InGaN platelets: synthesis and applications towards green and red

§Center for Analysis and Synthesis/nCHREM, Lund University, Box 124, S-221 00 Lund ...... L. Reine Wallenberg and B. Jonas Ohlsson joined result disc...
0 downloads 0 Views 938KB Size
Subscriber access provided by KEAN UNIV

Communication

InGaN platelets: synthesis and applications towards green and red light emitting diodes Zhaoxia Bi, Filip Lenrick, Jovana Colvin, Anders Gustafsson, Olof Hultin, Ali Nowzari, Taiping Lu, L Reine Wallenberg, Rainer Timm, Anders Mikkelsen, B. Jonas Ohlsson, Kristian Storm, Bo Monemar, and Lars Samuelson Nano Lett., Just Accepted Manuscript • DOI: 10.1021/acs.nanolett.8b04781 • Publication Date (Web): 02 Apr 2019 Downloaded from http://pubs.acs.org on April 2, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

InGaN platelets: synthesis and applications towards green and red light emitting diodes Zhaoxia Bi*,†, Filip Lenrick‡, Jovana Colvin‡, Anders Gustafsson†, Olof Hultin†,||, Ali Nowzari†, Taiping Lu†,||, Reine Wallenberg§, Rainer Timm‡, Anders Mikkelsen‡, B. Jonas Ohlsson†, Kristian Storm||, Bo Monemar† and Lars Samuelson† †Division

of Solid State Physics and NanoLund, Department of Physics, Lund University, Box

118, S-221 00 Lund, Sweden ‡Division

of Synchrotron Radiation Research and NanoLund, Department of Physics, Lund

University, Box 118, S-221 00 Lund, Sweden §Center

for Analysis and Synthesis/nCHREM, Lund University, Box 124, S-221 00 Lund,

Sweden ||RISE

Acreo AB, Lund, Sweden

KEYWORDS: InGaN, platelets, light emitting diodes, selective area growth, metal organic vapor phase epitaxy

ABSTRACT: In this work, we present a method to synthesize arrays of hexagonal InGaN submicron-platelets with a top c-plane area having an extension of a few hundred nanometers by selective area metal organic vapor phase epitaxy. The InGaN platelets were made by in-situ annealing of InGaN pyramids, whereby InGaN from the pyramid apex was thermally etched away, leaving a c-plane surface, while the inclined {1011} planes of the pyramids were intact. The as-formed c-planes, which are rough with islands in a size of a few tens nanometers, can be flattened with InGaN regrowth, showing single bilayer steps and high-quality optical properties (full width at half maximum of photoluminescence at room temperature: 107 meV for In0.09Ga0.91N and 151 meV for In0.18Ga0.82N). Such platelets offer surfaces having relaxed lattice constants thus enabling shifting the quantum well emission from blue (as when grown on GaN) to green and red. For single InGaN quantum wells grown on the c-plane of such InGaN platelets, a sharp interface between the quantum well and the barriers was observed. The emission energy from the quantum well, grown at the same conditions, was shifted from 2.17 eV on In0.09Ga0.91N platelets to 1.95 eV on In0.18Ga0.82N platelets, due to a thicker quantum well

ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

and reduced indium pulling effect on In0.18Ga0.82N platelets. Based on this method, prototype light emitting diodes were demonstrated with green emission on In0.09Ga0.91N platelets and the red emission on In0.18Ga0.82N platelets.

Blue GaN light emitting diodes (LEDs), typically comprising InGaN quantum wells (QWs) with indium contents around 15% on a GaN buffer layer, can be achieved with an external quantum efficiency higher than 80%.1,2 In order to realize white solid state lighting with nitride LEDs of all three primary colors, highly efficient green and red nitride LEDs are desirable.3 Micro-sized red nitride LEDs may have a unique potential for achieving full color microLED displays, due to the much lower surface recombination than the AlInGaP normally used for red LEDs. Meanwhile, nitrides have shorter minority carrier diffusion lengths, compared with the AlInGaP, also contributing to the less recombination at the surface.4,5 In order to reach the green and red colors, the indium content in the QWs needs to be increased at least to ~25% and ~35%, respectively.6-9 When grown on GaN, these green and red-emitting InGaN QWs incurs much higher compressive strain than the blue-emitting ones. The large strain has a direct impact on the crystal quality of the QWs, potentially introducing plastic crystal deformation during growth and then leading to dislocation formation.9-11 The strain also induces a piezoelectric polarization for GaN-based heterostructures oriented along [0001] direction, which causes an internal electric field in the InGaN QWs. This introduces the quantum confined Stark effect (QCSE), detrimental on optical properties of InGaN QWs since it spatially separates electron and hole wave functions in the QWs.1,10,12,13 The large strain in the QWs introduced by the high indium content increases the QCSE, which also contributes to the low efficiency of green and red nitride LEDs. One potential solution to the strain issue is to replace the conventional GaN buffer layer with an InGaN layer that is closer in indium content to the InGaN QWs.14-16 Less difference in the indium content means smaller lattice mismatch, leading to a reduced QCSE and less risk for plastic deformation via formation of crystal defects. A reduced QCSE leads to an increased overlap between electron and hole wave functions, and a much higher spontaneous emission rate can be achieved compared to the same InGaN QWs grown on a GaN buffer layer.15,16 Even et al. studied growth dependence of InGaN QWs on the indium content in InGaN pseudosubstrates and found an enhanced indium incorporation into the QWs due to reduced indium pulling effect.17 This enables growth of green and red InGaN QWs at a higher temperature for a higher quality.17,18 In addition, a higher hole concentration is expected in InGaN due to a

ACS Paragon Plus Environment

Page 2 of 20

Page 3 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

lower activation energy for the p-type doping, compared with GaN.19,20 Until now, the growth of thick InGaN layers has not been extensively studied due to the synthesis challenges arising from the phase separation and surface control.21,22 Pantzas et al. reported that insertions of thin GaN interlayers during InGaN growth can suppress the phase separation and improve the surface quality. However, the reported root mean square roughness of 3.1 nm still seems too high for the subsequent QW growth.23 To improve the growth control on the surface quality, bulk GaN substrates, with much lower dislocation density than GaN films grown on sapphire or other dissimilar substrates, were also used for the InGaN film growth.24,25 Senda et al. realized high quality InGaN film growth on grooved m-plane GaN films, on which InGaN QWs showed much better optical emission properties than the reference ones grown on a GaN buffer layer.26 Besides metal organic vapor phase epitaxy (MOVPE), halide vapor phase epitaxy and molecular beam epitaxy were also applied to grow thick InGaN films.27,28 In this work, we present a method to synthesize arrays of InGaN sub-micron-platelets by selective area MOVPE plus a re-formation process step that turns the pyramidal shape into a coriented platelet. The InGaN hexagonal platelets have a thickness of 100-200 nm and a top cplane with a size of a few hundred nanometers. The platelets are prepared by in-situ annealing of hexagonal c-oriented InGaN pyramids, defined by six {1011} planes. In the annealing step, the pyramids are etched down from the apex, resulting in the formation of a c-plane top. The rough c-plane formed by the annealing can be flattened by InGaN regrowth, showing single bilayer steps. The InGaN platelets show narrow photoluminescence (PL) peaks with full widths at half maximum (FWHM) of 107 meV for In0.09Ga0.91N and 151 meV for In0.18Ga0.82N. It was also found that the use of an InGaN buffer layer enhances the indium incorporation into the InGaN QW due to the reduced compressive strain, as discussed above. Based on this approach, we demonstrate prototype QW LEDs grown on such InGaN platelets, emitting green on In0.09Ga0.91N and red on In0.18Ga0.82N at room temperature. The growth procedures are shown in Figure 1 for the synthesis of the InGaN platelets and single QW LED structures on them. It starts from arrays of InGaN pyramids selectively grown on patterned (0001) GaN/Si substrates with SiN as a growth mask (Figure 1a). The InGaN pyramid growth by MOVPE was reported in an earlier work.29 After the growth of InGaN pyramids, in-situ annealing was conducted under NH3 ambient at a temperature of 1070 ºC. Without the NH3 during the annealing, the InGaN pyramids would decompose completely. With the NH3 protection, the InGaN pyramids decomposed from the top apex, leading to the formation of a top c-plane, while the six inclined {1011} facets of the pyramids were left intact.

ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

With continuing annealing, the c-plane increased in size with the etching moving downward as shown in Figure 1b. The c-plane surface of this bottom InGaN layer remained from the annealing was rough with islands in a size of a few tens nanometers as shown in Figure 2b and 2e. InGaN regrowth was done on the bottom InGaN layer to flatten the c-plane and enhance the quality of InGaN (Figure 1c), including an intermediate InGaN layer first and then a top InGaN layer. This was followed by the growth of an InGaN single QW on the c-plane (Figure 1d) and p-InGaN (Figure 1e) to realize a prototype LED structure based on such InGaN platelet templates.

Figure 1 Procedures to synthesize InGaN platelets and LED structures. (a) InGaN pyramids grown by selective area MOVPE, as reported in an earlier work.29 (b) Bottom InGaN layer with a top c-plane formed by etching down from the pyramid apex using in-situ annealing. (c) InGaN regrowth to flatten the rough c-plane formed in (b). An intermediate layer of InGaN is grown first with an indium content below 5%, as shown in Figure S1 in the Supporting Information. Then, a top InGaN layer is grown with indium contents similar to the bottom InGaN layer in (b). (d) InGaN single QW growth on the c-plane of the InGaN platelet templates. (e) A prototype LED structure with p-InGaN grown above the single QW, with definitions of all InGaN layers.

Figure 2 shows the as-grown InGaN pyramids (2a and 2d) and the morphology after the annealing (2b and 2e) for two indium contents of 10% and 17%. For both indium contents, the diagonal base size was about 630 nm and the same annealing behavior was observed. As mentioned above, a c-plane top surface was formed due to the etching from the top apex by the high temperature annealing. After the annealing, the remaining {1011} facets were still smooth and the diagonal base size showed no change compared with the as-grown InGaN pyramids.

ACS Paragon Plus Environment

Page 4 of 20

Page 5 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Surface of {1011} planes are N-terminated. N atoms at the surface are supposed to bond with H atoms cracked from NH3,30,31 which can be a reason for the stable {1011} planes during the annealing. PL spectra were measured at room temperature before and after the annealing, shown in Figure 2c and 2f. After annealing, the PL intensity was lower, but the peak positions did not present any visible shift, indicating similar indium content. Note that indium contents of 10% and 17% were obtained based on the positions of the PL peaks.32 This indium content could be overestimated since it is known that the PL emissions are usually from localized, indium-rich radiative centers.33 This will be discussed below together with the indium content measurement with energy dispersive x-ray spectroscopy (EDS).

Figure 2 InGaN pyramids before and after the in-situ annealing with two different indium contents. (a, d) SEM images of the as-grown InGaN pyramids. (b, e) SEM images of the InGaN pyramids after the annealing. (c, f) PL spectra of the InGaN pyramids before and after the annealing measured at room temperature. The indium contents of 10% and 17% were obtained based on these PL spectra.

ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 20

Regrowth of InGaN directly on the bottom InGaN layer remained from the annealing step, with PL emission at the same wavelength range, can flatten the c-plane surface. However, pits were observed after the regrowth on the c-plane. To suppress the pit formation, an intermediate layer of InGaN had to be grown directly after the annealing, as shown in Figure S1 (see the Supporting Information). This intermediate layer was grown with a low V/III ratio of 700, which is close to the limit for indium and gallium droplet formation. These conditions may lead to a metal-rich surface and longer adatom diffusion length, making it efficient to flatten the rough surface.28 The indium content in this layer is less than 5%. On top of this intermediate InGaN layer, a top InGaN layer was grown with the similar PL emission energy as the bottom InGaN layer. Figure 3 shows the growth results with two different indium contents of 9% and 18%. From the images by scanning electron microscopy (SEM) in Figure 3a and 3e, the c-plane surface is smooth for both samples. For the In0.09Ga0.91N platelets, no pits were observed. However, pits can occasionally be seen for the In0.18Ga0.82N platelets (such as the one indicated by the white arrow and shown by the inset in Figure 3e) most probably because of the large lattice mismatch to the underlying intermediate layer. Due to the extremely low growth rate of

Figure 3 InGaN platelets with indium contents of 9% and 18% formed by regrowth on the annealed InGaN pyramids. (a, e) SEM images of the InGaN platelets showing smooth top c-plane surface. Inset in (e) shows a pit indicated by the white arrow. (b, f) AFM height images of the top c-plane

surface.

Surface

steps

were

observed for both samples. Height profiles along the black lines in (b, f) are presented in (c, g), showing single bilayer steps on cplane. (d, h) PL spectra of the InGaN platelets measured at room temperature, based on which the indium contents of 9% and 18% were obtained.

ACS Paragon Plus Environment

Page 7 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

the inclined {1011} planes, the InGaN regrowth mainly took place on the top c-plane and this top InGaN layer is about 40-50 nm thick for both samples. Figure 3b and 3f show atomic force microscopy (AFM) images on the top c-planes. Surface steps were observed for both samples with the step width of about 70 nm. Based on the height profiles shown on Figure 3c and 3g, the observed step heights amount to 0.30±0.05 nm for In0.09Ga0.91N platelet and 0.32±0.07 nm for In0.18Ga0.82N one, corresponding to a single bilayer thickness on c-plane. The surface steps propagate along the 〈1120〉 direction (from corner to corner in platelets) with jagged step edges due to the alternative switch between type A (two dangling bonds per edge atom) and type B (one dangling bond per edge atom) step edges.34,35 This step orientation is believed to be related to the preferential nucleation at the corners between the top c-plane and two adjacent {1011} planes. This was observed during the intermediate layer growth as shown in Figure S1. PL measurements were conducted for both samples at room temperature (Figure 3d and 3h). Both PL spectra show similar peak positions as the original InGaN pyramids, but with much smaller FWHMs. The indium contents in Figure 3 were also obtained based on these two PL spectra.32 The results in Figure 3 show that InGaN platelets with a smooth top surface and good optical properties were prepared.

Figure 4 Single QW of InGaN grown at the same conditions on the platelets of In0.09Ga0.91N and In0.18Ga0.82N. (a, e) Cross-sectional HAADF-STEM images of the single QW samples. The white arrow

ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

in (e) indicates a dislocation. EDS measurements were conducted on the bottom InGaN layer (marked as 1) and the top InGaN layer (marked as 2) and the data are shown in Table 1. (b, f) are the magnified HAADF-STEM images close to the periphery, and (c, g) are high resolution TEM images recorded in the middle of the QWs. (d, h) CL spectra measured at 10 K, showing the emissions from the single QWs on both samples. Insets illustrate SEM images and corresponding monochromatic CL images with the emissions from the QW. CL images were recorded with an energy window of 1.80-2.5 eV in (d) and 1.85-2.35 eV in (h).

Two samples with a single InGaN QW were grown on such InGaN platelets with indium contents of 9% and 18%. On top of the QW, an InGaN cap layer was grown at the same conditions as the top InGaN layer and the thickness of the cap layer was around 20 nm. Figure 4 shows the cross-sectional transmission electron microscopy (TEM) images and optical measurements by cathodoluminescence (CL) at 10 K. Figure 4a and 4e are overview high-angle annular dark-field scanning TEM (HAADF-STEM) images of the QW on both In0.09Ga0.91N and In0.18Ga0.82N platelet templates. The single QW can be clearly seen as a white line at the top of the structures. In total, 5 platelets from each sample were characterized using TEM. For the QW on In0.09Ga0.91N, no dislocation was observed in the platelet structures. This was confirmed by STEM imaging at low-angle annular dark-field conditions, which is sensitive to crystallography information (see Figure S2 in the Supporting Information). However, dislocations were formed for the QW sample on In0.18Ga0.82N (indicated by the white arrow in Figure 4e), which originated from the interface between the intermediate layer and the bottom InGaN layer due to the larger lattice mismatch. This dislocation propagated to the top surface through the QW. The single QW is uniform in thickness with sharp interfaces, as shown in Figure 4c and 4g. The only exceptions are close to the periphery, where the QW becomes thinner until it disappears completely at the surface (Figure 4b and 4f). The thinner QW can reduce the carrier injection to the periphery due to the higher quantization energy, possibly leading to a reduced influence from surface recombination on the {1011} facets. Figure 4b and 4f do not show any clear QW growth on the inclined {1011} facets. The QW and the cap layer (and also p-InGaN as shown below) grow mainly on the c-plane, thereby rebuilding the pyramid shape. The slow growth rate on the {1011} facets can be explained by the above-mentioned Hpassivation which inhibits adsorption of Ga atoms to the {1011} facets.30,31 For all TEM specimens, the indium contents were measured with EDS on the bottom InGaN layers (marked as 1) and the top InGaN layers (marked as 2) in Figure 4a and 4e. Layers 1 and 2 were calibrated to have similar PL emission energy with separate growth for both InGaN platelets, which means that layers 1 and 2 are supposed to have similar indium contents for each

ACS Paragon Plus Environment

Page 8 of 20

Page 9 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

platelet sample. The indium contents obtained by PL and EDS for layers 1 and 2 are listed in Table 1 for both types of platelet samples. Note that the data from PL are based on the calibration samples, rather than the ones in this figure. The indium contents from the EDS measurements for layer 1 are much lower than the ones obtained from the PL data, especially for the In0.18Ga0.82N sample. PL and EDS show consistent indium content in layer 2 for both platelet samples. The obtained EDS data were based on the measurements on all platelets in the TEM analysis. Indium content fluctuation is well known for the InGaN growth, which can form localized, indium-rich radiative centers.33 Excited carriers can diffuse to such indium-rich centers and recombine there, leading to an overestimation of the indium content by PL. This could explain the large indium content discrepancy in layer 1 between PL and EDS data. The InGaN layer 2 show much better quality with reduced indium content fluctuation based on the consistent PL and EDS data.

Table 1 Indium content comparison between layers 1 (bottom InGaN layer) and 2 (top InGaN layer) based on the PL and EDS data. Note that indium contents obtained from PL are based on the calibration samples, rather than the ones in Figure 4.

Figure 4d and 4h show typical low temperature CL emission from single QWs of these two samples. The QWs on In0.09Ga0.91N and In0.18Ga0.82N platelets were grown at the same conditions, including the growth time. The emission peak was shifted from 2.17 eV (yellow) on the In0.09Ga0.91N platelet to 1.95 eV (red) on the In0.18Ga0.82N platelet. The c-plane size of In0.18Ga0.82N platelet before the QW growth is slightly smaller than on the In0.09Ga0.91N, resulting in a thicker QW on In0.18Ga0.82N (5.8±0.4 nm) than the one on In0.09Ga0.91N (4.5±0.2 nm). The thicker QW can lead to a red-shift of the emissions due to the QCSE.1 Meanwhile, the indium content in the QW grown on In0.18Ga0.82N platelets could be higher due to reduced indium pulling effect, compared with the QW grown on In0.09Ga0.91N platelets.17,18 This also contributes to the red-shift of the emissions from the QW on In0.18Ga0.82N platelets. The insets in Figure 4d and 4h are the monochromatic CL images of the QW emission (yellow or red depending on the emission wavelength) together with the SEM images. In general, the emission is distributed uniformly on the QW area. The dark line features in the CL images could be due

ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

to the step bunches formed during the intermediate layer growth as shown in Figure S1.24 CL spectra covering the emissions from the single QWs, underlying InGaN platelets and GaN buffer layer is shown in Figure S3 in the Supporting Information. Based on the results above, a p-InGaN layer was grown on top of the single QW and prototype LED devices were fabricated. A sketch of the LED devices is shown in Figure 5a. The n-type contact was made to the underlying GaN buffer layer and the p-type contact was made to the top p-InGaN layer. To electrically isolate the top contact from the QW and n-InGaN, two insulating layers were added: First, a passivating layer of Al2O3 was deposited by atomic layer deposition. Second, a planarizing polymer spacer was used to lift the top contact so that it only connects to the p-InGaN. The polymer spacer was also used as an etch mask to remove the Al2O3 on the p-InGaN. The p-type contact had a circular shape with a diameter of 360 µm, contacting approximately 1.2×105 platelets. Figure 5b and 5c illustrate the electroluminescence (EL) spectra under different current injection levels, showing green emission (2.30 eV) on the In0.09Ga0.91N platelets and red emission (1.98 eV) on the In0.18Ga0.82N platelets. The FWHMs of the spectra at a current density of 70 A/cm2 are 210 meV and 150 meV for the green and red LEDs, respectively. Note that the green and red single QWs were grown at different conditions here, possibly explaining the narrower EL spectra for the red LEDs, compared with the green one. Compared with the CL spectra in Figure 4d and 4h, EL spectra show smaller linewidths. This could be due to a strong carrier injection during the EL measurement, where the emissions are mainly from recombination at ground state while the localized states are saturated. Meanwhile, EL spectra were recorded from QWs on a large number of platelets, in contrast to the CL spectra from a QW on a single platelet. In this way, the EL peak can get magnified with emissions from much larger QW volume, also possibly leading to a smaller linewidth. In summary, we present a method to prepare relaxed InGaN templates offering high quality c-planes with an extension of a few hundred nanometers for long wavelength nitride LEDs. Arrays of such InGaN platelets were synthesized by annealing hexagonal c-oriented InGaN pyramids and subsequent InGaN regrowth. The InGaN platelets have high quality c-planes with single bilayer steps and also show good optical properties. Single QWs emitting long wavelength light were grown on InGaN platelets with a uniform thickness and sharp interfaces. Prototype electrically driven LEDs were demonstrated emitting green on In0.09Ga0.91N and red on In0.18Ga0.82N platelets with InGaN single QWs as the active emitting layer.

ACS Paragon Plus Environment

Page 10 of 20

Page 11 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Figure 5 Process and EL measurements of prototype InGaN platelet LEDs. (a) A schematic of InGaN platelet LEDs and related process for EL measurements. (b, c) EL spectra obtained at different current injection levels for the green LEDs grown on In0.09Ga0.91N platelets and the red ones on In0.18Ga0.82N platelets, respectively. The current density in both legends was calculated according to the actual QW area, rather than the contact area.

EXPERIMENTAL SECTION InGaN platelets were synthesized based on the InGaN pyramids which were grown by MOVPE (in a 3x2-inch close coupled showerhead reactor). Arrays of InGaN pyramids were grown selectively from 100 nm-large openings in a SiN mask deposited on a (0001) GaN/Si substrate by low pressure chemical vapor deposition, which was reported in an earlier work.29 The pitch was 1 µm. After the pyramid growth, in-situ annealing was performed at 1070 ºC to form a c-plane by thermally etching down of the pyramids from the top apex. NH3 flow was 9.5 standard liter per minute during the annealing to protect the inclined {1011} facets of the pyramids. Subsequently, InGaN platelets were made by InGaN regrowth on the as-formed bottom InGaN layer with a firstly grown intermediate layer (see Figure S1 in the Supporting Information). This layer was able to flatten the rough c-plane surface efficiently. The V/III ratio for the top InGaN layer growth after the intermediate layer was high to help the indium incorporation and obtain high quality material. It was 10000 for In0.09Ga0.91N and 4000 for

ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

In0.18Ga0.82N with the same flow of triethygallium (TEG). The much lower V/III ratio for In0.18Ga0.82N is due to its higher flow of trimethyindium (TMI). The TMI/(TMI+TEG) ratio was 0.7 for In0.09Ga0.91N growth and 0.9 for In0.18Ga0.82N growth, respectively. The growth temperature for the top InGaN layer was 770 oC for In0.09Ga0.91N and 755 oC for In0.18Ga0.82N and the growth time was 1800 s to obtain 40-50 nm thick top InGaN layers. SiH4 was used for n-type doping and the doping level was about 1 × 1018 /cm3. The top c-plane of the platelets was about 470 nm large diagonally. Then, a single InGaN QW and p-InGaN were grown on such InGaN platelets successively. The growth temperature for the QW is 675-700 oC. Thanks to the extremely low growth rate on the {1011} planes, practically the regrowth of InGaN layers mainly took place on the c-planes. Thickness of p-InGaN was about 200 nm. Bis(cyclopentadienyl)magnesium was used as the p-type doping and the Mg/Ga ratio in the gas phase was 0.8%. Optical properties were characterized by PL and CL. PL spectra were recorded at room temperature using a micro PL setup with a UV laser working at 375 nm. A 400 nm filter was used for the PL measurements except the one in figure 2c. The laser beam was focused into a spot with a size of several micrometers. Hyperspectral CL measurements were done at 10 K with an acceleration voltage of 5 kV and a current of 20 pA. SEM characterization was done in a Hitachi SU8010 Cold Field Emission SEM at an acceleration voltage of 15 kV. A 300 kV JEOL 3000F TEM was used to characterize the structural quality of the InGaN platelets and the QW in cross section. The imaging was performed under high-resolution and HAADF conditions. EDS measurements were done on thinned TEM specimens during the TEM characterization using a SDD detector from Oxford Instruments. The TEM specimens were prepared by an FEI Nova 600 dual beam FIB/SEM system. A Nanowizard AFM from JPK Instruments was used to characterize the c-plane surface of the InGaN platelets in intermittent contact mode. Processing of the LEDs on the platelets was done in 6 steps. 1) 30 nm of Al2O3 was deposited using atomic layer deposition to provide electrical insulation to the QW and the bottom n-InGaN from the p-InGaN contact. 2) A polymer spacer layer (S1805) was spin coated on the sample so that the LED structures were just barely covered. 3) The spacer layer was then etched down in a RIE O2 plasma process to remove the resist from the top c-plane. 4) The Al2O3 covering the top surface of the p-InGaN was etched away with buffer oxide etch. 5) A thin layer of metal (11 nm thick) was deposited through a shadow mask to form a contact to the p-InGaN. The contact was circular with a diameter of 360 µm. 6) A back n-type contact was made to the

ACS Paragon Plus Environment

Page 12 of 20

Page 13 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

n-GaN underlying the SiN growth mask by mechanically removal of the resist, Al2O3 and SiN outside the platelet arrays. These devices were characterized at room temperature with respect to their current-voltage characteristics and light output. EL spectra were acquired with a fiber spectrometer under different current injection levels.

AUTHOR INFORMATION Corresponding Author *E-mail: [email protected] Author contributions Zhaoxia Bi designed and performed all sample growth, SEM characterizations, PL measurements and wrote the main part of the manuscript. Filip Lenrick prepared TEM specimens with FIB/SEM and conducted all TEM characterizations. Jovana Colvin, Rainer Timm and Anders Mikkelsen conducted AFM measurements on the InGaN platelet top surface. Anders Gustafsson carried out the CL measurements. Olof Hultin, Ali Nowzari and Kristian Storm fabricated the platelet LED devices and carried out the EL measurements. Taiping Lu, L. Reine Wallenberg and B. Jonas Ohlsson joined result discussion. Bo Monemar and Lars Samuelson supervised all the experiments in this work. All co-authors contributed to the revision of this manuscript.

ACKNOWLEDGMENT This work was done within NanoLund and the authors would like to thank the supports from the European Union under the project “NWs4LIGHT” (grant No. 280773), from the Swedish Foundation for Strategic Research under the project “Energy-efficient LED-lighting Based on Nanowires” (grant No. EM11-0015) and from the Swedish Energy Agency under the project “Nanowire-based Light Emitting Diodes with Multiple Wavelengths in the Visible Range” and the EELYS-project “Ultra-efficient RGB-lighting Based on Nanowire Technology”. The work was also supported by the Swedish Research Council, by the Knut and Alice Wallenberg Foundation, and by VINNOVA.

SUPPORTING INFORMATION:

ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure S1 shows how the intermediate InGaN layer flattens the rough top c-plane which is formed by the high temperature annealing. This top c-plane after the intermediate layer growth was characterized with SEM and AFM. PL spectra were also recorded before and after the intermediate layer growth. Figure S2 shows a low-angle annular dark field scanning TEM image of a single QW on In0.09Ga0.91N platelet, illustrating the structure free of dislocations. Figure S3 shows the CL spectra with a broader photon energy range so that the emissions from underlying InGaN layer and GaN buffer layer can be presented together with the emissions from the single InGaN QW.

ACS Paragon Plus Environment

Page 14 of 20

Page 15 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

REFERENCES (1)

Damilano, B.; Gil, B. J. Phys. D: Appl. Phys. 2015, 48, 403001.

(2)

Narukawa, Y.; Ichikawa, M.; Sanga, D.; Sano, M.; Mukai, T. J. Phys. D: Appl. Phys. 2010, 43, 354002.

(3)

Karpov, S. Y. Light-emitting diodes for solid-state lighting: searching room for improvements. Proc. SPIE. 2016, 9768, 97680C.

(4)

Massoubre, D.; Xie, E.; Guilhabert, B.; Herrnsdorf, J.; Gu, E.; Watson, I. M.; Dawson, M. D. Semicond. Sci. Technol. 2014, 29, 015005.

(5)

Boroditsky, M.; Gontijo, I.; Jackson, M.; Vrijen, R.; Yablonovitch, E.; Krauss, T.; Bhat, R.; Krames, M. J. Appl. Phys. 2000, 87, 3497-3504.

(6)

Sekiguchi, H.; Kishino, K.; Kikuchi, A. Appl. Phys. Lett. 2010, 96, 231104.

(7)

Qi, Y. D.; Liang, H.; Wang, D.; Lu, Z. D.; Tang, W.; Lau, K. M. Appl. Phys. Lett. 2005, 86, 101903.

(8)

Tzou, A. J.; Lin, D. W.; Yu, C. R.; Li, Z. Y.; Liao, Y. K.; Lin, B. C.; Huang, J. K.; Lin, C. C.; Kao, T. S.; Kuo, H. C.; Chang, C. Y. Opt. Express, 2016, 24, 11387-11395.

(9)

Hwang, J-II.; Hashimoto, R.; Saito, S.; Nunoue, S. Appl. Phys. Express 2014, 7, 071003.

(10)

Ponce, F. A.; Bour, D. P. Nature 1997, 386, 351-359.

(11)

Wu, F.; Lin, Y. D.; Chakraborty, A.; Ohta, H.; DenBaars, S. P.; Nakamura, S.; Speck, J. S. Appl. Phys. Lett. 2010, 96, 231912.

(12)

Takeuchi, T.; Sota, S.; Katsuragawa, M.; Komori, M.; Takeuchi, H.; Amano, H.; Akasaki, I. Jpn, J. Appl. Phys. 1997, 36, L382-L385.

(13)

Saito, S.; Hashimoto, R.; Hwang, J.; Nunoue, S. Appl. Phys. Express 2013, 6, 111004.

(14)

Sharma, Y. K.; Towe, E. J. Appl. Phys. 2010, 107, 024516.

(15)

Zhang, J.; Tansu, N. J. Appl. Phys. 2011, 110, 113110.

(16)

Park, S. H.; Chung, T. H.; Baek, J. H.; Ahn, D. Jpn. J. Appl. Phys. 2015, 54, 022101.

(17)

Even, A.; Laval, G.; Ledoux, O.; Ferret, P.; Sotta, D.; Guiot, E.; Levy, F.; Robin, I. C.; Dussaigne, A. Appl. Phys. Lett. 2017, 110, 262103.

(18)

Daubler, J.; Passow, T.; Aidam, R.; Kohler, K.; Kirste, L.; Kunzer, M.; Wagner, J. Appl. Phys. Lett. 2014, 105, 111111.

(19)

Kumakura, K.; Makimoto, T.; Kobayashi, N. Jpn. J. Appl. Phys. 2000, 39, L337-L339.

(20)

Pantha, B. N.; Sedhain, A.; Li, J.; Lin, J. Y.; Jiang. H. X. Appl. Phys. Lett. 2009, 95, 261904.

ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(21)

Ho, I. H.; Stringfellow, G. B. Solid Phase Immiscibility in GaInN. Appl. Phys. Lett. 1996, 69, 2701-2703.

(22)

Tian, A.; Liu, J.; Zhang, L.; Jiang, L.; Ikeda, M.; Zhang, S.; Li, D.; Wen, P.; Cheng, Y.; Fan, X.; Yang, H. Appl. Phys. Lett. 2017, 111, 112102.

(23)

Pantzas, K.; Gmili, Y. E.; Dickerson, J.; Gautier, S.; Largeau, L.; Mauguin, O.; Patriarche, G.; Suresh, S.; Moudakir, T.; Bishop, C.; Ahaitouf, A.; Rivera, T.; Tanguy, C.; Voss, P.L.; Ougazzaden, A. J. Cryst. Growth 2013, 370, 57-62.

(24)

Butte, R.; Lahourcade, L.; Uzdavinys, T. K.; Callsen, G.; Mensi, M.; Glause, M.; Rossbach, G.; Martin, D.; Carlin, J. F.; Marcinkevicius, S.; Grandjean, N. Appl. Phys. Lett. 2018, 112, 032106.

(25)

Koslow, I. L.; Hardy, M. T.; Hsu, P. S.; Dang, P. Y.; Wu, F.; Romanov, A.; Wu, Y. R.; Young, E. C.; Nakamura, S.; Speck, J. S.; DenBaars, S. P. Appl. Phys. Lett. 2012, 101, 121106.

(26)

Senda, R.; Matsubara, T.; Iida, D.; Iwaya, M.; Kamiyama, S.; Amano, H.; Akasaki, I. Appl. Phys. Express 2009, 2, 061004.

(27)

Hirasaki, T.; Asano, K.; Banno, M.; Ishikawa, M.; Sakuma, F.; Murakami, H.; Kumagai, Y.; Koukitu, A. Jpn, J. Appl. Phys. 2014, 53, 05FL02.

(28)

Heying, B.; Averbeck, R.; Chen, L. F.; Haus, E.; Riechert, H.; Speck, J. S. J. Appl. Phys. 2000, 88, 1855-1860.

(29)

Bi, Z. X.; Gustafsson, A.; Lenrick, F.; Lindgren, D.; Hultin, O.; Wallenberg, L. R.; Ohlsson, B. J.; Monemar, B.; Samuelson, L. J. Appl. Phys. 2018, 123, 025102.

(30)

Feenstra, F. M.; Dong, Y.; Lee, C. D.; Northrup, J. E. J. Vac. Sci. Technol. B 2005, 23, 1174-1180.

(31)

Li, S. F.; Fuendling, S.; Wang, X.; Merzsch, S.; Al-Suleiman, M. A. M.; Wei, J. D.; Wehmann, H. –H.; Waag, A. Cryst. Growth Des. 2011, 11, 1573-1577.

(32)

Schubert, E. F. Light-emitting Diodes. Cambridge University Press, New York, 2006, pp223.

(33)

Chichibu, S. F.; Uedono, A.; Onuma, T.; Haskell, S. A.; Chakraborty, A.; Koyama, T.; Fini, P. T.; Keller, S.; DenBaars, S. P.; Speck, J. S.; Mishra, U. K.; Nakamura, S.; Yamaguchi, S.; Kamiyama, S.; Amano, H.; Akasaki, I.; Han, J.; Sota, T. Nat. Mater. 2006, 5, 810-816.

(34)

Xie, M. H.; Seutter, S. M.; Zhu, W. K.; Zheng, L. X.; Wu, H.; Tong, S. Y. Phys. Rev. Lett. 1999, 82, 2749-2752.

ACS Paragon Plus Environment

Page 16 of 20

Page 17 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

(35)

Zhou, K.; Ren, H.; Liu, J.; Ikeda, M.; Ma, Y.; Gao, S.; Tang, C.; Yang, H. Superlattices Microstruct. 2016, 100, 968-972.

ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

FIGURE CAPTIONS: Figure 1 Procedures to synthesize InGaN platelets and LED structures. (a) InGaN pyramids grown by selective area MOVPE, as reported in an earlier work.29 (b) Bottom InGaN layer with a top c-plane formed by etching down from the pyramid apex using in-situ annealing. (c) InGaN regrowth to flatten the rough c-plane formed in (b). An intermediate layer of InGaN is grown first with an indium content below 5%, as shown in Figure S1 in the Supporting Information. Then, a top InGaN layer is grown with indium contents similar to the bottom InGaN layer in (b). (d) InGaN single QW growth on the c-plane of the InGaN platelet templates. (e) A prototype LED structure with p-InGaN grown above the single QW, with definitions of all InGaN layers. Figure 2 InGaN pyramids before and after the in-situ annealing with two different indium contents. (a, d) SEM images of the as-grown InGaN pyramids. (b, e) SEM images of the InGaN pyramids after the annealing. (c, f) PL spectra of the InGaN pyramids before and after the annealing measured at room temperature. The indium contents of 10% and 17% were obtained based on these PL spectra. Figure 3 InGaN platelets with indium contents of 9% and 18% formed by regrowth on the annealed InGaN pyramids. (a, e) SEM images of the InGaN platelets showing smooth top c-plane surface. Inset in (e) shows a pit indicated by the white arrow. (b, f) AFM height images of the top c-plane surface. Surface steps were observed for both samples. Height profiles along the black lines in (b, f) are presented in (c, g), showing single bilayer steps on c-plane. (d, h) PL spectra of the InGaN platelets measured at room temperature, based on which the indium contents of 9% and 18% were obtained. Figure 4 Single QW of InGaN grown at the same conditions on the platelets of In0.09Ga0.91N and In0.18Ga0.82N. (a, e) Cross-sectional HAADF-STEM images of the single QW samples. The white arrow in (e) indicates a dislocation. EDS measurements were conducted on the bottom InGaN layer (marked as 1) and the top InGaN layer (marked as 2) and the data are shown in Table 1. (b, f) are the magnified HAADF-STEM images close to the periphery, and (c, g) are high resolution TEM images recorded in the middle of the QWs. (d, h) CL spectra measured at 10 K, showing the emissions from the single QWs on both samples. Insets illustrate SEM images and corresponding monochromatic CL images with the emissions from the QW. CL images were recorded with an energy window of 1.80-2.5 eV in (d) and 1.85-2.35 eV in (h). Figure 5 Process and EL measurements of prototype InGaN platelet LEDs. (a) A schematic of InGaN platelet LEDs and related process for EL measurements. (b, c) EL spectra obtained at different current injection levels for the green LEDs grown on In0.09Ga0.91N platelets and the red ones on In0.18Ga0.82N platelets, respectively. The current density in both legends was calculated according to the actual QW area, rather than the contact area.

ACS Paragon Plus Environment

Page 18 of 20

Page 19 of 20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Table 1 Indium content comparison between layers 1 (bottom InGaN layer) and 2 (top InGaN layer) based on the PL and EDS data. Note that indium contents obtained from PL are based on the calibration samples, rather than the ones in Figure 4.

ACS Paragon Plus Environment

Nano Letters 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For TOC only.

ACS Paragon Plus Environment

Page 20 of 20