BY 4.0 license Open Access Published by De Gruyter August 26, 2021

Emission characteristics of GaInN/GaN multiple quantum shell nanowire-based LEDs with different p-GaN growth conditions

Sae Katsuro, Weifang Lu ORCID logo, Kazuma Ito, Nanami Nakayama, Naoki Sone, Koji Okuno, Motoaki Iwaya, Tetsuya Takeuchi, Satoshi Kamiyama ORCID logo and Isamu Akasaki
From the journal Nanophotonics

Abstract

Improving current injection into r- and m-planes of nanowires (NWs) is essential to realizing efficient GaInN/GaN multiple quantum shell (MQS) NW-based light-emitting diodes (LEDs). Here, we present the effects of different p-GaN shell growth conditions on the emission characteristics of MQS NW-LEDs. Firstly, a comparison between cathodoluminescence (CL) and electroluminescence (EL) spectra indicates that the emission in NW-LEDs originates from the top region of the NWs. By growing thick p-GaN shells, the variable emission peak at around 600 nm and degradation of the light output of the NW-LEDs are elaborated, which is attributable to the localization of current in the c-plane region with various In-rich clusters and deep-level defects. Utilizing a high growth rate of p-GaN shell, an increased r-plane and a reduced c-plane region promote the deposition of indium tin oxide layer over the entire NW. Therefore, the current is effectively injected into both the r- and m-planes of the NW structures. Consequently, the light output and EL peak intensity of the NW-LEDs are enhanced by factors of 4.3 and 13.8, respectively, under an injection current of 100 mA. Furthermore, scanning transmission electron microscope images demonstrate the suppression of dislocations, triangular defects, and stacking faults at the apex of the p-GaN shell with a high growth rate. Therefore, localization of current injection in nonradiative recombination centers near the c-plane was also inhibited. Our results emphasize the possibility of realizing high efficacy in NW-LEDs via optimal p-GaN shell growth conditions, which is quite promising for application in the long-wavelength region.

1 Introduction

Over the past few decades, group III nitride semiconductors, with fundamental bandgaps covering a broad spectral range, have been developed for solid-state lighting and lasing technologies [1], [2], [3], [4]. Among them, GaInN plays a dominant role in the visible light region, despite a declining luminous efficiency with high InN fractions [5], [6], [7]. Nevertheless, efficiency degradation and blue-shift of the luminescence peak with increasing carrier injection are prevalent in conventional c-plane light-emitting diodes (LEDs). These impeding factors are commonly related to the polarization-induced piezoelectric field, known as the quantum-confined Stark effect (QCSE) [8, 9]. The strong QCSE phenomenon and high density of defects in GaInN/GaN multiple-quantum-wells make it difficult to achieve high-efficiency LEDs in the yellow–red region [10]. Core–shell GaInN/GaN nanowires (NWs) provide an alternative approach for achieving both high crystalline quality and high In content for NW-based LEDs (NW-LEDs) [11], [12], [13], [14]. In these structures, the nonpolar m-plane sidewall allows reducing the efficiency droop and blue-shift caused by the QCSE. Control of emission wavelength is indispensable for realizing GaInN/GaN multiple quantum shell (MQS) NW-LEDs that emit in the long-wavelength region. A wide range of emission colors is available by tunning the In incorporation in GaInN quantum wells with different pitches and diameters [15], [16], [17], [18], [19]. Also, monolithic integration of multiple NW patterns has been of great interest for the application in white and micro-LEDs since GaInN/GaN MQS manifest a much lower surface recombination rate and shorter carrier diffusion length than those of conventional AlInGaP materials [20], [21], [22].

Epitaxial growth of uniform n-GaN core and improvement in GaInN/GaN MQS active structures have been widely reported. For instance, continuous and pulsed supply modes of precursors were applied to grow uniform n-GaN core NWs on patterned templates by metalorganic chemical vapor deposition (MOCVD) [23], [24], [25]. Coaxial growth of GaInN/GaN MQS structures on n-GaN NWs has been systematically investigated to achieve high-performance NW-LEDs. High internal quantum efficiency (69%) was realized by inserting AlGaN spacers [26], [27], [28], [29]. Despite the progress in epitaxial growth, limited results on fabrication and electrical performance of MQS NW-LED chips have been presented [30, 31]. Currently, several issues such as the blue-shift tendency with increasing the injection current [32, 33], current localization at the c-plane apex region of NWs [12], and the red emission under low current injection [34, 35] need to be resolved. In addition, even though sputtering conditions of indium tin oxide (ITO) have been optimized to improve the current spreading over the entire NWs, the device fabrication process is still not sufficiently stable for uniform current injection in NW-LEDs [36]. Unlike planar c-plane LEDs, the shape of the p-GaN shell on MQS NWs is vital for conformal ITO deposition and uniform current injection. Detailed insight into the characteristics of MQS NW-LEDs is necessary to understand their performance and potential improvements with superior p-GaN shells. However, relevant investigations on p-GaN shell growth and its effects on the emission characteristics of NW-LED chips have not yet been reported.

In this work, we explored the emission characteristics of NW-LEDs with different p-GaN morphologies, aiming to improve the uniformity of ITO deposition and thereupon current injection into r- and m-planes of the NWs. The MQS active structures and n-GaN cores were prepared under identical growth conditions to minimize the probability of different In contents among the samples. NW-LEDs were fabricated and then characterized in terms of current–voltage curves, light output, and electroluminescence (EL) spectra. With optimal growth conditions, lateral growth near the bottom of the NWs was promoted, and there was a uniform current injection along the NWs. In addition, the emission features of the NW-LEDs were discussed based on the scanning transmission electron microscopy (STEM) results.

2 MOCVD growth and device fabrication

NW samples (with a size of 10 × 10 mm2) for LED processing were grown on patterned n-GaN/sapphire templates using the selective area growth method with MOCVD (SR 2000, TAIYO NIPPON SANSO Co., Shinagawa-ku, Tokyo, Japan) [18, 29, 37]. Regarding the preparation of templates, patterns in a triangular arrangement, with a diameter of 300 nm and a pitch of 1200 nm, were defined on the SiO2 mask layer by nanoimprint lithography technology. The GaN inside the whole patterns was exposed by inductively coupled plasma (ICP) etching (MV06-7001-0, ULVAC, Inc., Chigasaki City, Kanagawa, Japan). After n-GaN core NW growth [13], five pairs of GaInN/GaN MQS with AlGaN spacers were grown on the NWs using modified growth sequences described in our previous work [26]. Four NW samples were prepared, as listed in Table 1, to investigate the effect of the p-GaN shell on the properties of LED devices. An Mg-doped (∼3 × 1019 cm−3) GaN shell was grown on the MQS NWs, followed by a higher Mg-doped (∼6 × 1019 cm−3) p-GaN shell. Here, the higher Mg concentration was doped to promote the lateral growth rate of the p-GaN shell, ultimately embedding the NWs without voids and pits. Samples a, b, and c, described in Table 1, have second p-GaN shell growth times of 30 s, 10 min, and 20 min, respectively. Since the shape of the p-GaN shell is essential for the LED process, sample d was additionally prepared by increasing the trimethylgallium (TMG) flow rate to 228.04 μmol/min for the first p-GaN shell. Notably, the growth time of the first p-GaN shell in sample d was adjusted to 3 min to maintain the same thickness. Furthermore, under the same Mg/Ga ratio, the Mg doping concentration is also expected to be the same in the NW samples.

Table 1:

MOCVD growth parameters for nanowire samples.

Sample Temperature and growth time of n-core Temperature of GaInN in MQS TMG flow rate for 1st p-GaN shell growth Growth time for 1st p-GaN shell TMG flow rate for 2nd p-GaN shell growth Growth time for 2nd p-GaN shell
a 1135 °C, 70 s 750 °C 114.02 μmol/min 6 min 114.02 μmol/min 30 s
b 1135 °C, 70 s 750 °C 114.02 μmol/min 6 min 114.02 μmol/min 10 min
c 1135 °C, 70 s 750 °C 114.02 μmol/min 6 min 114.02 μmol/min 20 min
d 1135 °C, 70 s 750 °C 228.04 μmol/min 3 min 114.02 μmol/min 30 s

Before the NW-LED fabrication process, annealing for 30 min at 650 °C was performed to active the p-GaN shells by rapid thermal annealing (RTA, MR094017-0, ULVACRIKO, INC., Tsuzuki City, Yokohama, Japan), under the ambient nitrogen and oxygen. Schematic illustrations in Figure 1(A) show the fabrication process flow for the NW-LEDs. To remove NWs from the n- and p-electrodes, ultrasonic cleaning, and ICP etching were carried out on sample a and the rest of the samples (b, c, and d), respectively. For samples b, c, and d, vapor-deposited Ni film and resist coating were applied on the surface to fabricate flattened electrodes selectively. A 200-nm thick SiO2 film was then deposited on the p-electrode area to avoid the connection to the n-GaN bulk. Subsequently, the p-electrode and NWs in the mesa area were covered by an ITO current spreading layer with a thickness of 60 nm. Since the sputtered ITO layer intrinsically has high electrical resistance, thermal annealing was performed at 600 °C for 4 min under ambient nitrogen. Finally, Cr/Ni/Au films with thicknesses of 10, 20, and 200 nm were sequentially deposited on defined p- and n-electrodes by electron beam evaporation (MA08-3065, ULVAC, Inc., Chigasaki City, Kanagawa, Japan). Figure 1(B) shows a schematic configuration of a fabricated NW-LED in sample c.

Figure 1: 
Schematic diagram of the fabrication process for NW-LEDs and device structure in sample c.
(A) Schematic of the device fabrication flow for nanowire (NW)-based light-emitting diodes (LEDs) in the sample a: ① as-grown NW structures are actively annealed for 30 min under the ambient nitrogen and oxygen; ② NWs on the electrodes are removed by ultrasonic cleaning or dry etching; ③ n-GaN bulk at the n-electrode is exposed by ICP etching; ④ SiO2 is sputtered on the p-electrode; ⑤ indium tin oxide (ITO) is sputtered and annealed for current spreading in the mesa area; ⑥ Cr/Ni/Au films are deposited on the electrodes. (B) Schematic configuration of a fabricated NW-LED in sample c.

Figure 1:

Schematic diagram of the fabrication process for NW-LEDs and device structure in sample c.

(A) Schematic of the device fabrication flow for nanowire (NW)-based light-emitting diodes (LEDs) in the sample a: ① as-grown NW structures are actively annealed for 30 min under the ambient nitrogen and oxygen; ② NWs on the electrodes are removed by ultrasonic cleaning or dry etching; ③ n-GaN bulk at the n-electrode is exposed by ICP etching; ④ SiO2 is sputtered on the p-electrode; ⑤ indium tin oxide (ITO) is sputtered and annealed for current spreading in the mesa area; ⑥ Cr/Ni/Au films are deposited on the electrodes. (B) Schematic configuration of a fabricated NW-LED in sample c.

The morphologies of the NW structures were characterized by scanning electron microscope (SEM) (SU70, Hitachi High-Technologies Co., Minato City, Tokyo, Japan). The crystalline quality of p-GaN shells on the NWs was further inspected by a Hitachi HD2700 STEM system (Hitachi High Technologies Cor., Tokyo, Japan). Our previous study provided details on the preparation of STEM lamella and typical NW sample measurements [27]. The optical properties of the NW-LED samples were analyzed via EL (Ocean Optics Co., United States) and cathodoluminescence (CL) measurements (SEM-SU5000, Hitachi Co., Minato City, Tokyo, Japan). Moreover, a semiconductor parameter analyzer (4156C, Agilent Technology, Santa Clara CA, USA) equipped with a micro prober station was used for current–voltage–light output (I–V–L) characterizations. It should be mentioned that the light output and EL signal of the NW-LEDs was detected from the backside of the chips.

3 Results and discussion

3.1 Comparison of CL and EL measurements in core–shell NW-LEDs

CL and EL measurements of sample a were carried out for singular NWs and fabricated NW-LEDs, respectively, aiming to compare the emission features. Figure 2(A) shows a cross-sectional schematic diagram of one NW consisting of an n-GaN core, MQS, and p-GaN shells. The CL measurement positions on c-, r-, and m-planes are also indicated in the schematic diagram. Figure 2(B) shows CL spectra of NWs in the center and edge (∼2 mm to the periphery) regions of the sample, respectively, which were acquired at the c-plane at the apex region, r-plane, and top and bottom regions of the m-plane. The CL spectra of the n-GaN cores are also plotted as a reference. The emission peaks near the c-plane area are approximate to that of the n-GaN core. Such yellowish emission peaks are usually referred to as point defects involved during epitaxial growth, such as Ga vacancies or C substitutes in N-vacancies [38, 39]. The observed emission peak is featured with a clear redshift from the bottom of the m-plane (420 nm) to the r-plane (450 nm), as shown in Figure 2(B). Likewise, similar behavior is further confirmed in NWs near the edge area of the as-grown sample. Nevertheless, the r-plane peak wavelength at 470 nm is longer than that in the center region. Cross-sectional view SEM images of the NWs are shown in the insets of Figure 2(B). The height of NWs on the edge (1.25 µm) is shorter than that of the center area (1.53 µm). Therefore, the higher CL intensity of the m-plane at the edge area can be ascribed to thicker MQS active shells on the shorter NWs. The difference in height was attributable to the edge effect during n-GaN core growth at high temperature since precursors diffused from the adjacent region were limited due to the gap or step existing between the NW template and surrounding dummy wafers.

Figure 2: 
Cathodoluminescence (CL) spectra acquired at different areas of NWs in the sample a.
(A) Cross-sectional schematic diagram of an NW indicating the CL measurement points. (B) CL spectra acquired on the c-plane at the apex region, the r-plane, and top and bottom areas on the m-plane of NWs at the center and edge area of the sample a. Insets show the cross-sectional view of SEM images of the NWs. The heights of NWs in the center and edge areas of the sample are 1.53 and 1.25 µm, respectively.

Figure 2:

Cathodoluminescence (CL) spectra acquired at different areas of NWs in the sample a.

(A) Cross-sectional schematic diagram of an NW indicating the CL measurement points. (B) CL spectra acquired on the c-plane at the apex region, the r-plane, and top and bottom areas on the m-plane of NWs at the center and edge area of the sample a. Insets show the cross-sectional view of SEM images of the NWs. The heights of NWs in the center and edge areas of the sample are 1.53 and 1.25 µm, respectively.

After fabricating NW-LEDs in sample a, electronic characteristics and EL spectra were investigated, as shown in Figure 3. The I–V characteristics in Figure 3(A) show that the threshold voltage under 1 mA injection was 2.3 V. As for the I–L curves, the NW-LEDs started to emit light beyond 4.8 mA. EL spectra of the NW-LEDs, measured under different current injections, are plotted in Figure 3(B) for insight into the optical properties. It is confirmed that the EL peaks are stable as a function of current injection, revealing that the QCSE-induced blue-shift is suppressed in the NW-LEDs. According to the different CL emission peaks on different planes, it indicates that the EL emission is dominant by either r- or m-planes. The inset in Figure 3(B) shows the cross-sectional SEM images after coating with the ITO films. Compared with the CL spectra of the r-planes (∼470 nm), the EL peaks located at 485 nm are presumed to be associated with the r-planes. The difference between EL and CL peaks is ascribed to the presence of all light from many NWs on one chip for EL measurement, while only one emission point from one NW corresponds to the CL emission. Since the activation annealing temperature of p-GaN was lower than the MQS growth temperature, it is assumed that there was almost no damage to the MQS active layer. Therefore, CL and EL measurements are considered comparable here.

Figure 3: 
Current–voltage–light output (I–V–L) and the EL characterization of NW-LED in the sample a.
(A) Current–voltage–light output (I–V–L) characteristics of the NW-LEDs of the sample a. (B) EL spectra measured under different current injections. Cross-sectional SEM images of the NWs coated with ITO are shown in the inset.

Figure 3:

Current–voltage–light output (I–V–L) and the EL characterization of NW-LED in the sample a.

(A) Current–voltage–light output (I–V–L) characteristics of the NW-LEDs of the sample a. (B) EL spectra measured under different current injections. Cross-sectional SEM images of the NWs coated with ITO are shown in the inset.

3.2 Electrical properties of NW-LEDs with different p-GaN thicknesses

Because the NWs are three-dimensional structures, emission features in NW-LEDs usually depend on the current injection. To determine the EL peak wavelength, three NW samples a, b, and c were prepared with different p-GaN growth times of 30 s, 10 min, and 20 min, respectively. Figure 4 (Ai)—(Aiii) schematically illustrate the cross-sectional configuration of samples a, b, and c in form of different thicknesses of p-GaN shells. Corresponding cross-sectional SEM images of the as-grown samples are shown in Figure 4 (Bi)—(Biii). The contrast between the n-GaN core and p-GaN indicates that the MQS NWs are eventually embedded inside the thick p-GaN layer when the growth time of p-GaN is increased. Figure 4(C) shows the I–V characteristics of NW-LEDs in samples a, b, and c. For sample c, the extremely high operating voltage is due to high resistivity in the thick p-GaN shells, especially the second p-GaN shell with higher Mg doping. The resistance of NW-LEDs in sample b is only slightly higher than that in sample a, so the current was expected to spread into the mesa area.

Figure 4: 
Morphologies and electronic-optical properties of the NWs in samples a, b, and c.
(Ai)—(Aiii) schematic illustrations of the cross-sectional configuration of samples a, b, and c. (Bi)—(Biii) cross-sectional SEM images of the grown samples a, b, and c. (C) I–V and (D) I–L characteristics of the NW-LEDs fabricated from samples a, b, and c. The normalized I–L curves are plotted in the inset. (E) EL spectra acquired under different current injections for NW-LEDs from samples a, b, and c, respectively.

Figure 4:

Morphologies and electronic-optical properties of the NWs in samples a, b, and c.

(Ai)—(Aiii) schematic illustrations of the cross-sectional configuration of samples a, b, and c. (Bi)—(Biii) cross-sectional SEM images of the grown samples a, b, and c. (C) I–V and (D) I–L characteristics of the NW-LEDs fabricated from samples a, b, and c. The normalized I–L curves are plotted in the inset. (E) EL spectra acquired under different current injections for NW-LEDs from samples a, b, and c, respectively.

In contrast, the I–L characteristics in Figure 4(D) suggest that the NW-LEDs in both samples b and c suffered from low effective current injection to the MQS active region. As a result, the light output of the NW-LEDs in the sample a manifests about tenfold larger than that in samples b and c. Furthermore, since the current was concentrated only in the upper part of the NWs in sample c, thermal saturation was observed under high current injection in the I–L characteristics, as shown in the inset of Figure 4(D). The EL spectra measured under different current injections are plotted in Figure 4(E). The results clearly reveal degradation of EL intensity with an increase in p-GaN thickness with high Mg doping. Light emission in the long-wavelength region predominantly appears in samples b and c, which may be associated with the In-rich clusters and point defects inducing recombination near the c-plane MQS region [40]. In addition, the blue-shift of the EL peak wavelength in sample b as increasing the current injection is further considered to be related to the radiative recombination through In-rich clusters and various deep-level defects at c-plane [12, 41]. Therefore, there is a trade-off between completely embedding the NWs and reducing the current concentration in the upper part of the NWs. In this case, it is necessary to optimize further growth conditions for the embedded p-GaN layer, aiming to promote the current injected into the r- and m-planes of the NWs.

3.3 Effect of p-GaN shape on the optical properties of NW-LEDs

To compare the effect of p-GaN with sample a, NW sample d was prepared with a higher p-GaN growth rate via a TMG flow rate of 228.04 μmol/min. Under a high TMG flow rate, the lateral growth rate was expected to be enhanced, leading to a variation of the p-GaN shape. Figure 5(A) compares the I–V–L characteristics of sample d with those of sample a grown with a low TMG flow rate (114.02 μmol/min). Sample d shows a higher resistance characteristic, but the estimated resistivity approximates to that of sample a. Because the Mg/Ga ratio for p-GaN growth and the activation conditions were identical, the Mg doping concentration and resistivity after activation were expected to be similar. The I–V characteristics between nn electrodes in samples a and d are shown in Figure 5(B). An Ohmic contact was obtained in sample a while the resistance was extremely high for sample d, which behaved more like a Schottky contact. The SEM images of the electrode surface fabricated after using ultrasonic (sample a) or ICP dry etching (sample b) are shown in the insets of Figure 5(B). Because of the different etching rates for the specific shells, there were many burrs generated during the ICP etching in the electrode region, resulting in poor contact with the metallic alloys. Therefore, the increased resistance was attributable to the difference in the fabrication method of the electrodes. Regarding the I–L characteristics, the light output starts to ascend from 3.3 mA in the sample a and 0.75 mA in sample d owing to the improved growth of the p-GaN shell. Moreover, the light output of NW-LEDs in the sample d is almost 4.6 times stronger than that in sample a under the same injection current of 100 mA.

Figure 5: 

I–V–L curves, SEM images, and corresponding luminescence photos of LEDs in samples a and d.
(A) I–V characteristics and light output in samples a and d having different growth rates. (B) I–V characteristics of between n–n electrodes in samples a and d, respectively. SEM images of the electrode surface fabricated after ultrasonic or ICP dry etching are shown in the inset. (Ci)–(Cii) cross-sectional SEM images of NWs in samples a and d after the fabrication process. (Di)–(Dii) schematic images of the current injection along NWs in samples a and d. Panels (Ei) and (Eii) show the luminescence photos of LED chips in samples a and d, respectively.

Figure 5:

I–V–L curves, SEM images, and corresponding luminescence photos of LEDs in samples a and d.

(A) I–V characteristics and light output in samples a and d having different growth rates. (B) I–V characteristics of between nn electrodes in samples a and d, respectively. SEM images of the electrode surface fabricated after ultrasonic or ICP dry etching are shown in the inset. (Ci)–(Cii) cross-sectional SEM images of NWs in samples a and d after the fabrication process. (Di)–(Dii) schematic images of the current injection along NWs in samples a and d. Panels (Ei) and (Eii) show the luminescence photos of LED chips in samples a and d, respectively.

Figure 5(Ci–Cii) and (Di–Dii) show cross-sectional SEM images and schematic illustrations of current injection in samples a and d, respectively. The c-plane area of the p-GaN shell was suppressed in sample d, whereas the r-planes and the diameter of the NWs were clearly increased with a high growth rate of p-GaN. Such enhanced lateral growth is attributable to the sufficient supply of precursors that diffused from the top area and SiO2 mask layer. As a result, the ITO current spreading layer can uniformly cover the NWs and then promote current injection into the bottom part of the NWs. From the emission photos in Figure 5(Ei), red, orange, and yellow light with a few blue dots of emission were observed in the sample a with an increase in current injection. At low current, red light emission may be mainly from the c-plane area and, thereafter, the increased current reached the r- and m-planes. Since the emission photos were taken from the front side of the NW-LEDs, the color is dominated by the emission near the c-planes. In contrast, the emission in sample d immediately changed from green to blue light as the current increased to 25 mA, as shown in Figure 5(Eii). This indicates that the shape of the p-GaN shell inevitably affected the current injection distribution along the NWs, resulting in the blue-green emission.

The EL spectra under different current injections were acquired from the backside of the NW-LEDs in samples a and d, as shown in Figure 6(A). For comparison, EL measurements were simultaneously carried out under identical conditions. The emission peaks and integrated EL intensity of the samples are plotted in Figure 6(B) as a function of current injection. It is confirmed that the EL intensity in sample d increased up to 11 times or more compared with sample a. Furthermore, the integrated intensity manifests a 13.8-folds enhancement under the current injection of 100 mA. Besides the suppressed current concentration near the c-plane area, the observed enhancement is also attributable to the uniform ITO coating on the NWs, which promotes the current spreading.

Figure 6: 
Comparison of EL spectra in samples a and d.
(A) EL spectra of sample a and sample d measured under different current injections. (B) Peak wavelength and integrated EL intensity as a function of current injection.

Figure 6:

Comparison of EL spectra in samples a and d.

(A) EL spectra of sample a and sample d measured under different current injections. (B) Peak wavelength and integrated EL intensity as a function of current injection.

The emission peaks of both samples are located at around 485 nm, while a shoulder peak around 520 nm is observed in sample d. For sample a, the red–yellow emission due to the In-rich clusters and deep level defects in c-plane MQSs was predominant in the emission photographs [see Figure 5(Ei)], but was not clearly observed in the EL spectra. This is because the luminescence photographs were taken from the topside, while the EL signal was detected from the backside of the LED chips via optical fiber. Therefore, the red–yellow emission from the upper part of the NWs appeared visually in photographs, and the emission from the r- and m-planes of the NWs was predominantly detected in the EL measurements. Of note, the localization of current injection in the c-plane area is also the main reason for the lower emission intensity in sample a. In the case of the sample d, both the emission photographs and EL spectra were contributed by the emission from the r- and m-planes, on account of the considerably improved current injection. Nevertheless, at low current injection (10 mA), the EL peak at 520 nm is associated with the MQS emission at the c-plane with higher InN-fraction and point defects-related yellowish emission. Furthermore, no obvious QCSE related blue-shift occurs as the current increases to values larger than 25 mA because the EL spectra consist of the emission from semipolar r-planes and nonpolar m-planes.

STEM inspection near the apex region was performed for sample d to shed light on the structural features of the p-GaN with different growth rates (shapes). For comparison, the STEM results of the NW sample grown with a classical low growth rate of the p-GaN shell are also shown in Figure 7. In Figure 7(A), the apex region above the c-plane MQS of the NW in the reference sample contains black dots (indicated by purple triangles), stacking faults (indicated by blue arrows), and dislocations (indicated by yellow arrows). From the high magnification STEM image in Figure 7(C), the black dots are identified as triangular defects with an especially high density located near the dislocation. Such triangular defects are widely referred to as Mg clusters in p-GaN [42, 43]. With the increased growth rate, the formation of dislocations and triangular defects has been suppressed in the apex region of sample d, as shown in Figure 7(B) and (D). Likewise, a few stacking faults formed from the c-plane MQS and extended parallel to the c-plane. It is inferred that such stacking defects are easily formed due to higher In content and the rough surface of the c-plane MQS. The high In content in c-plane MQS is mainly attributable to the intrinsic feature of the selective-area-growth method [28], while the effect of surface energy with different growth front facets is negligible [44]. Since the triangular defects were considerably suppressed in the apex of the NWs [see Figure 7(D)], the nonradiative recombination rate via defect complexes consisting of Ga vacancies and multiple N-vacancies was also expected to be reduced [45]. For the case with a low growth rate of p-GaN shell, the injection current was preferentially localized at the c-plane region with In-rich clusters and deep-level defects, resulting in stronger red–yellow light emission, even under low current injection. Nevertheless, the luminous efficiency of defects is rather low due to the generation of heat or vibration of the lattice in addition to light emission. Therefore, it is necessary to suppress the formation of defects in MQS NWs by further optimizing the Mg doping concentration, the shape of the p-GaN shell, and the roughness of the c-plane MQS grown on NWs.

Figure 7: 
STEM images of the MQS and p-GaN shell at the c-plane apex region of NWs in reference and sample d.STEM images of the NW c-plane region grown by the (A) classical p-GaN growth method (ref) and with (B) high growth rate of p-GaN (sample d). High magnification STEM images of the p-GaN above c-plane MQS regions for the (C) reference and (D) sample d, respectively. The captured regions are indicated by dashed rectangles.

Figure 7:

STEM images of the MQS and p-GaN shell at the c-plane apex region of NWs in reference and sample d.STEM images of the NW c-plane region grown by the (A) classical p-GaN growth method (ref) and with (B) high growth rate of p-GaN (sample d). High magnification STEM images of the p-GaN above c-plane MQS regions for the (C) reference and (D) sample d, respectively. The captured regions are indicated by dashed rectangles.

4 Conclusion

We have systematically investigated GaInN/GaN MQS NW-LEDs by controlling the growth conditions of the p-GaN shell and investigating optical characteristics. The coincidence between CL and EL spectra indicated that the EL peaks located at 485 nm were presumably dominated by emission from the top side of the NWs. Different growth times of the p-GaN shell were applied to embed the MQS-NWs partially or fully. As a result, the EL intensity degraded with an increase in p-GaN thickness, and light emission in the long-wavelength region (∼600 nm) appeared. Also, the EL peak of NW-LEDs with thick p-GaN shells blue-shifted as the current injection increased, which was attributable to recombination through In-rich clusters and various deep level defects in the c-plane MQS region. Current injection in the NW-LEDs was improved by the change in the shape of the p-GaN shell, resulting in a considerable 13.8-fold enhancement compared to the reference sample. A high lateral growth rate promoted the formation of large r-planes and small c-planes of the p-GaN shell and, subsequently, conformal deposition of the ITO layer. Therefore, it suppressed the localization of current injection near the c-plane area, where defects were easily formed, resulting in strong blue-green emission. High-resolution STEM images verified the reduction of pyramid defects, threading dislocations, and stacking faults in the p-GaN shell around the c-plane MQS region. In this regard, the nonradiative recombination rate via defect complexes consisting of Ga vacancies and multiple N-vacancies should also be reduced. The results demonstrated a possible realization of highly efficient NW-LEDs in the long-wavelength region by further improving the crystallinity of p-GaN shells.


Corresponding author: Weifang Lu, Department of Materials Science and Engineering, Meijo University, 1-501 Shiogamaguchi, Tenpaku-ku, Nagoya, 468-8502, Japan, E-mail:
Sae Katsuro and Weifang Lu have contributed equally to this work.

Funding source: CREST

Award Identifier / Grant number: 16H06416

Funding source: JSPS

Award Identifier / Grant number: 15H02019

Award Identifier / Grant number: 16H06416

Award Identifier / Grant number: 17H01055

Funding source: MEXT

Award Identifier / Grant number: JPJ005357

  1. Author contributions: W.L. devised the experimental plan for this work and grew the nanowire samples. S. Katsuro fabricated the LED devices, measured chip performance, concluded the results, and wrote the first draft of the manuscript. W.L. analyzed the results, reprepared the figures, and re-wrote the manuscript. W.L. prepared the response to reviewers’ comments and revised the manuscript. K.I. contributed to the device fabrication and characterizations. N.N. contributed to the epitaxial growth. N.S. and K.O. supported the epitaxial growth and joined the discussion for p-GaN growth. S. Kamiyama participated in the discussion and revised the manuscript. S. Kamiyama, T.T., M.I., and I.A. conceived the original idea, contributed to the data analysis and supervised the project.

  2. Research funding: This work was financially supported by MEXT “Program for research and development of next-generation semiconductor to realize energy-saving society” [No. JPJ005357], MEXT “Private University Research Branding Project”, JSPS KAKENHI for Scientific Research A [No. 15H02019], JSPS KAKENHI for Scientific Research A [No. 17H01055], JSPS KAKENHI for Innovative Areas [No. 16H06416], and Japan Science and Technology CREST [No. 16815710].

  3. Conflict of interest statement: The authors declare no conflicts of interest regarding this article.

References

[1] I. Akasaki and H. Amano, “Breakthroughs in improving crystal quality of GaN and invention of the p–n junction blue-light-emitting diode,” Jpn. J. Appl. Phys., vol. 45, p. 9001, 2006, https://doi.org/10.1143/jjap.45.9001.Search in Google Scholar

[2] I. Akasaki and H. Amano, “Crystal growth and conductivity control of group III nitride semiconductors and their application to short wavelength light emitters,” Jpn. J. Appl. Phys., vol. 36, p. 5393, 1997, https://doi.org/10.1143/jjap.36.5393.Search in Google Scholar

[3] S. Nakamura, M. Senoh, N. Iwasa, and S.-I. Nagahama, “High-brightness InGaN blue, green and yellow light-emitting diodes with quantum well structures,” Jpn. J. Appl. Phys., vol. 34, p. L797, 1995, https://doi.org/10.1143/jjap.34.l797.Search in Google Scholar

[4] D. Feezell and S. Nakamura, “Invention, development, and status of the blue light-emitting diode, the enabler of solid-state lighting,” Compt. Rendus Phys., vol. 19, pp. 113–133, 2018, https://doi.org/10.1016/j.crhy.2017.12.001.Search in Google Scholar

[5] D. Iida, Z. Zhuang, P. Kirilenko, M. Velazquez-Rizo, and K. Ohkawa, “High-color-rendering-index phosphor-free InGaN-based white light-emitting diodes by carrier injection enhancement via V-pits,” Appl. Phys. Lett., vol. 117, p. 172103, 2020, https://doi.org/10.1063/5.0026017.Search in Google Scholar

[6] S. F. Chichibu, A. Uedono, T. Onuma, et al.., “Origin of defect-insensitive emission probability in In-containing (Al,In,Ga)N alloy semiconductors,” Nat. Mater., vol. 5, pp. 810–816, 2006, https://doi.org/10.1038/nmat1726.Search in Google Scholar

[7] F.-P. Massabuau, M. J. Davies, F. Oehler, et al.., “The impact of trench defects in InGaN/GaN light emitting diodes and implications for the “green gap” problem,” Appl. Phys. Lett., vol. 105, p. 112110, 2014, https://doi.org/10.1063/1.4896279.Search in Google Scholar

[8] T. Takeuchi, C. Wetzel, S. Yamaguchi, et al.., “Determination of piezoelectric fields in strained GaInN quantum wells using the quantum-confined Stark effect,” Appl. Phys. Lett., vol. 73, pp. 1691–1693, 1998, https://doi.org/10.1063/1.122247.Search in Google Scholar

[9] T. Takeuchi, S. Sota, M. Katsuragawa, et al.., “Quantum-confined Stark effect due to piezoelectric fields in GaInN strained quantum wells,” Jpn. J. Appl. Phys., vol. 36, p. L382, 1997, https://doi.org/10.1143/jjap.36.l382.Search in Google Scholar

[10] B. Damilano and B. Gil, “Yellow–red emission from (Ga,In)N heterostructures,” J. Phys. Appl. Phys., vol. 48, p. 403001, 2015, https://doi.org/10.1088/0022-3727/48/40/403001.Search in Google Scholar

[11] Y. H. Ra, R. Navamathavan, J. H. Park, and C. R. Lee, “Coaxial In(x)Ga(1−x)N/GaN multiple quantum well nanowire arrays on Si(111) substrate for high-performance light-emitting diodes,” Nano Lett., vol. 13, pp. 3506–3516, 2013, https://doi.org/10.1021/nl400906r.Search in Google Scholar

[12] Y. J. Hong, C. H. Lee, A. Yoon, et al.., “Visible-color-tunable light-emitting diodes,” Adv. Mater., vol. 23, pp. 3284–3288, 2011, https://doi.org/10.1002/adma.201100806.Search in Google Scholar

[13] S. Kamiyama, W. Lu, T. Takeuchi, M. Iwaya, and I. Akasaki, “Growth and characterization of core–shell structures consisting of GaN nanowire core and GaInN/GaN multi-quantum shell,” ECS J. Solid State Sci. Technol., vol. 9, p. 015007, 2019, https://doi.org/10.1149/2.0252001jss.Search in Google Scholar

[14] K. Kishino and S. Ishizawa, “Selective-area growth of GaN nanocolumns on Si(111) substrates for application to nanocolumn emitters with systematic analysis of dislocation filtering effect of nanocolumns,” Nanotechnology, vol. 26, p. 225602, 2015, https://doi.org/10.1088/0957-4484/26/22/225602.Search in Google Scholar

[15] H. Sekiguchi, K. Kishino, and A. Kikuchi, “Emission color control from blue to red with nanocolumn diameter of InGaN/GaN nanocolumn arrays grown on same substrate,” Appl. Phys. Lett., vol. 96, p. 231104, 2010, https://doi.org/10.1063/1.3443734.Search in Google Scholar

[16] K. Kishino, A. Yanagihara, K. Ikeda, and K. Yamano, “Monolithic integration of four-colour InGaN-based nanocolumn LEDs,” Electron. Lett., vol. 51, pp. 852–854, 2015, https://doi.org/10.1049/el.2015.0770.Search in Google Scholar

[17] S. D. Hersee, A. K. Rishinaramangalam, M. N. Fairchild, L. Zhang, and P. Varangis, “Threading defect elimination in GaN nanowires,” J. Mater. Res., vol. 26, p. 2293, 2011, https://doi.org/10.1557/jmr.2011.112.Search in Google Scholar

[18] W. Lu, K. Ito, N. Sone, et al.., “Color-tunable emission in coaxial GaInN/GaN multiple quantum shells grown on three-dimensional nanostructures,” Appl. Surf. Sci., vol. 539, p. 148279, 2021, https://doi.org/10.1016/j.apsusc.2020.148279.Search in Google Scholar

[19] K. Ito, W. Lu, N. Sone, et al.., “Development of monolithically grown coaxial GaInN/GaN multiple quantum shell nanowires by MOCVD,” Nanomaterials, vol. 10, p. 1354, 2020, https://doi.org/10.3390/nano10071354.Search in Google Scholar

[20] J. Lin and H. Jiang, “Development of microLED,” Appl. Phys. Lett., vol. 116, p. 100502, 2020, https://doi.org/10.1063/1.5145201.Search in Google Scholar

[21] T. Taki and M. Strassburg, “Visible LEDs: more than efficient light,” ECS J. Solid State Sci. Technol., vol. 9, p. 015017, 2019, https://doi.org/10.1149/2.0402001jss.Search in Google Scholar

[22] J. J. WiererJr and N. Tansu, “III‐Nitride micro‐LEDs for efficient emissive displays,” Laser Photonics Rev., vol. 13, p. 1900141, 2019, https://doi.org/10.1002/lpor.201900141.Search in Google Scholar

[23] P.-M. Coulon, B. Alloing, V. Brändli, D. Lefebvre, S. Chenot, and J. Zúñiga-Pérez, “Selective area growth of Ga-polar GaN nanowire arrays by continuous-flow MOVPE: a systematic study on the effect of growth conditions on the array properties,” Phys. Status Solidi (B), vol. 252, pp. 1096–1103, 2015, https://doi.org/10.1002/pssb.201451589.Search in Google Scholar

[24] X. Wang, S. Li, M. S. Mohajerani, et al.., “Continuous-flow MOVPE of Ga-polar GaN column arrays and core–shell LED structures,” Cryst. Growth Des., vol. 13, pp. 3475–3480, 2013, https://doi.org/10.1021/cg4003737.Search in Google Scholar

[25] S.-Y. Bae, K. Lekhal, H.-J. Lee, et al.., “Selective-area growth of doped GaN nanorods by pulsed-mode MOCVD: effect of Si and Mg dopants,” Phys. Status Solidi (B), vol. 254, p. 1600722, 2017, https://doi.org/10.1002/pssb.201600722.Search in Google Scholar

[26] W. Lu, Y. Miyamoto, R. Okuda, et al., “Correlation between optical and structural characteristics in coaxial GaInN/GaN multiple quantum shell nanowires with AlGaN spacers,” ACS Appl. Mater. Interfaces, vol. 12, no. 45, pp. 51082–51091, 2020. https://doi.org/10.1021/acsami.0c15366.Search in Google Scholar

[27] W. Lu, M. Terazawa, D.-P. Han, et al.., “Structural and optical impacts of AlGaN undershells on coaxial GaInN/GaN multiple-quantum-shells nanowires,” Nanophotonics, vol. 9, pp. 101–111, 2020.Search in Google Scholar

[28] W. Lu, N. Sone, N. Goto, et al.., “Effect of AlGaN undershell on the cathodoluminescence properties of coaxial GaInN/GaN multiple-quantum-shells nanowires,” Nanoscale, vol. 11, pp. 18746–18757, 2019, https://doi.org/10.1039/c9nr07271c.Search in Google Scholar

[29] W. Lu, N. Goto, H. Murakami, et al.., “Controlled synthesis of nonpolar GaInN/GaN multiple-quantum-shells on GaN nanowires by metal-organic chemical vapour deposition,” Appl. Surf. Sci., vol. 509, p. 145271, 2020, https://doi.org/10.1016/j.apsusc.2020.145271.Search in Google Scholar

[30] A. K. Rishinaramangalam, M. Nami, D. M. Shima, G. Balakrishnan, S. R. J. Brueck, and D. F. Feezell, “Reduction of reverse-leakage current in selective-area-grown GaN-based core–shell nanostructure LEDs using AlGaN layers,” Phys. Status Solidi (A), vol. 214, p. 1600776, 2017, https://doi.org/10.1002/pssa.201600776.Search in Google Scholar

[31] M. Nami, I. E. Stricklin, K. M. DaVico, et al.., “Carrier dynamics and electro-optical characterization of high-performance GaN/InGaN core–shell nanowire light-emitting diodes,” Sci. Rep., vol. 8, p. 501, 2018, https://doi.org/10.1038/s41598-017-18833-6.Search in Google Scholar

[32] C.-G. Tu, C.-H. Liao, Y.-F. Yao, et al.., “Regularly patterned non-polar InGaN/GaN quantum-well nanorod light-emitting diode array,” Opt. Express, vol. 22, p. A1799, 2014, https://doi.org/10.1364/oe.22.0a1799.Search in Google Scholar

[33] C. Mounir, T. Schimpke, G. Rossbach, A. Avramescu, M. Strassburg, and U. T. Schwarz, “Polarization-resolved micro-photoluminescence investigation of InGaN/GaN core–shell microrods,” J. Appl. Phys., vol. 121, p. 025701, 2017, https://doi.org/10.1063/1.4973899.Search in Google Scholar

[34] M. Musolino, D. Van Treeck, A. Tahraoui, et al.., “A physical model for the reverse leakage current in (In,Ga) N/GaN light-emitting diodes based on nanowires,” J. Appl. Phys., vol. 119, p. 044502, 2016, https://doi.org/10.1063/1.4940949.Search in Google Scholar

[35] Y. Huang, F. Yun, Y. Li, et al.., “Defect-induced color-tunable monolithic GaN-based light-emitting diodes,” Appl. Phys. Express, vol. 7, p. 102102, 2014, https://doi.org/10.7567/apex.7.102102.Search in Google Scholar

[36] N. Sone, A. Suzuki, H. Murakami, et al.., “Improved uniform current injection into core–shell‐type GaInN nanowire light‐emitting diodes by optimizing growth condition and indium‐tin‐oxide deposition,” Phys. Status Solidi (A), vol. 217, p. 1900715, 2020, https://doi.org/10.1002/pssa.201900715.Search in Google Scholar

[37] K. Choi, M. Arita, and Y. Arakawa, “Selective-area growth of thin GaN nanowires by MOCVD,” J. Cryst. Growth, vol. 357, pp. 58–61, 2012, https://doi.org/10.1016/j.jcrysgro.2012.07.025.Search in Google Scholar

[38] J. L. Lyons, A. Janotti, and C. G. Van de Walle, “Carbon impurities and the yellow luminescence in GaN,” Appl. Phys. Lett., vol. 97, p. 152108, 2010, https://doi.org/10.1063/1.3492841.Search in Google Scholar

[39] J. Neugebauer and C. G. Van de Walle, “Gallium vacancies and the yellow luminescence in GaN,” Appl. Phys. Lett., vol. 69, pp. 503–505, 1996, https://doi.org/10.1063/1.117767.Search in Google Scholar

[40] E. A. Evropeitsev, D. R. Kazanov, Y. Robin, et al., “State-of-the-art and prospects for intense red radiation from core–shell InGaN/GaN nanorods,” Sci. Rep., vol. 10, p. 19048, 2020. https://doi.org/10.1038/s41598-020-76042-0.Search in Google Scholar

[41] B. O. Jung, S. Y. Bae, S. Lee, et al.., “Emission characteristics of InGaN/GaN core-shell nanorods embedded in a 3D light-emitting diode,” Nanoscale Res. Lett., vol. 11, p. 215, 2016, https://doi.org/10.1186/s11671-016-1441-6.Search in Google Scholar

[42] S. Tomiya, S. Goto, M. Takeya, and M. Ikeda, “Structural defects in Mg-doped GaN and AlGaN grown by MOCVD,” MRS Online Proceedings Library Archive, vol. 743, 2002, https://doi.org/10.1557/proc-743-l12.7.Search in Google Scholar

[43] P. Vennéguès, M. Benaissa, B. Beaumont, et al.., “Pyramidal defects in metalorganic vapor phase epitaxial Mg doped GaN,” Appl. Phys. Lett., vol. 77, pp. 880–882, 2000, https://doi.org/10.1063/1.1306421.Search in Google Scholar

[44] M. A. Johar, H.-G. Song, A. Waseem, et al.., “Universal and scalable route to fabricate GaN nanowire-based LED on amorphous substrate by MOCVD,” Appl. Mater. Today, vol. 19, p. 100541, 2020, https://doi.org/10.1016/j.apmt.2019.100541.Search in Google Scholar

[45] S. F. Chichibu, K. Shima, K. Kojima, et al.., “Large electron capture-cross-section of the major nonradiative recombination centers in Mg-doped GaN epilayers grown on a GaN substrate,” Appl. Phys. Lett., vol. 112, p. 211901, 2018, https://doi.org/10.1063/1.5030645.Search in Google Scholar

Received: 2021-05-02
Accepted: 2021-08-10
Published Online: 2021-08-26

© 2021 Sae Katsuro et al., published by De Gruyter, Berlin/Boston

This work is licensed under the Creative Commons Attribution 4.0 International License.