Open Access
Add to BibSonomy
Abstract
The high performance of a light-emitting diode (LED) with the total p-type thickness as small as 38 nm is demonstrated. By increasing the Mg doping concentration in the p-AlGaN electron blocking layer through an Mg pre-flow process, the hole injection efficiency can be significantly enhanced. Based on this technique, the high LED performance can be maintained when the p-type layer thickness is significantly reduced. Then, the surface plasmon coupling effects, including the enhancement of internal quantum efficiency, increase in output intensity, reduction of efficiency droop, and increase of modulation bandwidth, among the thin p-type LED samples of different p-type thicknesses that are compared. These advantageous effects are stronger as the p-type layer becomes thinner. However, the dependencies of these effects on p-type layer thickness are different. With a circular mesa size of 10 μm in radius, through surface plasmon coupling, we achieve the record-high modulation bandwidth of 625.6 MHz among c-plane GaN-based LEDs.
© 2017 Optical Society of America
1. Introduction
Surface plasmon (SP) coupling in a light-emitting diode (LED) can lead to several advantages, including the enhancement of internal quantum efficiency (IQE), the reduction of the efficiency droop effect, and the increase of modulation bandwidth [1–5]. The enhancement of IQE is due to the increase of radiative recombination rate and hence the conversion from non-radiative recombination energy into radiative recombination [6, 7] in a quantum well (QW) under SP coupling. The reduction of the efficiency droop effect and the increase of modulation bandwidth in an SP-coupled LED are caused by the effective reduction of carrier density in a QW [8]. SP coupling relies on the coverage of the near field distribution of an SP mode onto a light emitter or a QW in an LED. Because the near field distribution of an SP mode decays very fast within one hundred nm, the distance between the metal nanostructure for inducing SP resonance and the light emitter must be small enough for effective SP coupling. Efforts have been made to reduce this distance without sacrificing the electrical property of an LED. To maintain a thick p-type layer (>150 nm) for a good current spreading effect, the structures of deep metal grating [4, 9], metal protrusion [5], and embedded metal nanoparticle (NP) [10, 11] have been proposed. However, the fabrications of such structures are complicated and costly. The simplest and most inexpensive metal nanostructure for fabricating an SP-coupled LED is the distribution of surface metal NPs, which can be fabricated through the procedures of metal deposition and thermal annealing [1–3]. Nevertheless, in this situation, we need a method to reduce the p-type thickness of an LED without significantly sacrificing its electrical and optical performances.
In a typical lateral LED structure, the p-type layer consists of a p-AlGaN electron blocking layer (EBL) for blocking injected electrons of high mobility, a p-GaN layer for spreading injected holes, and a thin p + -GaN layer for reducing the contact resistivity with the top transparent conductor or p-electrode. The hole injection efficiency is controlled by the hole tunneling capability through the p-AlGaN EBL and the current spreading effect in the p-GaN layer. Recently, this research team discovered a technique for significantly increasing the hole tunneling efficiency through the EBL by increasing the Mg doping concentration in the p-AlGaN EBL, particularly around the interface between the EBL and the QW structure, through an Mg pre-flow process [12]. The increased hole concentration in the EBL can screen the polarization field in this layer and reduce the potential barrier height of hole. Hence, the hole tunneling efficiency can be significantly enhanced and the overall LED emission efficiency is increased. With the enhanced hole tunneling efficiency, the factor of current spreading in the p-GaN layer becomes less crucial. In this situation, we can reduce the p-GaN thickness without significantly sacrificing the electrical property or the overall LED efficiency.
Although all the SP coupling effects are expected to become stronger when the distance between the metal nanostructure and the LED QWs is reduced, different SP coupling effects may have different variation trends. For instance, an LED performance parameter can be influenced by other factors in reducing the p-type thickness, besides the different SP-coupling strengths. Also, different SP-coupling effects may rely on different LED operation parameters, which are changed under SP coupling. For example, the enhancement of IQE is mainly controlled by the increase of radiative recombination rate through SP coupling. The reduction of efficiency droop is mainly determined by the decrease of the carrier density in the QWs. Then, the modulation bandwidth improvement is mainly caused by the increased rate of carrier density decay in the QWs. Although those LED operation parameters are related, the differences of their variations under SP coupling can lead to different behaviors of SP-coupling effects. These issues deserve investigations for exploring the pathways to further improve the SP-coupling effects.
In this paper, based on the aforementioned technique of enhancing the hole injection efficiency through the increase of Mg doping concentration in the p-AlGaN EBL, we compare the LED performances of three single-QW epitaxial structures with different reduced p-GaN thicknesses. In particular, we compare their SP coupling behaviors to see how the SP coupling effects vary with the distance between surface metal Ag NPs and the QW. The total p-type thickness is reduced from >150 nm in a conventional LED to 38-78 nm in the three LED epitaxial structures under study. It is found that without SP coupling, the LED emission efficiency increases with decreasing p-GaN thickness. With SP coupling, the SP coupling effects, including IQE enhancement, LED output intensity increase, efficiency droop reduction, and modulation bandwidth increase, become stronger with decreasing p-GaN thickness. Except IQE, the SP coupling effects increase more significantly when the p-GaN thickness decreases from 30 to 10 nm, when compared with the condition of p-GaN thickness decrease from 50 to 30 nm. A high modulation bandwidth of 625.6 MHz is achieved in the SP-coupled LED of 10 nm in p-GaN thickness. This is a new high record among c-plane GaN-based LEDs. In section 2 of this paper, the sample structures and their fabrication procedures are described. The performances of the LED samples under study are compared in section 3. Discussions about the LED performances are made in section 4. Finally, conclusions are drawn in section 5.
2. Sample structures and fabrication procedures
Three LED epitaxial structures are prepared with metalorganic chemical vapor deposition on c-plane sapphire substrate. In each epitaxial structure, as schematically shown in Figs. 1(a) and 1(b), after a thin un-doped GaN buffer layer, an n-GaN layer of ~2 µm and a single InGaN/GaN QW structure (emission wavelength around 465 nm) with ~3 nm in well thickness and ~15 nm in lower barrier thickness are deposited. The grown thickness of the upper barrier layer is ~25nm. Before the growth of the p-AlGaN EBL (~20% Al), an Mg pre-flow process is applied by pausing Ga and Al supplies with Mg flow at 220 sccm in Cp2Mg flow rate [12]. After the Mg pre-flow process, the p-AlGaN EBL of ~18 nm in thickness is grown with continuing Mg supply. Then, the Cp2Mg flow rate is increased to 280 sccm for the growth of a p-GaN layer. Finally, a 10-nm p + -GaN layer is deposited with 800 sccm in Cp2Mg flow rate. In LED epitaxial structures A-C, the p-GaN layer thicknesses are 50, 30, and 10 nm, respectively. The substrate temperature during the Mg pre-flow process and the growths of p-AlGaN, p-GaN, and p+-GaN layers is maintained at 970 °C. This temperature is high enough to thermally anneal the InGaN/GaN QW structure for changing its IQE [13]. Because the p-GaN layer thickness varies among the three LED epitaxial structures, their high-temperature growth durations are different. The high-temperature durations of epitaxial structures A-C are 881, 803, and 726 sec, respectively. It is noted that during the Mg pre-flow process, the residual NH3, which provides the nitrogen source for growing GaN or InGaN, is dissolved to produce hydrogen. Hydrogen can back-etch the GaN upper quantum barrier during the Mg pre-flow process [14]. Therefore, the thickness of the upper quantum barrier is estimated to be 20 nm after the growth of an LED epitaxial structure is completed, i.e., GaN of ~5 nm in thickness is back-etched.
Fig. 1 (a) and (b): Schematic demonstrations of the LED structures of the reference samples (A-R, B-R, and C-R) and SP-coupling samples (A-SP, B-SP, and C-SP), respectively.
With each LED epitaxial structure, a reference LED sample (A-R, B-R, or C-R) and an LED sample with surface Ag NPs for inducing SP coupling (A-SP, B-SP, or C-SP) are fabricated. A circular mesa with 10 µm in radius is fabricated for each of all the LED samples. Because of the small mesa size, the p-contact pad is fabricated outside the LED mesa. The p-contact pad is supported by a SiO2 layer, which has a low refractive index (~1.5) for minimizing the parasitic capacitance of the device [1]. To further improve current spreading on the mesa, 80% surface area of the mesa is covered by the p-contact. However, a current spreading layer is still needed below the p-contact for covering the rest 20% surface area of the mesa. For sample A-R, B-R, or C-R, the current spreading layer consists of Ni/Au of 5/5 nm [see Fig. 1(a)]. For sample A-SP, B-SP, or C-SP, a ~10-nm Ga-doped ZnO (GaZnO) layer is grown followed by the deposition of surface Ag NPs and then the coverage of a current spreading layer of Ti/Au (5/5 nm) [see Fig. 1(b)]. The GaZnO layer is deposited with molecule beam epitaxy at 250 °C in substrate temperature [15]. GaZnO is a highly conductive transparent oxide with the refractive index at ~1.8. Here, it also serves as an interlayer for blue-shifting the localized surface plasmon (LSP) resonance wavelength of the Ag NPs atop. The Ag NPs are formed by first depositing an Ag layer of ~2 nm and then thermally annealing at 250 °C for 30 min with ambient nitrogen [1–3]. Figure 2 shows the scanning electron microscopy (SEM) image of the surface Ag NPs on the GaZnO layer. Figure 3 shows the transmission spectra of samples A-SP, B-SP, and C-SP. Here, one can see that with a depression minimum at ~483 nm, the transmission spectra almost coincide with each other. This depression corresponds to the SP resonance of the Ag NPs. The vertical dashed-line indicates the QW emission wavelength around 465 nm, which is located in the range of strong SP resonance. It is noted that the results shown in Fig. 3 are obtained by using the transmission spectra of the epitaxial structures without surface Ag NP as measurement baselines such that the effects of QW absorption cannot be observed in Fig. 3. The LED samples are completed by depositing the p-contact of Ni/Au layers of 20/100 nm in thickness and the n-contact of Ti/Au layers also of 20/100 nm in thickness. In row 2 of Table 1, we list the thicknesses of the p-type layers of various samples. In row 3 of Table 1, the distances between the surface Ag NPs and the QW of samples A-SP, B-SP, and C-SP are shown. Such a distance includes the thicknesses of the upper quantum barrier ~20 nm, p-AlGaN (~18 nm), p-GaN, p + -GaN (~10 nm), and GaZnO layers (~10 nm). It decreases from 108 nm in sample A-SP to 88 nm in sample B-SP, and then to 68 nm in sample C-SP. In row 4 of Table 1, we show the high-temperature (970 °C) durations of various samples, which will affect their IQEs.
Fig. 3 Transmission spectra of samples A-SP, B-SP, and C-SP. The vertical dashed line indicates the QW emission wavelength around 465 nm.
Table 1. Structures and the performances of the LED samples under study.
3. Device performances
Post-thermal annealing of a QW at a temperature higher than its growth temperature with a certain duration can reorganize the indium-rich cluster structure for enhancing carrier localization and hence increase the IQE of the QW [13]. However, if the thermal annealing process is too long, the QW crystal quality can be degraded and the QW IQE can be reduced. The IQEs of the samples under study are shown in row 5 of Table 1. The IQE of a sample is obtained by taking the ratio of the photoluminescence (PL) intensity at 300 K over that at 10 K in temperature-dependent PL measurement. The IQE increases from 26.4% in sample A-R to 28.7% in sample B-R, and then to 31.5% in sample C-R. In other words, in the range between 726 and 881 sec in high-temperature duration, a shorter annealing duration leads to a higher IQE. With surface Ag NPs, the SP coupling with the QW results in a significant enhancement of IQE in each LED epitaxial structure. As shown in row 5 of Table 1, the IQE increases from 26.4% in sample A-R to 38.6% in sample A-SP (a 46% enhancement), from 28.7% in sample B-R to 49.4% in sample B-SP (a 72% enhancement), and from 31.5% in sample C-R to 57.3% in sample C-SP (a 82% enhancement). The enhancement ratios of IQE through SP coupling are shown inside the parentheses in row 5 of Table 1. Figure 4 shows the PL decay profiles of the six samples under study based on time-resolved PL (TRPL) measurement. The TRPL measurement is excited by the second-harmonic of a 780-nm fs Ti:sapphire laser. The PL signal is monitored by a photon counter. A PL decay profile is related to the decay of carrier density in the QW. It is noted that we fit the TRPL profiles in their early portions, i.e., between 0.5 and 2.5 ns, which are quite linear in the semi-log plotting, because certain previous studies have shown that this TRPL portion is closely related to radiative recombination [16]. Among the LED samples without SP coupling, the PL decay rate increases with decreasing p-GaN thickness. This trend is consistent with the variation of IQE among samples A-R, B-R, and C-R. With SP coupling, the PL decay rates become significantly larger. In row 6 of Table 1, we list the fitted PL decay times of the six LED samples. Here, the numbers inside the parentheses show the ratio of the square-root of inverse PL decay time of each SP-coupled LED sample with respect to the corresponding reference sample. With the same device structure, under the same injected carrier density, this ratio shows the enhancement of LED modulation bandwidth [1].
Figure 5 shows the normalized output intensity versus injected current density of each LED sample. Here, all the output intensities are normalized with respect to that of sample A-R at 3.185 kA/cm2 in injected current density. As shown in Fig. 5, the increasing output intensity from sample A-R to C-R is consistent with the variation trend of IQE. With SP coupling, the LED output intensity is significantly increased in each LED epitaxial structure. In row 7 of Table 1, we show the normalized output intensity of each sample at 3.185 kA/cm2 in injected current density. Again, the numbers inside the parentheses show the enhancement ratios after SP coupling is applied. The enhancements through SP coupling are 67, 73, and 91% in LED epitaxial structures of A-C, respectively. Figure 6 shows the relative wall-plug efficiencies as functions of injected current density of the six LED samples. All the efficiency values are normalized with respect to the maximum level of sample C-SP. Here, among samples A-R through C-R, the efficiency droop effect is weaker in a sample with a thinner p-GaN layer. With SP coupling, the efficiency droop effect is significantly reduced. The droop ranges of all the samples at 3.185 kA/cm2 in injected current density are shown in the row 8 of Table 1. The droop range of a sample is defined as the percentage of efficiency reduction from its maximum to the level at 3.185 kA/cm2 in injected current density. The numbers within the parentheses in row 8 of Table 1 show the reduction percentages of the droop ranges through SP coupling in the three LED epitaxial structures. It is noted that the injected current density for the maximum efficiencies in the samples under study (1 kA/cm2) is generally higher than what usually reported in literature. This is so because the mesa size of the used samples is smaller at 10 μm in radius. The smaller device size leads to a weaker heating effect and hence reduces the droop behavior caused by heating. Figure 7 shows the relations between injected current and applied voltage (I-V curves) of the six LED samples under study. Its insert shows the magnified I-V curves in the voltage range between 2 and 4 V. Here, one can see that the turn-on voltage slightly increases as p-GaN thickness is reduced. SP coupling does not significantly change the turn-on voltage. No significant current leakage can be observed up to −10 V in each LED sample. In row 9 of Table 1, we show the differential resistance levels of the six LED samples. Because of the small mesa size of the LED samples under study, the resistance levels are larger than that of a conventional LED of 300 μm x 300 μm in mesa size. With the transparent conductive GaZnO layer in an SP-coupled LED sample, the resistance level is slightly reduced, when compared with the corresponding reference sample.
Fig. 5 Normalized output intensity versus injected current density of each LED sample. All the output intensities are normalized with respect to that of sample A-R at 3.185 kA/cm2 in injected current density.
Fig. 6 Relative wall-plug efficiencies as functions of injected current density of the six LED samples. All the efficiency values are normalized with respect to the maximum level of sample C-SP.
Fig. 7 I-V curves of the six LED samples under study. The insert shows the magnified I-V curves in the voltage range between 2 and 4 V.
Figure 8 shows the variations of modulation bandwidth with injected current density in the six LED samples under study. Without SP coupling, modulation bandwidth increases slightly with decreasing p-GaN thickness. With SP coupling, modulation bandwidth increases significantly, when compared with the corresponding reference sample. With the thinnest p-GaN layer in sample C-SP, its modulation bandwidth can reach a level beyond 600 MHz when injected current density is high enough. In the bottom row of Table 1, we show the modulation bandwidths of the six LED samples at 9.555 kA/cm2 in injected current density. In sample C-SP, we can reach the modulation bandwidth of 625.6 MHz, which is believed to be the highest ever reported in a c-plane GaN-based surface-emitting LED (~100 MHz higher than our previous record of 528.8 MHz [1]). In the bottom row of Table 1, the numbers inside the parentheses show the ratios of modulation bandwidths in SP-coupled LEDs with respect to those of the corresponding reference samples. These ratios are very close to the numbers inside the parentheses in row 6 of Table 1 for individual SP-coupled LED samples.
Fig. 8 Variations of modulation bandwidth with injected current density in the six LED samples under study.
4. Discussions
By reducing the distance between the surface metal nanostructure and the top QW of an LED, its SP coupling effect can be enhanced. Although the thicknesses of the p-AlGaN EBL and p+-GaN layer can be somewhat reduced, they cannot be completely removed from an LED structure. To significantly reduce the p-type thickness for decreasing the aforementioned distance and hence enhancing SP coupling effect, we can only decrease the thickness of the p-GaN layer, which helps in current spreading in an LED. In a conventional LED, the decrease of p-GaN thickness usually leads to the increases of turn-on voltage and differential resistance. Based on the technique of increasing the hole concentration in the p-AlGaN EBL for screening the polarization field and hence reducing the hole potential barrier in this layer such that the hole tunneling or injection efficiency can be significantly enhanced, the turn-on voltage can be reduced and the overall emission efficiency of an LED is increased. In this situation, although a decrease of the p-GaN layer may still have the effects of increasing turn-on voltage and differential resistance, the significant enhancement of hole injection efficiency can compensate the performance degradation due to the decrease of p-GaN thickness. Hence, by reducing the p-GaN thickness from 50 nm in sample A-R to 10 nm in sample C-R, the wall-plug efficiency of LED is not reduced. As shown in Figs. 5 and 6, with the Mg pre-flow process, the LED emission efficiency increases with decreasing p-GaN thickness. Also, the efficiency droop effect is weaker as the p-GaN layer becomes thinner. The higher emission efficiency and weaker droop effect of an LED with a thinner p-GaN layer are attributed to the higher IQE, which is due to the shorter high-temperature annealing duration.
The reduced p-GaN thickness leads to the decreased distance between the surface Ag NPs and the QW in a used epitaxial structure. By reducing this distance from 108 nm in sample A-SP, to 88 nm in sample B-SP, and then to 68 nm in sample C-SP, the IQE enhancement percentage increases from 46, to 72, and then 82%. Also, the enhancement percentage of LED output intensity at 3.185 kA/cm2 in injected current density increases from 67, to 73, and then 91%. Meanwhile, the decrease percentage of droop range is increased from 41, to 42, and then 51%. In addition, the enhancement percentage of modulation bandwidth is increased from 48, to 52, and then 62%. According to Fermi’s golden rule, the emission efficiency of a light emitter is proportional to the square of the SP field strength if the dipole orientation and other conditions are fixed [17]. Based on the quasi-static field approximation, the SP resonance of a metal NP can be regarded as an equivalent dipole oscillation [18]. In this situation, the SP field strength decays with the third power of the distance between the metal NP and light emitter. Hence, the SP coupling effects are expected to be enhanced faster when the distance becomes smaller until a certain critical distance is reached, beyond which the SP coupling effect leads to emission quenching [19]. However, this critical distance is expected to be smaller than 20 nm, which is much smaller than the distance between Ag NPs and the QW in sample C-SP (68 nm). Therefore, the variation trend of SP coupling effect for enhancing emission can continue when the distance between Ag NPs and the QW is further reduced. In this regard, it should be noted that the distance dependence of the SP-QW coupling can be different from that of the SP-dipole coupling. In our experimental data, except IQE, the variation trends of the SP coupling effects coincide well with the theoretical prediction. The smaller increase percentage of IQE when the p-GaN thickness decreases from 30 to 10 nm, as compared with the increase percentage when the p-GaN thickness decreases from 50 to 30 nm, can be due to the larger intrinsic IQE in sample C-R. The increment of IQE through SP coupling in a QW decreases with increasing intrinsic IQE. It is noted that the non-radiative recombination rate in a QW, which is related to the defect density in the QW, is not affected by SP coupling. Since SP coupling can enhance the radiative recombination rate, the percentage of carrier energy consumed by non-radiative recombination is reduced and hence IQE is increased. In other words, the IQE enhancement through SP coupling relies on the transfer of the energy originally consumed by non-radiative recombination into radiative recombination. A higher intrinsic IQE implies a relatively smaller energy percentage consumed by non-radiative recombination and hence leads to a smaller SP-coupling induced IQE enhancement. However, other SP-coupling effects related to carrier density decrease in the QW are not directly influenced by the factor of IQE. In other words, SP coupling may not be useful for increasing the IQE of an LED with a very high intrinsic IQE. However, it is still useful for reducing the efficiency droop effect and increasing the modulation bandwidth. It is noted that SP coupling can also enhance the light extraction efficiency of an LED [20]. Therefore, the increase percentage of LED output intensity can be larger than that of its IQE. Regarding the reduction of efficiency droop range and the increase of modulation bandwidth, they are mainly related to the decrease of carrier density in the QW. Their improvement ranges are expected to increase faster when the p-GaN thickness is further reduced.
LED modulation bandwidth is an important factor for developing visible communications [21–24]. A laser diode or a super-luminescent diode (SLD) can have a higher modulation bandwidth due to the stimulated or amplified spontaneous emission and cavity resonance natures [25, 26]. However, due to the light directivity and higher cost of a laser diode or SLD, the application of lighting LEDs to visible communications is still attractive. A non-polar or semi-polar LED may have a larger modulation bandwidth due to the higher radiative recombination rate. However, so far, the performance of a non-polar or semi-polar device is still not as good as that of a c-plane LED unless expensive non-polar or semi-polar GaN substrate is used. For visible communication application, c-plane LED is still a good choice deserving further improvement. In a previous publication, this research team has demonstrated the implementation of an SP-coupled LED with the distance between the surface Ag NPs and the QW slightly larger than that of sample A-SP 108 nm to achieve the modulation bandwidth of 528.3 MHz [1]. In the current work, by reducing this distance to 68 nm, a new record-high modulation bandwidth of 625.6 MHz is achieved. It is noted that the SP resonance peak in the current work does not well coincide with the QW emission wavelength, as shown in Fig. 3. A careful adjustment of Ag NP size can blue-shift the SP resonance peak for further increasing the SP coupling strength at the designated QW emission wavelength (465 nm). In this situation, the modulation bandwidth can be further increased.
5. Conclusions
In summary, we have first demonstrated the high performances of thin p-type-layer LEDs. By increasing the Mg doping concentration in the p-AlGaN EBL through an Mg pre-flow process, the hole injection efficiency could be significantly enhanced. In this situation, the high LED performance could be maintained even with a very thin p-GaN layer. Then, we compared the SP coupling effects, including the enhancement of IQE, increase of output intensity, reduction of efficiency droop, and increase of modulation bandwidth, among the thin p-type LED samples of different p-GaN thicknesses. These effects were stronger as the p-GaN layer became thinner. However, the dependencies of these effects on p-type layer thickness differed. With SP coupling in an LED of only 10 nm in p-GaN thickness, we have achieved a new record-high modulation bandwidth of 625.6 MHz with a circular mesa size of 10 µm in radius among c-plane GaN-based LEDs.
Funding
Ministry of Science and Technology, Taiwan, China (MOST 104-2622-E-002-031-CC2, MOST 105-2221-E-002-159-MY3, and MOST 105-2221-E-002-118); National Taiwan University (105R89095A); U.S. Air Force Scientific Research Office (AOARD-14-4105).
References and links
1. C. H. Lin, C. G. Tu, Y. F. Yao, S. H. Chen, C. Y. Su, H. T. Chen, Y. W. Kiang, and C. C. Yang, “High modulation bandwidth of a light-emitting diode with surface plasmon coupling,” IEEE Trans. Electron Dev. 63(10), 3989–3995 (2016). [CrossRef]
2. C. H. Lin, C. Y. Su, E. Zhu, Y. F. Yao, C. Hsieh, C. G. Tu, H. T. Chen, Y. W. Kiang, and C. C. Yang, “Modulation behaviors of surface plasmon coupled light-emitting diode,” Opt. Express 23(6), 8150–8161 (2015). [CrossRef] [PubMed]
3. C. H. Lin, C. Y. Su, Y. Kuo, C. H. Chen, Y. F. Yao, P. Y. Shih, H. S. Chen, C. Hsieh, Y. W. Kiang, and C. C. Yang, “Further reduction of efficiency droop effect by adding a lower-index dielectric interlayer in a surface plasmon coupled blue light-emitting diode with surface metal nanoparticles,” Appl. Phys. Lett. 105(10), 101106 (2014). [CrossRef]
4. C. H. Lin, C. Hsieh, C. G. Tu, Y. Kuo, H. S. Chen, P. Y. Shih, C. H. Liao, Y. W. Kiang, C. C. Yang, C. H. Lai, G. R. He, J. H. Yeh, and T. C. Hsu, “Efficiency improvement of a vertical light-emitting diode through surface plasmon coupling and grating scattering,” Opt. Express 22(S3Suppl 3), A842–A856 (2014). [CrossRef] [PubMed]
5. H. S. Chen, C. P. Chen, Y. Kuo, W. H. Chou, C. H. Shen, Y. L. Jung, Y. W. Kiang, and C. C. Yang, “Surface plasmon coupled light-emitting diode with metal protrusions into p-GaN,” Appl. Phys. Lett. 102(4), 041108 (2013). [CrossRef]
6. E. M. Purcell, “Spontaneous emission probabilities at radio frequencies,” Phys. Rev. 69, 681 (1946).
7. Y. Kuo, W. Y. Chang, H. S. Chen, Y. W. Kiang, and C. C. Yang, “Surface plasmon coupling with a radiating dipole near an Ag nanoparticle embedded in GaN,” Appl. Phys. Lett. 102(16), 161103 (2013). [CrossRef]
8. Y. Kuo, C. H. Lin, H. S. Chen, C. Hsieh, C. G. Tu, P. Y. Shih, C. H. Chen, C. H. Liao, C. Y. Su, Y. F. Yao, H. T. Chen, Y. W. Kiang, and C. C. Yang, “Surface plasmon coupled light-emitting diode – experimental and numerical studies,” Jap. J. Appl. Phys. 54(2S), 02BD01 (2015). [CrossRef]
9. K. C. Shen, C. Y. Chen, H. L. Chen, C. F. Huang, Y. W. Kiang, C. C. Yang, and Y. J. Yang, “Enhanced and partially polarized output of a light-emitting diode with its InGaN/GaN quantum well coupled with surface plasmons on a metal grating,” Appl. Phys. Lett. 93(23), 231111 (2008). [CrossRef]
10. M. K. Kwon, J. Y. Kim, B. H. Kim, I. K. Park, C. Y. Cho, C. C. Byeon, and S. J. Park, “Surface-plasmon-enhanced light-emitting diodes,” Adv. Mater. 20(7), 1253–1257 (2008). [CrossRef]
11. C. Y. Cho, S. J. Lee, J. H. Song, S. H. Hong, S. M. Lee, Y. H. Cho, and S. J. Park, “Enhanced optical output power of green light-emitting diodes by surface plasmon of gold nanoparticles,” Appl. Phys. Lett. 98(5), 051106 (2011). [CrossRef]
12. C. Y. Su, C. G. Tu, W. H. Liu, C. H. Lin, Y. F. Yao, H. T. Chen, Y. R. Wu, Y. W. Kiang, and C. C. Yang, “Enhancing the hole injection efficiency of a light-emitting diode by increasing Mg doping in the p-AlGaN electron blocking layer,” IEEE Trans. Electron Dev. 64(8), 3226–3233 (2017). [CrossRef]
13. C. Y. Chen, C. Hsieh, C. H. Liao, W. L. Chung, H. T. Chen, W. Cao, W. M. Chang, H. S. Chen, Y. F. Yao, S. Y. Ting, Y. W. Kiang, C. C. Yang, and X. Hu, “Effects of overgrown p-layer on the emission characteristics of the InGaN/GaN quantum wells in a high-indium light-emitting diode,” Opt. Express 20(10), 11321–11335 (2012). [CrossRef] [PubMed]
14. Y. T. Moon, Y. Fu, F. Yun, S. Dogan, M. Mikkelson, D. Johnstone, and H. Morkoç, “A study of GaN regrowth on the micro-facetted GaN template formed by in-situ thermal etching,” Phys. Status Solidi 202(5), 718–721 (2005). [CrossRef]
15. Y. F. Yao, C. H. Lin, C. Hsieh, C. Y. Su, E. Zhu, S. Yang, C. M. Weng, M. Y. Su, M. C. Tsai, S. S. Wu, S. H. Chen, C. G. Tu, H. T. Chen, Y. W. Kiang, and C. C. Yang, “Multi-mechanism efficiency enhancement in growing Ga-doped ZnO as the transparent conductor on a light-emitting diode,” Opt. Express 23(25), 32274–32288 (2015). [CrossRef] [PubMed]
16. H. Y. Ryu, H. S. Kim, and J. I. Shim, “Rate equation analysis of efficiency droop in InGaN light-emitting diodes,” Appl. Phys. Lett. 95(8), 081114 (2009). [CrossRef]
17. A. Neogi, C. W. Lee, H. O. Everitt, T. Kuroda, A. Tackeuchi, and E. Yablonvitch, “Enhancement of spontaneous recombination rate in a quantum well by resonant surface plasmon coupling,” Phys. Rev. B 66(15), 153305 (2002). [CrossRef]
18. 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] [PubMed]
19. D. V. Guzatov, S. V. Vaschenko, V. V. Stankevich, A. Y. Lunevich, Y. F. Glukhov, and S. V. Gaponenko, “Plasmonic enhancement of molecular fluorescence near silver nanoparticles: Theory, modeling, and experiment,” J. Phys. Chem. C 116(19), 10723–10733 (2012). [CrossRef]
20. Y. Kuo, H. T. Chen, W. Y. Chang, H. S. Chen, C. C. Yang, and Y. W. Kiang, “Enhancements of the emission and light extraction of a radiating dipole coupled with localized surface plasmon induced on a surface metal nanoparticle in a light-emitting device,” Opt. Express 22(S1), A155–A166 (2014). [CrossRef] [PubMed]
21. J. J. D. McKendry, R. P. Green, A. E. Kelly, Z. Gong, B. Guilhabert, D. Massoubre, E. Gu, and M. D. Dawson, “High-speed visible light communications using individual pixels in a micro light-emitting diode array,” IEEE Photonics Technol. Lett. 22(18), 1346–1348 (2010). [CrossRef]
22. J. J. D. McKendry, D. Massoubre, S. Zhang, B. R. Rae, R. P. Green, E. Gu, R. K. Henderson, A. E. Kelly, and M. D. Dawson, “Visible-light communications using a CMOS-controlled micro-light-emitting-diode array,” J. Lightwave Technol. 30(11), 61–67 (2012). [CrossRef]
23. J. M. Wun, C. W. Lin, W. Chen, J. K. Sheu, C. L. Lin, Y. L. Li, J. E. Bowers, J. W. Shi, J. Vinogradov, R. Kruglov, and O. Ziemann, “GaN-based miniaturized cyan light-emitting diodes on a patterned sapphire substrate with improved fiber coupling for very high-speed plastic optical fiber communication,” IEEE Photonics J. 4(5), 1520–1529 (2012). [CrossRef]
24. S. Zhang, S. Watson, J. J. D. McKendry, D. Massoubre, A. Cogman, E. Gu, R. K. Henderson, A. E. Kelly, and M. D. Dawson, “1.5 Gbit/s multi-channel visible light communications using CMOS-controlled GaN-based LEDs,” J. Lightwave Technol. 31(8), 1211–1216 (2013). [CrossRef]
25. S. Watson, M. Tan, S. P. Najda, P. Perlin, M. Leszczynski, G. Targowski, S. Grzanka, and A. E. Kelly, “Visible light communications using a directly modulated 422 nm GaN laser diode,” Opt. Lett. 38(19), 3792–3794 (2013). [CrossRef] [PubMed]
26. C. Shen, C. Lee, T. K. Ng, S. Nakamura, J. S. Speck, S. P. DenBaars, A. Y. Alyamani, M. M. El-Desouki, and B. S. Ooi, “High-speed 405-nm superluminescent diode (SLD) with 807-MHz modulation bandwidth,” Opt. Express 24(18), 20281–20286 (2016). [CrossRef] [PubMed]
