Charge Injection and Energy Transfer of Surface-Engineered InP/ZnSe/ZnS Quantum Dots

Abstract

Surface passivation is a critical aspect of preventing surface oxidation and improving the emission properties of nanocrystal quantum dots (QDs). Recent studies have demonstrated the critical role of surface ligands in determining the performance of QD-based light-emitting diodes (QD-LEDs). Herein, the underlying mechanism by which the capping ligands of InP/ZnSe/ZnS QDs influence the brightness and lifetime of the QD-LEDs is investigated. The electrochemical results demonstrate that highly luminescent InP/ZnSe/ZnS QDs exhibit modulated charge injection depending on the length of the surface ligand chains: short alkyl chains on the ligands are favorable for charge transport to the QDs. In addition, the correlation between the spectroscopic and XRD analyses suggests that the length of the ligand chain tunes the ligand–ligand coupling strength, thereby controlling the inter-QD energy transfer dynamics. The present findings shed new light on the crucial role of surface ligands for InP/ZnSe/ZnS QD-LED applications.

1. Introduction

Environmentally benign indium phosphide (InP) quantum dots (QDs) have gained significant attention as a promising platform for commercializing quantum dot light-emitting diodes (QD-LEDs) due to their highly efficient, stable, and color-pure photoluminescence (PL) with a near-unity PL quantum yield and narrow-band emission [1,2,3,4,5,6]. However, the retention of these remarkable properties in electroluminescence (EL) remains challenging [7,8,9,10,11]. The PL and EL each involve the recombination of non-equilibrium charge carriers, but they differ in their charge generation processes. The PL occurs when a material is excited by an external light source, thereby generating non-equilibrium charge carriers that can recombine, resulting in the emission of light. However, EL occurs when a voltage is applied to a material, thereby creating a flow of non-equilibrium charge carriers, such as electrons and holes, that release energy in the form of light when they recombine. Consequently, efficient charge injection and transfer within the device are critical to achieving high EL efficiency in optoelectronic devices.

While organic capping ligands are generally used to enhance the colloidal stability and protect the QD surfaces against oxidation, surface ligands have also been reported to affect the performance of QD devices by modulating the charge carrier mobility, charge injection, and balance in the spatial distribution of electrons and holes [8,12,13,14,15,16,17,18]. For instance, the performance of the QD device is determined by the operando electrochemical reactions of the organic ligands, and the conductivity and quantum efficiency can be improved via chlorination of the QD surface, which engineers exciton confinement [17,18]. In addition, colloidal core–shell QDs have been shown to exhibit significantly improved QD-LED performance as the surface ligand chain length is reduced [19]. Conversely, QDs with ligands with long chains exhibit lower charge transfer rates due to the large surface transport barrier and the weakening of the charge coupling between QDs [20,21,22,23]. At the same time, the fluorescence resonance energy transfer (FRET) between the QDs is a major determinant of the QD-LED performance particularly the device lifetime [19,24,25]. Energy transfer between QDs can lead to nonradiative recombination, which decreases the device efficiency and shortens its lifetime. This effect is referred to as energy-transfer-induced quenching and is particularly problematic when the QDs are closely packed together. Therefore, controlling energy transfer processes in QD-LEDs is crucial to optimizing the devices’ performance and enhancing their lifetimes.

Hence, the present study examines the synthesis and electrochemical characterization of InP/ZnSe/ZnS QDs with various surface ligand lengths. Furthermore, the inter-QD electronic energy transfer properties are thoroughly probed via the emerging technique of fluorescence lifetime imaging microscopy (FLIM) by obtaining the PL lifetime and intensity with a scan area of 5 × 5 µm². Since the scanned image has an array of 512 by 512 pixels, the dimensions of each pixel are about 10 × 10 nm². The electrochemical results indicate that decreasing the ligand chain length facilitates charge injection due to a smaller energy barrier towards charge transport, thus highlighting the crucial role of the surface ligand in determining the efficiency of charge injection in the nanoparticles and, in turn, their performance in optoelectronic devices. Moreover, the FLIM results reveal that the energy transfer population and efficiency between neighboring QDs are proportional to the surface ligand chain lengths. In addition, the X-ray diffraction (XRD) patterns of InP/ZnSe/ZnS QD solids prove that the alkyl-chain-length-controlled interligand coupling is a strong driving force for FRET and that shorter alkyl ligands result in weaker interligand van der Waals interactions. Therefore, reducing the surface ligand length of the InP-based QDs improves the performance of the QD-LED by promoting effective charge injection and reducing FRET.

2. Materials and Methods

As shown in Scheme 1 and Figure 1A, QDs surrounded by oleic acid (OA), decanoic acid (DA), or hexanoic acid (HA) moieties were synthesized using the procedure reported by Hahm et al. with minor modifications [26]. The ligand-exchanged InP/ZnSe/ZnS QDs were prepared via post-synthetic methods to eliminate any experimental errors originating from batch-to-batch variations in the QD synthesis. The as-synthesized InP/ZnSe/ZnS QDs from the same batch were then used to prepare OA-InP, DA-InP, and HA-InP. The full details of starting materials, commercial sources, quantities and synthetic conditions used are provided in the Supplementary Materials.

In addition, the monodispersed QD-ordered OA-InP, DA-InP, and HA-InP films were fabricated by using an established procedure [27,28,29]. The details are provided in the Supplementary Materials. The effects of ligand length upon the energy transfer dynamics of the InP/ZnSe/ZnS QDs were further validated by using a lab-built FLIM setup to investigate the FRET phenomenon. In this setup, a pump beam was focused through a 100× oil-immersion objective onto a small spot of the sample, and an image was obtained based on the difference in the PL decay rate and intensity of the emitting sample. For the FLIM measurements, a 470 nm pulsed laser beam with a repetition frequency of 10 MHz was used as the excitation source, and the highly-ordered QD films were photoexcited under the condition of ⟨N_X⟩ >> 1.

Thermogravimetric analysis (TGA) revealed that 31% of the OA was replaced by DA in DA-InP, and 54% of the OA was replaced by HA in HA-InP (Figure 1B). Although previous studies have shown that the ligand chain length can impact the overlap of the electron–hole wavefunction [30,31,32], the three InP/ZnSe/ZnS QDs fabricated herein exhibited similar excitonic peak positions, emission energies, and PL quantum yields (Figure 1C). The successful ligand substitution is further demonstrated by the quantitative analysis of the decay curves of the excitonic bleach signal extracted from the transient absorption spectra (Figure S1), in which a fast decay component of a few tens of picoseconds can be attributed to charge transfer into the QD surface. These decay time constants are seen to increase as the ligand chain length of the QDs is extended.

3. Results and Discussion

3.1. Chronoamperometry

Considering that charge carriers are electrically generated in the QD-LEDs, the intrinsic charge injection properties of the various InP/Znse/ZnS QDs are examined and compared via chronoamperometry, an electrochemical technique that monitors the current from Faradaic processes occurring at the electrode as a function of time. Because electron migration is poorly modulated, and redox reactions proceed unevenly within QD films [33,34], electrochemical charge injection is performed in the QD solution. A schematic diagram of the charging process is shown in Figure 2A; redox reactions occur in the near-electrode region depending on the external potential. First, the QD energy levels obtained from the onset redox potentials are estimated from the cyclic voltammogram in Figure S3, and the measurement results are summarized in Table S1. The evolution of the current density with time in response to an applied potential is then measured via chronoamperometry, and the results are presented in Figure 2B. Next, the effects of the type of surface ligand on the current density near the onset potential (i.e., −1 and +1.5 V) are revealed in Figure 2C,D. A high current density implies effective charge injection into the corresponding QDs, as it is dependent on the concentration of the charged dots. Thus, the effect of the ligand chain length upon charge injection is seen to be more significant when there are fewer than 10 carbon atoms in the capping ligand (Figure 2E). Decreasing the number of carbon atoms in the ligand from 18 to 6 is seen to increase the integrated current density by a factor of 1.5. In addition, a decrease in ligand length is found to increase the current density at high applied voltages (Figure S4), thereby indicating that short ligands improve the charge injection efficiency by reducing the energy barrier toward charge transport in the InP/ZnSe/ZnS QDs.

3.2. Time-Resolved Photoluminescence of Quantum Dot Solids

The performance of the QD-LED is influenced by the interplay between the QDs; hence, it is important to understand the physics of the inter-QD electronic coupling during photoinduced energy conversion, as shown in Figure 3A [35,36,37,38]. Figure 3B−D show the steady-state PL spectra of the OA-InP, DA-InP, and HA-InP, respectively. The single QD emission was recorded with an exposure time of 1 s per frame (see the Supplementary Materials for details on single QD PL microscopy). Zero-dimensional QDs are characterized by the confinement of the electronic wavefunctions in all directions, producing a non-zero electronic density of states only at discrete energies. Thus, narrow PL spectral lines similar to those observed in atomic gases are expected [39]. However, even the spectra of a single QD show significant line broadening due to exciton–phonon coupling, spectral shifts, and exciton fine-structure splitting [40]. The PL spectral linewidth of an ensemble is also affected by inhomogeneous broadening due to particle-to-particle inhomogeneity. In OA-InP, DA-InP, and HA-InP, the slightly increased linewidth of the film compared to the solution suggests an interaction between the QDs. For the QD films, the spectral signature of energy transfer dynamics is revealed by a continuous red-shift of the emission peaks in the two-dimensional (2D) contour maps of the time-resolved PL spectra (Figure 3E−G), accompanied by a decrease in PL intensity [41,42]. Compared to the QD solution and disordered films, a fast-decaying component with a ~2.5 ns time constant appeared in the PL decay kinetics of the QD-ordered films, demonstrating the occurrence of FRET processes (Figures S6 and S7). Additionally, the PL lifetime of each QD array is seen to increase as the emission energy decreases (Figure S8), thereby indicating exciton flow [43]. In particular, the PL lifetime of the OA-InP is seen to increase steeply with the decrease in detection energy, whereas that of the DA-InP and HA-InP varies only modestly with emission energy (Figure S9). Meanwhile, the emission peaks of the OA-InP, DA-InP, and HA-InP arrays exhibit shifts of 30, 24, and 17 meV, respectively, toward low energy (Figure 3H−J). Based on these spectroscopic results, the energy transfer efficiency was expected to vary with ligand length. In general, energy transfer dynamics depend on the donor-to-acceptor separation distance based on the theoretical basis of FRET. However, the transmission electron microscopy (TEM) images reveal similar distances between adjacent QDs of 6.4, 6.3, and 6.2 nm for the OA-InP, DA-InP, and HA-InP arrays, respectively (Figure S5). In addition, the size dispersion is not responsible for the correlation between the FRET kinetics and the ligand chain length given that the QD size distributions were similar regardless of ligand length due to the post-synthetic methods.

3.3. Photoluminescence Lifetime Imaging Microscopic Measurements

To further validate the ligand length dependence of energy transfer dynamics in InP/ZnSe/ZnS QDs, FRET was thoroughly investigated using our lab-built FLIM setup. The FLIM technique provides an image based on the differences in the PL decay rates and intensities from an emitting sample. The 2D digital imaging process is shown schematically in Figure 4A, in which the outcome is an array of pixels defined by the PL decay distribution [44]. Here, each pixel contains lifetime data indicating the number of photons collected at various times after pulsed light excitation. In addition, the pixel-by-pixel PL decay curves provide a sensitive approach to measuring the FRET by quantifying the decrease in the PL lifetime even when the distance between the donor and acceptor QDs is less than 10 nm. The FLIM images of the OA-InP, DA-InP, and HA-InP (scan area: 5.5 × 5.5 µm²) are presented in Figure 4B, in which an increase in the PL lifetime is observed with the decrease in ligand length. Meanwhile, the PL decay traces, which represent the total number of photons collected at various times over the entire pixels, exhibit double-exponential decay with a fast-decaying component (τ₁) reflecting the inter-QD energy transfer processes and a slow-decaying component (τ₂) reflecting charge recombination (Figure 4C). We further analyzed the FLIM images with an array of 512 by 512 pixels by treating each pixel independently. Figure 4D shows the distribution of PL lifetimes for each pixel. The distribution of the energy transfer rates (${\mathsf{\tau }}_{\mathrm{FRET}}$) appears within a time window of 0–5 ns (the orange-shaded areas on the left-hand side of Figure 4D). Given that the FRET efficiency (E_f) is proportional to the FRET rate constant ($k_{\mathrm{FRET}}$) [45], the FRET efficiency can be compared using ${\mathsf{\tau }}_{\mathrm{FRET}}$. The average ${\mathsf{\tau }}_{\mathrm{FRET}}$ is seen to increase with the decrease in ligand chain length from 1.3 for the OA-InP to 1.6 and 2.3 ns for the DA-InP and HA-InP, respectively. These experimental results, along with the calculated FRET efficiencies in Table S2, demonstrate that reducing the length of the capping ligand suppresses FRET processes. In addition, the energy transfer populations could be estimated by A₁, where A₁ and A₂ are the amplitude of the fast- (τ₁) and slow- (τ₂) decaying components. Figure S10 depicts A₁ and A₂ for 512 by 512 pixels, from which we extracted the numerical results of A₁ over A₂ and visualized them with images in Figure 4E. From these results, the average A₁/A₂ ratios of the OA-InP, DA-InP, and HA-InP are found to be 26.0, 3.33, and 1.03, respectively, thereby indicating that the energy transfer population is significantly decreased as the ligand chain length decreases. Taken together, the spatiotemporal spectroscopic results confirm that the FRET efficiency and population are tuned by the ligand chain length of InP/ZnSe/ZnS QDs.

3.4. X-ray Diffraction Spectra and Ligand Ordering

The surface characteristics of the QD samples were investigated using XRD to understand the mechanism by which the ligand length influences inter-QD energy transfer dynamics. The XRD patterns of OA-InP, DA-InP, and HA-InP were experimentally measured and compared with the predicted peaks of zinc-blended InP, represented by blue bars (reference pattern: ISCD 1600543) in Figure 5A−C [46,47,48]. Here, Gaussian peaks were used as an approximation to deconvolute the XRD patterns. In addition to the peaks attributed to InP, an additional unassigned peak was observed around 2$\theta$ = 20°, indicated by the red arrow. Alivisatos et al. recently reported that this XRD peak around 2$\theta$ = 20° is attributed to aliphatic ligands [46].

Interestingly, as the number of carbon atoms in the capping ligand (i.e., the chain length) decrease, the intensity of the ligand peak relative to that of the (111) diffraction peak is seen to decrease, and the ligand peak is seen to shift toward a lower 2$\theta$ value (Figure S11). This indicates an increase in the interligand distance—corresponding to weakening van der Waals interactions—as the chain length increases, as shown schematically in Figure 5D. Taken together, these XRD results, along with the qualitative and quantitative interpretations of the spectroscopic features, reveal that the interligand interactions between neighboring QDs significantly affect the FRET efficiency and populations in the InP/ZnSe/ZnS QD films (Figure 5E). The comprehensive results suggest that ligand-length-controlled interligand coupling is an important factor in controlling the FRET process of the colloidal InP/ZnSe/ZnS QDs, which ultimately influences the performance of the QD-LEDs. Notably, previous research on InP/ZnSe/ZnS QD-LEDs has achieved the overall best performance (i.e., excellent EQE and the best lifetime) via ligand displacement from native OA to HA [19].

4. Conclusions

The present study investigated the effects of surface ligand upon the optoelectronic properties of InP/ZnSe/ZnS quantum dots (QDs) via electrochemical and time- and space-resolved spectroscopic methods. The results indicated that the surface ligands play crucial roles in both the charge injection and energy transfer processes, which are critical factors for the performance of the QD-based light-emitting diode (LED). Shorter ligands were shown to reduce the barrier towards charge injection, thus leading to an enhanced current flow. Additionally, the spatiotemporal spectroscopic results showed that the rate of inter-QD energy transfer decreased with the decrease in ligand length. From the XRD patterns, we suggested that the interligand van der Waals interactions became weaker with the decrease in the chain length, thus indicating that ligand-length-controlled interligand coupling is a key factor influencing the FRET process in the InP/ZnSe/ZnS QDs. These findings emphasize the importance of surface ligand engineering in optimizing the performance and stability of the QD-LEDs. By tuning the FRET suppression and charge transport promotion in the surface-engineered InP-based QDs, QD-LEDs can achieve high performance and prolonged operational stability.