Quantum dots (QDs) present attractive features of precise emission bandwidth, saturated emission, tunable emission wavelengths, and high quality production with low cost solution processing.^[1–5] Such promising features make QDs potential candidates for next-generation display technologies through fully functionalized quantum dot light emitting diodes (QD-LEDs).^[6–10] Since the demonstration of first QD-LED in 1994, continuous efforts have been made to improve their performance and the device is comparable to conventional organic light-emitting diode in some performances.^[2,3,11] In particular, the inverted device is suitable for the display due to the compatibility to the n-channel thin film transistor (TFT). The QD-LEDs still suffer low luminous efficiency and poor stability^[12,13] due to imbalanced transportation and injection of carriers,^[14] intrinsic photoluminescence (PL) quenching,^[15] and weak out-coupling.^[16,17]

Surface plasmons (SPs), which are collectively oscillating free electrons at the interface between metal and dielectric,^[8,18] are being extensively investigated^{[10,14,19,20]} and being preliminarily investigated in QD-LEDs.^[21] Enhancements such as in PL^[8,22] and internal quantum efficiency (IQE)^[19] of electroluminescent (EL) devices have been demonstrated. Furthermore, localized surface plasmons (LSPs) associated with noble metal nanostructures show resonance coupling to excitons, providing the enlargement of local electromagnetic field, which results in a rapid radiative emission rate by effective energy coupling from QDs to LSPs.^[8,23] Metal nanostructures have been previously made^[3] and incorporated by chemical or thermal vapor deposition of thin films with post annealing at high temperature,^[1] by spin coating,^[24] or by patterned templates.^[16] However, most of the methods to incorporate plasmonic metal nanostructures require a complicated process and they are sometimes incompatible with solution processed device applications.

Here in this paper, the direct sputtering of gold nanoparticles (Au NPs) on ITO substrate is proposed for improving the QD-LEDs. The generated Au NPs’s layer creates a plasmonic effect without damaging soft or thin underlying organic films,^[25] and thus contributing to not only increasing the work function of the electron transport layer, but also quickly injecting the electrons into QDs and rapidly coupling excitons to LSPs.

The corresponding SEM micrograph of optimized sputtered substrate after thermal annealing is shown in Figs. 1(a) and Fig. S1 (see Appendix A). It can be clearly seen that spherical Au NPs with an average size of 22 nm (20-s sputtering time) are uniformly distributed over the entire surface of ITO. The changes in the size of as-prepared Au NPs are found to be 12 nm, 13.5 nm, 22 nm, and 35 nm, corresponding to the variations of sputtering time. Figure 1(b) shows the size-dependent extinction characteristics of Au NPs on ITO substrates and the PL spectra of QD film. Size dependent absorption of Au NP film is observed and the absorption of non-annealed continuous sputtered metal film shows no plasmonic resonance absorption peak. The PL emission peak of QD at a wavelength of 518 nm matches perfectly with an absorption peak of 22-nm Au NP film, which indicates highly efficient coupling of excitons to the SPs.^[9]

Fig. 1.

	Figure Option ViewDownloadNew Window
	Fig. 1. (color online) (a) Schematic diagram of sputtering machine and SEM micrograph of Au NP film, (b) plots of absorption versus wavelength of different size Au NPs, and plot of PL intensity versus wavelength of CdSe QDs. The w/o means with and without something.

To incorporate the Au plasmonic layer into the electroluminescent quantum dot light emitting diode device, all-solution processed inverted device structure of ITO/Au NPs/ZnMgO/QDs/TFB/PEDOT:PSS/Al is used as shown in Fig. 2(a), where ZnMgO is used as an electron transport layer and also acts as a dielectric spacer layer between Au NP and QDs, which diminishes non-radiative quenching induced by SPs.^[26–29] Figure 2(b) depicts an energy level diagram, according to which, the PEDOT:PSS and TFB act as hole injecting and hole transport layers, respectively. Figure 2(c) shows the PL spectra of Au/ZnMgO/QDs and ZnMgO/QDs films, and the PL intensity for plasmonic enhanced QD film is enhanced by 1.7 fold as compared with that of pure QD films, signifying the enhanced spontaneous emission through SP coupling. In order to clarify the origin of PL intensity enhancement in Au NPs’ existence, we perform TRPL measurements for QD films coated over pure ZnMgO and Au NP/ZnMgO substrates respectively, the corresponding decay for both films are shown in Fig. 2(d). It has been revealed that excitons couple to SPs more rapidly than spontaneous recombination of excitons.^[30,31] By the incorporation of Au NPs the lifetime of PL should be reduced. Calculations depict the decay times to be 24.82 ns and 23.38 ns for normal quantum dot film and plasmonic quantum dot film, respectively. Reduction in the decay time is related to the increase in exciton coupling, resulting in enhanced quantum dot emission.^[32] This new form of recombination, in which electron–hole pairs (excitons) couple to vibrating electrons on the surface of metal to produce SPs instead of phonons or photons, accelerates the spontaneous radiative emission rate and internal quantum efficiencies.^[19,33]

Fig. 2.

	Figure Option ViewDownloadNew Window
	Fig. 2. (color online) (a) Illustration of device structure, (b) energy level diagram, (c) PL intensities of normal and plasmonic substrates, and (d) PL decay behavior of normal and plasmonic substrate.

Half normal/half plasmonic device is fabricated on the same substrate to diminish the systematic error among different batches of devices, which is shown in Fig. 3(a). A comparison of image between normal and plasmonic devices driven at 5 V is also presented. Figure S2 in Appendix A shows normalized EL and PL emissions at the same peak wavelength of 518 nm. Luminance–voltage, current density–voltage characteristics of the normal and plasmonic devices are displayed in Fig. 3(b). The plasmonic device displays the highest luminance of 20720 cd/cm², i.e., 1.41 folds larger than that of a normal device. A maximum current efficiency of 4.02 cd/A for a plasmonic device is obtained (Fig. 3(c)), which is improved by 1.29 folds from the normal device. The improvement of current efficiency can be ascribed to the internal field enhancement in the presence of Au NPs due to resonance coupling and fast injection of charge carriers, which ultimately leads to more balanced charge injection. As a result, the turn-on voltage of the plasmonic device is reduced from 3 V to 2.8 V.

Fig. 3.

	Figure Option ViewDownloadNew Window
	Fig. 3. (color online) (a) Optical image of half normal plasmonic device and half plasmonic device on the same substrate before being on and after lighting, (b) luminance and current density varying with applied voltage for simple and plasmonic devices, and (c) comparison of current efficiency versus current density between two devices.

The TREL is performed to elaborate the transportation of charge carriers under the effect of Au NPs (Fig. S3 in Appendix A). Transient evolution profiles of normalized electroluminescence (EL) are obtained under a 5-μs pulse with a pulse cycle of 1 ms. Due to low power per pulse, a relatively high applied voltage of 7.5 V is used to yield a measurable luminance of the device. It is observed that EL evolution profiles of normal and plasmonic devices each experience a delay period first, which is followed by a fast rise due to enhanced spontaneous emission and finally overtaken by a slow rise until it reaches to its maximum value. It is clearly seen from the comparison of EL evolution profiles between plasmonic and normal devices that the normal device takes a longer delay and shows a slower initial rise and experiences longer time lag in reaching the maximum EL than the plasmonic device.^[33] A mobility determination technique was used, which is based on a phenomenological numerical model.^[33] From the plots of normalized {[EL(t → ∞)−EL(t)/EL(t → ∞)]} versus time and EL versus time (Fig. 4(a)), delay time t_d is defined as the intersection between the horizontal axis and the first slope of the curve, initial rise time t₁ is the time where the first slope and the second slope of the curve intersect, and t₂ is the time when the EL intensity reaches 95% of its steady value.

Fig. 4.

	Figure Option ViewDownloadNew Window
	Fig. 4. (color online) (a) Plots of simulated normalized EL versus time from a phenomenological model, and (b) UPS spectrum of QD, ZnMgO, and Au NP enhanced ZnMgO film.

We assume that μ_h ≈ μ_e, and taking time measurements t_d, t₁, and t₂ from our experimental measurements we use the equation μ_e = L²/(V − V_bi)(t_d + t₁) to calculate the mobility of electrons,^[33] where V_bi is assumed to be built in the voltage of the device, taken from the difference in work function of ITO between Au NPs and ZnMgO NPs (1.04 V and 0.93 V for normal and plasmonic devices, respectively) and L is the thickness of the device (120 nm).

UPS spectra of ITO/ZnMgO and ITO/Au/ZnMgO are shown in Fig. 4(b). All the spectra are measured using He Iα photons (hυ = 21.22 eV). The lowest occupied molecular orbital (LOMO) energy levels calculated from UPS characteristics (Table S1 in Appendix A) show a 0.11-eV decrement for the ITO/Au/ZnMgO film as compared with that for the ITO/ZnMgO film. Therefore, the charge injection barrier between the ZnMgO and QDs is reduced with Au incorporation and such a reduction consequently contributes to the balanced transportation of carriers across the QDs’ layer. As a result, delay time t_d is shortened by 0.27 μs for a plasmonic device compared with for a normal device, manifesting faster injection of charge carriers from the charge transfer layers into QDs.

Table 1.

Table 1.

Table 1. Characteristic times td, t1, t2, and electron mobility μe. . Normalized EL td/μs t1/μs t2/μs μe/(cm2/V⋅s) W/O Au 0.77 0.84 2.05 1.39 × 10−5 With Au 0.50 0.65 1.65 1.91 × 10−5

Table 1. Characteristic times td, t1, t2, and electron mobility μe. .

The time for the initial rise, t₁ is shortened by 0.19 μs for the plasmonic device. This can be attributed to the efficient resonance coupling of excitons to SPs, which in turn speeds up the spontaneous emission by continuously balancing the flow of carriers in this region. Therefore, the emission process is also accelerated by resonance coupling to produce a brighter emission peak. The comparison of t₂ reveals that the plasmonic device reaches to 95% of its steady EL intensity 0.4 μs faster than the normal device. This consistent behavior of EL evolution holds under our assumption of μ_h ≈ μ_e across the device. Due to this balance in transportation of charge carriers, mobility enhancement from 1.39 × 10⁻⁵ cm²/V⋅s to 1.91 × 10⁻⁵ cm²/V⋅s for Au NPs contained in the device has been achieved. The I–V traces of the devices comply with the Mott–Gurney power law (Fig. S4 in Appendix A) with an Ohmic region at the lower bias and a space charge-limited current (SCLC) regime at the higher bias. For a plasmonic device, the Ohmic region is observed to be shorter than those for the normal devices, which signifies more rapid carrier transportation, efficient resonance coupling enhanced spontaneous emission, and reduced quenching of excitons under the influence of the plasmonic effect. Furthermore, electron-only devices with the structure of ITO/ZnMgO/QDs/Al and ITO/Au NPs/ZnMgO/QDs/Al are fabricated, and the corresponding current density of the latter with Au NPs is higher than that of the former (Fig. S5 in Appendix A), demonstrating an improved charge injection rate from the charge transport layer to QDs. Combined with the TREL results, the reduced time delay and efficient transportation result in the balance and smooth transportation of carriers in plasmonic devices.

In this research, we have elaborated SP-enhanced QD-LEDs using Au NPs. Enhancements in the device performance are attributed to the increased charge injection rates into QDs, balanced transportation of carriers across the device, and the strong resonance coupling of SPs and excitons, verified by TRPL, TREL, UPS, and phenomenological numerical results. Enhancement in carrier mobility from 1.39 × 10⁻⁵ cm²/V⋅s to 1.91 × 10⁻⁵ cm²/V⋅s verifies the improvement in transportation of charge carriers. Based on the Au incorporation, efficient green emission QD-LEDs each with a maximum luminance of 20700 cd/m² and a maximum current efficiency of 4.02 cd/A are achieved. These results demonstrate an effective method to use plasmonic metal nanostructures for achieving super bright and efficient solid-state light emitting devices.