InAlN Multiple Quantum Well Tubes - Nano Letters

Apr 25, 2017 - In Figure 4a, the SEM image illustrates again the tubelike character ...... G. ; Carlin , J.-F. ; Grandjean , N. III-Nitride Semiconduc...
0 downloads 0 Views 17MB Size
Subscriber access provided by SUNY DOWNSTATE

Communication

Thin-wall GaN/InAlN multiple quantum well tubes Christophe Durand, Jean-François Carlin, Catherine Bougerol, Bruno Gayral, Damien Salomon, Jean-Paul Barnes, Joel Eymery, Raphael Butte, and Nicolas Grandjean Nano Lett., Just Accepted Manuscript • DOI: 10.1021/acs.nanolett.6b04852 • Publication Date (Web): 25 Apr 2017 Downloaded from http://pubs.acs.org on May 1, 2017

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 free 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 accessible to all readers and 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.

Nano Letters 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 30

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

Thin-wall GaN/InAlN multiple quantum well tubes

Christophe Durand1,2, Jean-François Carlin3, Catherine Bougerol1,4, Bruno Gayral1,2, Damien Salomon5, Jean-Paul Barnes1,6, Joël Eymery1,2, Raphaël Butté3, Nicolas Grandjean3 1

2

"Nanophysique et semiconducteurs" group, CEA, INAC-PHELIQS, 17 Avenue des Martyrs, 38000 Grenoble, France 3

4

5

Université Grenoble Alpes, 38000 Grenoble, France

Institute of Physics, Ecole Polytechnique Fédérale de Lausanne (EPFL), CH-1015 Lausanne, Switzerland

"Nanophysique et semiconducteurs" group, CNRS, Institut Néel, 25 Avenue des Martyrs, 38000 Grenoble, France

European Synchrotron Radiation Facility, 71 Avenue des Martyrs, 38000 Grenoble, France 6

CEA, LETI, MINATEC Campus, 38000 Grenoble, France

ACS Paragon Plus Environment

1

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 2 of 30

Abstract Thin-wall tubes composed of nitride semiconductors (III-N compounds) based on GaN/InAlN multiple quantum wells (MQWs) are fabricated by metal-organic vapor-phase epitaxy in a simple and full III-N approach. The synthesis of such MQW-tubes is based on the growth of N-polar c-axis vertical GaN wires surrounded by a core-shell MQW heterostructure followed by in situ selective etching using controlled H2/NH3 annealing at 1010°C to remove the inner GaN wire part. After this process, well-defined MQW-based tubes having non-polar m-plane orientation exhibit UV light near 330 nm up to room temperature, consistent with the emission of GaN/InAlN MQWs. Partially etched tubes reveal a quantum-dot-like signature originating from nanosized GaN residuals present inside the tubes. The possibility to fabricate in a simple way thin-wall III-N tubes composed of an embedded MQW-based active region offering controllable optical emission properties constitutes an important step forward to develop new nitride devices such as emitters, detectors or sensors based on tube-like nanostructures.

Keywords : Nanotubes, Multiple quantum wells, Nitride semiconductors, MOVPE, Quantum dots, UV emission

ACS Paragon Plus Environment

2

Page 3 of 30

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

Over the past two decades, III-nitride (III-N) semiconductor nanostructures with different geometries have been extensively studied, namely stripes, wires or pyramids at the micro and nanometer scales to cite only a few, as an alternative to standard planar structures in view of reducing strain and defects, and enhancing/further tuning their optical properties [1,2,3]. Such nanostructures have even been functionalized with multiple quantum wells (MQWs) for the development of efficient light emitters on non- or semi-polar surfaces [4,5,6]. Novel light emitting devices based on stripes [7], pyramids [8,9 ] or nanowires [10,11 ] have been successfully demonstrated and are intensively studied thanks to the opportunity to grow III-N nanostructures on low-cost substrates while offering a low defect density [12,13,14,15,16]. For instance, the development of nanowire light-emitting diodes (LEDs) is based on the possibility to grow ordered c-axis oriented GaN nanowires on large Si substrates, and also to make coaxial InGaN/GaN multiple quantum wells (MQWs) on non-polar m-plane wire sidewalls increasing the active region area [17] with an expected higher efficiency in the case of thick QWs, due to the absence of quantum confined Stark effect [6,10,11]. Among all nanostructure types, the tube geometry has been scarcely investigated so far with III-N materials. However, similarly to the wire geometry, tube-like nanostructures could also be highly promising as they offer the following key-advantages: (1) fully relaxed materials since thin-walls can act as a freestanding surface that represents an ideal template preventing the formation of structural defects; (2) non-polar surfaces for tubes synthesized from c-axis grown GaN nanowires with sidewalls corresponding to (10-10) m-plane surfaces; (3) large surface-to-volume ratio taking into account the inner and the external tube surfaces and (4) enhanced light extraction in the case of vertical dense nanotube ensembles for which out-ofplane light is scattered from tube-to-tube while in-plane light is guided along the tube axis. Compared to wires, the larger surface to volume ratio of tubes covered with MQWs is an important asset for the realization of more efficient light-emitting devices. Additionally, UV-

ACS Paragon Plus Environment

3

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 4 of 30

LEDs based on GaN wires shall suffer from the presence of a GaN core with a lower bandgap than the active region as the former would lead to unwanted UV light absorption that would eventually affect the overall efficiency. The fabrication of tubes based on MQWs emitting UV light represents a direct way to overcome the previous issue and hence enhance the UV light extraction. Besides, such light-emitting III-N tubes of micrometer diameter could also be integrated in microfluidic platforms for chemical or biological detection. As far as the literature is concerned, we can mention the successful realization of hexagonal BN (hBN) nanotubes presenting a crystalline structure similar to that of carbon nanotubes [18]. It has generated an intense research on this class of nanostructures [19] without any demonstration of light emission despite deep UV emission observed for bulk hexagonal BN crystals [20,21]. Monolayers of hBN [22] and more recently GaN [23] have also been reported as promising building blocks for two-dimensional (2D) material systems without any evidence regarding their optical properties. So far, very few works focused on III-N nanotubes having the wurtzite crystal structure, even if the fabrication of such GaN and AlN nanotubes has been successfully reported in the literature back to 2003 [24,25]. Two reasons can be put forward: (i) the challenge in controlling the nanotube synthesis and (ii) the poor optical emission properties. The current approach reported to form GaN nanotubes is based on the growth of a GaN shell surrounding ZnO nanowire cores acting as a sacrificial template that can be subsequently etched by annealing. This epitaxial casting method requires the growth of two types of semiconductors with the possibility of dislocation formation [26], doping and chemical interdiffusion. This ZnO casting method has been used to grow coreshell InGaN/GaN MQWs on GaN tubes with a 600 nm-thick wall [27] and LEDs based on these ordered tube arrays have been demonstrated [ 28 , 29 , 30 , 31 ]. However, the actual development of those tube-devices is presently limited by the complexity of their fabrication, which requires the combination of two different materials systems (GaN and ZnO). In

ACS Paragon Plus Environment

4

Page 5 of 30

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

addition, the MQWs are grown on a thick GaN shell template to prevent defect formation resulting in thick-wall tubes, which look more like wires than tubes after the functionalization by MQW growth [27]. Beyond the ZnO casting approach, other methods have been reported to synthesize III-nitride thin-wall tubes (< 50 nm). The spontaneous formation of GaN or AlN nanotubes has been reported by chemical vapor deposition [32,33] and by nitridation of Ga2O3 powders [34,35]. Ordered arrays of GaN nanotubes on diamond grown by molecular beam epitaxy have recently been reported for the specific Ga-limited growth regime in the case of selective area growth [36]. AlN and InAlN nanotubes have also been obtained based on catalyzed-GaN or catalyzed-InN nanowires casting method, respectively [37,38]. In this latter case, after an external step that chemically removes the catalyst localized at the wire-top, the inner wire is selectively etched during a thermal annealing step. The optical properties of all these thin-wall tubes are usually not reported or are altered due to the presence of surface states [39]. Only tubes with thick enough walls (typically >100 nm) exhibit significant optical properties, as reported by Li et al. with lasing emission based on thick-wall GaN tubes fabricated using a top-down GaN etching process using ordered ring-like patterns [40]. Therefore, the issues related to the low optical quality of thin-wall tubes and their difficulties of fabrication strongly limit the development of such III-N tube-like nanostructures and require to be further addressed. In this work, we investigate a novel type of III-N tubes with a thin-wall only composed of MQW heterostructures acting as the active region, well known for their remarkable light emission properties. The tube wall is composed of GaN/InAlN MQWs having non-polar mplane orientation exhibiting UV light emission near 330 nm up to room temperature. The fabrication process is based on a full III-N metal-organic vapor-phase epitaxy (MOVPE) approach consisting in the growth of catalyst-free GaN nanowires with radial MQWs

ACS Paragon Plus Environment

5

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 30

followed by a second step, where the GaN core is selectively etched by an in situ controlled annealing step using a H2/NH3 mixture leading to high-quality thin MQW-based tubes. The structural and optical features of those MQW-tubes are studied and discussed by taking into account their specific fabrication process. The tube fabrication method is detailed hereafter. Self-assembled c-axis GaN wires were grown on c-plane sapphire substrates by MOVPE in a closed coupled showerhead reactor using a silane-based catalyst-free method detailed in Ref. 41 consisting in (i) pre-treatment of the sapphire surface (annealing and ammonia nitridation), (ii) in situ thin SiNx layer deposition acting as a partial mask and (iii) growth of GaN wires combining high flux of silane (200 nmol/min) and specific conditions, namely a low V/III ratio (20-50) using trimethylgallium and NH3 as III and V precursors, respectively, high temperature (1040°C) and high pressure (800 mbar), to enhance the vertical growth. The wire growth mechanism is based on spontaneous formation of a SiNx thin layer on the wire sidewalls due to the high silane flux in the gas phase. This passivation of the vertical facets acts as a mask that prevents any lateral growth [42,43]. In a second step, the wire growth is pursued without silane flux leading to GaN wires composed of a bottom passivated highly n-doped part (~10 µm, 2 × 1020 cm-3) [44] followed by an upper unintentionally doped (u-doped) section (~15 µm, n-doping at the level of 3 × 1018 cm-3) [45]. The as-grown wires on c-plane sapphire substrates having a typical length of 25-30 µm and a diameter about 0.8-1.7 µm are then transferred in a horizontal MOVPE reactor to grow core-shell InAlN heterostructures at 835°C under NH3 using both trimethylindium and trimethylaluminum for InAlN growth and trimethylgallium for GaN growth, as previously reported in Ref. 46. The heterostructures involve an InAlN shell followed by 9 MQWs composed of 2-nm-thick GaN wells and 5-nm-thick InAlN barriers. Arita et al. reported a high etching selectivity between GaN and AlGaN layers even for a low Al fraction (from 5%) with a simple in situ annealing step under H2/NH3 in MOVPE

ACS Paragon Plus Environment

6

Page 7 of 30

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

reactors for the fabrication of AlGaN membranes [47]. The same method has been employed for the tube formation that consists in a controlled annealing at 1010°C and 200 mbar using a H2/NH3 gas mixture (10 sccm / 5000 sccm) that selectively etches the GaN wires with respect to the InAlN-based heterostructures with an etching rate of about 3-5 µm/h depending on the wire diameter. For thick InAlN-based heterostructure shell covering, reactive-ion etching (RIE) under argon at 200 W (typically 150 nm/h for vertical RIE etching) is required to uncap the top facets of the vertical wires in order to proceed to the inner GaN etching. These wires grown on sapphire consist in an N-polar phase having –c axis growth direction with usually the presence of a residual Ga-polar phase having a pyramidal shape on the wire top facets [48,49,50]. We observe preferential RIE etching of the Ga-polar part that creates an aperture in the top facet of the wires in order to reach the inner GaN wire part. To demonstrate the feasibility of the etching selectivity between GaN and InAlN using the in situ H2/NH3 annealing to fabricate III-N tubes, we perform the deposition of an ultra-thin InAlN shell on as-grown self-assembled GaN wires with an In-content of 18% and two targeted shell thicknesses (tshell) of 5 and 0.5 nm, respectively. The growth of a thin smooth InAlN shell mainly occurs in the non-passivated part of the as-grown GaN wires, i.e. in the upper part of the wires corresponding to the u-doped part. A rough and non-uniform growth of InAlN is observed in the bottom n-doped part of the wires. In situ annealing under the H2/NH3 gas mixture is then performed during 4 hours on these coated as-grown wires without any external RIE etch process. Following the whole MOVPE process (including the wire epitaxy, the InAlN shell growth and the controlled annealing), scanning transmission electron microscopy (STEM) images taken at 20 kV shown in Fig. 1a clearly demonstrate the formation of well-defined tube nanostructures for both targeted thicknesses (tshell = 5 and 0.5 nm) having the same diameter than the wire template. The typical length of the tubes is a little bit shorter than the wire length, due to partial coverage of the thin InAlN shell. A degradation

ACS Paragon Plus Environment

7

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 8 of 30

in the tube morphology is observed under the dashed line shown in Fig. 1a that corresponds to the rough growth of the InAlN shell on the n-GaN part of passivated sidewalls. For thinner tubes, the rigidity is not completely ensured and deformation under e-beam radiation is observed during STEM imaging, which is not the case for 5 nm-thick tubes. Similarly, after the thermal annealing, the thinner tubes fall down on the surface substrate, whereas the 5 nmthick tubes stay vertical illustrating the effect of tube wall thickness on the structural tube rigidity. High-resolution transmission electron microscopy (HRTEM) has been performed on the 5 nm-thick InAlN tubes (see Fig. 1b). From high-angle annular dark-field (HAADF) image (top image) and the enlarged HRTEM images (middle image plus the insets), the tube wall thickness can be estimated to about 5-10 nm, a value which is in good agreement with the targeted amount of deposited InAlN. From the associated Fourier transforms (bottom left inset), the tubes appear as single crystals with the wurtzite phase with (10-10) m-planes as lateral facets. The composition analysis based on energy dispersive X-ray spectroscopy (EDX) (shown in supplementary information) reveals that the InAlN sidewalls of the tubes are nearly Ga free meaning that the inner etching process is a very efficient one. For 0.5 nm-thick tubes, the sidewalls correspond to pure AlN. The absence of In is consistent with the first steps of InAlN growth that starts with a thin AlN layer before the effective In incorporation [51]. A small amount of Ga cannot be totally excluded due to possible interdiffusion process between AlN and GaN. The TEM images taken at 200 kV shown in Fig. 1c indicate that the etching of the GaN inner core by the H2/NH3 gas phase starts in first place from the bottom part of the wires through small holes present in the non-uniform InAlN shell. Then, the etching front progresses toward the wire top, as shown by arrows in the TEM image (Fig. 1c) resulting in tubes opened at one end and closed at the other one. Thanks to the closed end, which insures some extra stiffness,

ACS Paragon Plus Environment

8

Page 9 of 30

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 nanotubes retain the facetted hexagonal morphology of the GaN wire templates. This bottom-up etching assumption is also confirmed by the direct observation of as-grown tubes collapsed on the substrate surface. Figure 1d depicts the etching mechanism in the case of ultra-thin InAlN shells (≤ 5 nm) that begins from the foot of the wire and evolves toward the top until the c-plane top surface of the wires is reached. The residual SiNx passivation of the n-part GaN wire does not prohibit the etching process from the sidewalls certainly due to a discontinuity in the SiNx layer in the n-part/u-part transition zone. This clearly demonstrates the efficiency of the in situ H2/NH3 annealing step for the etching of the GaN core with the possibility to achieve a nanowall thickness estimated to be as narrow as 1 ±0.5 nm (note that the effective thickness of 0.5 nm-thick tubes is assumed between 0.4 and 1.5 nm taking into account an error bar of 20 % in targeted thickness and a possible interdiffusion effect between the AlN shell and the GaN core forming a thin AlGaN layers). In comparison, a thickness of 10 nm is reported for nanowall network structures obtained by molecular beam epitaxy [52]. In order to get functionalized tubes with light emission properties, we extend this method to the fabrication of III-N tubes composed of Al-based MQW heterostructures. For that, the growth of a 5 nm-thick InAlN shell followed by a 9-period GaN(2 nm)/InAlN(5 nm) MQW is performed around the GaN wires with the core/shell geometry as reported in Ref. 46. After the RIE processing step to remove the wire top cap, a two-hour in situ H2/NH3 annealing step is performed in the MOVPE reactor. Figures 2a and 2b show typical STEM images of the MQW-tubes. A clear hexagonal tube shape is observed with a thickness of 70 ±5 nm consistent with the targeted width for the MQW structures. We observe an effective etching of the GaN wire in the upper part of the wire along about one third to one half of the wire length forming the MQW-tube part. Below, an intermediate zone is observed with partial GaN etching while the bottom part of the structure corresponds to the unetched GaN core wires. For these MQW structures, the etching mechanism is different compared to that of the

ACS Paragon Plus Environment

9

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 10 of 30

above-mentioned ultra-thin InAlN shells. Thanks to the wire top etching with RIE, the GaN wire core is accessible from the top surface (preferential etching of the Ga-polar phase favors the formation of localized holes on the top wire surface). Consequently, in situ MOVPE etching by H2/NH3 annealing starts from the wire top and progresses toward the bottom as illustrated in Fig. 2c. Depending on the wire length and the annealing duration, we can control the height of the GaN wire etching. Structural characterization of such MQW-tubes are presented in the Fig. 3. We performed nano synchrotron X-ray fluorescence (S-XRF) measurements on the ID16B beamline at the European Synchrotron Radiation Facility (ESRF) using a monochromatic beam at 29.6 keV (beam size 70×70 nm2) on single wires dispersed on kapton foils [53], and grazing incidence X-ray diffraction (GIXRD) using 9.5 keV photon energy available from the BM32 beamline. Typical S-XRF maps of the Ga and In Kα lines (at 9.251 keV and 24.210 keV, respectively) are shown in Fig. 3a and 3b before and after in situ annealing (the composition being colored by the degraded color from blue to red, combined with scanning electron microscopy (SEM) images of individual as-grown wires). In Fig. 3a corresponding to the situation without in situ annealing, the presence of Ga is detected all along the wire proving that the bulk wire is made of GaN. Mappings of In performed along the wire axis show a clear preferential localization of this species in the upper part of the wires, as commonly observed in the case of InGaN/GaN MQWs [6]. This is explained by local growth selectivity of In species on the upart of the GaN, which is not passivated by a SiNx thin layer that corresponds to the core/shell growth of GaN/InAlN MQWs, while the poor-quality growth occurs in the bottom part. When investigating the effect of in situ annealing such as shown in the SEM image of Fig. 3b, we presume that the GaN wire core is etched from the upper part of the wires, which is indeed supported by Ga chemical mapping. The Ga signal is only detected in the bottom part of the wires under the dashed line where the etching has not been completed. Some unetched GaN

ACS Paragon Plus Environment

10

Page 11 of 30

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

residues inside the tubes are also visible (highlighted by red arrows). Above the dashed line, a weaker Ga signal is also detected inside the tube wall that could be attributed to the presence of GaN wells. Furthermore, no difference is observed in the In mapping before and after in situ annealing, which can be safely interpreted as the signature of MQWs. These data evidence the etching selectivity between GaN and InAlN/GaN MQWs. Extra insights were subsequently obtained from dual-beam time-of-flight secondary ion mass spectrometry (ToFSIMS) analysis performed on single tubes mechanically dispersed on a reference substrate consisting in 5 InGaN/GaN MQWs grown on GaN/sapphire in order to detect the GaN/InAlN MQWs in the tube-wall. A bismuth liquid metal ion gun is used to obtain ToF spectra and a 500 eV oxygen beam is used for depth profiling corresponding to a depth and lateral resolution estimated to be 2 and 200 nm, respectively. The visualization of the structure is achieved by combining the 3D position of the scanned Bi ion beam with the time-of-flight data recorded at each pixel. Figure 3c shows the depth profile consisting in Al, In, Ga ion maps performed on a selected area involving a fragment of tube-wall lying on the reference substrate. Consequently, the top part corresponds to the tube-wall structure and the bottom part is related to the substrate. While the 5 InGaN wells are revealed in the In mapping for the reference substrate, we also observe 9 GaN wells in the GaN mapping, which are sandwiched between 10 InAlN barriers evidenced thanks to both Al and In mapping. This tube-wall structure corresponds to the initial core-shell heterostructure proving the high stability of GaN/InAlN MQWs even after a 2 hour-annealing step at 1040°C to form the tube nanostructures. Even if the InAlN growth requires a limited growth temperature (750-900°C), this work emphasizes that the InAlN alloy is stable at temperatures slightly larger than 1000°C, which is fully consistent with previous results reported in Refs. [ 54 , 55 ]. Interestingly, the sandwiched GaN wells between InAlN barriers are preserved during the in situ annealing, certainly due either to the difficulty to etch a so thin sandwiched layer or to a

ACS Paragon Plus Environment

11

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 12 of 30

limited Al interdiffusion (etching selectivity is effective even for an Al content as low as 5% in AlGaN layers [47]). Hence, S-XRF and ToF-SIMS analyses confirm the feasibility to fabricate MQW-tubes thanks to the high stability of InAlN-based MQWs at high temperature and the etching selectivity between the GaN core and the InAlN-based MQWs. In addition, GIXRD experiments have been conducted on as-grown core-shell wire assemblies before and after the in situ annealing step: samples labelled MQW-wires and MQW-tubes, respectively in Figs. 3d and 3e. A sample of GaN wires without MQW has also been measured to provide a reference labelled GaN wires. The grazing incidence angle was set to 0.15°, i.e. a value lower than the critical angle of refraction (around 0.24° at this energy) to strongly enhance the intensity diffracted by the wires/tubes with respect to the substrate surface. The samples were aligned with respect to the usual hexagonal c-plane sapphire

surface unit cell (a = 0.4758 nm, c = 1.2991 nm) to define an orientation matrix that uses (hkl)sap. Miller indices and sapphire reciprocal lattice units (rlu). This choice allows an accurate scanning of the reciprocal space. GaN wires are grown on the sapphire substrate as reported in Ref. 41 confirming the in-plane 30° rotation between the two unit cells, i.e. 1 0 1 0 ∥ [1 1 2 0]. and

[0 0 0 1 ] ∥ [0 0 0 1]. . Figure 3d shows in-plane

diffraction scans taken along the [h 0 0]sap. direction exhibiting the (h h 0)GaN reflection due to the epitaxial relationship and Figure 3d is related to [h h 0]sap. direction scans to measure the (h 0 0)GaN reflection. The usual in-plane interatomic distances corresponding to the (110) and (200) GaN interplanar distances were obtained for the three samples. As expected, they correspond to those of the GaN core, even for the MQW-tube sample because the GaN core is not completely etched at the base of the GaN wire. Moreover, an additional contribution is measured at larger rlu, which is attributed to the 0-order peak of the GaN/InAlN MQW shell. The broadening of this peak for the MQW-wire sample is wider and probably corresponds to a mixing of different contributions including InAlN overgrowth on the substrate. Following

ACS Paragon Plus Environment

12

Page 13 of 30

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 etching annealing step forming MQW-tubes, the (110) and (200) reflections evolve from a shoulder-like feature for the wires to a well-defined peak for the tubes (see vertical right black lines in Figs. 3d and 3e). It originates from the elastic equilibrium state of the MQW system due to the free boundaries conditions inherent to the tube-like geometry. From the direct measurement of the peak positions with respect to the GaN core reference, we can deduce a decrease in the average interplanar distances by about -0.44 % for these two plane families suggesting a smaller InAlN in-plane lattice parameter than that of bulk GaN. The interpretation in terms of composition inside the InAlN layer is made difficult because one must take into account the nature of the MQWs (thickness and composition) as well as anisotropic strain and relaxation. Nevertheless, GIXRD results suggest that the GaN quantum wells in the MQW-tube samples should be in tension to balance the average strain corresponding to the peak position. Complementary grazing incidence angle measurements performed on truncated rods along the l direction for (h k) values (see supplementary information) indicated by #S1, S2, S3 and S4 in Figs. 3d/3e confirm the smaller value of the average lattice parameter of the MQW system along the c-axis growth direction (decrease of 1.15% with tubes and -1.3% with wires for the measurements along (1 1 l)). Next, we focus on the optical emission properties of these MQW-tubes. The optical features of tubes dispersed on Si substrates were characterized both by cathodoluminescence (CL) and photoluminescence (PL) measurements. The electron beam voltage and current for CL mappings were set to 30 kV and 1 nA, respectively. A frequency-doubled continuous wave Ar+ laser excitation source emitting at 244 nm has been used to conduct PL experiments with a typical spot size of about 50 µm allowing us to probe about one hundred tubes. The PL and CL signals are sent to a monochromator coupled to a liquid-nitrogen cooled UV-enhanced charge-coupled device providing a spectral resolution of about 1 meV. Figure 4 shows optical features of MQW-tubes including CL and PL data taken at 5 K (Figs. 4a and 4b, respectively),

ACS Paragon Plus Environment

13

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 14 of 30

and temperature-dependent PL measurements (Fig. 4c). In Fig. 4a, the SEM image illustrates again the tube-like character that extends over about two third of the top wire part, whereas a residual unetched GaN section is present in the bottom part. The emission of MQW-tubes is exclusively observed in the tube part for the CL map measured at a wavelength of 332 nm corresponding to the expected UV emission of GaN/InAlN MQWs, previously measured on core-shell MQW GaN wires in Ref. 46. MQW emission is only observed in the upper part of the tube corresponding to the high-quality growth of core-shell MQWs on the unpassivated wire sidewalls. The emission occurring at 350 nm related to the GaN near-band edge (NBE) is actually observed only for the unetched wire part. GaN residuals are also visible in the inner tube surface as indicated by the black arrows in Fig. 4a. For the sake of clarity, Fig. 4b compares the PL emission of core-shell MQW wires before and after the in situ annealing step, i.e. before and after the GaN wire core etching labeled as MQW-wires and MQW-tubes, respectively. Both spectra exhibit two main contributions: a peak centered near 355 nm assigned to the NBE emission of GaN and a second peak near 330 nm ascribed to MQW emission. As already mentioned in Ref. 41, the NBE emission peak is composed of two contributions: the first one centered at 357 nm related to the u-doped GaN wire upper part and the second one at higher energy, 353 nm, corresponding to the n+-doped GaN wire bottom part, which is due to the band-filling effect (Burstein-Moss shift in the highly n-doped GaN section). These two contributions are clearly visible in the PL spectra of the core-shell MQW wires. On the contrary, the contribution at 355 nm has almost disappeared for the MQW-tubes owing to the inner wire etching located in the top part corresponding to the u-doped GaN. Consequently, the NBE emission of the GaN peak is positioned at 352 nm and is related to the residual unetched GaN core at the wires foot characterized with a high n-type doping level. The contribution at ~330 nm related to MQWs for tubes is shifted to shorter wavelength (327.4 vs. 329.8 nm) and exhibits a much broader full width at half maximum (260 vs. 100

ACS Paragon Plus Environment

14

Page 15 of 30

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

meV) compared to wires. The origin of this blue-shift and this broadening is attributed to the presence of GaN quantum dots (QDs) highlighted by micro-PL measurements performed on single tubes presented hereafter. Figure 4c shows temperature-dependent PL measurements taken from 10 to 300 K. The PL intensity ratio between room and low temperature for the MQW peak is around 2%, which is similar to the ratio reported on m-plane and c-plane GaN/InAlN MQWs [46,56,57]. We can therefore establish that the optical features of MQWs are completely preserved after the in situ etching step, consistent with the ToF-SIMS structural analysis. In order to gain further insights on the light emission features, micro-PL measurements have been performed at 5 K along single wires after the in situ annealing treatment to form MQWtubes. The wires were dispersed on pre-patterned Si substrates with position markers to identify individually each wire. The 244 nm continuous wave laser beam was focused to ~2 µm diameter spot by means of a UV microscope objective in order to make a scan along chosen single wires at an excitation power of 10 µW. Figure 5a displays the PL signal evolution along a typical single wire measured for 4 equidistant points labeled on the SEM image. The PL spectrum of point #1 exhibits NBE emission of u-doped GaN and MQW emission centered around 339 nm having a FWHM equal to 85 meV. The presence of GaN emission is likely due to GaN residuals. The PL spectrum taken at position #2 shows similar features with u-GaN NBE emission centered at the same wavelength and a contribution due to MQWs slightly blueshifted to 334 nm with comparable FWHM (100 meV). This blueshift of 5 nm is likely arising from a reduction in the GaN well thickness along the wire along the topdown direction, the latter being known to amount to 3% per micrometer as reported in Ref. 46. Calculations using the k p method correlating the well width to the energy transition [46] allow approximating the GaN well thickness at positions #1 and #2 to 2.3 and 1.9 nm, respectively. As the distance between points #1 and #2 is estimated to 6 ± 2 µm, the well

ACS Paragon Plus Environment

15

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 16 of 30

thickness is assumed to be reduced by 18 ± 6%, which is consistent with the predicted thickness value at point #2: 1.9 ± 0.1 nm. For the last measurement (point #4), only the NBE peak at 355 nm is visible on the PL spectrum, which is in agreement with the presence of the unetched n+-doped GaN core at the wires foot. At this position, MQW emission is not expected as already observed in the CL mappings due to the degraded core-shell MQW structure on passivated wire sidewalls. For point #3, we observe an intermediate zone with the same NBE emission due to the unetched n+-doped GaN core, and instead of the MQW emission, an ensemble of very sharp peaks is visible between 310 and 350 nm, as specifically shown in the inset of Fig. 5a. Those resolution-limited sharp lines on the high-energy side of the spectrum are likely the signature of GaN QDs having a large distribution of sizes. The presence of GaN QDs in the transition region between the MQW-tube and the core-shell GaN wire can basically be attributed to residual clusters inside the tubes. PL scans performed along the growth axis of many such wires show the same optical features: 1) a strong emission of MQWs in the tube part situated in the upper wire part, 2) an intermediate zone with partial GaN core etching exhibiting sharp QD-lines and 3) a NBE emission due to the unetched n+doped GaN located at the bottom of the wires. Figure 5b shows PL spectra measured at 3 different locations labeled #1, #2 and #3 in the SEM image inset for a random wire fragment corresponding to the intermediate etching zone. These three PL spectra highlight the presence of sharp peaks related to QDs, which are systematically observed for the wire/tube intermediate zone. The observed variations in the emission wavelength (300-360 nm) and the peak broadening (1-20 meV) are attributed to the large distribution in QD sizes inside the tube, which is related to an incomplete etching of the GaN core in the transition zone. The SEM image in transmission mode shown in the inset of Fig. 5b reveals the presence of unetched nanosized residuals inside the InAlN tubes substantiating the assumption that light emission originates from GaN residuals acting as uncapped QDs. As further proof, when the annealing

ACS Paragon Plus Environment

16

Page 17 of 30

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

step is long enough to totally complete the wire core etching, no QD signature is observed. Figure 5c corresponds to PL measurements performed on the same wire fragment that the one shown in Fig. 5b as a function of temperature in the 5-100 K range. The intensity of QD peaks emitting between 310 and 325 nm stays stable up to 40 K. It is followed by a gradual decrease in intensity and a weak broadening from 60 to 100 K, as previously reported for non-polar single GaN QDs [58]. This behavior is also accompanied by a slight redshift (about 6 meV) of the QD-peaks (see the dotted lines in Fig. 5c) related to the temperature-dependence of the band gap. The intensity reduction in the 60-100 K range can be attributed to non-radiative processes, which could be linked to an enhancement of the diffusion of photocreated carriers toward surface states. Note that even if the density and the size of the GaN QDs are not well controlled, these results illustrate the possibility to fabricate tubes with embedded QD. In conclusion, the fabrication of thin-wall III-N MQW-based tubes exhibiting light emission up to room temperature is achieved by a simple in situ MOVPE approach. It is based on the growth of GaN nanowires with core-shell GaN/InAlN MQWs followed by selective etching (H2/NH3 annealing) to remove the wire core. This fabrication method favors elastic relaxation of MQW system and integrally preserves their structural and optical emission properties. For uncompleted wire core etching, the tube/wire transition region exhibits sharp emission peaks over a large spectral range (310-350 nm) corresponding to GaN QDs originating from nanosized residuals inside the tubes. Interestingly, the present in situ etching approach can be extended to core-shell Al-based heterostructures provided the Al content is higher than 5% [47] to fabricate other types of MQW-tubes (such as GaN/AlGaN, AlGaN/AlGaN, AlGaN/AlN…) with the possibility to tune the emission wavelength by playing with the well composition or thickness. The selective area growth of GaN nanowires can provide an additional possibility to manage the density, height and size of tube arrays. This simple method to fabricate III-N nanowall tubes with an embedded tunable active region opens

ACS Paragon Plus Environment

17

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 18 of 30

routes for the development of novel tube-based devices ranging from emitters to chemical and biological sensors.

Supporting information: EDX measurements performed on single 5 nm-thick InAlN tubes and out-of-plane GIXRD measurements performed on MQW-tubes based on synchrotron source.

Acknowledgments The authors thank J. Dussaud for technical MOCVD support, Fabrice Donatini for powerful CL guidance and Kuntheak Kheng for fruitful discussions. We acknowledge the ESRF for provision of synchrotron beam and would like to thank the assistance of the CRG-IF BM32 and ID-16B beamlines. ToF-SIMS measurements were performed on the nanocharacterization platform (PFNC) of the CEA Grenoble. C. Durand acknowledges financial support from the program Franco-Swiss “Partenariats Hubert Curien (PHC) Germaine de Staël” and the French state funds ANR-10-LABX-51-01 (Labex LANEF du Programme d'Investissements d’Avenir).

ACS Paragon Plus Environment

18

Page 19 of 30

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 caption

Figure 1: Formation of InAlN tubes after H2/NH3 annealing that etches the GaN wire core. a) STEM images taken at 20 kV of dispersed InAlN tubes on holey carbon membranes having two targeted wall thicknesses (tshell) of 5 and 0.5 nm. b) TEM images taken at 200 kV of single 5 nm thick InAlN tubes dispersed on holey carbon membrane: HAADF image of a whole single InAlN tube (top image), enlarged HRTEM enlargement image of the same tube (middle image), enlarged TEM image and corresponding Fourier transform (insets) showing that the tube corresponds to a single crystal with the wurtzite structure. c) Complementary TEM images of dispersed 5 nm thick InAlN tubes showing the bottom-up etching front. d) Schematic of the etching mechanism by in situ H2/NH3 annealing for ultra-thin InAlN shells.

Figure 2: Formation of MQW-tubes after incomplete wire etching by in situ H2/NH3 annealing of GaN wires coated with radial GaN/InAlN MQWs (10-period). a,b) STEM images taken at 20 kV of dispersed MQW-tubes on holey carbon membrane having a tube wall thickness of 70 nm corresponding to the expected thickness of 10 MQWs. c) Schematic of the etching mechanism by in situ H2/NH3 annealing for the thick MQW shell geometry.

Figure 3: Structural characterization of MQW-tubes. a,b) SEM images and XRF mappings performed with synchrotron radiation using the Kα line at 9.251 keV for Ga and the Kα line at 24.210 keV for In (the composition is colored by the degraded color ranging from blue to red) for the as-grown wire before (namely MQW-wire) and after in situ annealing (namely MQWtube), respectively. c) ToF-SIMS experiment performed on a fragment of tube-wall lying on a reference substrate composed of 5 InGaN/GaN MQWs. A 3D depth profile consisting of Al, In and Ga maps is shown (note that horizontal and vertical scales are different). d,e) Grazing

ACS Paragon Plus Environment

19

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 20 of 30

incidence X-ray diffraction for assemblies of bare GaN wires, MQW-wires and MQW-tubes grown on c-plane sapphire substrates. Reciprocal lattice units correspond to the c-plane sapphire substrate lattice and the (110)GaN and (200)GaN peaks are shown in d) and e), respectively. Positions indicated by #S1, #S2, #S3 and #S4 are used to perform out-of-plane measurements at given (h k) positions (see supplementary information).

Figure 4: Optical light emission properties of MQW-tubes based on CL and PL measurements. a) SEM and CL mapping taken at 332 and 350 nm at 5 K measured on typical single MQW-tubes (the light emission stems from the dark regions). b) Comparison of PL spectra taken at 5 K measured on an ensemble of dispersed wires before and after in situ annealing labeled as MQW-wire (black curve) and MQW-tubes (red curve), respectively. c) Temperature-dependent PL spectra ranging from 10 to 300 K measured on a dispersed MQWtube ensemble.

Figure 5: Micro-PL spectra taken along the axis of single MQW-tubes. a) 4-point scan measured at 5 K taken along a typical single MQW-tube with an incomplete GaN core etching. Inset: SEM image showing the location of the measured points, where three different zones can be distinguished: the MQW-tube region (points #1 and #2), the transition region with QD signatures (point #3) and the GaN unetched wire (point #4). b) 3-point scan performed at 5 K on a wire fragment corresponding to the transition region exhibiting sharp peaks related to QDs. Top inset: SEM image with the location of the measured points. Bottom inset: SEM image taken in transmission mode showing the inner part of a single InAlN thin tube proving the presence of nanosized GaN residuals. c) Temperature-dependent PL spectra ranging from 5 to 150 K measured on the same wire fragment than the one shown in b) with QD emission peaks. The spectra are vertically shifted for clarity.

ACS Paragon Plus Environment

20

Page 21 of 30

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)

Zheleva, T. S.; Nam, O.; Bremser, M. D.; Davis, R. F. Appl. Phys. Lett. 1997, 71, 2472.

(2)

Colby, R.; Liang, Z.; Wildeson, I. H.; Ewoldt, D. A.; Sands, T. D.; Garcia, R. E.; Stach, E. A. Nano Lett. 2010, 10, 1568–1573.

(3)

Ramesh, V.; Kikuchi, A.; Kishino, K.; Funato, M.; Kawakami, Y. J. Appl. Phys. 2010, 107, 114303.

(4)

Feng, W.; Kuryatkov, V. V; Chandolu, A.; Song, D. Y.; Pandikunta, M.; Nikishin, S. A.; Holtz, M. J. Appl. Phys. 2008, 104, 103530.

(5)

Kim, T.; Kim, J.; Yang, M.; Lee, S.; Park, Y.; Chung, U.-I.; Cho, Y. Appl. Phys. Lett. 2010, 97, 241111.

(6)

Koester, R.; Hwang, J.; Salomon, D.; Chen, X.; Bougerol, C.; Barnes, J.; Le, D.; Dang, S.; Rigutti, L.; Bugallo, A. D. L.; Jacopin, G.; Tchernycheva, M.; Durand, C.; Eymery, J. Nano Lett. 2011, 11, 4839–4845.

(7)

Wunderer, T.; Brückner, P.; Neubert, B.; Scholz, F.; Feneberg, M.; Lipski, F.; Schirra, M.; Thonke, K. Appl. Phys. Lett. 2006, 89, 041121.

(8)

Wildeson, I. H.; Colby, R.; Ewoldt, D. A.; Liang, Z.; Zakharov, D. N.; Zaluzec, N. J.; García, E.; Stach, E. A.; Sands, T. D. J. Appl. Phys. 2010, 108, 044303.

(9)

Ko, Y.; Kim, J.; Gong, S.; Kim, J.; Kim, T.; Cho, Y. ACS Photonics 2015, 2, 515–520.

(10)

Li, S.; Waag, A. J. Appl. Phys. 2012, 111, 071101.

(11)

Monemar, B. ; Ohlsson, B.J. ; Gardner N.F. ; Samuelson L.; Semiconductors and Semimetals 2016, 94, 227-271

(12)

Salomon, D.; Dussaigne, A.; Lafossas, M.; Durand, C.; Bougerol, C.; Ferret, P.; Eymery, J. Nanoscale Res. Lett. 2013, 8, 61.

(13)

Bavencove, A.; Salomon, D.; Lafossas, M.; Martin, B.; Dussaigne, A.; Levy, F.; Andre, B.; Ferret, P.; Durand, C.; Eymery, J.; Dang, L. S.; Gilet, P. Electron. Lett. 2011, 47, 765-766.

(14)

Kumaresan, V.; Largeau, L.; Oehler, F.; Zhang, H.; Mauguin, O.; Glas, F.; Gogneau, N.; Tchernycheva, M.; Harmand, J.-C. Nanotechnology 2016, 27, 135602.

(15)

Brandt, O.; Grahn, H. T.; Geelhaar, L.; Riechert, H. Nano Lett. 2015, 15, 3743–3747.

(16)

Choi, J. H.; Zoulkarneev, A.; Kim, S. Il; Baik, C. W.; Yang, M. H.; Park, S. S.; Suh, H.; Kim, U. J.; Son, H. Bin; Lee, J. S.; Kim, M.; Kim, J. M.; Kim, K. Nat. Photonics 2011, 5, 763–769.

ACS Paragon Plus Environment

21

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 22 of 30

(17)

Jung, B. O.; Bae, S.-Y.; Kim, S. Y.; Lee, S.; Lee, J. Y.; Lee, D.-S.; Kato, Y.; Honda, Y.; Amano, H. Nano Energy 2015, 11, 294–303.

(18)

Chopra, N. G.; Luyken, R. J.; Cherrey, K.; Crespi, V. H.; Cohen, M. L.; Louie, S. G.; Zettl, A. Science 1995, 269, 966–967.

(19)

Golberg, D.; Bando, Y.; Tang, C. C.; Zhi, C. Y. Adv. Mater. 2007, 19, 2413.

(20)

Kubota, Y.; Watanabe, K.; Tsuda, O.; Taniguchi, T. Science. 2007, 317, 932–935.

(21)

Watanabe, K.; Taniguchi, T.; Kanda, H. Nat. Mater. 2004, 3, 404–409.

(22)

Gorbachev, R. V.; Riaz, I.; Nair, R. R.; Jalil, R.; Britnell, L.; Belle, B. D.; Hill, E. W.; Novoselov, K. S.; Watanabe, K.; Taniguchi, T.; Geim, A. K.; Blake, P. Small 2011, 7, 465–468.

(23)

Al Balushi, Z. Y.; Wang, K.; Ghosh, R. K.; Vilá, R. A.; Eichfeld, S. M.; Caldwell, J. D.; Qin, X.; Lin, Y.-C.; DeSario, P. A.; Stone, G.; Subramanian, S.; Paul, D. F.; Wallace, R. M.; Datta, S.; Redwing, J. M.; Robinson, J. A. Nat. Mater. 2016, 15, 1166.

(24)

Goldberger, J.; He, R.; Zhang, Y.; Lee, S. Nature 2003, 422, 599–602.

(25)

Wu, Q.; Hu, Z.; Wang, X.; Lu, Y.; Chen, X.; Xu, H.; Chen, Y. J. Am. Chem. Soc. 2003, 125, 10176–10177.

(26)

Yoon, A.; Park, J. Y.; Jeon, J.-M.; Cho, Y.; Park, J. B.; Yi, G.-C.; Oh, K. H.; Han, H. N.; Kim, M. Small 2013, 9, 2255–2259.

(27)

Heinz, D.; Fikry, M.; Aschenbrenner, T.; Schowalter, M.; Meisch, T.; Madel, M.; Huber, F.; Hocker, M.; Frey, M.; Tischer, I.; Neuschl, B.; Mehrtens, T.; Müller, K.; Rosenauer, A.; Hommel, D.; Thonke, K.; Scholz, F. Phys. Status Solidi C 2014, 11, 648–651.

(28)

Hong, Y. J.; Jeon, J.-M.; Kim, M.; Jeon, S.-R.; Park, K. H.; Yi, G.-C. New J. Phys. 2009, 11, 125021.

(29)

Lee, C. H.; Yoo, J.; Hong, Y. J.; Cho, J.; Kim, Y. J.; Jeon, S. R.; Baek, J. H.; Yi, G. C. Appl. Phys. Lett 2009, 94, 213101.

(30)

Lee, C. H.; Hong, Y. J.; Kim, Y. J.; Yoo, J.; Baek, H.; Jeon, S. R.; Lee, S. J.; Yi, G. C. IEEE J. Sel. Top. Quantum Electron. 2011, 17, 966–970.

(31)

Tchoe, Y.; Lee, C.-H.; Park, J. B.; Baek, H.; Chung, K.; Jo, J.; Kim, M.; Yi, G. ACS Nano 2016, 10, 3114–3120.

(32)

Hemmingsson, C.; Pozina, G.; Khromov, S.; Monemar, B. Nanotechnology 2011, 22, 085602.

ACS Paragon Plus Environment

22

Page 23 of 30

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

(33)

Li, H.; Ji, X.; Wu, Z.; Hu, H.; Yan, D.; Zhuo, R.; Feng, J.; Geng, B.; Yan, P. Appl. Phys. Express 2014, 7, 065003.

(34)

Liu, B.; Bando, Y.; Tang, C.; Shen, G.; Golberg, D.; Xu, F. Appl. Phys. Lett. 2006, 88, 093120.

(35)

Liu, B.; Bando, Y.; Wang, M.; Tang, C.; Mitome, M.; Golberg, D. Nanotechnology 2009, 20, 185705.

(36)

Schuster, F.; Hetzl, M.; Weiszer, S.; Garrido, J. A.; Magen, C.; Arbiol, J.; Stutzmann, M. Nano Lett. 2015, 15, 1773–1779.

(37)

Stan, G.; Ciobanu, C. V; Thayer, T. P.; Wang, G. T.; Creighton, J. R.; Purushotham, K. P.; Bendersky, L. A; Cook, R. F. Nanotechnology 2009, 20, 035706.

(38)

Li, H.; Zhao, G.; Kong, S.; Han, D.; Wei, H.; Wang, L.; Chen, Z.; Yang, S. Nanoscale 2015, 7, 16481.

(39)

Park, Y. S.; Lee, G.; Holmes, M. J.; Chan, C. C. S.; Reid, B. P. L.; Alexander-Webber, J. A.; Nicholas, R. J.; Taylor, R. A.; Kim, K. S.; Han, S. W.; Yang, W.; Jo, Y.; Kim, J.; Im, H. Nano Lett. 2015, 15, 4472–4476.

(40)

Li, C.; Liu, S.; Hurtado, A.; Wright, J. B.; Xu, H.; Luk, T. S.; Figiel, J. J.; Brener, I.; Brueck, S. R. J.; Wang, G. T. ACS Photonics 2015, 2, 1025–1029.

(41)

Koester, R.; Hwang, J. S.; Durand, C.; Dang, D. L. S.; Eymery, J. Nanotechnology 2010, 21, 015602.

(42)

Eymery, J.; Salomon, D.; Chen, X.; Durand, C. Process for Catalyst-Free Selective Growth on a Semiconductor Structure. Patent WO2012136665, 2012.

(43)

Tessarek, C.; Heilmann, M.; Butzen, E.; Haab, A.; Hardtdegen, H.; Dieker, C.; Spiecker, E.; Christiansen, S. Cryst. Growth Des. 2014, 14, 1486–1492.

(44)

Tchoulfian, P.; Donatini, F.; Levy, F.; Amstatt, B.; Dussaigne, A.; Ferret, P.; Bustarret, E.; Pernot, J. Appl. Phys. Lett. 2013, 103, 202101.

(45)

Tchoulfian, P.; Donatini, F.; Levy, F.; Dussaigne, A.; Ferret, P.; Pernot, J. Nano Lett. 2014, 14, 3491–3498.

(46)

Durand, C.; Bougerol, C.; Carlin, J.-F.; Rossbach, G.; Godel, F.; Eymery, J.; Jouneau, P.; Mukhtarova, A.; Butté, R.; Grandjean, N. ACS Photonics 2014, 1, 38–46.

(47)

Arita, M.; Kako, S.; Iwamoto, S.; Arakawa, Y. Appl. Phys. Express 2012, 5, 126502.

(48)

Chen, X. J.; Perillat-Merceroz, G.; Sam-Giao, D.; Durand, C.; Eymery, J. Appl. Phys. Lett. 2010, 97, 151909.

ACS Paragon Plus Environment

23

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 24 of 30

(49)

Coulon, P. M.; Mexis, M.; Teisseire, M.; Jublot, M.; Vennéguès, P.; Leroux, M.; Zuniga-Perez, J. J. Appl. Phys. 2014, 115, 153504.

(50)

Labat, S.; Richard, M.-I.; Dupraz, M.; Gailhanou, M.; Beutier, G.; Verdier, M.; Mastropietro, F.; Cornelius, T. W.; Schülli, T. U.; Eymery, J.; Thomas, O. ACS Nano 2015, 9, 9210–9216.

(51)

Cosendey, G.; Carlin, J.-F.; Kaufmann, N. A. K.; Butté, R.; Grandjean, N. Appl. Phys. Lett. 2011, 98, 181111.

(52)

Kesaria, M.; Shetty, S.; Shivaprasad, S. M. Cryst. Growth Des. 2011, 11, 4900–4903.

(53)

Martínez-Criado, G.; Villanova, J.; Tucoulou, R.; Salomon, D.; Suuronen, J.-P.; Labouré, S.; Guilloud, C.; Valls, V.; Barrett, R.; Gagliardini, E.; et al., J. Synchrotron Radiat. 2016, 23, 344–352.

(54)

Maier, D.; Alomari, M.; Grandjean, N.; Carlin, J.-F.; Diforte-Poisson, M. A.; Dua, C.; Delage, S.; Kohn, E. IEEE Electron Device Lett. 2012, 33, 985–987.

(55)

Medjdoub, F.; Carlin, J.-F.; Gonschorek, M.; Feltin, E.; Py, M. A.; Ducatteau, D.; Gaquière, C.; Grandjean, N.; Kohn, E. IEEE Electron Devices Meet. Tech. Dig., Dec. 11–13, 2006, 1–4.

(56)

Butté, R.; Carlin, J.-F.; Feltin, E.; Gonschorek, M.; Nicolay, S.; Christmann, G.; Simeonov, D.; Castiglia, A.; Dorsaz, J.; Buehlmann, H. J.; Christopoulos, S.; Baldassarri Höger von Högersthal, G.; Grundy, A. J. D.; Mosca, M.; Pinquier, C.; Py, M. A.; Demangeot, F.; Frandon, J.; Lagoudakis, P. G.; Baumberg, J. J.; Grandjean, N. J. Phys. D Appl. Phys. 2007, 40, 6328–6344.

(57)

Butté, R.; Cosendey, G. ; Lugani, L. ; Glauser, M. ; Castiglia, A.; Perillat-Merceroz, G. ; Carlin, J.-F.; Grandjean, N., Book « III-Nitride Semiconductors and their Modern Devices », p. 177 (Chap. 6), edited by B. Gil : Oxford University Press, Oxford, 2013.

(58)

Rol, F.; Founta, S.; Mariette, H.; Daudin, B.; Dang, L. S.; Bleuse, J.; Peyrade, D.; Gérard, J. M.; Gayral, B. Phys. Rev. B 2007, 75, 125306.

ACS Paragon Plus Environment

24

Page 25 of 30

Figure 1 a)

tshell=5 nm tshell=0.5 nm

b)

c)

500#nm#

10 nm 2 µm

1 µm

2 µm

10-10

d)

ACS Paragon Plus Environment

void

GaN wires

Sapphire (c-plane)

void

H2 NH3

0002

1 nm

In situ GaN etching

GaN wires

InAlN shell

GaN wires with 0.5-5 nm InAlN shell



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

Nano Letters

Nano Letters

Figure 2

Unetched GaN wire

MQWs

200#nm#

70#±5#nm# 500#nm#

H2 NH3

Sapphire (c-plane)

ACS Paragon Plus Environment

void

void

b)

c)

In situ GaN etching

GaN wires

Par7al etching with GaN residuals

GaN wires with core-shell MQWs

RIE etching

2 µm

GaN wires

a)

MQW-tube

10x GaN/InAlN MQWs

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

Page 26 of 30

Figure 3 SEM*

S$XRF*

TOF-SIMS mappings of tubewall fragment lying on reference substrate contain>ng 5 InGaN/GaN MQWs

MQW$tubes* A.er&etching& A.er&etching& A.er&etching& A.er&etching&

MQW$ tube)

In#Ka#

b) In#Ka# In#Ka#

Ga#Ka# In#Ka#

Ga#Ka# Ga#Ka#

Before&etching& Before&etching& MQW$wires* Before&etching& Before&etching& Ga#Ka#

a)

SEM*

S$XRF*

MQW$ tube)

Tube fragment

SEM*

S$XRF*

e)

S$XRF* ACS Paragon Plus Environment

Ga

1 µm

5x InGaN/GaN MQWs

non$ etched) GaN)

SEM*

In

Reference substrate

In#Ka#

d) MQW$tubes* A.er&etching& A.er&etching&

c)

Al

9x GaN/InAlN MQWs

non$ etched) GaN)

In#Ka#

Ga#Ka#

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 tching& 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47

Nano Letters

20 nm

Page 27 of 30

ToF-SIMS

Nano Letters

Figure 4

b) 5$μm$

SEM$ MQW$$emission$ CL$−$332$nm$

GaN$emission$ CL$−$350$nm$

5 K

4.4

1.0

4.2

5K

4

3.8

3.6

MQWs

3.4

3 (eV)

3.2

MQW-wires MQW-tubes

GaN

327.4 nm ~260 meV

0.5 329.8 nm ~100 meV

c) PL intensity (arb. u.)

a)

Normalized PL (arb. u.)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 10$μm$ 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

Page 28 of 30

4.5

4

(eV)

GaN MQWs

104

10 K

3

300 K

10

102

0.0 280

3.5

MQW-tubes 300

320

340

360

380

Wavelength (nm)

ACS Paragon Plus Environment

400

420

300

350

400

Wavelength (nm)

Page 29 of 30

Figure 5

3.8

3.4

Energy (eV) 4

3.95

3.9

3.85

3.8

SEM"

1 2 3 4

3.9

c)

(eV)

1""+" 2""+"

5K

10"μm"

Wavelength (nm)

4

3.8

380

3.6

3.4

5K

1 2 3

SEM"

1""""+" 2""""+"

20K

40K

(

360

be

340

#Tu

(eV)

320

5K

PL intensity (arb. u.)

4""+"

330 (nm)

320

Wire % QD s%

3""+"

300

b)

3.6

%

(eV)

3.8

3""""+"

QD

Normalized PL (arb. u.)

4

Tube

a)

PL intensity (arb. u.)

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

Nano Letters

60K

2"μm"

80K 100K 500"nm"

300

320

340

Wavelength (nm)

360

ACS Paragon Plus Environment

310

315

320

325

Wavelength (nm)

Nano Letters

TOC TOF-SIMS mappings of tubewall fragment lying on TOF-SIMS mappings of tubewall fragment lying on TOF-SIMS mappings of tubewall fragment lying on Tube-wall based on MQWs reference substrate contain>ng 5 InGaN/GaN MQWs reference substrate contain>ng 5 InGaN/GaN MQWs reference substrate contain>ng 5 InGaN/GaN MQWs tubes GaN/InAlN MQWs

200#nm#

9x GaN/InAlN 9x GaN/InAlN 9x GaN/InAlN MQWs MQWs MQWs

70#±5#nm# 500#nm#

Ga Ga Ga

20 nm

Tube Tube Tube fragment fragment fragment

In In In

20 nm

Al Al Al

1 µm 1 µm 1 µm

5K

320

Reference Reference Reference substrate substrate substrate

GaN

PL intensity20 nm (arb. units)

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

Page 30 of 30

360

Wavelength (nm)

5x InGaN/GaN 5x InGaN/GaN 5x InGaN/GaN MQWs MQWs MQWs

ACS Paragon Plus Environment