1. Introduction

InGaN light-emitting diodes (LEDs) can cover the entire range of visible wavelengths, thereby attracting significant attention for applications such as color displays and general lighting devices. The monolithic integration of emission wavelengths allows for full color displays with multicolor LEDs as individual pixels. Monolithic multi-wavelength or white LEDs have been demonstrated using two or three vertically integrated quantum wells in an LED structure on the same substrate; however, the limited carrier diffusion length causes difficulties with controlling the emission spectra [1,2]. In principle, the emission of multiple wavelengths from a single GaN-based chip could be achieved by vertically stacked emitting layers with different band gaps. To date, such a structure cannot be realized because of the lack of workable tunneling junctions [3]. Monolithic integration could be performed by laterally connected LEDs with different emitting layers, i.e., emission wavelengths; this set-up requires numerous runs of crystal growth on the same substrate. However, the in-plane modulation of the epitaxial structure using the selective area growth (SAG) technique is likewise a challenge. To overcome these inherent material problems and achieve multi-wavelength LEDs, InGaN/GaN multi-quantum wells (MQWs) were grown on a GaN template with microfacets using the SAG technique [4]. This technique is of particular interest because semi-polar/non-polar GaN planes can be formed on a c-plane polar substrate, thereby providing GaN films with different facets and growth rates that depend on the direction or shape of the opening mask. The growth mechanism of semi-polar/non-polar GaN can be described by gas phase diffusion and surface migration processes, as established by SAG [5]. The formation of different shapes of the semi-polar/non-polar GaN can be understood by considering the different growth rates in the directions of the a-, m-, and c-axes, which can be influenced by the metal-organic vapor-phase epitaxy (MOVPE) growth conditions [6–9]. The SAG of InGaN/GaN MQWs on a patterned c-plane GaN template caused the in-plane wavelength modulation, ranging from 371 nm to 438 nm [10]. For InGaN emissions at this range, the emission wavelength is apparently influenced by the thickness modulation of the InGaN quantum wells. Furthermore, SAG can be used to produce microstructures such as pyramids, quantum dots, and quantum wires that can produce high In-containing InGaN layers; the resulting multiple-wavelength emission can then be used to construct green and white LEDs [11–15]. Mask layers such as a SiO₂ film were selectively formed on the GaN templates before epitaxial growth to achieve SAG during growth. The SiO₂ mask layer is generally embedded under the regrown layers. The embedded insulating layer may produce LEDs with poor electrical properties, such as high series resistance. In this study, we demonstrated that InGaN/GaN MQWs emitted multiple wavelengths from microfacets grown on n-GaN templates with selective-area Si ion implantation. The dosage of Si ions used in this study was 1 × 10¹⁶ cm^–2 to selectively create shallow lattice-distortion areas on the n-GaN templates. The Si-implanted GaN (SIG) layer has low resistivity [16]. Therefore, this layer would not impede current conduction, unlike the aforementioned SiO₂ mask layer in GaN-based LEDs [15]. During growth of the InGaN/GaN LED structures on the SIG templates, the n-GaN epitaxial layer is initially grown on the implantation-free area, whereas the deposition of n-GaN on the Si-implanted regions did not occur because the surface layer of the implanted regions was amorphorized by high-dose implantation [16,17]. Selective growth occurred in the regrown GaN layer on the SIG templates, and the n-GaN bump array featured the formation of a series of truncated hexagonal pyramids (THP). Finally, the LED structure composed of MQWs was grown on the non-planar surface to create the emitting layers on the semi-polar facets and the c-plane facet.

2. Experiments

The Si-doped GaN (n-GaN) epitaxial layers used in this study were initially grown on c-faced (0001) sapphire substrates in a vertical metal-organic vapor-phase epitaxy (MOVPE) system. The SiO₂ and Al layers with thicknesses of 90 and 200 nm, respectively, were deposited in sequence on the wafers with a 3 μm-thick n-GaN epitaxial layer. Circular dots with diameters of 3 μm were defined by photolithography of the Al layer to create the selectively implanted regions. This SiO₂ layer leads to implanted Si ions that accumulate next to the surface of the n-GaN layer. The schematic mask layered structure is shown in Fig. 1(a). Before growing the LED epitaxial structures, Si-ion implantation at 70 keV with a dosage of 1 × 10¹⁶ cm^–2 was performed on the wafers to prepare the SIG templates. The mask layers produced selective implantation to form circular implantation-free area on the n-GaN layer, as shown in Fig. 1(b). The circular implantation-free area had diameter of 3 μm, and the spacing between the circular areas was 3 μm. After ion implantation, the SiO₂ layer was removed using the buffer oxide etchant. The SIG templates were then loaded into the MOVPE chamber to regrow epitaxial layers of the LED structure. The epitaxial layers included a 1.7 μm-thick Si-doped n-GaN layer grown at 1050 °C, a ten-pair In_0.2Ga_0.8N/GaN MQW structure grown at 750 °C, a 0.03 μm-thick Mg-doped p-Al_0.12Ga_0.85N electron blocking layer, and a 0.15 μm-thick Mg-doped p-GaN top contact layer grown at 950 °C. The schematic structure of the regrown layer is shown in the inset of Fig. 1(c). The carrier concentrations of the n-GaN and p-GaN layers were approximately 8 × 10¹⁸ cm^–3 and 5 × 10¹⁷ cm^–3, respectively. A heavily Si-doped n⁺-InGaN top layer was then grown on the p-GaN contact layer [18]. Its structural properties were characterized by microscopy; its optical and electrical properties were also discussed.

 

Fig. 1 Schematic illustrations of (a)Al and SiO₂ were served as shadow mask and ion-stopping layers, respectively.(b) an n-GaN template with selective-area Si implantation (c) LED structure with truncated hexagonal pyramid array.

Download Full Size | PDF

3. Results and discussions

A typical scanning electron microscope (SEM) image of an n-GaN layer regrown on the SIG template shows the THP array on the n-GaN layer [Fig. 2(a)]. The formation of THP could be attributed to the initial nucleation of GaN on the implantation-free regions rather than on the Si-implanted regions. That is, the GaN growth rate on the implanted regions was markedly lower than that on the implantation-free regions. The discrepancy in the growth rate could be attributed to fact that GaN epitaxial layer was difficult to grow on the amorphous surface layer caused by the high-dose implantation. In general, the crystal structure of GaN layer subjected to high doses and/or energy levels of ion bombardment produces an amorphous layer [16]. According to previous studies, the critical dose for the amorphorization of GaN was approximately ~10¹⁶ cm^–2, and the lattice constant of the Si-implanted GaN increases with the implantation dose and/or energy [16,19]. In this study, the implanted Si ions created a damage layer with a depth of approximately 50 nm from the surface, as indicated by the simulation results. A typical SEM image of a LED structure grown on the SIG templates is shown in Fig. 2(b). The THP array can be clearly observed on the regrown wafer after the growth of the LED structure because the total thickness of the LED epitaxial layers, including the MQW layers, was not enough to evenout the depressions between GaN THPs. As shown in Fig. 2(b), the LED structure grown on the GaN THP array was composed of the semi-polar and the c-plane (0001) facets. To determine the crystal planes of the semi-polar facets, cross-sectional SEM images were evaluated; the angles between the semi-polar and the c-plane facets in the images were measured. The semi-polar facets were determined to be $\left\{1\overline{1}01\right\}$ planes [20]. The average heights of the THPs were approximately twice as large as those of the reference samples, i.e., the epitaxial layers grown on implantation-free GaN templates. This difference was due to the loading effect during the SAG process. As shown in Figs. 2(a) and 2(b), the average spacing between THPs decreased from 2 μm to 0.2 μm after the regrowth of the LED structure. Therefore, the significant lateral growth reduced the size of the depression and/or allowed the neighboring inclined facets to reach each other. A typical cross-sectional image of an LED with a THP array in the $\left\{11\overline{2}0\right\}$ direction is shown in Fig. 3(a), which was obtained by transmission electron microscopy (TEM). The V-shaped depression indicated that the TEM image illustrated a local region between two THPs. A lateral dark line with a width of approximately 50 nm was observed under the V-shaped depression. This dark line was attributed to the Si-implantation-induced damaged layer. Its line width was consistent with the estimates of simulation using the TRIM (Transport of Ions in Matter) program. To investigate the evolution of dislocations (TDs) around a Si-implanted region, an enlarged TEM image was inspected at a local area around the said implanted region. As shown in Fig. 3(b), the dislocations were vertically extended from the implantation-free GaN template layer to the regrown surface layer. Aside from a relatively lower TD density, the TDs over the implanted region appeared bent. The TD bending was attributed to the occurrence of lateral growth during the regrowth of epitaxial layers over the implanted regions. The reduced TD density could be attributed to the Si-implantation-induced damaged layer, which had similar functions as the dielectric layer used in the epitaxial lateral overgrowth (ELOG) of GaN to interrupt the propagation of TDs along the growth direction from the GaN template layer to the regrown layers [21]. Considering the selected-area electron diffraction patterns determined around the implanted region, as shown in Fig. 3(c), one can see that the implanted GaN exhibited the amorphous phase, and the implanted region was not recovered to a single crystal phase after the regrowth. In addition, Fig. 3(c) also shows that the regrown GaN above the implanted region exhibited a single crystal phase.

 

Fig. 2 Typical SEM images of (a) an n-GaN layer regrown on the Si-implanted GaN templates (b) LED structure with truncated hexagonal pyramid array.

Download Full Size | PDF

 

Fig. 3 Typical cross-sectional TEM images taken (a) between two THPs near the valley in the $\left[11\overline{2}0\right]$direction (b) an enlarged inspection taken from a local area of Fig. 3(a) (c) selective-area electron diffraction patterns determined around the implanted region.

Download Full Size | PDF

Photographs of the LED wafer with the THP arrays (THP LEDs) under different injection currents are shown in Fig. 4(a). The indium balls were attached on the wafer to serve as the n- and p-type contact electrodes. The emitted light clearly changed from green to blue when the current was increased from 5 mA to 100 mA. The electroluminescence spectra of the THP LEDs driven under different currents are shown in Fig. 4(b). At a relatively low current of 5 mA, the THP LEDs produced broad spectra that peaked at approximately 500 nm (green peak).

 

Fig. 4 (a) Typical images taken from the THP LEDs under different current injection (b) typical EL spectra of the THP LEDs driven under different currents.

Download Full Size | PDF

The blue peak at approximately 450 nm emerged at higher energies when the injection currents were higher than or equal to 50 mA. The green peak was blue-shifted with the increasing injection current, as shown in Fig. 4(a). However, the blue peak was nearly independent of the injection current. This observation indicates that the two emission peaks may arise from different areas within each THP. To clarify the origin of the blue and green emissions within the pyramidal structure, spatially resolved cathodoluminescence (CL) measurements were performed under room temperature. Typical CL spectrum taken under an electron beam current of 300 pA and an accelerate voltage of 10 keV is shown in Fig. 5(a), whereas the mapped SEM image with a frame of 20 μm × 15 μm is shown in Fig. 5(b). The monochromatic emissive CL images at different peaks in Fig. 5(a) are individually presented in Figs. 5(c) to 5(f). As shown in Fig. 5(c), the monochromatic CL image at 365 nm presents the emissions from the lower sidewall facets and near the V-shaped depression; this phenomenon was attributed to the band-edge related emission of GaN [22,23]. The monochromatic CL image at 400 nm [Fig. 5(d)] shows the emissions mainly from the top facet of each THP; these emissions were attributed to the Mg-related deep-level emissions of p-GaN [24].

 

Fig. 5 (a)Typical CL spectrum taken under an electron beam current of 300 pA and an accelerate voltage of 10 keV (b) SEM mapped image with a frame of 20 × 15μm² (c) corresponded monochromatic emissive CL image at 365 nm (d) 400 nm (e) 433 nm (f) 500 nm.

Download Full Size | PDF

By contrast, Fig. 5(e) shows the monochromatic CL image at 433 nm, with emissions from the sidewall facets of THPs. Meanwhile, the CL image at 500 nm had emissions mainly from the V-shaped depressions between THPs, as shown in Fig. 5(f). The blue and green emissions were attributed to the InGaN/GaN MQWs grown at semi-polar facets of each THP and in the V-shaped depressions between THPs, respectively. As displayed in Fig. 5, the well thickness and/or the In content of the MQW structure grown on the semi-polar facets of each THP and in the V-shaped depressions between THPs were different. The local InGaN well thickness was determined by TEM for the semi-polar facets and the c-axis plane; these values were estimated to be approximately 2.2 and 3.5 nm, respectively. In addition to the difference on well thicknesses, the preliminary results indicated the in content in the semi-polar QWs was lower than in the c-axis plane. The quantitative difference in the well thickness (or growth rate) and the In content of the semi-polar facets and the c-axis plane is dependent on the growth conditions, such as temperature and pressure. The trends of the growth rate and the In incorporation efficiency observed in this study are consistent with a previous report by Funato et al. [25]. The facet control can be easily conducted by tuning the growth conditions during the regrowth of epitaxial layers. Therefore, the Si-implanted GaN template is a potential choice for fabricating GaN-based LEDs that emit multiple peaks by a single chip. Thus, white light emission may be achieved without the use of phosphors.

4. Conclusions

In summary, we have demonstrated the selective-area regrowth of Si-implanted GaN template layers to form the truncated hexagonal pyramid arrays for achieving non-planar InGaN/GaN MQWs in GaN-based LEDs. Contrary to the conventional designs that use a dielectric layer as the mask layer, the Si-implanted GaN template featuring a planar surface and the low-resistivity mask layer could prevent devices from high resistance. On the other hand, EL and CL measurements showed that the broad emission spectra with multiple peaks could be achieved using a set of MQWs grown at fixed conditions. The broad emission spectra were due to the spatial variation of the well thicknesses and the In content in MQWs on the semi-polar facets or c-planes.