1. Introduction

GaN and related materials are today widely used in opto-electronic devices, such as light emitting diodes (LEDs), laser diodes and photo-detectors with operation wavelengths going from visible to a deep UV spectral range [1, 2]. However, the material quality is still a major issue for thin film devices, which suffer from threading dislocations mainly induced by a lattice mismatch with the growth substrate, especially for nitride layer on Si [3]. Defects are inherent to any growth process, especially to III-Nitride materials which are mainly synthesized by heteroepitaxy due to the extremely high cost of native GaN substrates with low defect density and high structural and optical quality. Many physical mechanisms triggered by different defects are still not perfectly understood [4–6]. Knowing that the GaN crystal quality directly affects its light emitting properties, the suppression of below-gap states, that act as radiative or non-radiative recombination centers, is of crucial importance.

The defect-related yellow (YL) and green (GL) luminescence are often observed in GaN, the formation and transition energy of these defects in GaN is a topic of intense research [7–11]. Yet, the origin of these defect-induced luminescences, remain under debate. In 2-dimensional layers, etching experiments demonstrated a near independence of the visible intensity with the etch depth supporting the bulk nature of the defects [12], whereas other experiments using photo-voltage spectroscopy argued in favor of surface defects as being responsible for the visible band [13,14]. From literature, the main candidates for YL band are defects, either isolated carbon in a nitrogen site (C_N) [8], a carbon-oxygen complex (C_NO_N) [9], or other complexes such as V_Ga-3H or V_GaO_N -2H [15]. The GL band, observed in high-purity GaN, is most recently [16] associated with transitions from different charge levels of the same C_N defect responsible for YL band.

Changing from the 2D film to the nanowire morphology provides an elegant solution to the dislocation problem. Indeed, thanks to the efficient strain relaxation by the free lateral surface, dislocation-free nanowires (NWs) can be grown on highly mismatched substrates [17]. GaN NWs open a new window of applications presenting a number of potential advantages for nanophotonics and nanoelectronics. In particular, nitride NWs have been proposed as an active medium for high-efficiency 3D LEDs [18–20] and high sensitivity photodetectors [21].

The most widespread way to fabricate GaN NWs relies on epitaxial techniques such as Molecular Beam Epitaxy (MBE) or Metalorganic Vapor Phase Epitaxy (MOVPE) using catalytic, self-catalyzed or selective area growth modes [22,23]. Although these techniques provide the best control over the wire morphology and composition, they are expensive and present difficulties for up-scaling the sample size. For a large number of applications (e.g. light sensors for UV exposure monitoring) the development of a low cost synthesis technique is desirable. The Low Pressure Chemical Vapor Deposition (LPCVD) method, which allows a significant cost reduction due to a lower vacuum and a lower thermal budget, can also be used for GaN NW growth [24].

In this letter, we report on the synthesis and characterization of GaN NWs grown by catalyst-assisted LPCVD. We demonstrate that the NWs are single crystal with Wurtzite structure and present an intense near-band-edge (NBE) luminescence. YL and GL bands are also observed and are similar to the ones reported from bulk and thin film GaN samples. Thanks to their large surface to volume ratio, LPCVD-grown GaN NWs appear as excellent candidates to investigate the origin of YL and GL bands. The presence of high density of surface states tends to pin the Fermi level near the surface and bends the electronic bands that affects the recombination process [25]. Structural defects, such as dangling bonds or point defects, charge the surface states and are reported to influence the optical and transport properties of nanowires [26,27]. By passivating the surface of LPCVD-grown GaN NWs with a thin layer of Al₂O₃, we demonstrate a drastic reduction of the defect-related YL band while maintaining the GL band and the NBE emission unchanged. The defect responsible for the YL band is believed to be localized near the surface of GaN nanowires. The GL not being affected by NW passivation, it could be attributed to defects with bulk nature.

2. Methodology

The GaN nanowires are grown by LPCVD on (100) Si substrate using Germanium (Ge) as catalyst. Previously to the growth, Si substrate is coated with a 3 nm thick layer of Ge by electron-beam evaporation system (Edwards). The sample is heated at 750 °C in the LPCVD reactor to dewet the Ge coating and form nanoparticles that act as seeds for the GaN NWs. Growth happens at 750 °C for 30 min under a continuous flow of 100 sccm of NH₃, diluted in 150 sccm of H₂. A solid form of Gallium (Ga) (99.99% purity) was placed next to Si samples. The vapor pressure of Gallium (melting point at 29.76 °C), amount of Gallium vapor that exists in gaseous form around the Ga liquid source, increases with temperature and provides the Ga source for nanowire growth. Before the passivation process, the sample was cleaned using classical organic (Isopropanol Alcohol (IPA) and Acetone) and inorganic (Buffer Oxide Etchant (BOE)) solutions. GaN NWs have been passivated with a 5 nm thin shell of Al₂O₃ deposited by Atomic Layer Deposition (ALD) at 250 °C using a Cambridge Nanotech Savannah 100/200 system.

As-grown and passivated nanowire (NW) ensembles were characterized and analyzed by scanning electron microscopy (SEM; LEO 1550 Gemini, electron gun at 5 kV), transmission electron microscopy (TEM; Tecnai X-TWIN, electron gun at 200 keV FEG and probe current of 0.5 nA at 1 nm) and photoluminescence (PL) spectroscopy. The PL set-up consisted of a T64000 Horiba Jobin-Yvon high resolution spectrometer with a continuous Helium-Cadmium laser emitting at 325 nm with laser power of 1.76 mW. The PL mapping consists in a 532 nm 37 μW continuous wave laser (WITec focus innovations) with x100 microscope objective of numerical aperture NA=0.8 (OLYMPUS). The emission is filtered at 570 nm with a longpass filter. For PL mapping, the luminescent intensity integrated over the 580–630 nm range is taken for each point. The scanning is done with a resolution of 0.5 μm/step. The time-resolved set-up consisted on a Picosecond Pulsed Diode Laser EPL-375 from Edinburgh Instruments. This excitation source emits at 371 nm ± 5 nm, with pulse period of 50 ns, 70 ps wide and an average power of 0.15 mW. The decay was measured using a Hamamatsu H7422-40 photo-sensor associated with a single-grating emission monochromator of 1200 grooves/mm density.

3. Results and discussion

Fig. 1.a shows scanning electron microscope (SEM) images of GaN NWs grown by LPCVD. Nanowires grow in random directions following a Vapor-Liquid-Solid (VLS) mechanism catalyzed with Ge. It is worth noticing that the background is composed of Germanium seeds that have not resulted in NW growth. Most likely, the supersaturation was not reached for these droplets explaining the absence of the VLS growth. GaN NWs present a straight shape and show a single-crystal Wurtzite structure growing along the c-axis, as previously reported [24], see also the transmission electron microscope images in appendix (Fig. 4). As-grown nanowires are characterized by a spatial distribution of ∼10⁸ NWs/cm², an average length of 3.4 ± 0.9 μm and a diameter of 34 ± 7 nm, determined from SEM images. Fig. 1(b) shows the GaN NWs after passivation with Alumina Oxide deposition. The NWs remain straight. The SEM charging effects confirm the successful deposition of alumina oxide around the GaN nanowires. Moreover, the average diameter after passivation is 41 ± 6 nm. This thickening indicates that the GaN core as embedded into a ∼3.5 nm thick Al₂O₃ shell.

 

Fig. 1 (a) SEM image of Ge-catalyzed GaN nanowires. (b) Same NWs after coating with 5 nm Al₂O₃. (Inset) higher magnification images of individual nanowires showing the Ge rich tip. Scale bars in (a, b) and insets indicate 500 nm and 100 nm, respectively

Download Full Size | PDF

Figure 2 presents the PL characterization, at room temperature, of a ensemble of single-crystal GaN nanowires before (Fig. 2(a)) and after (Fig. 2(b)) passivation. Both nanowire ensembles were analyzed under the same excitation and detection conditions so that the emission intensities can be compared. PL emission of Ge (contained inside the droplet on top of nanowires) is localized from 0.6 to 0.9 eV and the Al₂O₃ shell (a wide band gap material with E_g = 6.6 eV) surrounding the core-GaN NWs, is transparent. The PL spectra of non-passivated GaN nanowires shows a NBE emission at 3.58 eV, characterized by a full width at half maximum (fwhm) of 450 meV. The spectrum is slightly asymmetric with a higher broadening at lower energies. A wide emission in the visible range, centered around 2.09 eV with a fwhm equal to 588 meV, is also observed. The BOE cleaning before passivation had no effect on the PL. And similar peaks are present in passivated GaN NWs. The NBE peak of GaN is observed at 3.58 eV with a sharper fwhm of 305 meV and a decrease by 3 % in intensity compared to non-passivated nanowires. We associate this slight decrease to a spatial variability of the nanowire ensemble since different areas were probed before and after passivation. The cleaning with BOE, before Al₂O₃ deposition, had no effect on the PL emission.

 

Fig. 2 Photoluminescence spectra of a forest of GaN nanowires before (left) and after (right) passivation excited with HeCd laser at 325 nm. The discontinued lines are Gaussian fits showing YL and GL band. (insets) Mapping of YL band emission from single nanowire excited with a continuous diode laser at 532 nm.

Download Full Size | PDF

The NBE peak is quite broad and strongly blue-shifted (for both as-grown and passivated NWs) compared to bulk undoped GaN. This shift is associated to the heavy n-doping of these nanowires (above 10¹⁹ cm⁻³). This high level doping most probably originates from the incorporation of Ge from the catalyst as Ge appears to be an efficient donor for GaN [28] (see Appendix A2 about GaN near-band-edge shift). The second peak in the visible range, is slightly shifted to higher energies (centered around 2.15 eV). We remark that its intensity decreases by 88 %, after passivation. The visible luminescence of these LPCVD-grown GaN nanowires is wide and higher in intensity than the NBE emission. Few reports show this PL signature in GaN NWs grown by the classical techniques, such as MOCVD [29] or Plasma Assited MBE. By contrast it is more commonly observed in thin film GaN [4].

Gaussian fitting of the asymmetric visible PL peak of passivated GaN NWs resulted in two distinct contributions: one centered at 2.09 eV and another at 2.40 eV (Fig. 2.b). The peak at 2.09 eV is commonly identified as a Yellow Luminescence, the one at 2.40 eV is associated to Green Luminescence. By considering the combination of these two peaks, the visible luminescence of as-grown nanowires can be attributed to a strong YL band emission (I_YL/I_NBE = 4.63), dominating a weak contribution of GL emission. By contrast, the visible luminescence of passivated NWs can be described as a combination of YL band (with a drastic decrease in its intensity I_YL/I_NBE = 0.60) and a GL band. The insets in Fig. 2 show the mapping of PL emission from single nanowire while excited with a laser diode at 532 nm (see Appendix A3 for more data). Thus, only the YL defect state is excited showing a strong and uniform distribution along the as-grown nanowire (Fig. 2(a)) while there is almost no YL emission from Al₂O₃ coated nanowire (Fig. 2(b)).

Time-resolved emission measurements show that after passivating the surface of the GaN NWs, the lifetime decay varies significantly for the YL-band (Fig. 3(a)). A 371 ± 5 nm pulsed laser was used to excite the defect-related states only. YL lifetime decay is observed as a single exponential decay before Al₂O₃ deposition (τ₁ = 0.23 ± 0.01 ns) while a slower component appears after passivation (τ₁ = 0.18 ± 0.01 ns and τ₂ = 5.90 ± 0.17 ns). τ₁ and τ₂ are the characteristic times of the bi-exponential decay (as defined in Eq. 1 of Appendix A4). The fast component in YL remains almost constant after passivating the surface, suggesting an emission from similar recombination state, but its contribution decreases, which concurs with the weakened YL emission observed earlier (Fig. 2), with Al₂O₃ layer most surface states may have been suppressed. However, after passivation the YL emission includes an additional slower component that may originate either from a volume recombination or a new defect could have been created by the Al₂O₃ deposition (interfacial defect states). From PL observations, this overall defect emission is still much weaker than before passivation.

 

Fig. 3 Lifetime decay curves of YL (564 nm) and GL (489 nm) emission from quasi-resonant excited with a pulsed laser at 371 nm ± 5 nm. The instrument response function is 0.35 ns. The time constants are obtained after deconvolution with the instrument response. The exponential decay fitting is shown with full lines.

Download Full Size | PDF

By comparison with the YL band, where the intensity of PL and decay curves are strongly affected by the passivation of NWs, the GL band is almost unchanged (see Appendix A4 for more data). The presence of high density of surface states tends to pin the Fermi level near the surface and bends the electronic bands which affects the recombination process. Growth of alumina oxide layer on the surface of core-GaN NWs induces a modification of the nature of surface charges, such as active defects. The poor effect of this passivation on GL decay curves is in favor of the selectivity of the process to YL related-defects. Thus, the selective suppression of the YL emission suggests YL peak is mostly related to structural defects located at or near the surface while GL-related defects are more uniformly distributed in the NW volume. This observation is consistent to the earlier photo-voltage spectroscopy [13] and temperature and excitation dependence experiments [14] on GaN layers.

 

Fig. 4 a) High resolution TEM image taken at the bottom of a non-passivated single-crystal GaN nanowire. Inset: The corresponding electron diffraction pattern (ED) b), c) and d) High Resolution TEM images of non-passivated single-crystal GaN nanowires.

Download Full Size | PDF

4. Conclusion

The optical properties of as-grown and aluminum-oxide passivated GaN nanowires synthesized by LPCVD have been analyzed. As-grown nanowires exhibit a strong visible luminescence. By analyzing the optical properties of the as-grown NWs and passivated with a thin Al₂O₃ shell, we observed that the visible luminescence results from two contributions: a yellow (2.09 eV) and green (2.40 eV) band. The passivation of NWs drastically reduces the yellow band while maintaining the green band and NBE emission. The results suggest the defects responsible for the YL band could be localized near the surface of the GaN nanowires and can be easily passivated by Al₂O₃ ALD while the defects causing GL are of bulk nature. LPCVD-grown GaN nanowires are cost-effective and their quality can be improved using a thin layer Al₂O₃ passivation of the surface.

Appendix

A1. Structural characterization

Ge-assisted catalytic growth by LPCVD method leads to the synthesis of GaN NWs with high crystalline structure as demonstrated by High Resolution Transmission Electron Microscopy (HRTEM) and Electron Diffraction (ED) characterization (Figure 4). The as-grown GaN NWs exhibit a single crystal structure growing along the c-axis, atomic planes are separated by 2.52 Å. The single-crystal structure and the [0001] growth direction are confirmed by the ED image in the inset of Fig. 4. TEM images of single nanowires in Fig. 4(b), 4(c) and 4(d) give insights on the morphology of Ge-catalyzed GaN nanowires. The surface of as-grown GaN NWs is atomically sharp with uniform lattice and no amorphous phase is observed.

A2. GaN near-band-edge shift

In both as-grown and passivated NWs, we observe that the band-edge peak is quite broad and largely shifted compared to bulk undoped GaN. Indeed, the photoluminescence (3.58 eV) is shifted by 140 meV compared to the bulk GaN at 300K (3.44 eV [30]). This shift could have been be attributed to the quantum confinement effects. However, both SEM and TEM characterizations confirmed the diameter of core-GaN nanowires (around 34 nm) to be much larger than the Bohr radius of GaN bulk exciton which has been reported to be 2.8 nm [31]. This Bohr radius value thus excludes the hypothesis that the shift may originate is originated from quantum confinement effects.

By contrast, the shift could be attributed to the high level of doping of our GaN NWs. Significant band-edge peak blue shift for heavily Si-doped Gallium Nitride layers [32], and nanowires [33,34] has been reported. In our case, GaN NWs were grown on Silicon substrate using Germanium as a catalyst. Since both Germanium and Silicon appear to be electron donors for GaN, the blue shift observed in our GaN NWs can be explained by the band filling effect. In fact, due to the high doping element concentration, electrons populate states inside the conduction band, pushing the Fermi level higher in energy and giving rise to indirect optical transitions (with non-conservation of k-vector allowed by many-body interactions). This band-filling hypothesis is supported by the observed high broadening and asymmetry of the PL peak, with a higher broadening on the low energy side, which is typical for momentum non-conserving transitions [35]. For bulk n-doped GaN, the emission first shifts to longer wavelengths, which is attributed to band gap renormalization due to Coulomb effects (up to approx. 10¹⁹ cm⁻³), and then blue-shifts due to band filling [32,36]. Therefore, the strong blue-shift of the NBE emission suggests that the nanowires are heavily n-doped. This high level doping most probably originates from the incorporation of Ge from the catalyst, Ge appearing as more efficient donors for GaN than Si [28].

A3. Mapping of YL band emission from single nanowire

For the mapping, GaN NWs samples were sonicated in a pure Isopropanol Alcohol (IPA) then dispersed on Si substrate. Figure 5(a) shows one single as-grown GaN nanowire with another smaller one, probably broken during sonication. The associated PL mapping (Fig. 5(b)) matches exactly with the GaN nanowire position, the emission seems uniformly distributed along the wire showing a similar distribution for the related defect. The comparison with Al₂O₃ coated NWs (Fig. 5(e)) shows the effect of passivation on YL band, the emission is drastically suppressed. We see an emission only at the bottom of two bonded wires (Fig. 5(e), marker ‘2’) while the nanowire near the marker ‘3’ doesn’t show any emission.

 

Fig. 5 Optical microscope image of dispersed nanowires on Si substrate of as-grown (a) and Al₂O₃ coated (d) NWs. Associated mapping of YL band emission from as-grown (b) and Al₂O₃ coated (e) NWs. Brightest emission from points 1 and 2 for single as-grown (c) and Al₂O₃ coated (f) nanowire, respectively. Excitated with a continuous diode laser at 532 nm

Download Full Size | PDF

A4. Lifetime decay

Figure 6 show the lifetime decay curves for several wavelengths from as-grown (Fig. 6(b)) and Al₂O₃ coated GaN nanowires (Fig. 6(b)). The qualitative observation when comparing these decays confirms the initial conclusion on the selectivity of passivation on the YL band. Indeed, for lifetime decay curves at 620, 564 and 557 nm (close to YL band) we see a strong effect of passivation on the nature of surface charges. The higher the detection energy we observe at (toward GL), the lesser is the effect of the passivation.

 

Fig. 6 Lifetime decay curves from quasi-resonant excitation of a forest of as-grown (a) and Al₂O₃ coated (b) GaN NWs. Excitated with a pulsed laser at 371 nm ± 5 nm.

Download Full Size | PDF

Exponential fit

For analysis, we fit the decay curves in Fig. 3 with a single or double exponential using the following equation:

(1)$$N(t)=N_1(0)\text{exp}\left(\frac{-t}{{\tau }_1}\right)+N_2(0)\text{exp}\left(\frac{-t}{{\tau }_2}\right)$$

where N(t) is the total number of population from all emission channels after t time. N₁(0) and N₂(0) are the initial population at t = 0, which have emission lifetimes for fast and slow emission channels of τ₁ and τ₂, respectively.

Funding

National Research Foundation (NRF) of Singapore (NRF-CRP12-2013-04); French “Agence Nationale de la Recherche” (ANR) through “GaNeX” (ANR-11-LABX-2014).

References and links

1. G. Fasol, “Room-Temperature Blue Gallium Nitride Laser Diode,” Science 272, 1751–1752 (1996). [CrossRef]  

2. S. Mohammad, A. Salvador, and H. Morkoc, “Emerging gallium nitride based devices,” Proceedings of the IEEE 83, 1306–1355 (1995). [CrossRef]  

3. A. Dadgar, “Sixteen years GaN on Si,” physica status solidi (b) 252, 1063–1068 (2015). [CrossRef]  

4. M. A. Reshchikov and H. Morkoç, “Luminescence properties of defects in GaN,” Journal of Applied Physics 97, 061301 (2005). [CrossRef]  

5. M. A. Reshchikov, J. D. McNamara, F. Zhang, M. Monavarian, A. Usikov, H. Helava, Y. Makarov, and H. Morkoç, “Zero-phonon line and fine structure of the yellow luminescence band in GaN,” Physical Review B - Condensed Matter and Materials Physics 94, 1–9 (2016). [CrossRef]  

6. P. Huang, H. Zong, J.-j. Shi, M. Zhang, X.-h. Jiang, H.-x. Zhong, Y.-m. Ding, Y.-p. He, J. Lu, and X.-d. Hu, “Origin of 3.45 eV Emission Line and Yellow Luminescence Band in GaN Nanowires: Surface Microwire and Defect,” ACS Nano 9, 9276–9283 (2015). [CrossRef]   [PubMed]  

7. S. Limpijumnong and C. Van de Walle, “Diffusivity of native defects in GaN,” Physical Review B 69, 035207 (2004). [CrossRef]  

8. J. L. Lyons, A. Janotti, and C. G. Van de Walle, “Carbon impurities and the yellow luminescence in GaN,” Applied Physics Letters 97, 152108 (2010). [CrossRef]  

9. D. O. Demchenko, I. C. Diallo, and M. A. Reshchikov, “Yellow Luminescence of Gallium Nitride Generated by Carbon Defect Complexes,” Physical Review Letters 110, 087404 (2013). [CrossRef]   [PubMed]  

10. M. A. Reshchikov, D. O. Demchenko, A. Usikov, H. Helava, and Y. Makarov, “Carbon defects as sources of the green and yellow luminescence bands in undoped GaN,” Physical Review B 90, 235203 (2014). [CrossRef]  

11. T. Mattila and R. M. Nieminen, “Point-defect complexes and broadband luminescence in GaN and AlN,” Physical Review B 55, 9571–9576 (1997). [CrossRef]  

12. D. Basak, M. Lachab, T. Nakanishi, and S. Sakai, “Effect of reactive ion etching on the yellow luminescence of GaN,” Applied Physics Letters 75, 3710 (1999). [CrossRef]  

13. I. Shalish, L. Kronik, G. Segal, Y. Shapira, M. Eizenberg, and J. Salzman, “Yellow luminescence and Fermi level pinning in GaN layers,” Applied Physics Letters 77, 987 (2000). [CrossRef]  

14. M. Reshchikov and R. Korotkov, “Analysis of the temperature and excitation intensity dependencies of photoluminescence in undoped GaN films,” Physical Review B 64, 1–11 (2001). [CrossRef]  

15. J. L. Lyons, A. Alkauskas, A. Janotti, and C. G. Van de Walle, “First-principles theory of acceptors in nitride semiconductors,” physica status solidi (b) 252, 900–908 (2015). [CrossRef]  

16. M. A. Reshchikov, D. O. Demchenko, J. D. McNamara, S. Fernández-Garrido, and R. Calarco, “Green luminescence in Mg-doped GaN,” Physical Review B - Condensed Matter and Materials Physics 90, 1–14 (2014). [CrossRef]  

17. F. Glas, “Critical dimensions for the plastic relaxation of strained axial heterostructures in free-standing nanowires,” Physical Review B 74, 121302 (2006). [CrossRef]  

18. C.-Y. Chen, G. Zhu, Y. Hu, J.-W. Yu, J. Song, K.-Y. Cheng, L.-H. Peng, L.-J. Chou, and Z. L. Wang, “Gallium Nitride Nanowire Based Nanogenerators and Light-Emitting Diodes,” ACS Nano 6, 5687–5692 (2012). [CrossRef]   [PubMed]  

19. X. Dai, A. Messanvi, H. Zhang, C. Durand, J. Eymery, C. Bougerol, F. H. Julien, and M. Tchernycheva, “Flexible Light-Emitting Diodes Based on Vertical Nitride Nanowires,” Nano Letters 15, 6958–6964 (2015). [CrossRef]   [PubMed]  

20. R. Yan, D. Gargas, and P. Yang, “Nanowire photonics,” Nature Photonics 3, 569–576 (2009). [CrossRef]  

21. F. González-Posada, R. Songmuang, M. Den Hertog, and E. Monroy, “Room-Temperature Photodetection Dynamics of Single GaN Nanowires,” Nano Letters 12, 172–176 (2012). [CrossRef]  

22. C. Chèze, L. Geelhaar, O. Brandt, W. M. Weber, H. Riechert, S. Münch, R. Rothemund, S. Reitzenstein, A. Forchel, T. Kehagias, P. Komninou, G. P. Dimitrakopulos, and T. Karakostas, “Direct comparison of catalyst-free and catalyst-induced GaN nanowires,” Nano Research 3, 528–536 (2010). [CrossRef]  

23. S.-K. Lee, H.-J. Choi, P. Pauzauskie, P. Yang, N.-K. Cho, H.-D. Park, E.-K. Suh, K.-Y. Lim, and H.-J. Lee, “Gallium nitride nanowires with a metal initiated metal-organic chemical vapor deposition (MOCVD) approach,” physica status solidi (b) 241, 2775–2778 (2004). [CrossRef]  

24. U. Saleem, H. Wang, D. Peyrot, A. Olivier, J. Zhang, P. Coquet, and S. L. G. Ng, “Germanium-catalyzed growth of single-crystal GaN nanowires,” Journal of Crystal Growth 439, 28–32 (2016). [CrossRef]  

25. L. Polenta, M. Rossi, A. Cavallini, R. Calarco, M. Marso, R. Meijers, T. Richter, T. Stoica, and H. Lüth, “Investigation on Localized States in GaN Nanowires,” ACS Nano 2, 287–292 (2008). [CrossRef]  

26. R. Calarco, T. Stoica, O. Brandt, and L. Geelhaar, “Surface-induced effects in GaN nanowires,” Journal of Materials Research 26, 2157–2168 (2011). [CrossRef]  

27. A. Armstrong, Q. Li, K. H. A. Bogart, Y. Lin, G. T. Wang, and A. A. Talin, “Deep level optical spectroscopy of GaN nanorods,” Journal of Applied Physics 106, 053712 (2009). [CrossRef]  

28. A. Ajay, J. Schörmann, M. Jiménez-Rodriguez, C. B. Lim, F. Walther, M. Rohnke, I. Mouton, L. Amichi, C. Bougerol, M. I. Den Hertog, M. Eickhoff, and E. Monroy, “Ge doping of GaN beyond the Mott transition,” Journal of Physics D49 (2016). [CrossRef]  

29. M. Julkarnain, T. Fukuda, N. Kamata, and Y. Arakawa, “A direct evidence of allocating yellow luminescence band in undoped GaN by two-wavelength excited photoluminescence,” Applied Physics Letters 107, 212102 (2015). [CrossRef]  

30. O. Madelung, Semiconductors - Group IV Elements and III–V Compounds, Data in Science and Technology (Springer Berlin Heidelberg, Berlin, Heidelberg, 1991).

31. P. Ramvall, S. Tanaka, S. Nomura, P. Riblet, and Y. Aoyagi, “Observation of confinement-dependent exciton binding energy of GaN quantum dots,” Applied Physics Letters 73, 1104 (1998). [CrossRef]  

32. M. Yoshikawa, M. Kunzer, J. Wagner, H. Obloh, P. Schlotter, R. Schmidt, N. Herres, and U. Kaufmann, “Band-gap renormalization and band filling in Si-doped GaN films studied by photoluminescence spectroscopy,” Journal of Applied Physics 86, 4400 (1999). [CrossRef]  

33. P. Tchoulfian, F. Donatini, F. Levy, B. Amstatt, P. Ferret, and J. Pernot, “High conductivity in Si-doped GaN wires,” Applied Physics Letters 102, 122116 (2013). [CrossRef]  

34. E. Oh, B. Woo Lee, S.-J. Shim, H.-J. Choi, B. Hee Son, Y. Hwan Ahn, and L. S. Dang, “Low temperature cathodoluminecence and electron beam induced current studies of single GaN nanowires,” Applied Physics Letters 100, 153110 (2012). [CrossRef]  

35. P. K. Kandaswamy, H. MacHhadani, Y. Kotsar, S. Sakr, A. Das, M. Tchernycheva, L. Rapenne, E. Sarigiannidou, F. H. Julien, and E. Monroy, “Effect of doping on the mid-infrared intersubband absorption in GaN/AlGaN superlattices grown on Si(111) templates,” Applied Physics Letters 96, 1–4 (2010). [CrossRef]  

36. F. Binet, J. Duboz, J. Off, and F. Scholz, “High-excitation photoluminescence in GaN: Hot-carrier effects and the Mott transition,” Physical Review B 60, 4715–4722 (1999). [CrossRef]