Open Access
Add to BibSonomy
Abstract
To improve the color conversion performance, we study the nanoscale-cavity effects on the emission efficiency of a colloidal quantum dot (QD) and the Förster resonance energy transfer (FRET) from quantum well (QW) into QD in a GaN porous structure (PS). For this study, we insert green-emitting QD (GQD) and red-emitting QD (RQD) into the fabricated PSs in a GaN template and a blue-emitting QW template, and investigate the behaviors of the photoluminescence (PL) decay times and the intensity ratios of blue, green, and red lights. In the PS samples fabricated on the GaN template, we observe the efficiency enhancements of QD emission and the FRET from GQD into RQD, when compared with the samples of surface QDs, which is attributed to the nanoscale-cavity effect. In the PS samples fabricated on the QW template, the FRET from QW into QD is also enhanced. The enhanced FRET and QD emission efficiencies in a PS result in an improved color conversion performance. Because of the anisotropic PS in the sample surface plane, the polarization dependencies of QD emission and FRET are observed.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Förster resonance energy transfer (FRET) is an effective near-field interaction for enhancing the color conversion from a shorter-wavelength energy donor into a longer-wavelength acceptor [1–5]. Normally, FRET is effective when the distance between donor and acceptor is shorter than a few tens nm [6]. In the application of FRET to a color conversion device, one of the challenging issues is the location for placing the color conversion acceptor, like colloidal quantum dot (QD), such that it is close to the energy donor, like a quantum well (QW) in a light-emitting diode (LED). The simplest approach is to place QDs on the top surface of the device. However, because the top p-type layer of an LED usually requires a thickness larger than 100 nm for maintaining a good electric performance, the distance between the QW and the top-surface QDs becomes too large to achieve an effective FRET process from the QW into QDs [7]. In this situation, placing QDs inside an LED structure is a reasonable method for shortening this distance. For such an implementation, the fabrication of a subsurface GaN porous structure (PS) through electrochemical etching (ECE) is a promising technique. In particular, such a nanoscale pore can induce a nanoscale-cavity effect for enhancing the emission of an inserted QD and the efficiency of FRET in the pore [8,9].
GaN can be wet-etched by acids, such as HNO3, or alkalis, such as KOH. In ECE, we can use the flow path of electric current to control the etching structure for forming subsurface extended pores in GaN. Such essentially lateral pores in GaN can be fabricated by placing the anode contact on an edge face of a GaN sample in an electrochemical system [10–14]. In such an etching process, certain surface holes can be formed on the top surface for the electrolyte to flow into the subsurface layer to be etched and for the gases produced in the chemical reactions to come out from the etched regions. In an ECE process, wet etching can occur in a highly conductive layer, such as a highly Si-doped n-type GaN layer, inside the GaN sample. Therefore, essentially lateral tube-like voids with a certain orientation can be formed inside a GaN sample. GaN and AlGaN PSs have been used for increasing LED crystal quality and light extraction [15–19], lifting a GaN layer or an LED structure from substrate [20–22], forming a distributed Bragg reflector [23,24], relaxing strain [18,19,25–27], embedding QDs in a vertical PS for display application [28], and fabricating lateral and vertical laser diodes [29–31].
A few advantages for improving the color conversion of QDs on an LED can be obtained by inserting QDs into such a subsurface PS. First, compared with those on the top surface, the QDs in the PS can be closer to the QWs of an LED such that the FRET from the QWs into QDs becomes more effective. Second, if we can also insert synthesized metal nanoparticles (NPs) into a subsurface PS, the distances between the metal NPs, QWs, and QDs can be shortened such that their surface plasmon (SP) couplings become stronger [4,5]. Third, with metal NPs and QDs inserted into a subsurface PS, the device top-surface becomes clear for other surface processes to enrich device function. Fourth, if a PS can be fabricated between the top surface and the QWs, which are SP-coupled with the SP resonance on surface metal NPs, the QDs inserted into the PS are located in the strong field region of the SP coupling [32,33]. In this situation, the strong electromagnetic field in this region can lead to the stronger QD absorption and hence a higher-efficiency color conversion process.
In this paper, we demonstrate one more advantageous feature in using such a GaN PS for enhancing the color conversion from a QW structure into the QDs inserted into the PS. With the QDs inside the PS, which consists of many nanoscale cavities, it is predicted that the efficiencies of QD emission and the FRET from nearby QWs into the QDs can be enhanced due to a nanoscale-cavity effect, leading to more effective color conversion. Generally speaking, the nanoscale-cavity effect is also caused by the Purcell effect, like the conventional far-field cavity resonance behavior [34]. However, the nanoscale-cavity effect here results from the near field redistribution in a nanoscale cavity. Its behavior is quite different from that of the conventional far-field cavity resonance [8,9]. Although the nanoscale-cavity effect varies with the cavity geometry, it is not sensitive to the change of cavity geometry because it does not behave like a far-field resonance cavity. In this paper, we insert QDs into the PSs fabricated in a GaN template and a QW template for studying the nanoscale-cavity effects on the enhancement behaviors of QD emission, FRET from a green-emitting QD (GQD) into a red-emitting QD (RQD), and FRET from the QW structure into QDs inside a PS through time-resolved and continuous photoluminescence (PL) measurements. We also show their polarization-dependent behaviors in such an anisotropic nanoscale-cavity structure. In section 2 of this paper, the sample structures under study and their fabrication procedures are described. Time-resolved PL (TRPL) results in the samples fabricated on the GaN template are shown section 3. Then, TRPL results in the samples fabricated on the QW template are reported in section 4. The results of continuous PL measurement in all samples are presented in section 5. Further discussions about the results are given in section 6. Finally, conclusions are drawn in section 7.
2. Sample structures and fabrication procedures
For fabricating GaN PSs, we prepare the GaN and InGaN/GaN QW templates with their structures illustrated in Figs. 1(a1) and 1(b1) (excluding the QDs on their top surfaces), respectively. The templates are grown on double-polished sapphire substrate with metalorganic chemical vapor deposition. As illustrated in Fig. 1(a1) for the GaN template, a 200-nm highly Si-doped GaN (n + -GaN) layer is grown on a ∼3-µm un-doped GaN (u-GaN) layer, and then capped with a ∼50-nm u-GaN layer. The growth temperatures for the n + -GaN and u-GaN layers are the same at 1040 °C. In the n + -GaN layer, the Si doping concentration is ∼2 × 1019 cm-3. As illustrated in Fig. 1(b1) for the QW template, after the growth of the 200-nm n + -GaN layer, a ∼50-nm n-GaN layer (Si doping concentration at ∼5 × 1018 cm-3) is deposited, followed by the 5-period InGaN/GaN QW structure. The growth temperature for the n-GaN layer is also 1040 °C. Those for the InGaN well layers and GaN barrier layers are 695 and 793 °C, respectively. The QW structure is capped by a u-GaN layer of ∼30 nm in thickness. PSs are to be fabricated in the n + -GaN layers. Based on the GaN and QW templates, we prepare four groups of sample for comparing the emission behaviors of inserted QDs. With the GaN (QW) template, two sample groups of GN/XXX-S (QW-XXX-S) and GN/XXX-P (QW/XXX-P) are fabricated. In the former group, no PS is formed and QDs are placed on the top surface of a sample. In the latter group, QDs are inserted into the PSs fabricated through ECE. In each sample group, XXX represents GQD, RQD, or GQD + RQD. Therefore, 12 samples in total are prepared for comparison in this study. In Figs. 1(a1)-1(b2), we schematically illustrate the structures of the samples with XXX = GQD + RQD.
Fig. 1. (a1): Schematic illustration of the structures for samples GN/GQD-S, GN/RQD-S, and GN/GQD + RQD-S with QDs on the top surface of a GaN template containing a 200-nm n + -GaN layer. (a2): Schematic illustration of the structures for samples GN/GQD-P, GN/RQD-P, and GN/GQD + RQD-P with QDs inserted into the PS fabricated in the n + -GaN layer of the GaN template. (b1): Schematic illustration of the structures for samples QW/GQD-S, QW/RQD-S, and QW/GQD + RQD-S with QDs on the top surface of a QW template containing a 200-nm n + -GaN layer and 5 periods of blue-emitting QW. (b2): Schematic illustration of the structures for samples QW/GQD-P, QW/RQD-P, and QW/GQD + RQD-P with QDs inserted into the PS fabricated in the n + -GaN layer of the QW template.
In the following, only a few samples are used to explain the sample structures and fabrication procedures. Figures 2(a1)-2(c1) show the photographs of samples QW-GQD-P, QW-RQD-P, and QW-GQD + RQD-P, respectively, after PS fabrication through ECE. The ECE conditions include 15 V in applied voltage, 5 wt% in HNO3 concentration as the electrolyte, and 330 s in ECE duration. In the ECE process, the sample is electrically connected with an indium ball as the anode and a Pt wire is used as the cathode. The sample is partly immersed in the electrolyte with the electrolyte surface level indicated by the blue dashed lines in Figs. 2(a1)-2(c1). PSs can be formed in the sample portions immersed in the electrolyte, in which the sample color turns into greenish yellow. The black continuous lines in Figs. 2(a1)-2(c1) indicate roughly the locations of sample cutting for cross-sectional scanning electron microscopy (SEM) observation. For polarization-dependent optical measurements, we define the perpendicular (⊥) and parallel (//) polarization directions with respect to the electric current direction, as shown in Fig. 2(a1). It is expected that the formed tube-like pores extend essentially along the current flow direction, i.e., the //-polarization direction.
Fig. 2. (a1)-(c1): Photographs of samples QW/GQD-P, QW/RQD-P, and QW/GQD + RQD-P, respectively, after PS fabrication. The blue dashed lines show the surface levels of the electrolyte in ECE, below which PSs are fabricated in the samples. A black line roughly indicates the cutting location for breaking a sample into two pieces. In part (a1), the designated directions of the //- and ⊥-polarization are shown. (a2)-(c2): Photographs of samples QW/GQD-P, QW/RQD-P, and QW/GQD + RQD-P (the portions below the cutting lines), respectively, after QD insertion but before surface cleaning (SC). (a3)-(c3): Photographs of samples QW/GQD-P, QW/RQD-P, and QW/GQD + RQD-P (the lower pieces), respectively, after SC.
Figure 3(a) shows the plane-view SEM image of sample QW-RQD-P after PS fabrication. Here, surface holes of a few tens nm in size can be observed. Those surface holes are formed during the ECE process, through which electrolyte can flow into the n + -GaN layer for etching and the produced gases (N2 and O2) can come out. Those surface nano-holes do not affect our study on the emission behaviors of surface or inserted QDs. Figure 3(b) shows the cross-sectional SEM image of sample QW-RQD-P after PS fabrication. We can see a PS in the originally n + -GaN layer. The cross-sectional pore size is ∼80 nm. For inserting QDs into a PS, we fabricate a linear array of 300-µm surface hole with depth at ∼450 nm near the edge of the sample close to the cutting line shown in Fig. 2(b1). After a droplet of ethanol solution containing QDs is applied to the sample surface, it can flow into the PS through the micron-scale holes. For spreading the solution laterally in the PS, the sample is spun with the speed at 300 rpm for 30 min. To focus on the emission behaviors of the QDs inserted into the subsurface PS, we sweep the surface QDs with wet cotton swabs to clean the sample surface. The sample top surface is then carefully examined with SEM to make sure that few QD exists on the surface. CdZnSeS/ZnS GQDs and RQDs are purchased from Taiwan Nanocrystals Inc. Hsinchu, Taiwan. They are capped with an amphiphilic polymer, i.e., poly(isobutylene-alt-maleic anhydride), and hence are negatively charged with zeta potentials between -40 and -50 mV [35]. Including the capped amphiphilic polymer, the size of a GQD or RQD is estimated to be in the range of 8-10 nm. When either GQD or RQD is to be inserted into a PS, a 0.01-mL droplet of the ethanol solution of QD (concentration at 10 mg/mL) is applied to the sample surface near the aforementioned surface micron-scale holes. When both GQD and RQD are to be inserted into a PS, a 0.01-mL droplet of the ethanol solution of GQD plus RQD (total concentration at 20 mg/mL) is applied to the sample surface. When QDs are to be applied to a sample surface, the same QD solution volume of the same concentration is drop-casted onto the surface. Figure 3(c) shows the plane-view SEM image of sample QW-RQD-P after QD insertion but before surface cleaning (SC). We can see a thick layer of QD on the sample surface. Figure 3(d) shows the plane-view SEM image of the same sample after SC. By comparing with Fig. 3(a), we can see that the sample surface is as clean as that before QD insertion. Figures 2(a2)-2(c2) show the photographs of samples QW-GQD-P, QW-RQD-P, and QW-GQD + RQD-P, respectively, after QD insertion but before SC. Figures 2(a3)-2(c3) show the photographs of the three samples after SC. Here, with QDs inside the PSs, the sample colors become darker than those in Figs. 2(a1)-2(c1).
Fig. 3. (a): [(b):] Plane-view (Cross-sectional) SEM image of sample QW/RQD-P after PS fabrication. (c): Plane-view SEM image of sample QW/RQD-P after QD insertion but before SC. (d): Plane-view SEM image of sample QW/RQD-P after SC.
Figure 4(a) shows the dark-field transmission electron microscopy (TEM) image of a PS before QD insertion. Here, we can clearly see the PS in the originally n + -GaN layer. Above the PS, the five horizontal bright thick lines correspond to the five QWs. If a QD exits near the upper boundary of the PS layer, its distance from the bottom QW is only ∼50 nm, short enough for effective FRET from the QW into QD. Also, the distance between surface QDs and the top QW is only ∼30 nm, again small enough for effective FRET. Figure 4(b) shows the mapping image of energy-dispersive X-ray spectroscopy (EDX) in the TEM observation of sample QW-RQD-P. Here, the green, orange, pink, and red dots show the distributions of Cd, Zn, S, and Se atoms, respectively. This EDX mapping image confirms the successful insertion of QDs into the PS. Because the coated polymer on the QD surface disturbs the electron diffraction of the QD semiconductor lattice, it is difficult to observe a particle-like image for a QD. The foggy regions in Fig. 4(b) correspond to the distributions of QD.
Fig. 4. (a): Cross-sectional TEM image of sample QW/RQD-P [the portion above the cutting line in Fig. 2(b1)] after PS fabrication. (b): EDX image of sample QW/RQD-P showing the distributions of QD composition elements in the PS.
Time-resolved PL measurement is excited by the second-harmonic (390 nm in wavelength and ∼1.5 mW in power) of a femtosecond Ti:sapphire laser (Coherent, VERDI-8W). The signals are monitored with a photon counter of 2.5 ps in temporal resolution (Acton research Corporation, SpectraPro 2150i). The method for calibrating the decay time of a PL decay profile has been reported in a previous publication [36]. The measurement of continuous PL is excited by an InGaN laser diode of 405 nm in wavelength and 6 mW in output power. The spectroscopic signals are detected by an Ocean Optics spectrometer. A polarization-dependent measurement is implemented by rotating the sample to align the excitation laser polarization with either //- or ⊥-polarization direction.
3. Time-resolved photoluminescence results in the samples fabricated on the GaN template
Figure 5(a) shows the normalized PL decay profiles of green light in samples GN/GQD-S, GN/GQD + RQD-S, GN/GQD-P (//- and ⊥-polarization), and GN/GQD + RQD-P (//- and ⊥-polarization). When QDs are placed on the sample top surface, the FRET from GQD into RQD leads to the increase of green-light decay rate in sample GN/GQD + RQD-S, when compared with sample GN/GQD-S. Then, when GQDs are inserted into the PS in sample GN/GQD-P, in either polarization, the green-light decay rate is significantly increased, when compared with that in sample GN/GQD-S. This decay-rate increase is attributed to the nanoscale-cavity effect, which can enhance the emission efficiency of a QD in a PS. Between the two polarizations, the decay rate of the ⊥-polarization is slightly higher than that of the //-polarization. This is so because the effective cavity size in the ⊥-polarization is smaller than that in the //-polarization. Then, when both GQDs and RQDs are inserted into the PS in sample GN/GQD + RQD-P, the FRET from GQD into RQD makes the green-light decay rate dramatically increased, when compared with that of sample GN/GQD + RQD-S. This dramatic increase implies that the FRET is stronger when QDs are inserted into a PS due to the nanoscale-cavity effect. In Fig. 5(a), one can also see that the decay rate of sample GN/GQD + RQD-P in the ⊥-polarization is slightly higher than that in the //-polarization. Figure 5(b) shows the normalized PL decay profiles of red light in samples GN/RQD-S, GN/GQD + RQD-S, GN/RQD-P (//- and ⊥-polarization), and GN/GQD + RQD-P (//- and ⊥-polarization). Here, although a few decay profiles are close to each other, we can still see that the red-light decay rate becomes lower in sample GN/GQD + RQD-S due to the FRET from GQD into RQD, when compared with sample GN/RQD-S. With RQDs inserted into the PS in sample GN/RQD-P, the red-light decay rate increases due to the nanoscale-cavity effect, when compared with that in sample GN/RQD-S. Again, the decay rate in the ⊥-polarization is slightly higher than that in the //-polarization in sample GN/RQD-P. Then, in sample GN/GQD + RQD-P, due to the strong FRET from GQD into RQD, the red-light decay rate becomes significantly lower. It is noted that the red-light decay rate in the ⊥-polarization is slightly lower than that in the //-polarization. These results confirm that in a PS, the nanoscale-cavity effect enhances the FRET efficiency.
Fig. 5. (a): Normalized green-light PL decay profiles of samples GN/GQD-S, GN/GQD + RQD-S, GN/GQD-P (//- and ⊥-polarization), and GN/GQD + RQD-P (//- and ⊥-polarization). (b): Normalized red-light PL decay profiles of samples GN/RQD-S, GN/GQD + RQD-S, GN/RQD-P (//- and ⊥-polarization), and GN/GQD + RQD-P (//- and ⊥-polarization).
Rows 2-7 of Table 1 show the green- and red-light PL decay times of those samples fabricated on the GaN template. We can see that the green-light decay time decreases from 5.78 ns in sample GN/GQD-S to 5.39 ns in sample GN/GQD + RQD-S due to FRET. Then, with GQDs inside the PS in sample GN/GQD-P, the nanoscale-cavity effect leads to the reduction of green-light decay time to 4.96 (4.88) ns in the //- (⊥-) polarization. In sample GN/GQD + RQD-P, the green-light decay time is further reduced to 3.52 (3.43) ns in the //- (⊥-) polarization due to the stronger FRET inside the PS. On the other hand, with FRET on the sample surface, the red-light decay time increases from 8.95 ns in sample GN/RQD-S to 9.35 ns in sample GN/GQD + RQD-S. When compared with that in sample GN/RQD-S, the nanoscale-cavity effect makes the red-light decay time in sample GN/RQD-P reduced to 8.46 (8.34) ns in the //- (⊥-) polarization. Then, through the stronger FRET inside the PS, the red-light decay time of sample GN/GQD + RQD-P is increased to 10.88 (10.92) ns in the //- (⊥-) polarization. The numbers inside the parentheses in Table 1 show the FRET efficiencies in the two GQD + RQD samples. The FRET efficiency, η, is defined as η = 1- τDA/τD [37]. Here, τDA (τD) is the PL decay time of the energy donor when the acceptor is present (absent). The green-light decay times in samples GN/GQD + RQD-S and GN/GQD + RQD-P correspond to the τDA values in the FRETs from GQD into RQD. The PL decay times in samples GN/GQD-S and GN/GQD-P can be used as the corresponding τD values. From Table 1, we can see that the FRET efficiency is increased from 6.75% in sample GN/GQD + RQD-S to 29.03 (29.71) % in the //- (⊥-) polarization of sample GN/GQD + RQD-P. Such results clearly show that the nanoscale-cavity effect can enhance the FRET efficiency. The FRET efficiency in the ⊥-polarization is indeed higher than that in the //-polarization.
Table 1. PL decay times of the green and red lights in the samples under study, including those in the //- and ⊥- polarization for the samples with PSs. The numbers inside the parentheses show the corresponding FRET efficiencies
4. Time-resolved photoluminescence results in the samples fabricated on the quantum-well template
Figure 6 shows the normalized PL spectra of the blue-emitting QW structure in sample QW/RQD-P before and after PS fabrication at 10 and 300 K, normalized with respect to the peak intensities of individual spectra at 10 K. The internal quantum efficiency (IQE) of a QW structure is defined as the ratio of the integrated intensity at 300 K over that at 10 K. The IQEs of this QW structure before (B) and after (A) PS fabrication are 44.7 and 52.3%, respectively. Similar measurements are undertaken for all other samples fabricated with the QW template. The QW IQEs in those samples are shown inside the curly brackets in rows 2 and 3 of Table 2. Here, we can see that after PS fabrication, the QW IQEs are all increased by 8-12%. Such an IQE increase is caused by the strain relaxation in the QW structure after PS fabrication [18,19,25–27], leading to the reduction of the quantum-confined Stark effect (QCSE) and hence the enhancement of QW emission efficiency. With the reduced QCSE, we may expect the blue shifts of the PL spectral peaks after PS fabrication in Fig. 6. However, no clear blue shift is observed. It is speculated that the PS causes strain relaxation in a QW, leading to a larger lattice constant in the InGaN well layer and hence a smaller bandgap, which results in a red shift in QW emission for compensating the blue shift due to the weaker QCSE. It is noted that the intrinsic IQEs of those samples vary slightly from sample to sample. However, such variations do not affect the conclusions of this study. It is also noted that no QW IQE result is available after QD application. This is so because with QDs in the sample, the measured QW PL intensities are modified by QD absorptions, which can be different at 10 and 300 K, such that the integrated intensity ratio becomes unreliable for IQE evaluation. In this situation, we use the data of PL decay time for understanding the QW emission behavior of a sample.
Fig. 6. Normalized continuous blue-light PL spectra of sample QW/RQD-P at 10 and 300 K before and after PS fabrication. The IQE increases from 44.7% before (B) PS fabrication to 52.3% after (A) PS fabrication.
Table 2. PL decay times of the blue light in those samples fabricated with the QW template at different sample fabrication stages. The numbers inside the parentheses show the corresponding FRET efficiencies. The numbers inside the curly brackets show the corresponding IQE values
In Fig. 7(a), we show the PL decay profiles of the QW structures in the samples with surface QDs before (intrinsic) and after QD application. We can see that after QD application, the QW PL decay rates are all increased due to the FRET from the QWs into surface QDs. Figure 7(b) shows the PL decay profiles of the QW structure in sample QW/GQD + RQD-P at different fabrication stages. We can see that after PS fabrication, the QW PL decay rate is increased that is consistent with the aforementioned IQE increase after PS fabrication. Then, after QD insertion and SC, the QW PL decay rate is dramatically increased in either polarization, indicating the strong FRET from the QWs into the QDs inserted into the PS. In particular, because both GQD and RQD of the same concentration (10 mg/mL) are inserted into the PS in this sample, the FRETs from the QW structure into both GQD and RQD make the QW PL decay rate significantly increased. It is noted that the QW PL decay rate in the //-polarization is higher than that in the ⊥-polarization. The QW PL decay behaviors at different fabrication stages in other PS samples based on the QW template are similar to those shown in Fig. 7(b). Without demonstrating their decay profiles, we will discuss their decay time variations as shown in Table 2.
Fig. 7. (a): Normalized blue-light PL decay profiles of samples QW/GQD-S, QW/RQD-S, and QW/GQD + RQD-S before (intrinsic) and after QD application. (b): Normalized blue-light PL decay profiles of sample QW/GQD + RQD-P before (intrinsic) and after PS fabrication, and after QD insertion and SC (//- and ⊥-polarization).
Figure 8(a) shows the green-light PL decay profiles of samples QW-GQD-S, QW-GQD + RQD-S, QW-GQD-P (//- and ⊥- polarization), and QW-GQD + RQD-P (//- and ⊥- polarization). The variation trends among the samples in Fig. 8(a) are similar to those of the corresponding samples in Fig. 5(a). However, because of the FRET from the QW structure into QDs, either on the sample top surface or in the PS, the relative decay rates among those samples in Fig. 8(a) are different. In either sample QW-GQD-P or QW-GQD + RQD-P, the decay rate in the ⊥-polarization is slightly higher than that in the //-polarization. Figure 8(b) shows the red-light PL decay profiles of samples QW-RQD-S, QW-GQD + RQD-S, QW-RQD-P (//- and ⊥- polarization), and QW-GQD + RQD-P (//- and ⊥- polarization). Again, the variation trends among those samples in Fig. 8(b) are similar to those of the corresponding samples in Fig. 5(b). In rows 8-13 of Table 1, we show the green- and red-light PL decay times of those samples based on the QW template. The longer green-light PL decay time in sample QW-GQD-S (6.48 ns), when compared with that in sample GN/GQD-S (5.78 ns), is due to the FRET from the QW structure into the GQDs on the sample surface. Based on the same reason, the red-light PL decay time in sample QW-RQD-S (9.58 ns) is longer than that in sample GN/RQD-S (8.95 ns). Then, in sample QW-GQD + RQD-S, besides the FRETs from the QW structure into the GQDs and RQDs on the top surface, the FRET from GQD into RQD makes the green-light decay time (4.46 ns) shorter and red-light decay time (10.42 ns) even longer. We can see that in either polarization, the green-light decay time in sample QW-GQD-P is shorter than that in sample QW-GQD-S although it is longer than those in samples GN/GQD-S and GN/GQD-P. In sample QW/GQD-P, the GQD in the PS receives energy from the QW structure through FRET for elongating its decay time. However, the nanoscale-cavity effect leads to a higher emission efficiency and hence a relatively shorter PL decay time. The behavior of the red-light decay time in sample QW/RQD-P, in comparison with samples QW/RQD-S, GN/RQD-S, and GN/RQD-P, is similar to that of the green-light decay time in sample QW/GQD-P. Next, by comparing sample QW/GQD + RQD-P with sample GN/GQD + RQD-P, we can see that the green- and red-light decay times in sample QW/GQD + RQD-P are relatively longer due to the FRET from the QW structure into the QDs inserted into the PS. It is noted that in sample QW/GQD-P, QW/RQD-P, or QW/GQD + RQD-P, the green- or red-light decay time in the ⊥- polarization is always shorter than that in the //-polarization. Such a variation trend is also true in sample GN/GQD-P, GN/RQD-P, or GN/GQD + RQD-P, except the red-light decay time in sample GN/GQD + RQD-P. In this sample, the red-light decay time in the ⊥- polarization is longer than that in the //-polarization. Although the decay time difference is small, the result implies certain physical insights. Inside the PS of sample GN/GQD + RQD-P, the FRET from GQD into RQD can be more enhanced in the ⊥- polarization such that its green- (red-) light decay time is relatively shorter (longer). In sample QW/GQD + RQD-P, although the FRET from GQD into RQD is effective, the FRETs from the QW structure into GQD and RQD also play important roles. From the results of samples QW/GQD-P and QW/RQD-P, we may conclude that the FRET from the QW structure into either GQD or RQD in the PS is relatively weaker in the ⊥- polarization. Therefore, the red-light decay time in the ⊥- polarization becomes shorter than that in the //-polarization in sample QW/GQD + RQD-P. However, the results of QW PL decay time, as shown in rows 5 and 6 of Table 2, tell us a more complicated story. Here, in sample QW/GQD-P, the blue-light decay time in the //-polarization is relatively longer and hence the FRET efficiency is lower, when compared with those in the ⊥- polarization. Nevertheless, the variation trends in samples QW/RQD-P and QW/GQD + RQD-P are opposite. It is noted that in an FRET process, the donor (QW) emission decay time or FRET efficiency describes the energy loss rate of the donor. If no energy turns into heat during the FRET process, it corresponds to the energy receiving rate of the acceptor(s) (QDs). However, the green- and red-light decay times in Table 1 are related to not only the energy receiving rates of the QDs, but also their emission efficiencies. Therefore, the variation behaviors of the decay times in Table 1 can be different from those in Table 2. In this regard, the nanoscale-cavity effect plays a crucial role. In sample QW/GQD + RQD-P, the FRET from GQD into RQD makes the energy transfer scenarios even more complicated. More discussions on this issue will be made in section 6.
Fig. 8. (a): Normalized green-light PL decay profiles of samples QW/GQD-S, QW/GQD + RQD-S, QW/GQD-P (//- and ⊥-polarization), and QW/GQD + RQD-P (//- and ⊥-polarization). (b): Normalized red-light PL decay profiles of samples QW/RQD-S, QW/GQD + RQD-S, QW/RQD-P (//- and ⊥-polarization), and QW/GQD + RQD-P (//- and ⊥-polarization).
5. Continuous photoluminescence results
Figure 9(a) shows the normalized continuous-PL spectra of samples GN/GQD + RQD-S and GN/GQD + RQD-P in the two polarizations. The spectrum of sample GN/GQD + RQD-S is normalized with respect to its green peak intensity. The spectra in the two polarizations of sample GN/GQD + RQD-P are normalized with respect to its green peak intensity in the ⊥- polarization. We can see that the normalized red intensities of sample GN/GQD + RQD-P are significantly higher than that of GN/GQD + RQD-S, indicating that the FRET from GQD into RQD is stronger in the PS that confirms the nanoscale-cavity effect. One can also see that either green or red intensity of the ⊥- polarization is higher than that of the //-polarization in sample GN/GQD + RQD-P, confirming the stronger nanoscale-cavity effect in the ⊥- polarization. In Fig. 9(a), the relatively weaker green intensity, when compared with red intensity, can have two attributions. First, the FRET transfers energy from GQD into RQD. Second, the emission efficiency of the purchased GQD is lower than that of RQD. This attribution to the QD emission efficiency difference is supported by the generally shorter PL decay time of GQD, as shown in Table 1. In particular, the green-light decay time in sample GN/GQD-S is 5.78 ns, which is significantly shorter than the red-light decay time in sample GN/RQD-S (8.95 ns). The shorter GQD decay time can be due to the weaker quantum confinement in GQD. The difference in emission efficiency between GQD and RQD does not affect the research targets or conclusions in this paper.
Fig. 9. (a): Normalized continuous PL spectra of samples GN/GQD + RQD-S and GN/GQD + RQD-P (//- and ⊥-polarization). (b): Normalized continuous PL spectra of samples QW/GQD-P, QW/RQD-P, and QW/GQD + RQD-P, all with the //- and ⊥-polarization.
Figure 9(b) shows the normalized continuous-PL spectra of samples QW/GQD-P, QW/RQD-P, and QW/GQD + RQD-P in the two polarizations. In each sample, the spectra are normalized with respect to the blue peak intensity in the ⊥- polarization. Here, we can see that the blue-light peak wavelength varies from sample to sample. This behavior is caused by the non-uniformity of QW emission wavelength over the 2-inch wafer. Such a QW emission-wavelength variation does not affect the conclusions in this paper. In Fig. 9(b), we can also see the relatively weaker GQD emission, when compared with RQD emission. In each sample, for either green or red light, the intensity in the ⊥- polarization is always higher than that in the //-polarization. However, for blue light, that in the //-polarization is relatively higher. Table 3 shows the color ratios and polarization ratios (PRs) of integrated intensity in continuous PL measurement for those samples with color conversion. The color ratios include the ratios of green over blue (G/B), red over blue (R/B), and red over green (R/G). The PR of a light color is defined as the integrated intensity in the ⊥- polarization over that in the //-polarization. The wavelengths used for separating the three color components in spectrum are 510 and 580 nm. In Table 3, we first notice that in either polarization, the R/G ratio in sample GN/GQD + RQD-P is higher than that in sample GN/GQD + RQD-S, again confirming that the FRET from GQD into RQD is stronger in a PS. Also, the higher R/G ratio in the ⊥- polarization confirms the stronger FRET in this polarization. Among samples QW/GQD-S, QW/RQD-S, and QW/GQD + RQD-S, the FRET from GQD into RQD reduces not only the G/B ratio but also the R/B ratio in sample QW/GQD + RQD-S. This result implies that the cascading FRET from QW into GQD and then into RQD does not necessarily lead to a higher color conversion efficiency from blue into red. Next, among samples QW/GQD-P, QW/RQD-P, QW/GQD + RQD-P, we can see that all the G/B and R/B ratios in the ⊥- polarization are higher than the corresponding values in the //-polarization. It is noted that we cannot compare the color ratios between the samples with surface QDs and those with QDs inserted into PSs because the QD density on the surface is significantly higher than that inside a PS. For green and red lights, we can see that the PRs in the samples with the QW structure are generally higher than the corresponding values in the samples based on the GaN template. This variation trend indicates that the color conversion from QW into QD is also polarization dependent. In this regard, the larger-than-unity green- and red-light PRs correspond to the smaller-than-unity blue-light PRs, as shown in Table 3.
Table 3. Color ratios (G/B, R/B, and R/G) and polarization ratios (PRs) in the samples under study
6. Discussions
Basically, three factors control the efficiency of a color conversion process, including the field intensity produced by the energy donor at the position of the acceptor, the absorption and emission efficiencies of the acceptor. In other words, we are concerned about the amount of transferred power from the donor into the emission of the acceptor. However, the definition of the FRET efficiency is only related to the decay times of donor emission (with the presence and absence of acceptor) [37]. Therefore, the FRET efficiency does not include the factor of acceptor emission efficiency. Although a shorter decay time of donor emission at the presence of acceptor or a higher FRET efficiency implies more energy absorbed by the acceptor, the decay time of acceptor emission is not necessarily elongated. The decay time of acceptor emission is still affected by its emission efficiency, which can be influenced by its surrounding structure. If the acceptor absorption efficiency is not influenced by the surrounding structure, a nanoscale cavity, like a PS in this study, can affect an FRET process through the changes of the field intensity produced by the donor at the position of the acceptor. In this situation, even though the FRET efficiency is enhanced to supply the acceptor with more energy in a PS, the nanoscale-cavity effect can increase the acceptor emission efficiency and hence reduce its decay time. For instance, as shown in Table 2 for sample QW/GQD-P, the FRET efficiency in the ⊥- polarization is higher, when compared with the //-polarization. In other words, the energy transfer from the QWs into the GQDs in the PS is more effective such that the GQD emission decay time can be more elongated in the ⊥- polarization. However, as shown in row 11 of Table 1 for sample QW/GQD-P, the green-light decay time in the ⊥- polarization (6.01 ns) is shorter than that of the //-polarization (6.11 ns).
The reduction of QD emission decay time caused by the enhanced QD emission efficiency in a nanoscale cavity can also be confirmed by the continuous PL intensity results, as shown in Table 3. Here, all the color ratios in the ⊥- polarization are higher than those in the //-polarization except the R/G ratio in sample QW/GQD + RQD-P, in which the FRET from GQD into RQD is also involved and hence it is difficult to interpret the result based on a simple model. In Table 3, we can also see that all the green- and red-light PRs are larger than unity while all the blue-light PRs are consistently lower than unity. It is noted that the continuous PL observations include not only the effects of near-field interactions (FRET and nanoscale-cavity effect), but also the contributions of far-field absorption-reemission processes. Those far-field processes cannot be monitored through the variation of PL decay time. However, they are included in the continuous-PL measurement, as shown in the data of Figs. 9(a) and 9(b), and Table 3. However, it is difficult to differentiate the contributions from the near-field and far-field processes.
It is noted that among different samples under study, it is difficult to control the QD density either on the top surface or in a PS. Therefore, in this paper, we can only show the results weakly dependent on QD density, including PL decay time, color ratio, and polarization ratio, from which we can observe the effects of nanoscale cavity on QD emission, the FRET from GQD into RQD, and the FRET from QW into QD. Regarding the nanoscale-cavity effect on QD emission efficiency, we can only show the reduction of the PL decay time for the QDs inserted into a PS. No result of QD emission intensity enhancement can be obtained due to the difficulties of systematic measurement and QD density control. However, because we do not produce any new non-radiative recombination channel in the process of inserting QDs into a PS, we believe that the reduction of QD emission decay time is caused by the enhancement of its radiative recombination and hence the increase of its emission efficiency.
Although the difference in PL decay time or PL intensity between the two polarizations is small, their systematic variations provide us with the confidence about the physical mechanism we propose for interpreting the polarization-dependent behaviors. The weak polarization dependence is due to the slant extension of a tube-like pore in a PS, as illustrated in the TEM images of Figs. 10(a) and 10(b). These images are obtained after the ECE process of an epitaxial sample with its structure the same as that illustrated in Fig. 1(a1) except that the n + -GaN layer is thicker (300 nm). The planes of the TEM images in Figs. 10(a) and 10(b) contain the directions parallel with and perpendicular to the applied current direction, respectively, as shown in Fig. 2(a1). The directions of the //- and ⊥- polarization are shown in Figs. 10(a) and 10(b). In Fig. 10(a), we can see that the tube-like pores start from a surface nano-hole [see Figs 3(a) and 3(d)] and extend in both lateral and depth directions to form a pattern like mountain ridges. Therefore, the size of a pore projected along the ⊥- polarization is smaller than that along the //-polarization. The smaller pore size along the ⊥- polarization leads to a stronger nanoscale-cavity effect. However, because of the slant pore extension, the difference in pore size between the two polarizations is small, resulting in the weak polarization dependence in emission behavior. When the u-GaN capping layer above the n + -GaN layer is not very thick, like the case in this study, before subsurface etching in an ECE process, nanoscale surface holes are formed through chemically etching (by the ECE electrolyte) the weak GaN structures, such as V-shaped pits or threading dislocations. Such surface nano-holes guide the electrolyte to the n + -GaN layer for subsurface ECE at a location where injected current can reach. The surface nano-holes can also serve as the outlet of the gases (H2 and N2) produced in the ECE reaction.
Fig. 10. (a): Cross-sectional TEM image of a PS sample with an n + -GaN layer of 300 nm in thickness. The image plane contains the designated //-polarization direction. (b): Cross-sectional TEM image of the same PS sample with its image plane containing the designated ⊥-polarization direction.
7. Conclusions
In summary, by inserting GQD and RQD into the fabricated PSs in the GaN and QW templates, we have studied the behaviors of the PL decay times and intensity ratios of blue, green, and red lights. In the samples based on the GaN template, we observed the enhancements of FRET efficiency when QDs were inserted into the PSs, showing that the nanoscale-cavity effect could increase FRET efficiency. From the observed reductions of QD PL decay time, it was believed that the nanoscale-cavity effect could also increase the emission efficiency of a QD inserted into a PS. In the samples based on the QW template, the FRET from a QW into a QD was also enhanced when the QD was inserted into a PS. Due to the anisotropic PS in the sample surface plane, the polarization dependent behaviors of QD emission and FRET were observed. The enhancements of QD emission and FRET with QDs in a PS are useful for improving the color conversion efficiency from QW into QD for display application. Finally, the efficiencies of FRET and acceptor emission can be further enhanced by introducing SP coupling to a system with QW and PS. The SP coupling can be induced through the insertion of synthesized metal NPs into the PS or depositing metal NPs onto the device surface.
Funding
Ministry of Science and Technology, Taiwan (MOST 109-2221-E-002-194, MOST 110-2221-E-002-131, MOST 111-2221-E-002-073).
Disclosures
The authors declare that there are no conflicts of interest related to this article.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. S. Chanyawadee, P. G. Lagoudakis, R. T. Harley, M. D. B. Charlton, D. V. Talapin, H. W. Huang, and C. H. Lin, “Increased color-conversion efficiency in hybrid light-emitting diodes utilizing non-radiative energy transfer,” Adv. Mater. 22(5), 602–606 (2010). [CrossRef]
2. H. V. Demira, S. Nizamoglub, T. Erdemb, E. Mutluguna, N. Gaponikc, and A. Eychmuller, “Quantum dot integrated LEDs using photonic and excitonic color conversion,” Nano Today 6(6), 632–647 (2011). [CrossRef]
3. F. Zhang, J. Liu, G. You, C. Zhang, S. E. Mohney, M. J. Park, J. S. Kwak, Y. Wang, D. D. Koleske, and J. Xu, “Nonradiative energy transfer between colloidal quantum dot-phosphors and nanopillar nitride LEDs,” Opt. Express 20(S2), A333–A339 (2012). [CrossRef]
4. C. C. Ni, S. Y. Kuo, Z. H. Li, S. H. Wu, R. N. Wu, C. Y. Chen, and C. C. Yang, “Förster resonance energy transfer in surface plasmon coupled color conversion processes of colloidal quantum dots,” Opt. Express 29(3), 4067–4081 (2021). [CrossRef]
5. Y. P. Chen, C. C. Ni, R. N. Wu, S. Y. Kuo, Y. C. Su, Y. Y. Huang, J. W. Chen, Y. C. Hsu, S. H. Wu, C. Y. Chen, P. H. Wu, Y. W. Kiang, and C. C. Yang, “Combined effects of surface plasmon coupling and Förster resonance energy transfer on the light color conversion behaviors of colloidal quantum dots on an InGaN/GaN quantum-well nanodisk structure,” Nanotechnology 32(13), 135206 (2021). [CrossRef]
6. T. Förster, “Energy transport and fluorescence,” Naturwissenschaften (1913-2014) 33(6), 166–175 (1946). [CrossRef]
7. C. Y. Su, C. H. Lin, Y. F. Yao, W. H. Liu, M. Y. Su, H. C. Chiang, M. C. Tsai, C. G. Tu, H. T. Chen, Y. W. Kiang, and C. C. Yang, “Dependencies of surface plasmon coupling effects on the p-GaN thickness of a thin-p-type light-emitting diode,” Opt. Express 25(18), 21526–21536 (2017). [CrossRef]
8. Y. Kuo, Y. J. Lu, C. Y. Shih, and C. C. Yang, “Simulation study on the enhancement of resonance energy transfer through surface plasmon coupling in a GaN porous structure,” Opt. Express 29(26), 43182–43192 (2021). [CrossRef]
9. Y. Y. Huang, Z. H. Li, Y. C. Lai, J. C. Chen, S. H. Wu, S. Yang, Y. Kuo, C. C. Yang, T. C. Hsu, and C. L. Lee, “Nanoscale-cavity enhancement of color conversion with colloidal quantum dots embedded in the surface nano-holes of a blue-emitting light-emitting diode,” Opt. Express 30(17), 31322–31335 (2022). [CrossRef]
10. D. Chen, H. Xiao, and J. Han, “Nanopores in GaN by electrochemical anodization in hydrofluoric acid formation and mechanism,” J. Appl. Phys. 112(6), 064303 (2012). [CrossRef]
11. P. H. Griffin and R. A. Oliver, “Porous nitride semiconductors reviewed,” J. Phys. D: Appl. Phys. 53(38), 383002 (2020). [CrossRef]
12. M. J. Schwab, D. Chen, J. Han, and L. D. Pfefferle, “Aligned mesopore arrays in GaN by anodic etching and photoelectrochemical surface etching,” J. Phys. Chem. C 117(33), 16890–16895 (2013). [CrossRef]
13. M. J. Schwab, J. Han, and L. D. Pfefferle, “Neutral anodic etching of GaN for vertical or crystallographic alignment,” Appl. Phys. Lett. 106(24), 241603 (2015). [CrossRef]
14. W. J. Tseng, D. H. van Dorp, R. R. Lieten, P. M. Vereecken, and G. Borghs, “Anodic etching of n-GaN epilayer into porous GaN and its photoelectrochemical properties,” J. Phys. Chem. C 118(51), 29492–29498 (2014). [CrossRef]
15. W. J. Hsu, K. T. Chen, W. C. Huang, C. J. Wu, J. J. Dai, S. H. Chen, and C. F. Lin, “InGaN light emitting diodes with a nanopipe layer formed from the GaN epitaxial layer,” Opt. Express 24(11), 11601–11610 (2016). [CrossRef]
16. Y. Li, C. Wang, Y. Zhang, P. Hu, S. Zhang, M. Du, X. Su, Q. Li, and F. Yun, “Analysis of TM/TE mode enhancement and droop reduction by a nanoporous n-AlGaN underlayer in a 290 nm UV-LED,” Photon. Res. 8(6), 806–811 (2020). [CrossRef]
17. C. B. Soh, C. B. Tay, R. J. N. Tan, A. P. Vajpeyi, I. P. Seetoh, K. K. Ansah-Antwi, and S. J. Chua, “Nanopore morphology in porous GaN template and its effect on the LEDs emission,” J. Phys. D: Appl. Phys. 46(36), 365102 (2013). [CrossRef]
18. C. Wurm, H. Collins, N. Hatui, W. Li, S. Pasayat, R. Hamwey, K. Sun, I. Sayed, K. Khan, E. Ahmadi, S. Keller, and U. Mishra, “Demonstration of device-quality 60% relaxed In0.2Ga0.8N on porous GaN pseudo-substrates grown by PAMBE,” J. Appl. Phys. (Melville, NY, U. S.) 131(1), 015701 (2022). [CrossRef]
19. S. S. Pasayat, C. Gupta, M. S. Wong, R. Ley, M. J. Gordon, S. P. DenBaars, S. Nakamura, S. Keller, and U. Mishra, “Demonstration of ultra-small (<10 µm) 632 nm red InGaN micro-LEDs with useful on-wafer external quantum efficiency (>0.2%) for mini-displays,” Appl. Phys. Express 14(1), 011004 (2021). [CrossRef]
20. S. Huang, Y. Zhang, B. Leung, G. Yuan, G. Wang, H. Jiang, Y. Fan, Q. Sun, J. Wang, K. Xu, and J. Han, “Mechanical properties of nanoporous GaN and its application for separation and transfer of GaN thin films,” ACS Appl. Mater. Interfaces 5(21), 11074–11079 (2013). [CrossRef]
21. Y. Zhang, Q. Sun, B. Leung, J. Simon, M. L. Lee, and J. Han, “The fabrication of large-area, free-standing GaN by a novel nanoetching process,” Nanotechnology 22(4), 045603 (2011). [CrossRef]
22. J. H. Kang, M. Ebaid, J. K. Lee, T. Jeong, and S. W. Ryu, “Fabrication of vertical light emitting diode based on thermal deformation of nanoporous GaN and removable mechanical supporter,” ACS Appl. Mater. Interfaces 6(11), 8683–8687 (2014). [CrossRef]
23. C. Zhang, S. H. Park, D. Chen, D. W. Lin, W. Xiong, H. C. Kuo, C. F. Lin, H. Cao, and J. Han, “Mesoporous GaN for photonic engineering highly reflective GaN mirrors as an example,” ACS Photonics 2(7), 980–986 (2015). [CrossRef]
24. C. F. Lin, Y. T. Zhang, C. J. Wang, Y. Y. Chen, G. Y. Shiu, Y. Ke, and J. Han, “InGaN resonant microcavity with n+-porous-GaN/p+-GaN tunneling junction,” IEEE Electron Device Lett. 42(11), 1631–1633 (2021). [CrossRef]
25. A. Najar, M. Gerland, and M. Jouiad, “Porosity-induced relaxation of strains in GaN layers studied by means of microindentation and optical spectroscopy,” J. Appl. Phys. 111(9), 093513 (2012). [CrossRef]
26. S. S. Pasayat, N. Hatui, W. Li, C. Gupta, S. Nakamura, S. P. Denbaars, S. Keller, and U. Mishra, “Method of growing elastically relaxed crack-free AlGaN on GaN as substrates for ultra-wide bandgap devices using porous GaN,” Appl. Phys. Lett. 117(6), 062102 (2020). [CrossRef]
27. S. S. Pasayat, C. Gupta, M. S. Wong, Y. Wang, S. Nakamura, S. P. Denbaars, S. Keller, and U. Mishra, “Growth of strain-relaxed InGaN on micrometersized patterned compliant GaN pseudo-substrates,” Appl. Phys. Lett. 116(11), 111101 (2020). [CrossRef]
28. J. H. Kang, B. Li, T. Zhao, M. Ali Johar, C. C. Lin, Y. H. Fang, W. H. Kuo, K. L. Liang, S. Hu, S. W. Ryu, and J. Han, “RGB arrays for micro-light-emitting diode applications using nanoporous GaN embedded with quantum dots,” ACS Appl. Mater. Interfaces 12(27), 30890–30895 (2020). [CrossRef]
29. R. Anderson, D. Cohen, H. Zhang, E. Trageser, N. Palmquist, S. Nakamura, and S. DenBaars, “Nano-porous GaN cladding and scattering loss in edge emitting laser diodes,” Opt. Express 30(2), 2759–2767 (2022). [CrossRef]
30. Y. Ke, C. J. Wang, G. Y. Shiu, Y. Y. Chen, Y. S. Lin, H. Chen, J. Han, and C. F. Lin, “Polarization properties of InGaN vertical-cavity surface-emitting laser with pipe distributed Bragg reflector,” IEEE Trans. Electron Devices 69(1), 201–204 (2022). [CrossRef]
31. R. T. Elafandy, J. H. Kang, C. Mi, T. K. Kim, J. S. Kwak, and J. Han, “Study and application of birefringent nanoporous GaN in the polarization control of blue vertical-cavity surface-emitting lasers,” ACS Photonics 8(4), 1041–1047 (2021). [CrossRef]
32. Y. Kuo, S. Y. Ting, C. H. Liao, J. J. Huang, C. Y. Chen, C. Hsieh, Y. C. Lu, C. Y. Chen, K. C. Shen, C. F. Lu, D. M. Yeh, J. Y. Wang, W. H. Chuang, Y. W. Kiang, and C. C. Yang, “Surface plasmon coupling with radiating dipole for enhancing the emission efficiency of a light-emitting diode,” Opt. Express 19(S4), A914–A929 (2011). [CrossRef]
33. Y. Kuo, W. Y. Chang, C. H. Lin, C. C. Yang, and Y. W. Kiang, “Evaluating the blue-shift behaviors of the surface plasmon coupling of an embedded light emitter with a surface Ag nanoparticle by adding a dielectric interlayer or coating,” Opt. Express 23(24), 30709–30720 (2015). [CrossRef]
34. E. M. Purcell, “Spontaneous emission probabilities at radio frequencies,” Phys. Rev. 69, 839 (1946). [CrossRef]
35. C. A. J. Lin, R. A. Sperling, J. K. Li, T. A. Yang, P. Y. Li, M. Zanella, W. H. Chang, and W. J. Parak, “Design of an amphiphilic polymer for nanoparticle coating and functionalization,” Small 4(3), 334–341 (2008). [CrossRef]
36. C. Y. Chen, C. C. Ni, R. N. Wu, S. Y. Kuo, C. H. Li, Y. W. Kiang, and C. C. Yang, “Surface plasmon coupling effects on the Förster resonance energy transfer from quantum dot into rhodamine 6 G,” Nanotechnology 32(29), 295202 (2021). [CrossRef]
37. P. Ghenuche, M. Mivelle, J. de Torres, S. B. Moparthi, H. Rigneault, N. F. V. Hulst, M. F. G. Parajó, and J. Wenger, “Matching nanoantenna field confinement to FRET distances enhances Förster energy transfer rates,” Nano Lett. 15(9), 6193–6201 (2015). [CrossRef]
