Molecular weight-dependent control of interfacial intermixing and dopant aggregation: a design principle for efficient solution-processed OLEDs

Abstract

Solution processed organic light emitting diodes (s-OLEDs) are promising for low cost, large area, and flexible displays, yet their external quantum efficiencies (EQEs) and operational lifetimes still lag far behind those of vacuum deposited devices owing to inefficient hole transport, interfacial intermixing, and poorly controlled dopant aggregation in the emissive layer. Here, we identify the molecular weight of the polymeric hole transport layer (HTL) as a powerful design parameter for simultaneously tuning dopant aggregation, interfacial intermixing, and carrier transport in green phosphorescent s-OLEDs. Three poly(9,9-dioctylfluorene-co-N-(4-butylphenyl)diphenylamine) (TFB) HTLs with number-average molecular weights of 51k, 60k, and 130k were incorporated into Ir(mppy)₃:CBP green emissive-layer devices and systematically investigated using atomic force microscopy, UV-visible absorption, electroluminescence, and J–V–luminance measurements. We show that the dopant aggregation state is not an undesirable byproduct of solution processing but a critical variable that must be optimized: complete suppression of aggregation by the highly solvent-resistant TFB_130k leads to a low maximum EQE (∼4.1%), severe EQE roll-off, strong parasitic TFB emission, and a shortened lifetime (∼5 h), whereas a moderate level of dopant–dopant interaction and aggregation realized with TFB_51k yields a maximum EQE of ∼9.9% with negligible roll-off at 1000 cd m⁻² and an extended lifetime of ∼36 h. These trends arise from molecular weight dependent interfacial intermixing at the TFB/CBP boundary: the relatively low solvent resistance of TFB_51k forms a graded intermixed junction that smooths hole injection and confines the recombination zone within the Ir(mppy)₃:CBP bulk, while the sharp interface of TFB_130k drives hole accumulation and interfacial recombination on TFB. Our results establish that engineering an optimal dopant aggregation window and controlled interfacial intermixing via HTL molecular weight design is essential for achieving high efficiency and reliable solution processed green OLEDs.

---

Introduction

Solution processed organic light emitting diodes (s-OLEDs) have garnered widespread attention as next generation display and lighting technologies because they can overcome the intrinsic limitations of vacuum deposited devices such as high capital expenditure and process complexity.^1–5 While OLEDs offer inherent advantages such as a high contrast ratio, wide viewing angles, fast response times, and mechanical flexibility, their practical application is often bottlenecked by intrinsic operational instability and severe efficiency roll off at high luminance. Solution processing enables low-cost fabrication and large area uniform coating; it is also compatible with roll-to-roll manufacturing and employs digital printing techniques such as inkjet printing, among other advantages. These diverse processing pathways reduce production costs and facilitate the fabrication of systems such as flexible displays and large area lighting panels, which are difficult to realize economically via vacuum deposition. With the current accelerating transition toward next generation display and lighting technologies, s-OLEDs are one of the most realistic and promising routes to commercialization for display and lighting technologies. However, their current real-time performance does not meet these expectations. Multilayer OLEDs fabricated via vacuum deposition have achieved external quantum efficiencies (EQEs) above 80%, whereas solution processed OLEDs remain at 20–30%.^6–12 This substantial gap in structural efficiency (more than a factor of three) is not merely attributable to device architecture or the quality of deposition materials. It results from the cumulative effect of fundamental issues inherent to the chemical and physical nature of solution processing. Therefore, a molecular level design strategy that explicitly accounts for the unique physicochemical constraints of solution processing must be proposed to address this gap beyond their isolated issues.^13–18

Interfacial intermixing is the most direct limitation of spin coating. During deposition, organic solvents partially dissolve the hole-transport layer (HTL) and blur the interface with the emissive layer (EML). The hole electron recombination zone consequently becomes spatially delocalized, reducing Förster energy-transfer efficiency and fundamentally limiting emission performance.^19–26 The inefficient hole transport pathway within HTL is another intrinsic constraint. As organic molecules in HTL do not form a rigid lattice, hole transport proceeds via intermolecular hopping. In addition, irregular intermolecular distances in low molecular weight (M_w) or weakly entangled HTLs lead to high operating voltages, thereby reducing carrier mobility. This causes hole accumulation and a vicious cycle of enhanced nonradiative recombination. Hole transport efficiency must thus be improved to reduce operating voltage and ensure long-term, reliable device operation. In addition, dopant aggregation within the host matrix during solvent drying is a performance limiting factor, and excessive aggregation distorts dopant–dopant Dexter energy transfer and promotes nonradiative recombination.^27–36 In contrast, complete dispersion can deteriorate dopant color purity and wavelength stability. These contradictory effects indicate the existence of an optimal dopant aggregation. These issues can be simultaneously addressed via the molecular level design of HTL. In solution processed stacks, the HTL partially intermixes with the subsequently coated CBP host, forming an intermixed TFB/CBP region that acts as an effective substrate for dopant aggregation in the emissive layer. Although such interfacial intermixing is typically regarded as an unavoidable defect of spin coating that broadens the recombination zone and deteriorates emission performance, we instead treat it as a tunable design parameter whose magnitude can be optimized via the solvent resistance and M_w of the TFB HTL.

In this framework, low M_w TFB with weaker solvent resistance deliberately allows partial dissolution during GEML spin coating and forms a graded TFB/CBP junction, whereas high M_w TFB with strong solvent resistance preserves a sharp interface, leading to fundamentally different charge accumulation and recombination behaviors.

Within this intermixed region, the molecular weight dependent chain entanglement and packing of TFB are expected to alter the local polarity and nanoscale morphology, which in turn influenced dopant aggregation and hole injections across the HTL/EML interface. This indicates that the M_w of the HTL is a powerful design parameter for tuning the local environment of dopants, rather than simply maximizing their dispersion. Previous studies have shown that increasing the molecular weight of polymeric semiconductors generally promotes chain entanglement and film densification, which can lower surface energy and suppress solvent penetration.^37–46 In our system, the observed increase in absorption intensity and the small variations in RMS roughness with M_w are in line with this trend, but should be regarded as indirect rather than direct evidence of entanglement.

However, at excessively high M_w, a different type of trade-off emerges. The extreme nonpolarity of the HTL surface reduces its chemical affinity with dopant molecules in the overlying EML and drives the system toward an over dispersed dopant state in which dopant–dopant interactions are strongly suppressed. While such a dispersed state effectively mitigates interfacial intermixing, it simultaneously destabilizes host–dopant energy transfer, enhances parasitic emission, and degrades device reproducibility and operational stability. Thus, the M_w of the HTL is a key control variable for balancing interfacial intermixing and the dopant-aggregation degree in the emissive layer. Low M_w (e.g., 51k) favor strong dopant–dopant interactions and efficient hole injection but cause severe interfacial intermixing, whereas high M_w (e.g., 130k) effectively suppress intermixing but can over-disperse dopants and increase the hole-injection barrier.

On this basis, the following central hypothesis is proposed herein: complete suppression of dopant aggregation degrades device performance and an optimal level of dopant aggregation, which can be achieved within a specific range of M_w, maximizes the optical and electrical performance of the device. This viewpoint fundamentally differs from existing methods that aim to maximize dopant dispersion. Three M_w of poly(9,9-dioctylfluorene-co-N-(4-butylphenyl)diphenylamine) (TFB), namely 51k, 60k, and 130k, were employed to systematically analyze the correlations among thin film morphology, dopant aggregation, and device performance and test the hypothesis. Atomic force microscopy (AFM) and UV-visible spectroscopy (UV-vis) were used to quantitatively evaluate thin-film morphology and dopant aggregation as a function of M_w and to elucidate their relationships with EQE, operational lifetime, and color purity. We also reported that device performance cannot be accurately evaluated based solely on turn-on voltage (V_on). A reduced V_on has been used as a key figure of merit for enhancing the HTL performance; however, the long-term operation of actual devices is determined by carrier-transport efficiency under steady-state conditions (operating voltage, V_op) and the shape of the J–V curve, particularly its linearity and hysteretic behavior. This shift in perspective is crucial for the reliable design and practical implementation of s-OLEDs. An optimal level of dopant aggregation was achieved with a TFB M_w of 51k, realizing a high EQE (9.93%) and excellent operational lifetime (36 h). This indicated that the central challenge in s-OLEDs lies not merely in controlling the dopant dispersion degree but in balancing dopant–dopant interactions and host–dopant energy transfer. Therefore, AFM, UV-vis spectroscopy, and electroluminescence characterization were employed to elucidate the interrelationships among the M_w of TFB, dopant aggregation, hole–transport mechanisms, and device reliability.

Experimental

Materials

Poly(3,4-ethylenedioxythiophene)-poly(styrenesulfonate) (PEDOT:PSS, Clevios™ P VP AI 4083) was purchased from Heraeus Co. and mixed with isopropyl alcohol (IPA) for use as the hole-injection layer (HIL). TFB with three different M_w (51k, 60k, and 130k) was obtained from OSM Co., in South Korea and used as the hole-transport layer (HTL). The green emissive layer (GEML) was prepared by blending tris[2-(p-tolyl)pyridine]iridium(III) [Ir(mppy)₃] as the dopant with 4,4′-bis(N-carbazolyl)-1,1′-biphenyl (CBP) as the host material; both were purchased from OSM Co., in South Korea. 2,2′,2″-(1,3,5-Phenylene)tris(1-phenyl-1H-benzimidazole) (TPBi) serving as the electron transport layer (ETL), and lithium fluoride (LiF) as the electron injection layer (EIL), were also purchased from OSM Co., in South Korea. Aluminum (Al) was purchased from iTASCO in South Korea and was used for the cathodes.

Device fabrication

The substrate was prepared by patterning indium tin oxide (ITO) with a thickness of 150 nm on glass. The ITO pattern was designed with the emitting area of 1.5 mm × 1.5 mm. Prior to solution processing, the ITO-patterned glass substrate was cleaned via ultrasonic washing in acetone and IPA for 10 min each. Then, the substrate was thoroughly dried in an oven at 100 °C, and its surface was treated in a UV-ozone cleaner (UV-30S) for 15 min. PEDOT : PSS : IPA (1 : 1 by volume) was spin-coated onto the patterned ITO substrate at 2000 rpm for 30 s and annealed in a vacuum oven at 100 °C for 30 min. After annealing, the substrate was transferred into a nitrogen-filled glovebox. TFB dissolved in p-xylene (15 mg mL⁻¹) was spin-coated at 4000 rpm for 30 s and annealed at 150 °C for 30 min. The GEML was prepared by spin-coating a CBP : Ir(mppy)₃ (host : dopant = 95 : 5 wt%) solution dissolved in toluene (10 mg mL⁻¹) at 4000 rpm for 30 s, followed by annealing at 100 °C for 30 min. All processes, except the PEDOT:PSS & IPA (HIL) deposition were performed within a nitrogen-filled glovebox. Then, TPBi, LiF, and Al were deposited via thermal vacuum evaporation at pressures below 3 × 10⁻⁷ Torr. Following Al deposition, device encapsulation was performed within the nitrogen-filled glovebox.

Instruments and measurements

Interfacial intermixing between the HTL and EML was evaluated via UV-vis spectrophotometry (UV-2600i, Shimadzu Scientific Korea Co.). The surface characteristics and thickness of TFB thin films were characterized via AFM (NX7, Park System Co.) and using a three-dimensional surface profiler (NV-1800, NanoSystem Co., Ltd.), respectively. The electrical and optical performances were evaluated using a source meter (2400, Keithley Instruments, LLC) and a luminance meter (PR655), respectively, and J–V–luminance (J–V–L) characteristics were analyzed. The operational lifetime was assessed via Lifetime Tester (M6000 Plus, McScience Inc.).

Results and discussion

The target thickness of the HTL/EML stacked thin film, determined from optical and device-design considerations, was approximately 70 nm. Under ideal conditions without interfacial intermixing, the total thickness of the stacked films should increase in proportion to the HTL thickness; however, during solution processing, severe interfacial intermixing causes the actual HTL/EML thickness to saturate in the range of 60–75 nm, far below the theoretical values at high TFB concentrations. Fig. 1 compares the UV-vis absorption spectra of the TFB thin film (15 mg mL⁻¹) before and after toluene rinsing, indicating that the thin film dissolved partially and its thickness decreased after rinsing.

Fig. 1 UV-visible absorption spectra of pristine and toluene-rinsed TFB thin films at different concentrations (12, 15, 20, 25 mg mL⁻¹).

The extent of interfacial intermixing and optimal processing conditions were quantitatively evaluated at different TFB concentrations of 12, 15, 20, and 25 mg mL⁻¹, and the corresponding changes in thin film thickness was investigated. The thickness of a single TFB thin film and GEML prepared at each concentration was measured using a three dimensional profiler. Then, the thickness of HTL/EML stacked structures was re-measured. The thicknesses of single TFB thin films were 45, 55, 70, and 115 nm at solution concentrations of 12, 15, 20, and 25 mg mL⁻¹, respectively, whereas that of a single GEML was ∼50 nm. Under ideal conditions with no interfacial intermixing, the expected thicknesses of HTL/EML stacked structures were 95, 105, 120, and 165 nm, respectively, at the concentrations. However, the actual thicknesses were only 60, 70, 75, and 70 nm, respectively. According to the following eqn (1), the thickness loss rates were 36.84%, 33.34%, 37.5%, and 57.58% at concentrations of 12, 15, 20, and 25 mg mL⁻¹, respectively.

Severe interfacial intermixing was observed at a concentration of 25 mg mL⁻¹, and more than half of the theoretical thickness was lost; this indicated significant instability in HTL thickness and hole transport characteristics during subsequent device fabrication. In contrast, at a TFB concentration of 15 mg mL⁻¹, the measured stacked thickness was 70 nm, which closely matches the designed target (∼70 nm) and exhibits the lowest thickness loss rate among the four conditions (Table 1). Therefore, a TFB concentration of 15 mg mL⁻¹ was selected as the baseline HTL condition for subsequent UV-vis analysis and device characterizations.

Table 1 Quantitative analysis of thickness reduction and loss rate of HTL/EML stacked structure as a function of TFB concentrations

HTL concentration (mg ml^−1)	Thickness (nm)	Theoretical HTL/EML thickness (nm)	Measurement HTL/EML thickness (nm)	Reduction rate (%)
12	45	95	60	36.84
15	55	105	70	33.34
20	70	120	75	37.5
25	115	165	70	57.58

---

Fig. 2(a) compares the UV-vis absorption spectra of TFB thin films with M_w of 51k, 60k, and 130k, all having a comparable thickness (within an experimental error margin of ±5 nm) prepared from 15 mg mL⁻¹ solutions, revealing major absorption peaks at ∼380–400 nm; the trends show that variations in the M_w of TFB does not alter the structure of the π-conjugated backbone.^47–56 Notably, the peak absorption intensity increases by ∼58% from 0.38 for 51k to 0.60 for 130k, arising from higher solution viscosity causing slight thickness increases (∼±5 nm) and packing density enhancement (not conjugation length extension). The increase in absorption intensity is the result of the combined effect of a slight increase in film thickness due to the rise in solution viscosity caused by high M_w, and the strengthening of denser packing and intermolecular π–π stacking interactions between molecular chains.

Fig. 2 UV-vis absorption and AFM morphology of TFB Thin Films (51k, 60k, 130k) from 15 mg mL⁻¹ Solutions (∼50 nm thickness). (a) Normalized spectra showing unchanged π–π* peaks at 380–400 nm, ∼58% intensity rise (0.38 → 0.60) due to packing density/π–π stacking enhancement (no conjugation length change). (b) TFB 51k AFM height (RMS: 0.422 nm): baseline chain structure. (c) TFB 60k AFM height (RMS: 0.486 nm): enhanced chain entanglement. (d) TFB 130k AFM height (RMS: 0.498 nm): maximum densification correlating with peak absorption.

Fig. 2(b–d) presents AFM height images and 3D morphology of TFB thin films, showing molecular-weight-dependent changes in surface topography that accompany the gradual increase in UV-vis absorption intensity. These morphological changes, together with the absorption enhancement, are consistent with a higher packing density and more pronounced chain overlap in higher-molecular-weight TFB films at the HTL interface, although chain entanglement is not directly quantified in this study.

Fig. 3(a) compares absorption spectra of pristine GEML (G) and GEMLs stacked on TFB thin films (G_51k, G_60k, G_130k), where G exhibits characteristic ligand-centered (LC) π–π* transition (250–280 nm), ¹MLCT (290–310 nm), and ³MLCT (320–360 nm).^57–65 In G_51k, LC absorption weakens relative to G while ¹MLCT remains positionally/intensity-stable, but ³MLCT red-shifts with increased intensity; G_60k shows LC strengthening, minor ¹MLCT gains, and persistent ³MLCT red-shift/enhancement; G_130k displays further LC/¹MLCT intensification without positional shifts, with ³MLCT recovering pristine position/intensity. These spectral variations, together with the AFM roughness evolution, suggest a progressive reduction in dopant–dopant interactions and aggregation-related morphological non-uniformity as the TFB M_w increases, although the dopant distribution is not directly resolved in this work. Fig. 3(b–f) displays corresponding GEML AFM height images, with RMS surface roughness decreasing from 37.1 nm (G_51k) to 29.5 nm (G_60k) and 23.6 nm (G_130k) a range mirroring dopant aggregate size/height non-uniformity.^66–70 High G_51k roughness arises from TFB_51k elevated interfacial polarity (Fig. 2(b)), driving non-uniform aggregation that compresses ppy ligands and attenuates LC intensity; despite extreme 37.1 nm roughness indicating large domains, underlying TFB_51k blocks leakage currents, localizing excitons in aggregates for high EQE/stability. Conversely, low G_130k roughness confirms uniform dispersion in TFB_130k's low-polarity environment (Fig. 2(d)), eliminating dopant–dopant interactions, stabilizing MLCT states, and enhancing LC intensity via ligand freedom. Changes in Ir(mppy)₃ absorption characteristics with varying TFB M_w demonstrate effective modulation of dopant aggregation through molecular-weight-dependent local polarity and packing in the intermixed TFB/CBP region, corroborated by GEML roughness reductions and TFB morphological gains (Fig. 2 and 3). The ³MLCT red shifts in G_51k/G_60k confirm polarized environments and strong dopant interactions matching high roughness, while G_130k pristine like spectrum indicates isolated dopant states aligned with flat morphology (Fig. 3(e)). Compared to pure GEML (Fig. 3(f)), G_130k shows residual aggregation non-uniformity yet recovered LC intensity; minimal ¹MLCT changes affirm its aggregation insensitivity. Overall, TFB M_w acts as a powerful design parameter for tuning dopant states and polar environments, crucial for s-OLED optimization beyond simple dispersion maximization.

Fig. 3 UV-vis absorption spectra and AFM morphology of GEML on TFB HTLs (51k, 60k, 130k). (a) Normalized spectra of pristine GEML (black) and stacked GEMLs: LC weakens → strengthens, ³MLCT red-shifts (G_51k/G_60k) → recovers (G_130k) indicating aggregation suppression. (b) G_51k AFM height (RMS: 37.1 nm): largest dopant aggregates from high TFB_51k polarity. (c) G_60k AFM height (RMS: 29.5 nm): reduced aggregation with moderate TFB_60k entanglement. (d) G_130k AFM height (RMS: 23.6 nm): uniform dispersion from TFB_130k densification. (e) Pristine GEML AFM height (RMS: 0.38 nm): smoothest baseline morphology.

Fig. 4(a) shows the structure and energy level diagram of the hole only device (HOD) used for analyzing the hole transport behavior. To suppress electron injection, a MoO₃ layer was inserted between the EML and the cathode, enabling pure analysis of hole transport characteristics. The J–V characteristics in Fig. 4(b) show that at an applied voltage of 1 V, the current densities for TFB_51k, 60k, and 130k are approximately 2–3, 0.5–1, and 0.1–0.2 mA cm⁻², respectively. This trend indicates progressively hindered hole transport in the low voltage regime. Enhanced π–π coupling at the PEDOT:PSS/TFB interface for higher M_w TFB, arising from denser packing, can favor initial charge transfer, but this benefit may diminish at very low bias where injection is not yet current limiting. In addition, changes in interfacial energetics at the PEDOT:PSS/TFB interface, including a possible increase in the effective hole injection barrier, together with packing induced modulation of the local microenvironment, can influence the hole injection efficiency. Reduced hole mobility within the bulk TFB layer due to deeper chain entanglement, which impedes intra chain hopping pathways in high M_w conjugated polymers, may further suppress the current. The observed current suppression is therefore interpreted as a combined effect of modified interfacial energetics and molecular-weight-dependent chain entanglement that hinders overall hole transport.

Fig. 4 Hole-only device structure and hole transport characteristics. (a) Device structure and corresponding energy level diagram. (b) Current density–voltage characteristics showing molecular weight-dependent hole injection in TFB_51k, 60k, and 130k.

In summary, the molecular weight dependent suppression of low voltage current in HODs reflects an interplay between interfacial energetics and bulk transport properties. While higher M_w TFB enhances π–π coupling at the PEDOT:PSS/TFB interface and lowers V_on, it simultaneously introduces transport impediments via chain entanglement and may raise injection barriers through packing induced changes in the interfacial energetics. These electrical trends indicate that TFB M_w is a powerful design parameter for tuning the balance between efficient hole injection and bulk transport in HODs.

The normalized electroluminescence spectra in Fig. 5(a) and the associated CIE 1931 trajectories in Fig. 5(b–d) provide direct optical signatures of how parasitic emission and color purity evolve with TFB M_w. All three devices share a main green emission peak at ∼530 nm with similar nominal color coordinates of (0.32, 0.61), yet the relative intensity of the TFB parasitic emission band at 420 nm differs significantly among devices. For the TFB_51k OLED, the CIE coordinates remain tightly clustered in the green region from low to high luminance and the 420 nm parasitic emission stays at only ∼3–5% of the green peak, indicating extremely efficient host-to-dopant Förster transfer and negligible parasitic loss across the operating range. In the TFB_60k device, the CIE coordinate starts between cyan and green at low luminance and gradually shifts toward green as luminance increases, implying that parasitic TFB emission contributes ∼15–20% at low brightness and progressively diminishes as Ir(mppy)₃ emission becomes dominant. The TFB_130k device exhibits the most severe color shift: its CIE coordinates begin in the Cyan region and move toward green over a much larger range, consistent with strong parasitic emission of ∼40–50% at low luminance and an 8–15-fold enhancement relative to TFB_51k. The pronounced 420 nm parasitic emission in the TFB_130k device cannot be explained solely by a collapse of host–dopant Förster transfer, because in that scenario the CBP host (∼380 nm) would be expected to dominate the residual emission rather than TFB. Instead, the emergence of an intense TFB band indicates that the main recombination zone has shifted away from the Ir(mppy)₃:CBP bulk and toward the TFB/CBP interface, where holes accumulate under steady-state bias. The severe chain entanglement and high solvent resistance of TFB_130k likely create a relatively sharp HTL/EML interface that hinders smooth hole transport through the GEML. Under steady-state bias, this situation is consistent with hole accumulation at the TFB/CBP boundary and an increased contribution of recombination on TFB molecules, which manifests as enhanced 420 nm parasitic emission, although other effects such as local thickness variations or microcavity modulation cannot be fully excluded. This interfacial recombination pathway efficiently generates TFB excitons and amplifies the 420 nm parasitic emission, while simultaneously depleting the exciton population available for green Ir(mppy)₃ emission. Together, these EL and color coordinate data show that color purity evolution is sensitive indicator of luminance dependent on parasitic emission and host–dopant energy transfer stability. The invariant color purity of TFB_51k indicates that the host–dopant MLCT state remains consistently stabilized over the full operating range, whereas the pronounced shifts in TFB_60k and TFB_130k devices reflect a growing contribution of TFB emission at low luminance that distorts the spectrum toward cyan. This behavior is not merely an optical curiosity but is physically equivalent to the EQE roll-off mechanism: as parasitic emission increases and green dopant emission is suppressed, fewer excitons are converted into useful photons, directly lowering EQE and shortening device lifetime.

Fig. 5 Electroluminescence Spectra and CIE 1931 Chromaticity of Green PhOLEDs with TFB HTLs of Different Molecular Weights (51k, 60k, 130k). (a) Normalized EL spectra of green PhOLEDs with TFB_51k (blue), TFB_60k (green), TFB_130k (orange) HTLs: common green peak (∼530 nm) but rising TFB parasitic emission at 420 nm, reflecting interface recombination shift. (b) CIE 1931 trajectory for TFB_51k device: stable green coordinates (0.32, 0.61), minimal ∼3–5% parasitic loss. (c) CIE 1931 trajectory for TFB_60k device: cyan-to-green shift, ∼15–20% low-luminance parasitic contribution. (d) CIE 1931 trajectory for TFB_130k device: largest cyan to green shift, ∼40–50% parasitic emission at low luminance.

Fig. 6(a and b) shows the device structure and the corresponding energy level diagram of green OLEDs containing TFB with different M_w as the HTL. These devices were fabricated with a stacked structure of ITO/PEDOT:PSS/TFB/Ir(mppy)₃:CBP/TPBi/LiF/Al. The energy level diagram shown in Fig. 6(b) confirms that the highest occupied molecular orbital-lowest unoccupied molecular orbital (HOMO–LUMO) levels of each functional layer are optimized for efficient carrier injection and transport.

Fig. 6 Molecular-weight-dependent electrical performance, efficiency, and operational lifetime of green OLEDs employing TFB as the HTL. (a) Device structure (b) associated energy level diagram. (c) J–V–L characteristics. (d) Current and power efficiency. (e) External quantum efficiency (f) operational lifetime curves (at 1000 cd m⁻²) and LT₅₀ values for devices.

The J–V–L characteristics (Fig. 6(c)) quantitatively reveal how the TFB M_w used as the HTL governs the electrical and optical behavior of the green OLEDs. The TFB_51k device exhibits a turn-on voltage (V_on, at 10 cd m⁻²) of ∼4.39 V and an operating voltage (V_op, at 1000 cd m⁻²) of ∼6.13 V, whereas TFB_60k and TFB_130k show reduced V_on values of ∼4.21 V and ∼4.01 V together with lower V_op values of ∼5.55 V and ∼5.59 V, respectively, with TFB_130k giving the lowest V_on among the three devices. TFB_51k reaches a maximum luminance of ∼38 000 cd m⁻², while TFB_60k exhibits a steeper current-density increase between 4–6 V, a slightly higher maximum luminance, and TFB_130k shows an even steeper J–V slope with a maximum luminance comparable to that of TFB_51k. These results indicate that V_on and V_op probe different physical processes, consistent with previous OLED J–V–L analyses. V_on is mainly governed by interfacial π–π coupling and energy-level alignment at the PEDOT:PSS/TFB interface, whereas V_op reflects the combined effects of bulk transport, recombination, and overall device resistance under high luminance operation. With increasing TFB M_w, enhanced structural disorder and deeper chain entanglement impede hole transport in the HTL, which is manifested as a transition from the gentle J–V slope of TFB_51k to the extremely steep slope of TFB_130k. This transport impedance introduces additional voltage loss as holes traverse the TFB and EML and promote carrier accumulation under steady-state bias. In the TFB_130k device, the combination of the lowest V_on, a relatively low V_op, and the steepest J–V slope signifies that hole injection at the anode/HTL interface is efficient, but carrier transport through the TFB and into the EML is strongly hindered. In contrast, the TFB_51k device, although it has a higher V_on and relatively high V_op, displays a gentle J–V slope indicative of efficient charge transport and reduced series resistance throughout the active region. Overall, Fig. 6(c) demonstrates that TFB_130k provides the lowest V_on but suffers from pronounced transport limitations, whereas TFB_51k offers more favorable steady-state transport characteristics, emphasizing that V_op and the J–V curve shape must be evaluated together with V_on when assessing HTL performance.

For comparison, we also fabricated a control OLED without the TFB HTL (ITO/PEDOT:PSS/Ir(mppy)₃:CBP/TPBi/LiF/Al) under identical processing conditions (Fig. S1 and Table S1). The HTL-free device shows a much higher operating voltage (V_op = 9.26 V at 1000 cd m⁻²) and markedly lower efficiencies, with a maximum EQE of 3.54% (12.26 cd A⁻¹, 3.67 lm W⁻¹) and an EQE of 2.6% (9.02 cd A⁻¹, 3.05 lm W⁻¹) at 1000 cd m⁻². In contrast, the TFB_51k OLED reaches a maximum EQE of 9.93% and maintains 9.92% at 1000 cd m⁻² with CE/PE values above 31 cd A⁻¹ and 16 lm W⁻¹. These results demonstrate that introducing a TFB HTL dramatically improves charge-injection balance and photon-generation efficiency, and that the molecular-weight optimization discussed here builds on this essential HTL function rather than being reproducible in HTL-free architectures.

The luminance-dependent current efficiency (CE) and power efficiency (PE) characteristics in Fig. 6(d) further clarify the effect of TFB M_w on device operation. The TFB_51k OLED shows exceptionally flat CE/PE curves, with a maximum CE of ∼31.45 cd A⁻¹ and PE of ∼16.74 lm W⁻¹ that remain nearly unchanged at 1000 cd m⁻² (31.44 cd A⁻¹ and 16.13 lm W⁻¹, respectively), indicating that energy conversion efficiency is preserved from low to high luminance. This stability suggests that host–dopant energy transfer remains robust and that carrier accumulation is effectively suppressed. By contrast, the TFB_60k device exhibits a ∼12% CE decrease from 23.20 to 20.47 cd A⁻¹ at 1000 cd m⁻², and the TFB_130k device shows a much larger ∼34% drop from 12.41 to 8.19 cd A⁻¹, accompanied by a ∼17.3% reduction in PE, indicating progressive carrier accumulation and nonradiative recombination at high luminance for the higher M_w HTLs. These trends demonstrate that the TFB_51k device maintains stable energy conversion efficiency under practical operating conditions, whereas TFB_60k and TFB_130k devices suffer from efficiency degradation due to unbalanced carrier transport and recombination.

The EQE versus luminance curves in Fig. 6(e) highlights the central role of dopant aggregation control. The TFB_51k OLED reaches a maximum EQE of ∼9.93% and maintains ∼9.92% at 1000 cd m⁻², exhibiting virtually no EQE roll-off and indicating highly stabilized photon generation and recombination efficiency. The TFB_60k device shows a lower maximum EQE of ∼7.46% that decreases to ∼6.64% at 1000 cd m⁻², while the TFB_130k device performs markedly worse, with a maximum EQE of ∼4.13% dropping to ∼2.88% at 1000 cd m⁻²; this corresponds to ∼58.4% lower peak EQE and more than ∼70% lower EQE at 1000 cd m⁻² relative to TFB_51k. The superior EQE stability of TFB_51k indicates that dopant molecules maintain an optimal spatial arrangement that supports strong dopant–dopant Dexter interactions and efficient host–dopant Förster transfer, allowing a large fraction of excitons to be converted into light across the entire operating range. In TFB_60k and TFB_130k devices, increased spacing, and weakened coupling between dopants due to partial or extreme dispersion reduce MLCT stabilization and Förster efficiency, leading directly to lower EQE and more pronounced roll-off. The approximately two-fold difference in maximum EQE and ∼3.5-fold difference at 1000 cd m⁻² between TFB_51k and TFB_130k strongly underscore that fine control of dopant aggregation through HTL M_w optimization is critical for maximizing light-output quantum efficiency.

Fig. 6(f) shows the operational lifetime characteristics of green OLEDs incorporating TFB with different M_w, plotted as relative luminance (L/L₀) versus time under an initial luminance of 1000 cd m⁻². TFB_51k OLED maintains ∼50% of its initial luminance after 36 h (LT₅₀ = 36 h), whereas TFB_60k and TFB_130k devices reach 50% L/L₀ after ∼10 h and ∼5 h, respectively. The L/L₀ curves indicate that the degradation dynamics are strongly governed by the stability of the dopant aggregation state. TFB_51k OLED exhibits a gentle luminance decay up to LT₅₀, reflecting that optimized dopant aggregation stabilizes the host–dopant MLCT state and the exciton generation annihilation pathways. In contrast, TFB_130k OLED shows a steep initial decrease in L/L₀, where extreme dopant dispersion gives rise to unstable energy transfer pathways, severe carrier accumulation, and strong exciton quenching, leading to rapid luminance loss. TFB_60k OLED displays intermediate behavior between these two extremes. Consequently, the devices reach LT₅₀ at markedly different times (36 h for TFB_51k, 10 h for TFB_60k, and 5 h for TFB_130k), highlighting that the dopant aggregation state critically determines the rate of operational degradation and the luminance half-lifetime. To quantify the practical impact of Mw on device operation, the principal performance metrics are summarized in Table 2.

To further test whether decreasing the TFB molecular weight below 51k could yield additional improvements, we also evaluated devices incorporating a lower M_w TFB (M_w ≈ 42k; Fig. S2 and Table S2). The 42k device shows a lower maximum EQE (8.17%) and reduced current and power efficiencies (25.53 cd A⁻¹ and 12.85 lm W⁻¹) compared with TFB_51k (9.93%, 31.41 cd A⁻¹, and 16.74 lm W⁻¹), while V_op remains comparable (6.32 V at 1000 cd m⁻²). These results indicate that further decreasing M_w below 51k does not improve, and in fact slightly degrades, the overall trade-off between efficiency and transport stability.

While a slight increase in thickness due to higher solution viscosity may contribute to enhanced optical absorption, the dramatic changes in device performance such as the 70% loss in EQE, the 8–15-fold increase in parasitic emission, and the significant reduction in operational lifetime cannot be explained by thickness variations alone. These results instead point to fundamental shifts in charge transport and energy transfer mechanisms driven by the TFB M_w. A similar optimization principle operates at the TFB/CBP interface through molecular-weight-dependent interfacial intermixing. In the TFB_51k devices, the relatively low solvent resistance of the low-M_w HTL allows controlled partial dissolution and swelling during Ir(mppy)₃:CBP spin coating, spontaneously forming an intermixed TFB/CBP region that behaves as a graded junction for hole injection. This graded interlayer smooths the energy and potential landscape at the HTL/EML boundary, spreading the electric field and preventing abrupt accumulation of hole at a single, atomically sharp interface. As a result, the recombination zone remains confined within the Ir(mppy)₃:CBP bulk, supporting stable green emission with negligible parasitic TFB contribution and an extended operational lifetime. By contrast, the highly entangled and strongly solvent resistant TFB_130k preserves an almost intact HTL during GEML deposition and produces a sharp TFB/CBP interface with minimal intermixing, so injected holes tend to pile up at this boundary under steady-state operation. Electrons that traverse the over dispersed EML can then directly recombine with the accumulated holes on TFB at the interface, efficiently generating intense 420 nm parasitic emission and accelerating luminance decay despite the smoother GEML morphology. These observations indicate that interfacial intermixing in solution-processed stacks should not be regarded as a purely detrimental artifact but as a tunable optimization parameter.

In this work, the term solvent resistance denotes the effective ability of the TFB HTL to withstand dissolution and swelling in the toluene GEML solvent under the specific spin coating conditions employed here, as inferred from the thickness loss and interfacial intermixing of the HTL/EML stack. We note that solvent resistance is not independently quantified as an intrinsic material parameter for each M_w, and a more systematic rinsing study would be required for that purpose. Low M_w TFB (51k) exhibits weaker solvent resistance and can be partially dissolved or swollen during GEML deposition, whereas high M_w TFB (130k) remains almost intact and preserves its original film morphology. For TFB 51k, this partial dissolution spontaneously generates a relatively thick intermixed TFB/CBP region, forming a graded HTL/EML junction. The graded junction acts as a buffer that smooths the hole injection barrier and suppresses abrupt hole accumulation at the interface, thereby enabling a softer electric field profile, mitigating parasitic TFB emission, and ultimately improving device efficiency and operational lifetime. In contrast, the strongly solvent resistant TFB 130k hardly dissolves, resulting in a sharp HTL/EML interface where holes readily pile up at the TFB/CBP boundary. This knife edge interface promotes interfacial recombination on TFB, which manifests as pronounced blue parasitic emission and accelerated luminance decay despite the suppressed dopant aggregation.

Although high M_w polymers with strong solvent resistance are generally regarded as beneficial because they effectively suppress interfacial intermixing, our results reveal a paradoxical trend: the moderately soluble low M_w TFB (51k), which allows controlled intermixing with the GEML, yields superior efficiency and stability compared with the highly solvent resistant 130k analogue. Therefore, interfacial intermixing in solution processed stacks should not be treated as an inevitable defect to be eliminated but rather as an optimized design parameter. An appropriate degree of intermixing, tuned via the M_w of the HTL, is essential for forming a graded junction that balances hole accumulation and energy transfer and thereby maximizes device performance.

Table 2 Performance of green OLEDs

Device	Von/Vop [V]	CE/PE/EQE [cd A^−1 lm^−1 W^−1/%]
		Maximum	At 1000 cd m^−2
51k	4.39/6.13	31.4/16.74/9.93	31.44/16.13/9.92
60k	4.21/5.55	23.20/11.62/7.46	20.47/11.59/6.64
130k	4.01/5.59	12.41/5.55/4.13	8.19/4.59/2.88

---

Conclusion

In this study, we demonstrated that the M_w of the polymeric hole-transport layer TFB is a decisive design parameter for simultaneously controlling dopant aggregation, interfacial intermixing, and carrier transport in solution-processed green phosphorescent OLEDs. By systematically varying the number-average M_w of TFB (51k, 60k, and 130k), we revealed that the dopant aggregation state in the Ir(mppy)₃:CBP emissive layer is not a simple byproduct of solution processing but a critical variable that must be quantitatively optimized rather than minimized. A dopant aggregation regime realized with TFB_51k, which lies between the strongly clustered and nearly fully dispersed limits, yielded a high maximum EQE of 9.9 with virtually no EQE roll-off at 1000 cd m⁻² and an extended operational lifetime (LT₅₀) of 36 h, whereas the seemingly more favorable “fully dispersed” state promoted by TFB_130k resulted in a much lower EQE (∼4.1%), severe roll-off, strong parasitic TFB emission, and a drastically shortened lifetime (∼5 h).

AFM and UV-vis analyses indicate that increasing the TFB M_w densifies the HTL film and smooths the overlying GEML, in a manner consistent with a reduced contribution of dopant–dopant interactions and a diminished MLCT red shift of Ir(mppy)₃. However, this morphological homogenization at 130k was accompanied by a steep J–V slope, enhanced carrier accumulation, and unstable host–dopant energy transfer, evidencing that extreme dopant dispersion undermines both efficiency and stability. In contrast, the low-M_n TFB_51k established an intermixed TFB/CBP region with higher interfacial polarity that supports controlled dopant clustering and balanced Dexter/Förster transfer pathways, thereby stabilizing the MLCT state and maintaining robust green emission over the full operating range.

Our results further show that interfacial intermixing at the HTL/EML boundary, often regarded as an unavoidable defect of spin coating, can be transformed into a beneficial graded junction when appropriately tuned via the solvent resistance of TFB. The relatively low solvent resistance of TFB_51k allows controlled partial dissolution during GEML spin coating, generating a graded TFB/CBP interlayer that smooths the hole-injection barrier, suppresses abrupt hole accumulation, and confines the recombination zone within the Ir(mppy)₃:CBP bulk. By contrast, the highly solvent-resistant TFB_130k preserves a sharp interface with minimal intermixing, driving holes to pile up at the TFB/CBP boundary and shifting the main recombination zone toward TFB, which manifests as intense 420 nm parasitic emission and accelerated luminance decay.

Collectively, these findings establish that optimal device performance in solution-processed green OLEDs is achieved not by eliminating dopant aggregation and interfacial intermixing, but by engineering an optimal aggregation window and a controlled intermixed junction through HTL molecular-weight design. The design concept proposed here tuning the M_w of polymeric HTLs to balance dopant aggregation, local polarity, interfacial intermixing, and steady-state carrier transport-highlights a promising route toward high-efficiency and reliable solution-processed OLED architectures. While our conclusions are based on the TFB/Ir(mppy)₃:CBP system at a fixed dopant concentration and solvent condition, we anticipate that similar molecular-weight-based strategies may be extendable to other material combinations, which will be an important subject of future studies.

Author contributions

S. H. J. performed the overall experiments, material synthesis, data analysis, and manuscript preparation. J. Y. J. and J. H. P. contributed to the theoretical analysis. All authors discussed the results and contributed to the final manuscript.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting the findings of this study, including electrical characteristics, efficiency data and EL spectra of green OLEDs with and without the TFB hole-transport layer, are available within the article and its supplementary information (SI).

Supplementary information is available. See DOI: https://doi.org/10.1039/d6cp01272h.

Acknowledgements

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. RS-2023-00222166).

References

1. J. Y. Woo, M. H. Park, S. H. Jeong, Y. H. Kim, B. Kim and T. W. Lee, et al., Advances in solution-processed OLEDs and their prospects for use in displays, Adv. Mater., 2023, 35(43), 2207454.
2. Q. Su, H. Zhang and S. Chen, Flexible and tandem quantum-dot light-emitting diodes with individually addressable red/green/blue emission, npj Flexible Electron., 2021, 5(1), 8.
3. B. Geffroy, P. Le Roy and C. Prat, Organic light-emitting diode (OLED) technology: materials, devices and display technologies, Polym. Int., 2006, 55(6), 572–582.
4. J. Bauri, R. B. Choudhary and G. Mandal, Recent advances in efficient emissive materials-based OLED applications: a review, J. Mater. Sci., 2021, 56(34), 18837–18866.
5. A. Salehi, X. Fu, D. H. Shin and F. So, Recent advances in OLED optical design, Adv. Funct. Mater., 2019, 29(15), 1808803.
6. C. W. Tang and S. A. VanSlyke, Organic electroluminescent diodes, Appl. Phys. Lett., 1987, 51(12), 913–915.
7. A. R. Duggal, C. M. Heller, J. J. Shiang, J. Liu and L. N. Lewis, Solution-processed organic light-emitting diodes for lighting, J. Disp. Technol., 2007, 3(2), 184–192.
8. R. H. Friend, R. Gymer, A. Holmes, J. Burroughes, R. Marks and C. Taliani, et al., Electroluminescence in conjugated polymers, Nature, 1999, 397(6715), 121–128.
9. B. W. D’Andrade and S. R. Forrest, White organic light-emitting devices for solid-state lighting, Adv. Mater., 2004, 16(18), 1585–1595.
10. H. Sasabe and J. Kido, Multifunctional materials in high-performance OLEDs: challenges for solid-state lighting, Chem. Mater., 2011, 23(3), 621–630.
11. J. Y. Lee, Charge trapping host structure for high efficiency in phosphorescent organic light-emitting diodes, J. Inf. Disp., 2008, 9(2), 14–17.
12. M. Hosseinnezhad, S. Nasiri, V. Nutalapati, K. Gharanjig and A. M. Arabi, A review of the application of organic dyes based on naphthalimide in optical and electrical devices, Prog. Color, Color. Coat., 2024, 17(4), 417–433.
13. C. Zhong, C. Duan, F. Huang, H. Wu and Y. Cao, Materials and devices toward fully solution processable organic light-emitting diodes, Chem. Mater., 2011, 23(3), 326–340.
14. L. Duan, L. Hou, T.-W. Lee, J. Qiao, D. Zhang and G. Dong, et al., Solution processable small molecules for organic light-emitting diodes, J. Mater. Chem., 2010, 20(31), 6392–6407.
15. K.-H. Kim, S.-Y. Huh, S.-m. Seo and H. H. Lee, Solution-based formation of multilayers of small molecules for organic light emitting diodes, Appl. Phys. Lett., 2008, 92(9), 093307.
16. L. Hou, L. Duan, J. Qiao, D. Zhang, G. Dong and L. Wang, et al., Efficient solution-processed small-molecule single emitting layer electrophosphorescent white light-emitting diodes, Org. Electron., 2010, 11(8), 1344–1350.
17. L. Hou, L. Duan, J. Qiao, W. Li, D. Zhang and Y. Qiu, Efficient single layer solution-processed blue-emitting electrophosphorescent devices based on a small-molecule host, Appl. Phys. Lett., 2008, 92(26), 263301.
18. C. Zhang, R. Yang, J. Wang, Y. Guo, B. Zhao and M. Zhao, et al., Nb₂CT_x-MXene Nanosheets for Organic Light-Emitting Diodes Achieving Extremely Outstanding Electroluminescence Performance and Stability, Adv. Funct. Mater., 2025, e16858.
19. N. C. Erickson and R. J. Holmes, Investigating the Role of Emissive Layer Architecture on the Exciton Recombination Zone in Organic Light-Emitting Devices, Adv. Funct. Mater., 2013, 23(41), 5190–5198.
20. V. Ahmad, J. Sobus, M. Greenberg, A. Shukla, B. Philippa and A. Pivrikas, et al., Charge and exciton dynamics of OLEDs under high voltage nanosecond pulse: towards injection lasing, Nat. Commun., 2020, 11(1), 4310.
21. T. N. Le and M. C. Suh, Spectral Stability Improvement by Controlling the Spatial Distribution of Excitons in Solution-processed Organic Light-emitting Diodes, J. Semicond. Technol. Sci., 2024, 24(2), 150–157.
22. F. Zhao, Y. Wei, H. Xu, D. Chen, T. Ahamad and S. Alshehri, et al., Spatial exciton allocation strategy with reduced energy loss for high-efficiency fluorescent/phosphorescent hybrid white organic light-emitting diodes, Mater. Horiz., 2017, 4(4), 641–648.
23. U. Deori, G. P. Nanda, C. Murawski and P. Rajamalli, A perspective on next-generation hyperfluorescent organic light-emitting diodes, Chem. Sci., 2024, 15(43), 17739–17759.
24. A. J. Sneyd, D. Beljonne and A. Rao, A new frontier in exciton transport: transient delocalization, J. Phys. Chem. Lett., 2022, 13(29), 6820–6830.
25. D. Balzer and I. Kassal, Even a little delocalization produces large kinetic enhancements of charge-separation efficiency in organic photovoltaics, Sci. Adv., 2022, 8(31), eabl9692.
26. H. Nakanotani, T. Furukawa, K. Morimoto and C. Adachi, Long-range coupling of electron-hole pairs in spatially separated organic donor-acceptor layers, Sci. Adv., 2016, 2(2), e1501470.
27. S.-W. Feng, M.-C. Shih, C. Huang and C.-T. Chung, Impacts of dopant concentration on the carrier transport and recombination dynamics in organic light emitting diodes, Thin Solid Films, 2009, 517(8), 2719–2723.
28. Y. Kawamura, J. Brooks, J. J. Brown, H. Sasabe and C. Adachi, Intermolecular Interaction and a Concentration-Quenching Mechanism <? format? > of Phosphorescent Ir(III) Complexes in a Solid Film, Phys. Rev. Lett., 2006, 96(1), 017404.
29. Y.-Y. Noh, C.-L. Lee, J.-J. Kim and K. Yase, Energy transfer and device performance in phosphorescent dye doped polymer light emitting diodes, J. Chem. Phys., 2003, 118(6), 2853–2864.
30. J. W. Kim, S. I. You, N. H. Kim, J.-A. Yoon, K. W. Cheah and F. R. Zhu, et al., Study of sequential dexter energy transfer in high efficient phosphorescent white organic light-emitting diodes with single emissive layer, Sci. Rep., 2014, 4(1), 7009.
31. H. Liao, H.-F. Meng, S. Horng, W. Lee, J. Yang and C. Liu, et al., Triplet exciton energy transfer in polyfluorene doped with heavy metal complexes studied using photoluminescence and photoinduced absorption, Phys. Rev. B: Condens. Matter Mater. Phys., 2006, 74(24), 245211.
32. Y. Tian, H. Wang, Y. Man, N. Zhang, J. Zhang and Y. Li, et al., Weaving host matrices with intermolecular hydrogen bonds for high-efficiency white thermally activated delayed fluorescence, Chem. Sci., 2021, 12(43), 14519–14530.
33. N. Haase, A. Danos, C. Pflumm, P. Stachelek, W. Brütting and A. P. Monkman, Are the rates of dexter transfer in TADF hyperfluorescence systems optically accessible?, Mater. Horiz., 2021, 8(6), 1805–1815.
34. P. Niyonkuru, R. A. Bennett, M. J. Zachman and J. D. Zimmerman, Effect of molecular permanent dipole moment on guest aggregation and exciton quenching in phosphorescent organic light emitting diodes, J. Chem. Phys., 2024, 160(24), 244304.
35. K. S. Yook and J. Y. Lee, Solvent effect on device performances of small molecule based solution processed blue phosphorescent organic light-emitting diodes using aromatic and alcohol solvents, J. Lumin., 2014, 146, 512–514.
36. A. A. Shoustikov, Y. You and M. E. Thompson, Electroluminescence color tuning by dye doping in organic light-emitting diodes, IEEE J. Sel. Top. Quantum Electron., 1998, 4(1), 3–13.
37. L. Si, M. V. Massa, K. Dalnoki-Veress, H. R. Brown and R. A. Jones, Chain entanglement in thin freestanding polymer films, Phys. Rev. Lett., 2005, 94(12), 127801.
38. D. C. Kong, M. H. Yang, X. S. Zhang, Z. C. Du, Q. Fu and X. Q. Gao, et al., Control of polymer properties by entanglement: a review, Macromol. Mater. Eng., 2021, 306(12), 2100536.
39. F. Wang, Z. Jiang, X. Lin, C. Zhang, K. Tanaka and B. Zuo, et al., Suppressed chain entanglement induced by thickness of ultrathin polystyrene films, Macromolecules, 2021, 54(8), 3735–3743.
40. Z. He, Z. Zhang and S. Bi, Tailoring the molecular weight of polymer additives for organic semiconductors, Mater. Adv., 2022, 3(4), 1953–1973.
41. H.-F. Xiang, Z.-X. Xu, V. Roy, C.-M. Che and P. Lai, Method for measurement of the density of thin films of small organic molecules, Rev. Sci. Instrum., 2007, 78(3), 034104.
42. T. Kida, T. Kimura, A. Eno, K. Janchai, M. Yamaguchi and Y. Otsuki, et al., Effect of ultra-high-molecular-weight molecular chains on the morphology, crystallization, and mechanical properties of polypropylene, Polymers, 2021, 13(23), 4222.
43. R. K. Bay, T. Zhang, S. Shimomura, M. Ilton, K. Tanaka and R. A. Riggleman, et al., Decoupling the impact of entanglements and mobility on the failure properties of ultrathin polymer films, Macromolecules, 2022, 55(19), 8505–8513.
44. M. Safandowska, C. Makarewicz, A. Rozanski and R. Idczak, Diminishment the gas permeability of polyethylene by “densification” of the amorphous regions, Sci. Rep., 2023, 13(1), 19838.
45. J. C. Moreira and N. R. Demarquette, Influence of temperature, molecular weight, and molecular weight dispersity on the surface tension of PS, PP, and PE. I. Experimental, J. Appl. Polym. Sci., 2001, 82(8), 1907–1920.
46. O. Brazil, H. Ö. Özer, B. Watts, J. B. Pethica and G. L. Cross, Densification of a polymer glass under high-pressure shear flow, Phys. Rev. B, 2022, 106(6), L060103.
47. L. Si, M. V. Massa, K. Dalnoki-Veress, H. R. Brown and R. A. Jones, Chain entanglement in thin freestanding polymer films, Phys. Rev. Lett., 2005, 94(12), 127801.
48. C. W. Cone, R. R. Cheng, D. E. Makarov and D. A. Vanden Bout, Molecular weight effect on the formation of β phase poly (9, 9′-dioctylfluorene) in dilute solutions, J. Phys. Chem. B, 2011, 115(43), 12380–12385.
49. F. Laquai, A. K. Mishra, M. R. Ribas, A. Petrozza, J. Jacob and L. Akcelrud, et al., Photophysical Properties of a Series of Poly (ladder-type phenylene) s, Adv. Funct. Mater., 2007, 17(16), 3231–3240.
50. S. Hayashi, S.-i Yamamoto and T. Koizumi, Effects of molecular weight on the optical and electrochemical properties of EDOT-based π-conjugated polymers, Sci. Rep., 2017, 7(1), 1078.
51. P. V. D. Scheer, T. V. D. Laar and J. Sprakel, Chain length-dependent luminescence in acceptor-doped conjugated polymers, Sci. Rep., 2019, 9(1), 11217.
52. M. Tong, S. Cho, J. T. Rogers, K. Schmidt, B. B. Hsu and D. Moses, et al., Higher molecular weight leads to improved photoresponsivity, charge transport and interfacial ordering in a narrow bandgap semiconducting polymer, Adv. Funct. Mater., 2010, 20(22), 3959–3965.
53. F. C. Spano, Modeling disorder in polymer aggregates: The optical spectroscopy of regioregular poly (3-hexylthiophene) thin films. The, J. Chem. Phys., 2005, 122(23), 234701.
54. K. Koynov, A. Bahtiar, T. Ahn, R. M. Cordeiro, H.-H. Hörhold and C. Bubeck, Molecular weight dependence of chain orientation and optical constants of thin films of the conjugated polymer MEH-PPV, Macromolecules, 2006, 39(25), 8692–8698.
55. D. Kasama, R. Takata, H. Kajii and Y. Ohmori, Optical property of poly (9, 9-dioctylfluorene) gel with β phase and application to polymer light emitting diode, Thin Solid Films, 2009, 518(2), 559–562.
56. J. C. Faria, A. J. Campbell and M. A. McLachlan, Fluorene copolymer bilayers for emission colour tuning in inverted hybrid light emitting diodes, J. Mater. Chem. C, 2015, 3(19), 4945–4953.
57. Y. Wu, C. Yang, J. Liu, M. Zhang, W. Liu and W. Li, et al., Phosphorescent [3+ 2+ 1] coordinated Ir(III) cyano complexes for achieving efficient phosphors and their application in OLED devices, Chem. Sci., 2021, 12(30), 10165–10178.
58. N. K. Zulkarnaen, N. M. Ali, S. Mustafar, A. Kamari and N. H. A. Saedah, Synthesis and characterisation of cationic iridium(III) complex with phenanthroline-based ancillary ligand, Malays. J. Anal. Sci., 2025, 29, 1–10.
59. V. A. Krylova, P. I. Djurovich, J. W. Aronson, R. Haiges, M. T. Whited and M. E. Thompson, Structural and photophysical studies of phosphorescent three-coordinate copper(I) complexes supported by an N-heterocyclic carbene ligand, Organometallics, 2012, 31(22), 7983–7993.
60. M. A. Kinzhalov, E. V. Grachova and K. V. Luzyanin, Tuning the luminescence of transition metal complexes with acyclic diaminocarbene ligands, Inorg. Chem. Front., 2022, 9(3), 417–439.
61. R. M. Jay, S. Eckert, B. E. Van Kuiken, M. Ochmann, M. Hantschmann and A. A. Cordones, et al., Following metal-to-ligand charge-transfer dynamics with ligand and spin specificity using femtosecond resonant inelastic X-ray scattering at the nitrogen K-edge, J. Phys. Chem. Lett., 2021, 12(28), 6676–6683.
62. K. P. Zanoni, B. K. Kariyazaki, A. Ito, M. K. Brennaman, T. J. Meyer and N. Y. Murakami Iha, Blue-green iridium(III) emitter and comprehensive photophysical elucidation of heteroleptic cyclometalated iridium(III) complexes, Inorg. Chem., 2014, 53(8), 4089–4099.
63. S. Yoon and T. S. Teets, Red to near-infrared phosphorescent Ir(III) complexes with electron-rich chelating ligands, Chem. Commun., 2021, 57(16), 1975–1988.
64. R. Fukumoto, E. Sakuda, K. Omoto, S. Horiuchi, Y. Arikawa and K. Umakoshi, Increased phosphorescence of cyclometalated Ir(iii) complex with a planar triarylborane moiety upon fluoride anion binding, Chem. Lett., 2024, 53(12), upae237.
65. Z. A. Siddique, Y. Yamamoto, T. Ohno and K. Nozaki, Structure-dependent photophysical properties of singlet and triplet metal-to-ligand charge transfer states in copper(I) bis (diimine) compounds, Inorg. Chem., 2003, 42(20), 6366–6378.
66. I. E. Jacobs, E. W. Aasen, J. L. Oliveira, T. N. Fonseca, J. D. Roehling and J. Li, et al., Comparison of solution-mixed and sequentially processed P3HT: F4TCNQ films: effect of doping-induced aggregation on film morphology, J. Mater. Chem. C, 2016, 4(16), 3454–3466.
67. E. J. Juarez-Perez, M. R. Leyden, S. Wang, L. K. Ono, Z. Hawash and Y. Qi, Role of the dopants on the morphological and transport properties of spiro-MeOTAD hole transport layer, Chem. Mater., 2016, 28(16), 5702–5709.
68. J. Lee, W. H. Jung, K. Lee, Y. Bae, M. Choi and J. Oh, et al., Improved performance of all-solution-processed quantum dot light-emitting diodes with TFB/PVK double-hole transport layers using 1, 2-dichloroethane, Nanoscale Adv., 2025, 7(11), 3449–3455.
69. K.-W. Chang, Y.-P. Hsieh, C.-C. Ting, Y.-H. Su and M. Hofmann, Increasing the doping efficiency by surface energy control for ultra-transparent graphene conductors, Sci. Rep., 2017, 7(1), 9052.
70. A. Beena Unni, R. Winkler, D. M. Duarte, K. Chat and K. Adrjanowicz, Influence of surface roughness on the dynamics and crystallization of vapor-deposited thin films, J. Phys. Chem. B, 2022, 126(40), 8072–8079.

---