). Red circles indicate QD particles, and density profile is overlaid for comparison. XRR patterns of (e) HT-12 single film and (f) ZnMgO single film, analysed with 1-layer model. Insets are the obtained density profiles.

From the above results, the analysis model for the HT polymer/QD/ZnMgO film was set as a six-layered structure: (1) HT polymer, (2) interface, (3) step 1 of QD, (4) QD inter step, (5) step 2 of QD, and (6) ZnMgO, as shown in Fig. 6a. The interface layer was introduced because a density difference between the HT polymer and QD is large. Using this analysis model, the complex reflection pattern for the HT-12/QD/ZnMgO stacked film (Fig. 6b) could be reproduced. The obtained density profile (Fig. 6c) showed good agreement with the TEM image (Fig. 6d), which indicates the validity of XRR for interface-structure analysis.

Fig. 6 XRR analysis for the HT-12/QD/ZnMgO stacked film. (a) Analytical structure model consisting of 6 layers. (b) XRR pattern analysed with the 6-layer model (a). (c) Obtained density profile. (d) TEM image of the HT-12/QD/ZnMgO stacked film.

Relationship between hole injection and interface structure

Fig. 7a shows the density profiles of the HT-12 single film and HT-12/QD/ZnMgO stacked films, where the process temperature of ZnMgO was varied. Table 3 lists the obtained parameters. The interface structure of the HT-12/QD/ZnMgO stacked film fabricated at the same process temperature as QD-LED fabrication is discussed. The process temperature of ZnMgO is 140 °C. The thickness of HT-12 in the stacked film decreased compared with that in the single film. The HT-12 roughness was larger for the stacked film than that for the single film, which means that the HT-12 surface structure and HT-12/QD interface structure are different. In other words, the QD particles are not simply stacking on the HT-12 surface. Since the HT-12 film did not dissolve in the QD solvent, the decrease in HT-12 thickness indicates interfacial mixing due to QD particles sinking into the HT-12 layer. Due to this sinking, a large contact area between the HT-12 and QD particles can be formed. Accordingly, the decrease in HT-12 thickness and large contact area are factors inducing hole injection.

Fig. 7 Density profiles of the HT polymer single film and HT polymer/QD/ZnMgO stacked films consisting of (a) HT-12 and (b) HT-4′. The baking temperature of ZnMgO was varied at 120, 140, and 180 °C. Inset is the density profiles focused on QD.

Table 3 Film parameters obtained from XRR analysis

Baking temperature of ZnMgO		HT-12 (Tg = 152 °C)				HT-4′ (Tg = 146 °C)
		HT polymer	HT polymer/QD/ZnMgO			HT polymer	HT polymer/QD/ZnMgO
			120 °C	140 °C	180 °C		120 °C	140 °C	180 °C
QD	Density2 (g cm^−3)		3.43	3.40	3.32		3.42	3.44	3.32
	Density1 (g cm^−3)		3.43	3.46	3.44		3.43	3.36	3.27
	Thickness (nm)		19.1	19.9	21.9		22.3	22.1	22.3
HT polymer	Roughness (nm)	0.8	1.2	1.2	1.2	0.6	1.1	1.2	1.2
	Density (g cm^−3)	1.06	1.08	1.08	1.06	1.10	1.07	1.06	1.07
	Thickness (nm)	25.5	20.8	20.2	17.8	27.3	26.7	26.6	26.4

---

We modelled the stacking process to explain the mechanism of QD particle sinking, as shown in Fig. 8a: (1) QD solution casting; QD solution is coated on HT polymer, (2) spin coating; the solvent (octane) is removed, and (3) baking; the QD particles are aligned by the interaction with the HT polymer and an interface is formed. Fig. 8b shows an enlarged view of the contact area between HT-12 and a QD particle. During the process from solvent removal to baking (steps 2 and 3), QD particles approach close to the HT-12 due to a strong interaction. Then, QD particles sink into the HT-12 layer. During this process, the molecular vibration and movement at the HT-12 surface also assist QD particles in approaching.

Fig. 8 (a) Schematic images of QD stacking on HT polymer at 3 steps: QD solution casting, spin coating, and baking. (b) and (c) Enlarged images of contact between the HT polymer and QD, consisting of HT-12 and HT-4′, respectively.

The temperature dependence of the interface structure was then investigated. The baking temperature of ZnMgO was varied: 120, 140, and 180 °C. Although the difference was slight at 120 and 140 °C, a decrease in HT-12 film thickness was clearly observed at 180 °C, which was above T_g. This indicates the progressive interfacial mixing at higher temperature. This is consistent with hole injection increasing in accordance with baking temperature (refer back to Fig. 4a). An increase in QD thickness was also observed. The QD density on the ZnMgO side decreased, while that on the HT-12 side remained unchanged, as shown in the inset of Fig. 7a. The densities are listed as QD density 2 and 1 in Table 3, respectively. These changes in the interface structure caused by high-temperature baking can be explained as follows. Above T_g, the molecular motion of the bulk accelerates QD sinking, resulting in a decrease in the HT-12 thickness. QD particles on the HT-12 side sink as they are due to a strong interaction, so the QD density is maintained. However, QD particles on the ZnMgO side cannot follow the sinking; therefore, the QD film stretches, which results in a decrease in the QD density and increase in QD thickness. These differences in interface structure are schematically shown as stacking images in Fig. 9 ((a): 120 °C and (b): 180 °C) with corresponding density profiles. In summary, the strong interaction induces QD sinking (= interfacial mixing), which results in the decrease in HT-12 thickness and large contact area, which is the interface favourable for hole injection. Increasing baking temperature progresses the interfacial mixing and increases hole injection.

Fig. 9 QD stacking images at the HT polymer/QD interface with the corresponding density profiles, (a) and (b) for the film consisting of HT-12, fabricated at 120 and 180 °C, (c) and (d) for the film consisting of HT-4′, fabricated at 120 and 180 °C, respectively. As a guide for the eye, QD layers are indicated with broken lines.

Next, we discuss the interface structure when the interaction is weak, as shown in Fig. 7b (see Table 3 for parameters). There was no change in HT-4′ film thickness between the HT-4′ single film and HT-4′/QD/ZnMgO stacked film, indicating no QD sinking. There is no favourable factors for hole injection, which is consistent with the suppressed hole injection (refer back to Fig. 3). Similarly to HT-12, the stacking process is modelled in Fig. 8c. During the process from solvent removal to baking (Steps 2 and 3), the QD particles did not approach HT-4′. The QD particles did not sink, so no interfacial mixing occurred. In addition, HT-4′ thickness did not change, and interfacial mixing did not progress when the baking temperature increased. This is also consistent with the fact that hole injection is not temperature dependent (refer back to Fig. 4b). Although the HT-4′ thickness did not change, QD density decreased overall as the baking temperature increased. Refer to the inset of Fig. 7b. No change in HT-4′ thickness even above T_g should be due to the QD particles not approaching closer to HT-4′ due to weak interaction. That is, QD particles do not sink; thus, interfacial mixing does not occur. Only the QD arrangement is disturbed by the molecular motion/vibration, thus reducing QD density. These differences in interface structure are schematically shown as stacking images in Fig. 9c and d with corresponding density profiles, similar to HT-12. In summary, when the interaction is weak, interfacial mixing does not occur and there is no favourable factor for hole injection. Since interfacial mixing does not occur at higher temperatures, hole injection shows no temperature dependence.

To summarize, when the interaction is strong, QD particles sink into the HT polymer layer, forming an interface favourable for hole injection. Furthermore, high-temperature treatment promotes interfacial mixing and hole injection. If the interaction is weak, interfacial mixing does not occur and there is no factor favourable to hole injection.

QD-LED performance

Finally, the QD-LED performance is shown in Fig. 10 and Fig. S8 (ESI†) and Table 4. The operation voltages were almost the same. Luminance increased in the order of alkyl-chain length, except for HT-14. The reason HT-14 showed lower luminance is probably due to its relatively lower T_g. The T_g of 141 °C is almost the same as that for the ZnMgO treatment temperature of 140 °C, suggesting that somewhat structural disorder has occurred. The maximum external quantum efficiencies (EQEs) for HT-6 and 8 were higher than those for HT-4′, 12 and 14, as shown in Fig. 10b. Since HT-12 and 14 showed enhanced hole injection, the recombination zone should be extended to the ZnMgO side, resulting in exciton quenching by ZnMgO.^36,37 It is also possible that ZnMgO penetrating the QD layer caused the EQE decrease because the T_g of HT polymer is close to the ZnMgO treatment temperature. For HT-6 and 8, which showed suppressed hole injection, the recombination zone should be closer to the HT polymer side, therefore, the quenching effect by ZnMgO will be small. The T_g is high enough that the impact of ZnMgO penetration is considered small. However, HT-4′ showed lower EQE despite its suppressed hole injection. The recombination zone is considered close to the HT polymer side, however, the effect of ZnMgO penetration due to the low T_g is strong, which will result in the low EQE. Fig. 10c shows lifetime measurement. The time required for 5 and 20% decay at the initial luminance of 4500 cd m⁻², respectively LT95 and LT80, were applied as representatives of lifetime. For LT80, the lifetimes of HT-12 and 14, with enhanced hole injection, were clearly longer than those for HT-4′, 6, and 8, with suppressed hole injection. For LT95, the lifetimes of HT-12 and 14 were also longer than those of HT-4′, 6, and 8, though the difference was small. Accordingly, the effects of interaction strength and thermal property on device performance can be distinguished and summarized as follows. Comparing devices with similar T_g (HT-4′, 12, and 14), the EQEs were similar while the device lifetime showed a correlation with hole injection, that is, interaction strength. Comparing devices with similar interaction strength (HT-4′ and 6), the device lifetimes were similar while the EQE was lower for HT-4′ with lower T_g. These results indicate that EQE is dominated by T_g and lifetime is dominated by interaction strength.

Fig. 10 QD-LED performance. (a) Current density (left axis) and luminance (right axis) vs. voltage profiles. (b) EQE voltage profiles. (c) Lifetime measurements at an initial luminance of 4500 cd m⁻². Inset is the initial stage of lifetime measurements.

Table 4 QD-LED performance

	HT-4′	HT-6	HT-8	HT-12	HT-14
EQE-max (%)	12.4	15.1	14.9	12.7	12.8
LT95 (hours)	7.0	1.3	4.3	10.0	16.9
LT80 (hours)	50.2	44.9	55.7	111.2	112.1

---

A longer alkyl chain of HT polymers, which shows stronger interaction with QDs, leads to enhanced hole injection, hence longer device lifetime, however, a low T_g causes penetration of ZnMgO into the QD layer, resulting in EQE lowering. Therefore, it is effective for higher device performance to strengthen the interaction while maintaining a T_g sufficiently higher than the device-fabrication temperature.

Conclusions

From the viewpoint of the interaction between HT polymers and QDs, we synthesized HT polymers with various alkyl chains to investigate the relationship of alkyl chains with a HT polymer/QD interface structure and hole injection. We clarified that longer alkyl chains interact more strongly with QDs and show enhanced hole injection, hence longer QD-LED device lifetime. The reason a strong interaction causes enhanced hole injection was revealed from the interface structure analysis. A strong interaction induces sinking of QD particles, leading to interfacial mixing, thus enhanced hole injection. Furthermore, high baking temperature results in progressive interfacial mixing and enhanced hole injection. While the interaction is weak, QD particles do not sink, and a separate interface forms, resulting in suppressed hole injection. In addition, interfacial mixing does not occur even at high baking temperature. Even if the HOMO energy, hole mobility, and thermal properties are the same, stronger interaction leads to enhanced hole injection. The above findings were revealed by establishing a technique to use density profiles obtained from XRR. Thus, we illustrated and quantified the HT polymer/QD interface structure, demonstrating that interfacial analysis with XRR is extremely effective.

Author contributions

The manuscript was written with contributions from all authors. K. Tsutsumi: conceptualization, data curation, formal analysis, investigation, methodology, visualization, writing – original draft. Y. Konishi: data curation, formal analysis, investigation, writing – review & editing. T. Motoyama: conceptualization, investigation, methodology. F. Kato: investigation, resources, visualization, writing – original draft, writing – review & editing. W. Sotoyama: investigation, writing – original draft, writing – review & editing. N. Suganuma: investigation, writing – review & editing. M. Tsuji: investigation, resources. T. Fujiyama: resources. K. Furuta: resources. N. Ishii: resources. T. Kikuchi: project administration, writing – review & editing. D.-Y. Chung: investigation, resources. H. Kwon: resources. Y.-H. Won: resources. T. Yagi: supervision, writing – review & editing.

Data availability

The data supporting this article have been included as part of the ESI.† More detailed data are available from the corresponding authors.

Conflicts of interest

There are no conflicts to declare.

Notes and references

1. E. Jang and H. Jang, Chem. Rev., 2023, 123, 4663–4692.
2. S. Y. Bang, Y. H. Suh, X. B. Fan, D. W. Shin, S. Lee, H. W. Choi, T. H. Lee, J. Yang, S. Zhan, W. Harden-Chaters, C. Samarakoon, L. G. Occhipinti, S. D. Han, S. M. Jung and J. M. Kim, Nanoscale Horiz., 2021, 6, 68–77.
3. Q. Yuan, T. Wang, P. Yu, H. Zhang, H. Zhang and W. Ji, Org. Electron., 2021, 90, 106086.
4. M. K. Choi, J. Yang, T. Hyeon and D. H. Kim, npj Flexible Electron., 2018, 2, 10.
5. F. Chen, Q. Lin, H. Shen and A. Tang, Mater. Chem. Front., 2020, 4, 1340–1365.
6. C.-Y. Han, S.-H. Lee, S.-W. Song, S.-Y. Yoon, J.-H. Jo, D.-Y. Jo, H.-M. Kim, B.-J. Lee, H.-S. Kim and H. Yang, ACS Energy Lett., 2020, 5, 1568–1576.
7. S. Kim, J.-A. Kim, T. Kim, H. Chung, S. Park, S.-M. Choi, H.-M. Kim, D.-Y. Chung and E. Jang, Chem. Mater., 2020, 32, 5200–5207.
8. E. P. Jang, C. Y. Han, S. W. Lim, J. H. Jo, D. Y. Jo, S. H. Lee, S. Y. Yoon and H. Yang, ACS Appl. Mater. Interfaces, 2019, 11, 46062–46069.
9. Y. H. Won, O. Cho, T. Kim, D. Y. Chung, T. Kim, H. Chung, H. Jang, J. Lee, D. Kim and E. Jang, Nature, 2019, 575, 634–651.
10. E. Jang, Y. Kim, Y.-H. Won, H. Jang and S.-M. Choi, ACS Energy Lett., 2020, 5, 1316–1327.
11. J. Fan, C. Han, G. Yang, B. Song, R. Xu, C. Xiang, T. Zhang and L. Qian, Adv. Mater., 2024, 36, 2312948.
12. Y. Kim, S. Ham, H. Jang, J. H. Min, H. Chung, J. Lee, D. Kim and E. Jang, ACS Appl. Nano Mater., 2019, 2, 1496–1504.
13. J. H. Chang, P. Park, H. Jung, B. G. Jeong, D. Hahm, G. Nabamine, J. Ko, J. Cho, L. A. Padilha, D. C. Lee, C. Lee, K. Char and W. K. Bae, ACS Nano, 2018, 12, 10231–10239.
14. W. Cao, C. Xiang, Y. Yang, Q. Chen, L. Chen, X. Yan and L. Qian, Nat. Commun., 2018, 9, 2608.
15. P. Gao, J. Wang, L. Wang, D. Wang, W. Peng, S. Zou, Y. Mo and Y. Zhang, Org. Electron., 2021, 92, 106138.
16. J. Song, O. Wang, H. Shen, Q. Lin, Z. Li, L. Wang, X. Zhang and L. S. Li, Adv. Funct. Mater., 2019, 29, 1808377.
17. G. Zaiats, S. Ikeda, S. Kinge and P. V. Kamat, ACS Appl. Mater. Interfaces, 2017, 9, 30741–30745.
18. S. Chen, W. Cao, T. Liu, S. W. Tsang, Y. Yang, X. Yan and L. Qian, Nat. Commun., 2019, 10, 765.
19. T. Davidson-Hall and H. Aziz, Nanoscale, 2019, 11, 8310–8318.
20. T. Davidson-Hall and H. Aziz, ACS Appl. Mater. Interfaces, 2020, 12, 16782–16791.
21. L. Zheng, G. Zhai, Y. Zhang, X. Jin, L. Gao, Z. Yun, Y. Miao, H. Wang, Y. Wu and B. Xu, Superlattices Microstruct., 2020, 140, 106460.
22. P. Tang, L. Xie, X. Xiong, C. Wei, W. Zhao, M. Chen, J. Zhuang, W. Su and Z. Cui, ACS Appl. Mater. Interfaces, 2020, 12, 13087–13095.
23. A. Neuhold, H. Brandner, S. J. Ausserlechner, S. Lorbek, M. Neuschitzer, E. Zojer, C. Teichert and R. Resel, Org. Electron., 2013, 14, 479–487.
24. A. Singth, A. Mathur, D. Pal, A. Sengupta, R. Singh and S. Chattopadhyay, Mater. Today Proc., 2019, 18, 1517–1523.
25. M. Wlodek, A. Slastanova, L. J. Fox, N. Taylor, O. Bikondoa, M. Szuwarzynski, M. Kokasinska-Sojka, P. Warszynski and W. H. Briscoe, J. Colloid Interface Sci., 2020, 562, 409–417.
26. H. Moon, C. Lee, W. Lee, J. Kim and H. Chae, Adv. Mater., 2019, 31, 1804294.
27. K. Tahata, S. Furukawa, H. Uji-i, T. Uchino, T. Ichikawa, J. Zhang, W. Mamdouh, M. Sonoda, F. C. D. I. Schryver, S. D. Feyter and Y. Tobe, J. Am. Chem. Soc., 2006, 128, 16613–16625.
28. K. Kawabata, M. Saito, N. Takemura, I. Osaka and K. Takimiya, Polym. J., 2017, 49, 169–176.
29. L. Qian, Y. Zheng, J. Xue and P. H. Holloway, Nat. Photonics, 2011, 5, 543–548.
30. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, G. A. Petersson, H. Nakatsuji, et al.Gaussian 16, Revision C.01, Gaussian Inc., Wallingford, CT, 2019.
31. A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652.
32. P. C. Hariharan and J. A. Pople, Theor. Chim. Acta, 1973, 28, 213–222.
33. S. Grimme, S. Ehrlich and L. Goerigk, J. Comput. Chem., 2011, 32, 1456–1465.
34. E. H. Kisi and M. Elcombe, Acta Crystallogr., 1989, C45, 1867–1870.
35. J. Lian, L. Ying, W. Yang, J. Peng and Y. Cao, J. Mater. Chem. C, 2017, 21, 5096–5101.
36. B. Liu, L. Lan, Y. Liu, H. Tao, H. Li, H. Xu, J. Zou, M. Xu, L. Wang, J. Peng and Y. Cao, Org. Electron., 2019, 74, 144–151.
37. X. Luo, S. He, D. Chen, G. Sun, J. Zeng, X. Zhu, W. Jin, X. Lu, Y. Hao and Y. Jin, J. Phys. Chem. Lett., 2024, 15, 6722.

---

Footnotes

† Electronic supplementary information (ESI) available: Optimized structure of QD/OA and QD/OA + HT-14; experimental details for the synthetic protocol; synthesis scheme of the related compounds; NMR spectra of the related compounds; synthesis results of HT polymers; photoelectron yield spectra of polymers; differential scanning calorimetry charts of polymers; absorption spectra of HT-14 before and after octane immersion; current density voltage characteristics of HODs without QDs; and luminance–current density characteristics of QD-LEDs. See DOI: https://doi.org/10.1039/d5tc00847f

‡ Electron blocking layer: 9-([1,1′-Biphenyl]-3-yl)-9′-([1,1′-biphenyl]-4-yl)-9H,9′H-3,3′-bicarbazole (36 nm)/HAT-CN (10 nm)

---