Core–shell GaN/AlGaN nanowires grown by selective area epitaxy

Abstract

GaN/AlGaN core–shell nanowires with various Al compositions have been grown on GaN nanowire array using selective area metal organic chemical vapor deposition technique. Growth of the AlGaN shell using pure N₂ carrier gas resulted in a smooth surface for the nonpolar m-plane sidewalls with superior optical properties, whereas, growth using a mixed N₂/H₂ carrier gas resulted in a striated surface similar to the commonly observed morphology in the growth of nonpolar III-nitrides. The Al compositions in the AlGaN shells are found to be less than the gas phase input ratio. The systematic reduction in efficiency of Al incorporation in the AlGaN shells with increasing the Al molar flow in the gas phase is attributed to geometric loss, strain-limited Al incorporation, and increased gas phase parasitic reactions. Defect-related luminescence has been observed for AlGaN shells with Al content ≥ 30% and the origin of the defect luminescence has been determined as the (V_III-2O_N)¹⁻ complex. Microstructural analysis of the AlGaN shells revealed that the dominant defects are partial dislocations. Growth of the nonpolar m-plane Al_xGa_1−xN/Al_yGa_1−yN quantum wells on the sidewalls of the GaN nanowires produced arrays with excellent morphology and optical emission, which demonstrated the viability of such a growth scheme for large area efficient ultraviolet LEDs as well as for next generation ultraviolet micro-LEDs.

---

New concepts

The concept of core–shell GaN/AlGaN nanowires has an unprecedented potential in ultraviolet (UV) light emitting diodes (LEDs) to overcome the existing performance issues of UV light sources. The efficiencies of current state of the art AlGaN-based solid-state UV LEDs, grown on planar c-plane substrates are limited to a modest 20% or below, while those of blue LEDs have surpassed 80%. The central concept of core–shell GaN/AlGaN nanowires is to resolve multiple issues simultaneously rather than solving one at a time. Specifically, the high threading dislocation density and its deleterious effect on the internal quantum efficiency, the low intensity of the emitted UV light from the c-plane due to transverse magnetic polarization, and poor light extraction efficiency due to small light-escape cones from existing planar UV LEDs. A proof of concept study demonstrating the achievement of UV emission from Al_xGa_1−xN/Al_yGa_1−yN multiple quantum wells on core–shell GaN/AlGaN nanowires that are free of threading dislocations, where the emission is from the nonpolar m-plane, and the issue of small light-escape cones in the conventional planar UV LEDs have been resolved, is presented. These results are important for the scalable production of nanowire-based UV LEDs.

Introduction

AlGaN with a tunable bandgap from 3.4 to 6.2 eV, is an important semiconductor alloy for applications in ultraviolet (UV) light sources and detectors. However, AlGaN heterostructures that are grown on the polar c-plane encounter the quantum-confined Stark effect (QCSE) due to spontaneous and piezoelectric polarizations, resulting in a redshifted emission with reduced internal quantum efficiency (IQE) due to the spatial separation of the electron and hole wavefunctions. In spite of the deleterious effect of the QCSE due to spontaneous and piezoelectric polarizations, Simon et al. reported an unconventional doping phenomenon called polarization-induced doping, applicable to semiconductor crystals with inherent spontaneous and piezoelectric polarizations.¹ Polarization-induced doping in AlGaN is achieved by incorporating an Al-graded AlGaN layer in the polar c-direction, whereby, the polarization-induced charge accumulated at an interface (two dimensional electron or hole gas) is spread over to a three dimensional region, resulting in an increased carrier concentration. Using an Al-graded p-AlGaN polarization-induced doped layer combined with excimer laser annealing, Khan et al. achieved a hole concentration of 2 × 10¹⁶ cm⁻³,² and incorporating polarization-induced doping in a UV LED structure with an emission wavelength of 304 nm resulted in an improved EQE of 8.2% compared to 6.4% for p-AlGaN without polarization-induced doping.³ Polarization-induced doping is limited to growth in the polar c-direction and this phenomenon does not occur in m- and a-nonpolar oriented AlGaN growth due to the absence of polarizations in these directions. Further, reordering of the topmost valence band from heavy hole (HH) to crystal field split-off hole (CH) band in c-plane AlGaN, depending on the Al-content, strain, and quantum size effect leads to polarization switching from transverse electric (TE, E ⊥ c) to transverse magnetic (TM, E ‖ c) causing inefficient light emission along the surface normal for c-plane based devices.⁴

The issues of QCSE and polarization switching pertaining to the polar c-plane AlGaN resulting in low efficiency can be alleviated by using nonpolar m-, and a-plane orientations. Planar GaN/AlGaN quantum wells (QWs) grown on the nonpolar m-plane have demonstrated QCSE-free emission with improved performance.⁵ Light emission from QWs on nonpolar AlGaN is always in-plane polarized and does not suffer from polarization switching as in the case of c-plane AlGaN, hence, efficient UV LEDs can be achieved over the entire Al composition range.^6,7 The anisotropic in-plane stress for multiple quantum wells (MQWs) on nonpolar planes provides the additional advantage to achieve a polarized light source to further enhance the efficiency of LEDs and lasers.⁸ Nonpolar m-plane InGaN vertical-cavity surface-emitting lasers emitting at 405 nm with a 100% polarization ratio,⁹ and a-plane AlN LED emitting at 210 nm with a polarization ratio as high as 90% and a 25 times higher emission intensity compared to the c-plane AlN LED have been reported as a testament to the advantages of nonpolar orientations.¹⁰

The highest external quantum efficiency (EQE) of 20.3% for a planar AlGaN-based UV LED was reported by Takano et al. in 2017 and this record still holds today.¹¹ Recent reports have shown that UV emission from AlGaN can be further enhanced by using AlGaN nanostructures.^12–14 Additionally, a regular arrangement of nanowires as an array is also beneficial for enhancing light extraction efficiency, which is a major bottleneck in planer c-plane AlGaN LEDs.¹⁵ In this context, selective area grown GaN nanowires, free of threading dislocations, can provide an excellent platform for the growth of nonpolar AlGaN shell QWs. Lateral growth of AlGaN in the form of a shell around a GaN nanowire is also commensurate with the conventional lateral epitaxial overgrowth technique known for reducing defects and improving crystalline quality. Shatalov et al. have highlighted the importance of reducing threading dislocations to improve IQE in AlGaN-based UV LEDs.¹⁶ They have achieved a significant increase in IQE from 20% to over 55% by reducing the threading dislocation density, resulting in a UV LED with an EQE of 10.4%. In comparison to the planar heteroepitaxy, the nanowire heterostructure has a higher degree of elastic relaxation and enables growth of a thicker shell as well as a higher alloy composition without the formation of surface cracks.^17–19 Using nanostructures, Mi et al. have achieved a commendable IQE of 80% in AlN nanowires, demonstrating the advantage of nanowires.²⁰ The large surface area to volume ratio of AlGaN nanostructures makes them suitable for incorporating efficient QWs. AlGaN QWs on nonpolar orientations are also known to have enhanced confinement potentials leading to a faster transition of electrons resulting in enhanced UV emission.²¹

Despite the high potential of AlGaN nanostructures for improving the performance of UV LEDs, there is a lack of detailed investigations on the growth of GaN/AlGaN core–shell structures. Here we report the growth of AlGaN shells on GaN nanowires using metal organic chemical vapor deposition (MOCVD). We carried out a systematic investigation on the surface morphology, Al incorporation, optical properties, nature of point defects and microstructural defects in these nanowires. Consequently, a nanowire array with nonpolar m-plane Al_xGa_1−xN/Al_yGa_1−yN MQWs has been grown, demonstrating the potential implication for scalable growth of nanowire-based UV LEDs using MOCVD.

Experimental methods

The GaN/AlGaN core–shell nanowires were grown using an Aixtron 3 × 2′′ close-coupled showerhead MOCVD. Prior to the growth of the AlGaN shell, selective area growth of the GaN nanowires was carried out on a 4 μm-thick c-plane GaN pseudo-substrate (GaN on a 2 inch c-plane sapphire substrate). Substrate preparation for the selective area growth was carried out by depositing a 15 nm-thick SiO_x layer on the substrate using an atomic layer deposition system (Picosun SUNALE). It was then followed by spin-coating a 100 nm-thick electron beam resist (mr-PosEBR, micro resist technology GmbH) layer on top. A water-based anti-charging agent for electron beam lithography (DisCharge, DisChem Inc., USA) was then coated on top of the electron beam resist. Circular holes of 100 nm diameter were defined in a hexagonal array using electron beam lithography (Raith 150) and subsequent development of the electron beam resist in n-amyl acetate. The developed pattern was then transferred to the substrate through reactive ion etching of the SiO_x layer using CHF₃. The substrate was then cleaned using acetone, isopropyl alcohol and deionized water. Additionally, oxygen plasma treatment was also carried out to clean any residual organic contaminants. The 2-inch GaN pseudo-substrate was then diced into 10 × 10 mm² pieces using a DAD321 dicing machine (Disco, Japan). A final cleaning of the substrate was carried out again using acetone, isopropyl alcohol and deionized water before loading into the MOCVD.

Selective area growth of GaN nanowires was carried out using triethylgallium (TEGa) and ammonia precursors in a mixed carrier gas of N₂ + H₂. Growth of the GaN nanowires was conducted at a temperature of ∼956 °C for 25 minutes. Further details on selective area growth of GaN nanowires has been reported elsewhere.²² GaN nanowire growth was immediately followed by AlGaN shell growth without any growth interruption by introducing trimethylaluminum (TMAl) and reducing the TEGa flow rate. During the AlGaN growth, the temperature was ramped down to ∼925 °C to promote lateral growth and a trimethylindium flow of 5 μmol min⁻¹ was added to promote adatom mobility. Unlike the N₂ + H₂ mixed carrier gas used during GaN nanowire growth, AlGaN growth was carried out in a pure N₂ carrier gas, unless specified otherwise.

The morphology of the GaN/AlGaN core–shell structures was analyzed using a scanning electron microscope (SEM, FEI Verios 460). Energy dispersive X-ray spectroscopy (EDS, Oxford) and room temperature cathodoluminescence (CL, Gatan MonoCL4 Elite) measurements were carried out using the same SEM. The electron transparent lamellae of the nanowires for the transmission electron microscopy studies were prepared using a focused ion beam (FIB, Zeiss Crossbeam 550). The microstructural characterization of the nanowires was carried out using a transmission electron microscope (TEM, JEOL 2100F FEGTEM) operating at an accelerating voltage of 200 kV.

Results and discussions

AlGaN shells with different Al compositions were grown on GaN nanowire sidewalls using MOCVD. Table 1 lists the samples, gas-phase Al percentage, flow rates of precursors, and the V/III used. The V/III ratio during the growth of the AlGaN shells was maintained at about 35.5 as listed in Table 1.

Table 1 Sample IDs, gas-phase Al percentage, corresponding flow rates of TMAl, TEGa, NH₃, and V/III ratio used for the growth of different samples

Sample	Al	TMAl		TEGa		TMAl + TEGa	NH3	V/III
	(%)	(sccm)	(μmol min^−1)	(sccm)	(μmol min^−1)	(μmol min^−1)	(sccm)
Al-15	15.0	3	2.64	44	14.99	17.63	14	35.4
Al-20	20.1	4	3.52	41	13.97	17.49	14	35.7
Al-30	30.1	6	5.28	36	12.26	17.55	14	35.6
Al-40	40.0	8	7.04	31	10.56	17.60	14	35.5
Al-50	49.8	10	8.80	26	8.86	17.66	14	35.4
Al-60	60.8	12	10.56	20	6.81	17.38	14	35.9

---

Morphology with carrier gas

Initially, we carried out the growth of the AlGaN shells in a mixed carrier gas composed of 6N₂ + 4H₂, adopting the similar mixed carrier gas used for the growth of the GaN nanowires. The AlGaN shell grown in this mixed carrier gas shows striations on the m-plane sidewalls as can be seen in Fig. 1a. Undulated or striated surfaces have been commonly observed in the conventional planar homoepitaxial growth of nonpolar m-plane GaN and AlN.^23–29 In such planar homoepitaxial growth, the undulations have been reduced by using miscut substrates in the case of GaN,^23,30,31 and by using a high growth temperature (>1350 °C) in the case of AlN.^29,32 However, in the heteroepitaxy of m-plane AlGaN on GaN nanowire sidewalls, a preferential miscut is not defined for the nanowire sidewalls and the high temperature growth required in the case of AlN risks thermal decomposition of the GaN nanowires. Nonetheless, a smooth sidewall surface of nonpolar m-plane AlGaN is a prerequisite for obtaining a sharp interface if the radial heterostructures are to be implemented in practical applications.

Fig. 1 SEM images of AlGaN on GaN nanowires. (a) AlGaN growth using a 6N₂ + 4H₂ carrier gas. (b) Panchromatic CL of a single nanowire from (a). (c) AlGaN growth using a pure N₂ carrier gas. (d) Panchromatic CL of a single nanowire from (c). (e–g) SEM images of samples Al-15, Al-30, and Al-50, respectively, grown in a pure N₂ carrier gas.

We have achieved AlGaN shells with smooth sidewalls when a pure N₂ carrier gas was used as shown in Fig. 1c. Barry et al. have recently reported the presence of a heavily undulated surface in the planar homoepitaxial growth of m-plane GaN using a H₂ carrier gas, whereas, growth in a N₂ carrier gas promoted the formation of atomically flat surface resulting in m-plane GaN with reduced oxygen, carbon, and silicon impurities.³³ Earlier work on using pure N₂ as a carrier gas have also demonstrated significantly improved homogeneity, layer purity, and crystal quality in Al-containing ternary AlGaAs alloys,³⁴ and superior optical properties in a planar AlN epitaxial layer.³⁵ Similar to our growth of GaN nanowires in the presence of H₂ followed by AlGaN grown in a pure N₂ carrier gas, Yamaguchi et al. reported a planar GaN grown in a H₂ carrier gas followed by a GaN/Al_0.17Ga_0.83N MQW grown in a N₂ carrier gas which had had superior crystalline and optical properties.³⁶ A recent study on the effect of H₂ and N₂ carrier gasses for the growth of a planar AlGaN layer in a GaN/AlGaN high-electron mobility transistor also revealed significantly improved surface morphology, interface, and crystalline quality for AlGaN grown in a N₂ carrier gas, resulting in excellent two-dimensional electron gas mobility exceeding 2000 cm² V⁻¹ s⁻¹.³⁷

The improvement in surface morphology of AlGaN by using a N₂ carrier gas can be attributed to the reduced disparity in the diffusion lengths of Al and Ga adatoms. In the growth of semiconductor alloys, the adatom mobility difference poses a significant issue as it could lead to substantial decomposition, stoichiometric issues, step-bunching and surface faceting.³⁸Fig. 2a and b show the schematic diagrams of AlGaN deposition using N₂ + H₂ and pure N₂ carrier gasses, respectively. In the case of the N₂ + H₂ carrier gas, a large disparity in the Al and Ga adatom diffusion lengths, as indicated by the blue arrows, results in a striated surface (Fig. 2a). Hence, it is imperative to minimize the disparity of the adatom mobilities during the growth of semiconductor alloys. Considering the diffusion lengths of the Al and Ga adatoms are 31–44 nm,³⁹ and 300–800 nm,⁴⁰ respectively, the difference between the Al and Ga adatom diffusion lengths can be minimized either by increasing the Al adatom mobility or by decreasing the Ga adatom mobility. An improved interface and surface morphology of AlGaN grown in the N₂ carrier gas has been attributed to significant enhancement of the Al adatom mobility in the N₂ carrier gas.³⁷ A Monte Carlo simulation has also shown that the Al adatom mobility is increased in the N₂ carrier gas compared to the H₂ carrier gas.⁴¹ Thus, the increased mobility of the Al adatom in the N₂ carrier gas is expected to reduce the disparity in the adatom mobilities between Ga and Al. Further, it has also been reported that the diffusion length of Ga is reduced by about 4 times if N₂ is used as the carrier gas instead of H₂.⁴² Hence, using a pure N₂ carrier gas proves to be effective in minimizing the difference between the Al and Ga adatom diffusion lengths, as illustrated by the blue arrows (Fig. 2b), resulting in smooth sidewalls of the AlGaN nanowires.

Fig. 2 Schematic diagrams of AlGaN deposition (a) using a N₂ + H₂ carrier gas showing a striated surface due to the large disparity in diffusion lengths of the Al and Ga adatoms as indicated by the blue arrows, (b) using a pure N₂ carrier gas showing the smooth sidewall surface due to the reduced disparity in the diffusion lengths of the Al and Ga adatoms.

Panchromatic CL images of the AlGaN shells grown in 6N₂ + 4H₂ and pure N₂ carrier gasses are shown as Fig. 1b and d, respectively. It is evident that the AlGaN grown in a pure N₂ carrier gas shows brighter luminescence compared to the ones grown using a mixed carrier gas of 6N₂ + 4H₂. Near-band-edge (NBE) emission was observed only for the AlGaN grown using a pure N₂ carrier gas and is attributed to the better crystalline quality and reduced impurity incorporation for the AlGaN grown in a N₂ carrier gas. In the panchromatic CL image of Fig. 1d, most parts of the nanowire show uniform brightness; however, differences in brightness could be observed near the top region of the nanowire possibly due to local composition fluctuations. Nonetheless, with the evidence of superior surface and optical quality for the AlGaN grown in a N₂ carrier gas compared to the ones grown using a N₂ + H₂ carrier gas, further growths of the AlGaN shells were carried out in pure a N₂ carrier gas. Fig. 1e–g show the SEM images of the GaN/AlGaN core–shell nanowires with gas phase Al compositions (i.e. TMAl/(TMAl + TEGa)) of 15, 30 and 50%, respectively. The diameters of these GaN/AlGaN nanowires are 563 ± 55, 504 ± 37, and 433 ± 37 nm for Al-15, Al-30, and Al-50, respectively, with corresponding heights of 1263 ± 37, 1210 ± 27, and 1659 ± 21 nm. Thus, a uniform array of AlGaN shells with different Al compositions can be grown reproducibly on GaN nanowires using MOCVD.

AlGaN alloy composition: SEM-EDS

The elemental composition of the AlGaN shells grown on GaN nanowires was analyzed using an SEM-EDS system on samples Al-15, Al-30, Al-50, and Al-60. In order to avoid the erroneous underestimation of Al composition due to the underlying GaN core, SEM-EDS analysis were carried out on the thin lamella prepared for the TEM analysis. Fig. 3a and b shows the elemental composition of the AlGaN shells along the height (line scan of the AlGaN shell regions on the lamella) and across the diameter of the lamella, respectively. The Al composition of the AlGaN shell is found to increase along the height in all of the nanowires studied due to the predominant mechanism of vapor-phase-diffusion during the growth of AlGaN on GaN nanowires. The top region receives a higher Al flux from the vapor-phase while the Al content in the vapor-phase decreases towards the bottom of the nanowire. Since Al has a higher sticking coefficient than Ga, this leads to preferential incorporation of Al at the top of nanowires, thereby gradually depleting the Al in the vapor-phase towards the bottom of nanowires resulting in a decreased Al incorporation. Analysis of the Al composition variation along the height of nanowires by fitting the data with the least square method determined that samples Al-15, Al-30, Al-50, and Al-60 have slopes of 7.73, 5.36, 5.53, and 7.98, respectively, thereby revealing that the Al composition changes by about 5–8% for every 1 μm of height in these samples. In the case of sample Al-60, the solid phase Al composition changes from ∼47% at the bottom to ∼54% at the top and hence a crossover is observed. The compositional nonuniformity along the height of the nanowire may therefore be limited by using relatively short nanowires (∼1 μm). Fig. 3b shows the Al composition across the diameter of the GaN/AlGaN lamella, revealing successful incorporation of the AlGaN shells around the GaN nanowire core.

Fig. 3 EDS elemental analysis of the lamella of the AlGaN nanowires of different Al compositions along the (a) height, and (b) diameter.

Table 2 lists the Al compositions determined from SEM-EDS for samples Al-15, Al-30, Al-50, and Al-60. The Al composition, Al/(Al + Ga), determined from SEM-EDS shows a lower Al content than those expected from the gas-phase, TMAl/(TMAl + TEGa), ratio used during the AlGaN shell growth. The data on the solid phase AlN fraction incorporated in the AlGaN grown by MOCVD scatters over a wide range due to the complex interdependence on the growth temperature, reactor pressure, gas-phase TMAl/III ratio, and V/III ratio.^43–46 Possible reasons for the reduced Al incorporation in the solid phase with respect to that of the gas-phase in our experiments will be discussed later.

Table 2 List of samples, gas phase Al content and solid phase Al content determined by SEM-EDS and SEM-CL

Sample	Al content (%)
	Gas phase	SEM-EDS	SEM-CL
Al-15	15.0	11.7 ± 2.2	10.3
Al-20	20.1	—	15.0
Al-30	30.1	23.1 ± 2.5	22.0
Al-40	40.0	—	30.5
Al-50	49.8	39.8 ± 1.8	40.2
Al-60	60.8	50.5 ± 2.3	—

---

AlGaN alloy composition: SEM-CL

The room-temperature cathodoluminescence (CL) measurement results on samples Al-15, Al-20, Al-30, Al-40, and Al-50 are shown in Fig. 4a. A systematic shift of the NBE emission is observed with the change in Al composition. The NBE peak is observed at 3.58 eV for the sample Al-10 and increases to 4.14 eV for the sample Al-50. The dashed line in Fig. 4a indicates the NBE emission from the GaN nanowire core. Based on the NBE emission from the AlGaN shells, the Al content of the samples are determined by considering that the m-plane AlGaN shells grown on the m-plane GaN sidewalls are pseudomorphically strained on the GaN nanowire core. GaN has a larger lattice constant than Al(Ga)N, therefore, the AlGaN shell experiences a tensile strain, resulting in a redshifted emission. Hence, we calculated the NBE emission energy for m-plane AlGaN grown on GaN using the lattice parameters of GaN and AlN,⁴⁷ their bandgap energies, elastic constants,⁴⁸ and deformation potentials.⁴⁹ The details of the calculation and the parameters used are included in the ESI.† Briefly, the NBE emission energy of AlGaN grown on GaN is lower than that expected from simple extrapolation using Vegard's law due to the presence of tensile strain. In conventional c-plane oriented growth of AlGaN on GaN, AlGaN experiences a biaxial strain due to the mismatch of the lattice parameter c alone; however, in the case of m-plane AlGaN grown on GaN, there exists an anisotropic tensile strain due to a mismatch in the lattice parameters c and a, along the [0001] and [112̄0] directions, respectively, as shown in Fig. 5b. The calculated NBE emission of AlGaN on c-plane GaN is shown with a green line in Fig. 4b, revealing a redshift due to tensile strain compared to the unstrained NBE emission (blue line). Similarly, the calculated NBE emission of AlGaN on m-plane GaN is shown with a red line in Fig. 4b. The solid phase Al composition of the AlGaN shells are determined from the calculated bandgap of strained AlGaN on m-plane GaN (red line of Fig. 4b) by plotting the experimentally determined NBE emission energies represented by black triangles in Fig. 4b. The solid phase Al compositions of the AlGaN shells determined from SEM-CL are also listed in Table 2 alongside the composition determined from SEM-EDS. The Al compositions determined using SEM-EDS and SEM-CL match fairly well, and the determination of the Al composition using the CL measurements by plotting the experimental data against the calculated bandgap suggests that the AlGaN shells are in a strained state.

Fig. 4 (a) Cathodoluminescence spectra of GaN/AlGaN nanowire samples with different Al-compositions as indicated. (b) Calculated bandgap energy for strain-free AlGaN (blue line), AlGaN strained on c-plane GaN (green line), and AlGaN strained on m-plane GaN (red line). The black triangular data points correspond to the NBE emission energies of GaN/AlGaN nanowires obtained from cathodoluminescence. (c) Deep acceptor levels in AlGaN classified as (V_III)³⁻, (V_III complex)²⁻, and (V_III complex)¹⁻ with data from Nepal et al.⁵⁰ Current data from the samples Al-40 and Al-50, with solid phase Al compositions of ∼30 and 40%, respectively, are shown as red diamonds.

Fig. 5 (a) Reduced Al incorporation in the AlGaN shell due to geometric loss and strain + parasitic loss. (b) Schematic representation of the GaN/AlGaN core–shell nanowire showing the anisotropic strain due to mismatch in the c and a lattice constants.

Cathodoluminescence: defect luminescence

Samples Al-15 to Al-30 show dominant NBE emissions only as can be seen in Fig. 4a. However, defect-related broad luminescence peaks appeared in addition to the NBE emission for the higher Al containing samples of Al-40 and Al-50 with solid phase Al compositions of ∼30 and 40%, respectively, as can be seen in Fig. 4a. Hoshi et al. have reported similar broad luminescence peaks in m-plane AlGaN grown on a freestanding GaN substrate for an Al content ≥25% and suggested the origin to be a DX centre.⁵¹ Further, these broad peaks indicate the presence of two peaks, where one peak is centered around 3.56 eV as indicated by the vertical dashed line, and another is centered near 3.64 eV as indicated by the red diamonds in Fig. 4a. The first peak at 3.56 eV is assigned to the GaN nanowire core which was confirmed after further measurement (Fig. 9). The second emission peak at 3.64 eV is assigned to an unintentional impurity level. In order to determine the type of this impurity emission, we calculated the acceptor energy level, E_A, according to Nepal et al. as,⁵⁰

E_A = E_g(x) − E_impurity − E_D + E_V (1)

Here, E_g(x) is the bandgap of Al_xGa_1−xN, E_impurity is the photon energy of the defect peak measured from CL, and E_D and E_V are the energy levels of the shallow donor and valence band, respectively. The acceptor energy levels, E_A, calculated for the samples Al-40 and Al-50 with solid phase Al contents of ∼30 and 40%, respectively, are plotted as red diamonds in Fig. 4c. From the trend of the pinned defect energy levels shown in Fig. 4c, the location of these acceptor impurity levels in the forbidden gap suggests that the defects in the samples Al-40 and Al-50 are a (V_III-complex)¹⁻. Oxygen-related defects in GaN,⁵² and AlN,⁵³ show luminescence at 2.8 and 4.7 eV, respectively, which are located at the ends of the horizontal line indicating the (V_III-complex)¹⁻ in Fig. 4c. Hence, the defect luminescence near 3.64 eV in the GaN/AlGaN core–shell nanowire can be assigned to a (V_III-2O_N)¹⁻ defect complex formed by combination of a metal vacancy with a pair of substitutional oxygens at the N site.

Low Al incorporation

The solid phase Al compositions of the AlGaN shells, determined by both SEM-EDS and SEM-CL are lower than the values of the input gas phase ratio and are listed in Table 2. Non-planar growth and epitaxial lateral overgrowth experiments of AlGaN involving non-planar structures have consistently reported lower Al incorporation on the sidewall facets.^54–57 Lower Al incorporation on the sidewall facets has been attributed to different growth rates in the vertical and lateral directions, modification of the vapor phase Al and Ga adatom diffusion by the non-planar structures, and the difference in surface diffusion on the top and sidewall facets.⁵⁸ On the other hand, AlGaN growth on co-loaded planar horizontal substrates of different crystal orientations concluded that Al incorporation is comparable on polar (0001), semipolar (101̄3) and (112̄2), and nonpolar (101̄0) and (112̄0) planes.⁵⁹ Hence, we infer from these two distinct results that Al-incorporation on non-horizontal facets is reduced, and the geometrical non-planarity is the primary cause of low Al incorporation on the sidewall facets. Considering that a similar geometry of GaN nanowire arrays have been employed during the growth of the AlGaN shells, the reduced Al incorporation due to non-planar geometry is expected to be constant. Therefore, in Fig. 5a which shows the solid phase Al composition against the gas phase, we draw a hypothetical dashed-line parallel to the experimentally determined Al composition, and the gap between this hypothetical line and the experimentally determined Al composition is called the “geometric loss” as indicated in Fig. 5a.

Furthermore, the nonpolar m-plane AlGaN shells grown on GaN nanowire sidewalls experience an anisotropic tensile strain as shown in Fig. 5b, which is known to reduce Al incorporation.⁶⁰ The uniform Al composition, away from the heterointerface, is also determined by the residual strain in the heteroepitaxy of AlGaN.⁶¹ In addition to strain, the gas phase parasitic reaction increases with the flow rate of TMAl and leads to decreased Al incorporation efficiency at higher TMAl flow rates.⁶² Thus, the increased strain with higher Al-content in the heterostructure and gas-phase parasitic reaction with a higher TMAl flow rate manifests in a non-monotonic reduction of Al incorporation. With reference to our hypothetical line, a further deviation in Al incorporation with increasing TMAl flow rate can be seen in Fig. 5a due to increased strain and parasitic reaction and is indicated as “strain + parasitic loss”. Hence, the lower incorporation of Al in the solid phase GaN/AlGaN nanowire as compared to the gas phase molar ratio is primarily attributed to geometric loss. However, the gas-phase parasitic reaction of TMAl and strain limited Al incorporation could become a limiting factor to obtain AlGaN shell with higher Al compositions.

Transmission electron microscopy (TEM) studies

Cracks or dense misfit dislocations often appear in the heteroepitaxial growth of AlGaN on GaN in order to partially relieve the tensile strain resulting from lattice mismatch.⁶³ Further, in the case of nonpolar III-nitride heteroepitaxy, a large number of stacking faults and related partial dislocations (PDs) are also known to be present.^64–66 In order to verify the nature of the defects in our nanowires, we carried out a microstructural analysis of the AlGaN shells by transmission electron microscopy. Fig. 6 shows a high-resolution transmission electron microscopy (HRTEM) image of sample Al-15 with the corresponding Bragg filtered images and strain maps.

Fig. 6 Transmission electron microscopy and related images of sample Al-15. The c-axis is oriented horizontally in these images. (a) HRTEM image of the AlGaN shell. (b) Fast Fourier transform image highlighting the diffraction spots (0002), (11̄00). (c and d) Bragg filtered images of (a) using the diffraction spots (0002) and (11̄00), respectively. (e and f) In-plane and out-of-plane strain maps of the AlGaN shell (a).

Edge, screw, and mixed dislocations are the major defects in the c-plane oriented growth of III-nitrides, whereas, stacking faults terminated by partial dislocations (PDs) have been identified as the predominant defects in nonpolar m- and a-plane oriented growths.⁶⁷ Stacking faults and PDs in crystals can be discerned by analyzing the HRTEM images through their Bragg filtered images. Fig. 6a shows a HRTEM image of the m-plane AlGaN shell (sample Al-15) grown on a GaN nanowire sidewall, captured along the [112̄0] zone axis. The crystallographic orientation indicated in Fig. 6a is the same for Fig. 6a–f. The diffraction pattern obtained by fast Fourier transform of the HRTEM image is shown in Fig. 6b. The (0002) and (11̄00) diffraction spots used for the dislocation analysis and strain maps are indicated by the red and green circles, respectively. Fig. 6c and d show the Bragg filtered images of the HRTEM image, generated using the (0002) and (11̄00) diffraction spots, respectively, and they reveal the presence of three types of PDs, viz. Frank–Shockley, Shockley, and Frank PDs. Frank–Shockley PDs have a Burgers vector, b = 1/6 〈202̄3〉 which is composed of 1/2 〈0001〉 with an associated 1/3 〈11̄00〉 slip, while, Shockley and Frank PDs have only Burgers vectors of 1/3 〈11̄00〉 and 1/2 〈0001〉, respectively. The Frank–Shockley, Shockley, and Frank PDs are indicated with red, yellow and blue arrows, respectively, in Fig. 6. Hence, the dominant type of defects observed here in the m-plane AlGaN shell are PDs and not basal plane stacking faults (BSFs). Following the criteria that the formation of BSFs and dislocations in GaN/AlN heterostructures are enhanced in a nitrogen-rich growth environment,⁶⁸ Shao et al. achieved homogeneous nonpolar m-plane GaN/AlGaN superlattices using metal-rich growth conditions and characterized the defects as short BSFs with lengths in the range 2–15 nm bounded by Frank–Shockley or Frank PDs.⁶⁹ More recently, Moneta et al. reported a BSF-free 100 nm-thick InGaN layer grown on GaN under low N-flux growth conditions.⁷⁰ In our case, the growth of m-plane AlGaN on GaN nanowire sidewalls were carried out under a low NH₃ environment with a flow rate of 14 sccm and a V/III ratio of ∼35.5 (Table 1). It may be noted that, the pyrolysis of NH₃ is inefficient during the growth of the III-nitrides and only 26% of the NH₃ is cracked even at a high temperature of 1200 °C.⁷¹ We have earlier identified that a V/III ratio of 13 to 74 is akin to a metal-rich (or low N) regime and is suitable for GaN nanowire growth in our MOCVD.²² Hence, our growth condition of AlGaN in low NH₃ ambient conditions is considered to be the reason that BSFs have not been observed while PDs remain the dominant defect in the nonpolar m-plane AlGaN grown on GaN nanowire sidewalls under a low V/III ratio.

The in-plane and out-of-plane strain maps, calculated using geometric phase analysis (GPA)⁷² are shown as Fig. 6e and f, respectively. Local strain components determined from GPA can be used as a visual aid to locate the dislocation cores from strain maps. A sudden change in strain from being compressive to tensile or vice versa at the dislocation core results in a high contrast node (yellow for tensile and blue for compressive strain) as shown in Fig. 6e and f. Dislocation cores located by strain maps correspond to those identified by Bragg filtered images, and hence, the formation of PDs for the AlGaN shell with different Al compositions may be visualized using strain maps from GPA. An increase in the formation of PDs with increasing Al content in the AlGaN shells, as expected, due to the increased lattice mismatch between GaN and AlGaN is shown in Fig. S1 in the ESI† for samples Al-30 and Al-50.

Al_xGa_1−xN/Al_yGa_1−yN multiple quantum wells

With our results of the m-plane GaN/AlGaN core–shell nanowires, we proceeded to explore the growth of Al_xGa_1−xN/Al_yGa_1−yN multiple quantum wells (MQWs) on a GaN nanowire array with gas-phase Al-compositions of x = 17.3 and y = 46.3%. Fig. 7a shows the high-angle annular dark-field (HAADF)–scanning tunneling electron microscope (STEM) image, also referred to as a Z-contrast image, of the Al_xGa_1−xN/Al_yGa_1−yN MQW incorporated on the sidewalls of a GaN nanowire. The GaN nanowire core in the middle appears bright owing to the higher atomic number (Z) of Ga while the AlGaN barriers appear darker due to the effective lower atomic number of AlGaN. Incorporation of three Al_xGa_1−xN/Al_yGa_1−yN MQWs is clearly visible from the HAADF-STEM image. Further, a smooth surface, free of cracks and macro-steps is also evident for the m-plane AlGaN sidewalls. On the other hand, a rough surface due to parasitic polycrystalline AlGaN deposition can be seen on the SiO_x mask. A low-magnification TEM image of the MQW region, marked with a dashed square in Fig. 7a, is shown as Fig. 7b with the Al_xGa_1−xN wells and Al_yGa_1−yN barriers identified. A high resolution TEM image of the region indicated by a dashed square in Fig. 7b is shown as Fig. 7c highlighting the successful incorporation of the m-plane Al_xGa_1−xN/Al_yGa_1−yN MQW with a sharp interface. The quantum well region, grown for a duration of 4 minutes, is about 10 nm as evident from Fig. 7c, and the quantum well width may be modulated by changing the growth duration.

Fig. 7 (a) HAADF-STEM image of the Al_xGa_1−xN/Al_yGa_1−yN MQW on the sidewalls of a GaN nanowire. (b) Low-magnification TEM image of the MQW region. (c) HRTEM image of an Al_xGa_1−xN/Al_yGa_1−yN double heterostructure.

EDS mapping of the Al_xGa_1−xN/Al_yGa_1−yN MQWs was carried out using TEM at higher resolution and are shown in Fig. 8a–c for nitrogen (red), aluminum (green), and gallium (blue), respectively. A uniform distribution of nitrogen throughout the region (Fig. 8a), as well as clear indications of different Al compositions (Fig. 8b), and Ga compositions (Fig. 8c) in quantum wells with respect to the barriers are evident. Fig. 8d shows the HAADF-STEM image of the region used for elemental mapping, a region similar to the one shown in Fig. 7b. A composite elemental mapping image of nitrogen (red), aluminum (green), and gallium (red) overlaid with X-ray photon counts is shown in Fig. 8e. The plot clearly shows that the MQW is indeed composed of Al_xGa_1−xN/Al_yGa_1−yN with a relatively uniform distribution of Al and Ga in the quantum wells and barrier regions. Hence, it is demonstrated that the growth of the nonpolar m-plane Al_xGa_1−xN/Al_yGa_1−yN MQWs on GaN nanowire sidewalls is feasible and could prove to be a potential candidate for nanostructured ultraviolet light sources and detectors.

Fig. 8 (a–c) EDS elemental maps of nitrogen, aluminum, and gallium respectively. (d) HAADF-STEM image of the Al_xGa_1−xN/Al_yGa_1−yN MQWs on the sidewall of a GaN nanowire. (e) Superimposed image of the elemental map with X-ray photon counts of the Al_xGa_1−xN/Al_yGa_1−yN MQWs.

Finally, in order to analyze the optical characteristics of the Al_xGa_1−xN/Al_yGa_1−yN MQW, one such nanowire was partially milled using a focused ion beam from one side to expose part of the GaN nanowire core as shown in Fig. 9a. CL mapping was carried out at room temperature on this MQW with a partially exposed GaN nanowire core. Electron beam conditions of 5 kV and 13 nA with a mapping pixel size of 30 nm and an exposure duration of 200 ms at each pixel was used for mapping. Distinct emissions from the GaN core (Fig. 9b) and the MQW region on the sidewalls (Fig. 9c) are clearly evident from the CL mapping. Emission from the GaN nanowire core is centered around a peak wavelength of 348 nm (3.56 eV), while the emission from the MQW is centered around a peak wavelength of 332 nm as shown in Fig. 9d. The successful demonstration of UV emission from the MQW on the sidewalls of GaN/AlGaN nanowires entails further investigations in terms of IQE and the optical polarization ratio, these studies will be carried out in future work. Nonetheless, calculations by Yamaguchi,⁶ as well as by Wang and Wu,⁷³ which are supported by experimental evidence of Banal et al.,⁷⁴ confirm that the UV light emitted from the nonpolar m-plane AlGaN MQW has an electric field, E, in the condition E ‖ c-axis. In other words, it means that the light emission from the m-plane surface propagating along the m-axis has an E ⊥ m-axis (since m ⊥ c), thus resulting in predominantly TE emission. Fig. 9e shows a top-view SEM image of a 45 μm × 45 μm array of nanowires with the Al_xGa_1−xN/Al_yGa_1−yN MQWs incorporated on the sidewall of the GaN nanowires. The SEM image reveals a few defects in the array in the form of missing and coalesced nanowires, in an otherwise largely uniform array. Fig. 9f shows the corresponding panchromatic CL image obtained with the electron beam conditions of 5 kV and 400 nA revealing bright emission from the sidewalls of the Al_xGa_1−xN/Al_yGa_1−yN MQW array, while the substrate (outside the array) without nanowires appears dark. Thus, the prospect for an efficient UV light source using the nonpolar m-plane Al_xGa_1−xN/Al_yGa_1−yN MQWs incorporated on GaN nanowire sidewalls is clearly demonstrated over an area of 45 μm × 45 μm with the potential for large area UV LEDs as well as for next generation micro-LEDs.

Fig. 9 (a) SEM image of a partially milled Al_xGa_1−xN/Al_yGa_1−yN MQW on GaN nanowire. (b) CL mapping showing the 348 nm emission from a partially exposed GaN nanowire core. (c) CL mapping showing the 332 nm emission from MQWs on nonpolar m-plane sidewalls. (d) CL spectra of the GaN nanowire core and the MQW. (e) Top-view SEM image of the Al_xGa_1−xN/Al_yGa_1−yN MQW incorporated on a GaN nanowire array. (f) Panchromatic CL image of (e).

Conclusions

We have investigated the growth of GaN/AlGaN nanowire arrays in the form of a core–shell structure using selective area epitaxy. Striations, commonly observed on the surface of nonpolar III-nitrides were present when the growth was conducted in presence of H₂ in the carrier gas. The issue of a rough surface on m-plane AlGaN due to striations, which could potentially lead to non-uniform alloy compositions and quantum wells with a wavy interface, was resolved by growing the AlGaN shell in a pure N₂ carrier gas. The smooth surface achieved on the m-plane AlGaN sidewalls using the N₂ carrier gas is attributed to reduced disparity in the diffusion lengths of the Al and Ga adatoms. The alloy compositions of the AlGaN shells determined through EDS and CL correlated well, albeit indicating a reduced Al incorporation in the solid phase compared to the gas phase input ratio. The reduced Al-incorporation is attributed to the vertical orientation of the growth surface, while, strain and a gas phase parasitic reaction of TMAl become severe with increasing the Al content in the AlGaN shell and the TMAl flow rate. A higher Al content (≥ 30%) in the m-plane AlGaN shells showed defect-related luminescence, and such defects have been identified as (V_III-2O_N)¹⁻ complexes. Microstructural characterization through TEM studies of the m-plane AlGaN shells revealed that the dominant type of defects are partial dislocations, and the formation of partial dislocations increased with Al-content in the AlGaN shell to counteract the increased tensile strain. Nonpolar m-plane Al_xGa_1−xN/Al_yGa_1−yN multiple quantum wells with emission in the ultraviolet wavelength of 332 nm have been successfully incorporated on the sidewalls of the GaN nanowires. Uniform growth and emission from the AlGaN MQWs on a GaN nanowire array over an area of 45 × 45 μm² was achieved. The demonstrated growth scheme for the nonpolar m-plane Al_xGa_1−xN/Al_yGa_1−yN multiple quantum wells on GaN nanowires using MOCVD is anticipated to provide a path for scalable large area efficient UV LEDs and next generation UV micro-LEDs.

Author contributions

S. A., H. H. T., and C. J. conceptualized the idea. S. A. carried out the investigation, formal analysis and writing the original draft. F. K. carried out the TEM investigation. M. L., H. H. T., and C. J. supervised the project. All authors contributed to the writing – review and editing of the manuscript.

Conflicts of interest

The authors declare no conflicts of interest.

Acknowledgements

We acknowledge the Australian Research Council for financial support. We thank Dr (Jenny) Nian Jiang for fruitful discussions, and Mr James Cotsell for providing training and access to the Disco dicing machine. Access to the epitaxial growth and fabrication facilities used in this work is enabled through the Australian National Fabrication Facility, ACT node.

References

1. J. Simon, V. Protasenko, C. Lian, H. Xing and D. Jena, Science, 2010, 327, 60–64.
2. M. A. Khan, J. P. Bermundo, Y. Ishikawa, H. Ikenoue, S. Fujikawa, E. Matsuura, Y. Kashima, N. Maeda, M. Jo and H. Hirayama, Nanotechnology, 2021, 32, 055702.
3. M. A. Khan, N. Maeda, J. Yun, M. Jo, Y. Yamada and H. Hirayama, Sci. Rep., 2022, 12, 2591.
4. Z. Bryan, I. Bryan, S. Mita, J. Tweedie, Z. Sitar and R. Collazo, Appl. Phys. Lett., 2015, 106, 232101.
5. P. Waltereit, O. Brandt, A. Trampert, H. T. Grahn, J. Menniger, M. Ramsteiner, M. Reiche and K. H. Ploog, Nature, 2000, 406, 865–868.
6. A. Atsushi Yamaguchi, Phys. Status Solidi C, 2008, 5, 2364–2366.
7. A. A. Yamaguchi, Appl. Phys. Lett., 2010, 96, 151911.
8. E. Matioli, S. Brinkley, K. M. Kelchner, Y.-L. Hu, S. Nakamura, S. DenBaars, J. Speck and C. Weisbuch, Light: Sci. Appl., 2012, 1, e22.
9. C. O. Holder, J. T. Leonard, R. M. Farrell, D. A. Cohen, B. Yonkee, J. S. Speck, S. P. DenBaars, S. Nakamura and D. F. Feezell, Appl. Phys. Lett., 2014, 105, 031111.
10. Y. Taniyasu and M. Kasu, Appl. Phys. Lett., 2010, 96, 221110.
11. T. Takano, T. Mino, J. Sakai, N. Noguchi, K. Tsubaki and H. Hirayama, Appl. Phys. Express, 2017, 10, 031002.
12. J. Kim, U. Choi, J. Pyeon, B. So and O. Nam, Sci. Rep., 2018, 8, 935.
13. X. Liu, K. Mashooq, T. Szkopek and Z. Mi, IEEE Photonics J., 2018, 10, 1–11.
14. M. Djavid and Z. Mi, Appl. Phys. Lett., 2016, 108, 051102.
15. J. Dai, B. Liu, Z. Zhuang, G. He, T. Zhi, T. Tao, Q. Xu, Y. Li, H. Ge, Z. Xie and R. Zhang, Nanotechnology, 2017, 28, 385205.
16. M. Shatalov, W. Sun, A. Lunev, X. Hu, A. Dobrinsky, Y. Bilenko, J. Yang, M. Shur, R. Gaska, C. Moe, G. Garrett and M. Wraback, Appl. Phys. Express, 2012, 5, 082101.
17. F. Glas, in Semicond. Semimet., ed. A. F. I. Morral, S. A. Dayeh and C. Jagadish, Elsevier, 2015, vol. 93, pp. 79–123.
18. S. Raychaudhuri and E. T. Yu, J. Appl. Phys., 2006, 99, 114308.
19. M. V. Nazarenko, N. V. Sibirev, K. Wei Ng, F. Ren, W. Son Ko, V. G. Dubrovskii and C. Chang-Hasnain, J. Appl. Phys., 2013, 113, 104311.
20. Z. Mi, S. Zhao, A. Connie and M. H. T. Dastjerdi, Proc. SPIE, 2015, 9373, 937306.
21. L. Chen, W. Lin, H. Wang, J. Li and J. Kang, Light: Sci. Appl., 2020, 9, 104.
22. S. Adhikari, M. Lysevych, C. Jagadish and H. H. Tan, Cryst. Growth Des., 2022, 22, 5345–5353.
23. A. Hirai, Z. Jia, M. C. Schmidt, R. M. Farrell, S. P. DenBaars, S. Nakamura, J. S. Speck and K. Fujito, Appl. Phys. Lett., 2007, 91, 191906.
24. H. Yamada, K. Iso, M. Saito, K. Fujito, S. P. DenBaars, J. S. Speck and S. Nakamura, Jpn. J. Appl. Phys., 2007, 46, L1117–L1119.
25. T. Wernicke, S. Ploch, V. Hoffmann, A. Knauer, M. Weyers and M. Kneissl, Phys. Status Solidi B, 2011, 248, 574–577.
26. R. M. Farrell, D. A. Haeger, X. Chen, C. S. Gallinat, R. W. Davis, M. Cornish, K. Fujito, S. Keller, S. P. DenBaars, S. Nakamura and J. S. Speck, Appl. Phys. Lett., 2010, 96, 231907.
27. H. Yamada, H. Chonan, T. Takahashi and M. Shimizu, Jpn. J. Appl. Phys., 2018, 57, 04FG01.
28. A. Tanaka, O. Barry, K. Nagamatsu, J. Matsushita, M. Deki, Y. Ando, M. Kushimoto, S. Nitta, Y. Honda and H. Amano, Phys. Status Solidi A, 2017, 214, 1600829.
29. M. Bobea Graziano, I. Bryan, Z. Bryan, R. Kirste, J. Tweedie, R. Collazo and Z. Sitar, J. Cryst. Growth, 2019, 507, 389–394.
30. R. M. Farrell, D. A. Haeger, K. Fujito, S. P. DenBaars, S. Nakamura and J. S. Speck, J. Appl. Phys., 2013, 113, 063504.
31. A. Tanaka, Y. Ando, K. Nagamatsu, M. Deki, H. Cheong, B. Ousmane, M. Kushimoto, S. Nitta, Y. Honda and H. Amano, Phys. Status Solidi A, 2018, 215, 1700645.
32. I. Bryan, Z. Bryan, M. Bobea, L. Hussey, R. Kirste, R. Collazo and Z. Sitar, J. Appl. Phys., 2014, 116, 133517.
33. O. I. Barry, K. Lekhal, S.-Y. Bae, H.-J. Lee, M. Pristovsek, Y. Honda and H. Amano, Phys. Status Solidi RRL, 2018, 12, 1800124.
34. H. Hardtdegen and P. Giannoules, III-Vs Rev., 1995, 8, 34–39.
35. A. Kakanakova-Georgieva, G. K. Gueorguiev, S. Stafström, L. Hultman and E. Janzén, Chem. Phys. Lett., 2006, 431, 346–351.
36. S. Yamaguchi, M. Kariya, M. Kosaki, Y. Yukawa, S. Nitta, H. Amano and I. Akasaki, J. Appl. Phys., 2001, 89, 7820–7824.
37. K. Narang, R. Khan, A. Pandey, V. K. Singh, R. K. Bag, M. V. G. Padmavati, R. Tyagi and R. Singh, Mater. Res. Bull., 2022, 153, 111875.
38. P. Venezuela and J. Tersoff, Phys. Rev. B: Condens. Matter Mater. Phys., 1998, 58, 10871–10874.
39. R. G. Banal, M. Funato and Y. Kawakami, Phys. Status Solidi C, 2009, 6, 599–602.
40. T. Narita, Y. Honda, M. Yamaguchi and N. Sawaki, Phys. Status Solidi B, 2006, 243, 1665–1668.
41. Y. Kangawa, T. Akiyama, T. Ito, K. Shiraishi and T. Nakayama, Materials, 2013, 6, 3309.
42. M. M. Rozhavskaya, W. V. Lundin, S. I. Troshkov, A. F. Tsatsulnikov and V. G. Dubrovskii, Phys. Status Solidi A, 2015, 212, 851–854.
43. L. Tang, B. Tang, H. Zhang and Y. Yuan, ECS J. Solid State Sci. Technol., 2020, 9, 024009.
44. W. V. Lundin, A. E. Nikolaev, M. M. Rozhavskaya, E. E. Zavarin, A. V. Sakharov, S. I. Troshkov, M. A. Yagovkina and A. F. Tsatsulnikov, J. Cryst. Growth, 2013, 370, 7–11.
45. D. V. Dinh, S. N. Alam and P. J. Parbrook, J. Cryst. Growth, 2016, 435, 12–18.
46. G. S. Huang, H. H. Yao, H. C. Kuo and S. C. Wang, Mater. Sci. Eng. B, 2007, 136, 29–32.
47. I. Vurgaftman and J. R. Meyer, J. Appl. Phys., 2003, 94, 3675–3696.
48. A. F. Wright, J. Appl. Phys., 1997, 82, 2833–2839.
49. Q. Yan, P. Rinke, A. Janotti, M. Scheffler and C. G. Van de Walle, Phys. Rev. B: Condens. Matter Mater. Phys., 2014, 90, 125118.
50. N. Nepal, M. L. Nakarmi, J. Y. Lin and H. X. Jiang, Appl. Phys. Lett., 2006, 89, 092107.
51. T. Hoshi, K. Hazu, K. Ohshita, M. Kagaya, T. Onuma, K. Fujito, H. Namita and S. F. Chichibu, Appl. Phys. Lett., 2009, 94, 071910.
52. H. C. Yang, T. Y. Lin and Y. F. Chen, Phys. Rev. B: Condens. Matter Mater. Phys., 2000, 62, 12593–12596.
53. W.-Y. Wang, P. Jin, G.-P. Liu, W. Li, B. Liu, X.-F. Liu and Z.-G. Wang, Chin. Phys. B, 2014, 23, 087810.
54. F. Mehnke, A. M. Fischer, Z. Xu, H. Bouchard, T. Detchprohm, S.-C. Shen, F. A. Ponce and R. D. Dupuis, J. Appl. Phys., 2022, 131, 073103.
55. V. Kueller, A. Knauer, F. Brunner, U. Zeimer, H. Rodriguez, M. Kneissl and M. Weyers, J. Cryst. Growth, 2011, 315, 200–203.
56. D. M. Follstaedt, A. A. Allerman, S. R. Lee, J. R. Michael, K. H. A. Bogart, M. H. Crawford and N. A. Missert, J. Cryst. Growth, 2008, 310, 766–776.
57. A. Bell, R. Liu, U. K. Parasuraman, F. A. Ponce, S. Kamiyama, H. Amano and I. Akasaki, Appl. Phys. Lett., 2004, 85, 3417–3419.
58. A. Ishibashi, H. Murotani, T. Yokogawa and Y. Yamada, Jpn. J. Appl. Phys., 2012, 51, 035604.
59. D. V. Dinh, N. Hu, Y. Honda, H. Amano and M. Pristovsek, Sci. Rep., 2019, 9, 15802.
60. V. Jindal, J. Grandusky, M. Jamil, N. Tripathi, B. Thiel, F. Shahedipour-Sandvik, J. Balch and S. LeBoeuf, Phys. E, 2008, 40, 478–483.
61. C. He, Z. Qin, F. Xu, L. Zhang, J. Wang, M. Hou, S. Zhang, X. Wang, W. Ge and B. Shen, Sci. Rep., 2016, 6, 25124.
62. M. E. Coltrin, J. Randall Creighton and C. C. Mitchell, J. Cryst. Growth, 2006, 287, 566–571.
63. J. A. Floro, D. M. Follstaedt, P. Provencio, S. J. Hearne and S. R. Lee, J. Appl. Phys., 2004, 96, 7087–7094.
64. V. Philippe, B. Zahia and G. Tobias, Jpn. J. Appl. Phys., 2007, 46, 4089.
65. M. D. Craven, S. H. Lim, F. Wu, J. S. Speck and S. P. DenBaars, Appl. Phys. Lett., 2002, 81, 469–471.
66. D. N. Zakharov, Z. Liliental-Weber, B. Wagner, Z. J. Reitmeier, E. A. Preble and R. F. Davis, Phys. Rev. B: Condens. Matter Mater. Phys., 2005, 71, 235334.
67. Z. Liliental-Weber, J. Jasiński and D. Zakharov, Opto-Electron. Rev., 2004, 12, 339–346.
68. A. Bourret, C. Adelmann, B. Daudin, J.-L. Rouvière, G. Feuillet and G. Mula, Phys. Rev. B: Condens. Matter Mater. Phys., 2001, 63, 245307.
69. J. Shao, D. N. Zakharov, C. Edmunds, O. Malis and M. J. Manfra, Appl. Phys. Lett., 2013, 103, 232103.
70. J. Moneta, E. Grzanka, H. Turski, C. Skierbiszewski and J. Smalc-Koziorowska, Semicond. Sci. Technol., 2020, 35, 034003.
71. Z. Ye, S. Nitta, K. Nagamatsu, N. Fujimoto, M. Kushimoto, M. Deki, A. Tanaka, Y. Honda, M. Pristovsek and H. Amano, J. Cryst. Growth, 2019, 516, 63–66.
72. M. J. Hÿtch, E. Snoeck and R. Kilaas, Ultramicroscopy, 1998, 74, 131–146.
73. C.-P. Wang and Y.-R. Wu, J. Appl. Phys., 2012, 112, 033104.
74. R. G. Banal, Y. Taniyasu and H. Yamamoto, Appl. Phys. Lett., 2014, 105, 053104.

---

Footnote

† Electronic supplementary information (ESI) available: Bandgap calculation for strained AlGaN on c- and m-plane GaN. HRTEM and strain-map images of samples Al-30 and Al-50. See DOI: https://doi.org/10.1039/d2nh00500j

---