Open Access
Add to BibSonomyAbstract
Self-assembled GaN rods were grown on sapphire by metal-organic vapor phase epitaxy using a simple two-step method that relies first on a nitridation step followed by GaN epitaxy. The mask-free rods formed without any additional catalyst. Most of the vertically aligned rods exhibit a regular hexagonal shape with sharp edges and smooth sidewall facets. Cathodo- and microphotoluminescence investigations were carried out on single GaN rods. Whispering gallery modes with quality factors greater than 4000 were measured demonstrating the high morphological and optical quality of the self-assembled GaN rods.
© 2013 Optical Society of America
1. Introduction
Microcavity systems such as photonic crystal cavities, micropillars and microdisks of high quality factor Q (Q = E/ΔE) permit the observation of quantum-electrodynamic effects [1, 2]. The group III-nitride material system has the potential to study these light-matter interactions at room temperature (RT) [3]. Polariton lasing from a GaN microcavity at RT could already be shown [4]. However, the realization of such a structure consisting of distributed Bragg reflector (DBR) below und above the GaN microcavity is highly demanding. An alternative to the time-consuming growth/deposition of DBRs is the use of microdisks [5]. Here, the light wave circulates within the disk and is totally reflected at the inner boundary. After one circulation the wave can interfere with itself resulting in whispering gallery modes (WGMs) [1]. To avoid the tedious processing of disks e.g. into a bulk material, the anisotropic growth of wurtzite crystals can be exploited to form self-assembled, hexagonally shaped rods such as e.g. faceted ZnO structures [6, 7]. In the GaN system, self-organized hexagonal rod structures have been fabricated using selective area growth by molecular beam epitaxy (MBE) [8]. However, mask processing and the MBE growth are time-consuming.
This letter will present a fast and simple two-step growth method to achieve self-assembled hexagonal GaN rod structures by metal-organic vapor phase epitaxy (MOVPE). No additional processing was needed to form these rods. High Q-factor hexagonal WGMs could be measured at RT by cathodo- and photoluminescence pointing out the high morphological and optical quality of the GaN rod structures.
2. Experimental
The investigated samples were grown using an Aixtron 200RF horizontal flow MOVPE reactor. As a group III source trimethylgallium (TMG), as a group V source ammonia (NH3), as n-type dopant silane (SiH4), and hydrogen (H2) as a carrier gas were used. The total reactor flux and pressure was set to ∼4 standard liter per minute and 100 mbar, respectively. All samples were grown on 2” c-plane sapphire substrates at temperatures measured with a thermocouple inside the susceptor. As there is no setup to measure the surface temperature, the thermocouple temperature will be given as a measure for the used growth conditions. The real surface temperature is estimated to be roughly 50–100 °C lower.
The two-step growth method to form hexagonal GaN rod structures consists of first a nitridation step and second the GaN growth process. Nitridation of c-oriented sapphire substrate is carried out at 1200 °C under H2 atmosphere and a high NH3 flow of 1500 sccm for 10 min. The influence of this step on the GaN structures was already reported in Ref. [9]. The nitridation forms a thin AlN layer on the sapphire surface with nitrogen polarity [10]. The polarity is a crucial parameter and determines whether rod or pyramidal growth takes place for low V/III ratio GaN deposition [11, 12]. In the second step Si-doped GaN is deposited for 5 min at 1150 °C with a reduced NH3 flux of 50 sccm. The V/III ratio and the Ga/Si ratio was set to 6 and 24100, respectively. Using such a low V/III ratio the presence of Si has a strong influence leading to vertical growth [13]. At the end of the growth process, the sample was cooled down to RT supplying NH3. We emphasize that no additional metal catalyst was used for rod growth in order not to deteriorate the optical and structural properties [14].
Scanning electron microscopy (SEM) and RT cathodoluminescence (CL) measurements were perfomed utilizing a combined system consisting of a Hitachi S4800 and a Gatan MonoCL setup. For CL measurements an acceleration voltage of 5 keV and a beam current of 10 μA was applied for excitation resulting in the excitation of both GaN near band edge emission and yellow defect luminescence. The SEM and CL measurements were performed at a tilt angle of 60° between the surface normal and the electron beam in order to reduce charging effects. CL of single rods were recorded either exciting the total area of the rod with the electron beam (25 area scans per second) or with a fixed (not scanning) electron beam at different positions on the rod. RT microphotoluminescence (μ-PL) measurements were done with a Horiba LabRam HR800 spectrometer in the backscattering configuration. An excitation laser emitting at 457 nm was focused through a 100× objective (numerical aperture 0.9) in the middle of the GaN rods. The diameter of the incident focused laser beam and the laser power onto the sample’s surface were ∼ 0.7 μm and ∼ 5.82 mW, respectively. Since the energy of the laser is below the band gap of GaN, the near band edge emission is not excited as in the case of CL. Instead, GaN is transparent for the used laser wavelength providing spectral information about the yellow defect luminescence of the whole rod.
3. Results and discussion
An ensemble of GaN rods is shown in Fig. 1. All rods are vertically aligned with respect to the sample surface. Most of the rods have a regular or an almost regular hexagonal shape. Only a few are non-regular or consist of two rods. The inner diameter of the regular rods is between 3 μm and 4 μm. The height variation is large as visible by flat disk-like structures as well as pronounced rods with a height up to 6 μm. The rod density is of the order of 105–106 cm−2. Although no antisurfactant mask was used (such as e.g. SiO2 or SiNx) there is no deposition of a GaN layer in between the rods on the sapphire surface. These results are showing that a masking procedure, as it is e.g. applied in Ref. [13], is not necessary to achieve vertical GaN rods/wires. Moreover, our experiments and other works show that pre-structured substrates as well as in-situ or ex-situ deposited masks lead to non-regular hexagonal rod shapes and the absence of WGMs [12, 13]. In this context, our mask-free growth procedure presented herein represents a significant technological advancement over pre-existing growth methods.
Fig. 1 SEM image of an ensemble of self-assembled GaN rods vertically aligned on a sapphire substrate.
A detailed view of a single GaN rod in Fig. 2 is showing smooth m-plane sidewalls and a c-plane top facet. The edges appear to be sharp. The CL spectrum A in Fig. 3 was measured exciting the total area of the rod in Fig. 2 with the electron beam. It reveals a typical GaN CL spectrum with GaN near band edge emission at 3.4 eV and a broad yellow defect luminescence band at 2.2 eV. Additionally, a weak emission is visible at 2.9 eV. The origin of this peak remains unclear, but might be attributed to surface states [15]. A slight modulation of the yellow defect band is already visible due to the appearance of optical modes as discussed in the next paragraph.
Fig. 2 SEM image of a single GaN rod exhibiting a regular hexagonal shape with smooth facets and well defined edges.
Fig. 3 CL spectra of the rod shown in Fig. 2. Spectrum A was recorded during 125 scans of the electron beam over the whole rod area. The inset is showing the position of the fixed electron beam for spectra B–F. The red dashed lines indicate the calculated TM WGMs with mode numbers from 37 to 49. The Q-factor of one selected WGM is labeled. The full width of half maximum of the GaN near band edge emission of graph A was determined to 155 meV.
CL measurements in Fig. 3 with a fixed (not scanning) electron beam of the single GaN rod shown in Fig. 2 have been performed at the center of the top facet and at four sidewall positions. Such a CL measurement procedure allows us to locally excite and study WGMs in the micrometer-sized GaN resonators. Exposing the top facet leads to spectrum B with weak, broad and non-periodic modes. Fixing the electron beam at the sidewall facets reveals a complete different mode structure. Periodic modes are dominating the spectra C–F with sharp modes being present in spectra E–F. Top-to-bottom or opposite sidewall Fabry-Perot oscillations as well as quasi-WGMs (reflections at three sidewall facets) would lead to larger spectral mode distance due to smaller optical path and are therefore not considered to be responsible for these modes [8]. The spectral mode distance is in agreement if hexagonal WGMs are assumed [6]. Hexagonal WGMs have been calculated for the most dominating modes occurring in both spectra E and F using the following equation deduced from a simple plane wave model [6, 16]:
Here, Ri = 1883 nm is the inner radius measured by SEM, λmode is the wavelength of the mode, n the refractive index, N the mode number, β = n for transversal electric (TE) and β = 1/n for transversal magnetic (TM) polarized modes. TEN modes are close to TMN+1 (∼1–2 nm) and are assumed to have broader linewidth compared to TM modes [6]. Therefore, TE modes are more difficult to be resolved. From equation 1 we can note that the position of a mode and the spacing between two peaks depend on the diameter of the rod. Small differences of the cavity diameter would lead to distinct changes of the mode position and spacing. The red dashed lines in Fig. 3 indicate the TM modes with N from 37 to 49. Good agreement between spectral positions of the calculated and measured modes in graph E and F is achieved. However, to obtain these results the refractive index n, used from Ref. [17], was multiplied by a factor of 0.9103. Such a low refractive index was also reported in Ref. [18] but its physical origin remains to be clearified. Possible reasons could be the limits of the applied simple plane wave model valid for Ri ≫ λ and/or a carrier induced refractive index change due to high doping level in the rods [6, 13, 19, 20]. Indeed, the full width of half maximum of the GaN near band edge emission of graph A measured to 155 meV indicates an n-doping concentration above 3 · 1019 cm−3 compared to the data presented in Ref. [21]. Not all the modes visible in the spectra can be assigned to TM WGMs. Furthermore, the intensity and spectral position of WGMs varies if the rod is excited at different positions as displayed in Fig. 3. It is expected that depending on the spatial exposure either TM, TE and/or higher radial WGMs are excited [22]. The hexagonal WGM with a Q-factor of 443 was achieved for this rod at a wavelength of 536.2 nm when exposing it with the electron beam at the upper part of the left sidewall (spectrum E).Even higher Q-factors were achieved using μ-PL because the entire rod is excited resulting in light being created uniformly inside the rod and not only locally as in the CL experiments. The optical image of another rod is shown in the inset of Fig. 4. Its inner radius cannot be exactly measured from the optical image but it is estimated to be 1.698 μm based on the best agreement between experimental and simulated results. The corresponding μ-PL spectra excited with a 457 nm laser beam focused in the middle of the rod is presented in Fig. 4. Following the same reasoning as in the previous paragraph, we identified TM modes with mode numbers from 29 to 42 marked by red dashed lines in Fig. 4. Many other sharp modes superimposed on the yellow defect luminescence are visible in the spectrum being attributed to the appearance of TE and higher radial WGMs [22]. The mode structure of hexagonally shaped cavities has been discussed in detail by several authors e.g. Ref. [16, 23]. A detailed classification of each mode observed for each investigated rod requires extensive calculations for the large mode numbers in question. This aspect is not in the scope of this paper and will be addressed in a future work. Several WGMs have a Q-factor above 1000. Q = 4177 was measured at 538.1 nm. Thus, the main result is that our GaN rods fabricated by a simple MOVPE growth process show similar high Q-factors such as GaN microdisks e.g. presented in Ref. [24].
Fig. 4 μ-PL spectrum of a single GaN rod. The yellow defect band is superimposed by a large amount of WGMs. TM modes with mode numbers from 29 to 42 are marked. Q-factors of WGMs with a value larger than 1000 are labeled. A Q-factor of 4177 was measured at 2.30390 eV (538.1 nm). A detailed view of this mode with a full width at half maximum of 551.6 meV is displayed in the right inset. The left inset is showing an optical image of the GaN rod. The scale bar corresponds to 4 μm.
CL and μ-PL measurements have been performed on the same GaN rod in order to study the correlation between the CL and μ-PL spectra. The spectra along with the SEM image of the studied rod are displayed in Fig. 5. The μ-PL spectrum measured at normal incidence on the top facet is compared with three CL spectra recorded with a fixed electron beam at different positions on the rod: top facet, left sidewall facet and right sidewall facet. There is no correlation between the μ-PL and the CL spectra and even the correlation between the CL spectra is rather low as already shown in Fig. 3. The reason for this discrepancy might be due to the different excitation volumes. While the entire rod is excited with the laser beam in μ-PL measurements, there is only local excitation with the electron beam in CL experiments. With an acceleration voltage of the electron beam of 5 keV the penetration depth of electrons into GaN is 150 nm. Thus, light is only created in this small volume in CL as compared to the entire rod in μ-PL studies. This may explain the occurrence of denser peaks on the μ-PL spectra as shown in Figs. 4 and 5. Depending on the spatial exposure different modes are excited as visible in Fig. 5.
Fig. 5 Comparison between CL and PL spectra recorded from the same rod. Graph A is the μ-PL spectrum of the rod. Graph B, C and D are recorded with a fixed electron beam at the right sidewall facet, left sidewall facet and top facet, respectively. The inset shows the SEM image of the rod.
Moreover, we have taken μ-PL measurements on the same rod under different excitation laser powers (not shown). Although we varied the power up to two orders of magnitude, no changes in the mode structure were observed.
We expect that lasing from these structures can be achieved in case of non-polar InGaN quantum wells or quantum dots deposited on the m-plane facets of the GaN rods used as an efficient optically active region [5]. This would allow the study of quantum-electrodynamic effects such as strong coupling between the excitonic mode and the cavity mode [1]. Additionally, they can be integrated in optoelectronic devices such as light emitting diodes and solar cells after the corresponding doping [25]. Furthermore, due to the appearance of high Q-factor WGMs we think that these rods can also be utilized for the realization of all-optical biological and chemical sensing devices [26].
There are some practical issues to be solved before the technological implementation of the GaN hexagonal resonators in real optical devices. This includes the realization of an ordered array of regular hexagonal shaped rods for scale integration, collection of the in-plane light emission via a fibre [27], monomode operation, and minimization of the optical coupling to the underlying substrate and the optical volume via diameter variation of the rod [28]. Based on the existing literature, we believe that these points can be addressed but at present they are beyond the scope of this work.
4. Conclusion
The paper reports about a simple and fast method to grow catalyst-free, self-assembled GaN rods on sapphire by MOVPE. No complex growth procedure and no additional processing was applied. Most of the rods exhibit a highly regular hexagonal shape with smooth sidewall facets and sharp edges. Compared to the present microdisk and micropillar fabrication processes, our method provides a simple, cost- and time-effective alternative to produce such structures. CL and μ-PL investigations were carried out and WGMs with high Q-factors up to 4177 were measured. Thus, these GaN rods can be used as hexagonal WGM resonators for fundamental investigations of quantum-electrodynamic effects as well as for lasing and sensing applications.
Acknowledgment
This project has been founded by the ROD SOL project within the 7th Framework of the European Union and by the German Research Society DFG within the project number FOR 1616.
References and links
1. K. J. Vahala, “Optical microcavities,” Nature 424, 839–846 (2003). [CrossRef] [PubMed]
2. G. Khitrova, H. M. Gibbs, M. Kira, S. W. Koch, and A. Scherer, “Vacuum Rabi splitting in semiconductors,” Nature Phys. 2, 81–90 (2006). [CrossRef]
3. G. Malpuech, A. Di Carlo, A. V. Kavokin, J. J. Baumberg, M. Zamfirescu, and P. Lugli, “Room-temperature polariton lasers based on GaN microcavities,” Appl. Phys. Lett. 81, 412–414 (2002). [CrossRef]
4. S. Christopoulos, G. Baldassarri Höger von Högersthal, A. J. D. Grundy, P. G. Lagoudakis, A. V. Kavokin, J. J. Baumberg, G. Christmann, R. Butté, E. Feltin, J.-F. Carlin, and N. Grandjean, “Room-temperature polariton lasing in semiconductor microcavities,” Phys. Rev. Lett. 98, 126405 (2007). [CrossRef] [PubMed]
5. A. C. Tamboli, E. D. Haberer, R. Sharma, K. H. Lee, S. Nakamura, and E. L. Hu, “Room-temperature continuous wave lasing in GaN/InGaN microdisks,” Nature Photon. 1, 61–64 (2007). [CrossRef]
6. T. Nobis, E. M. Kaidashev, A. Rahm, M. Lorenz, and M. Grundmann, “Whispering gallery modes in nanosized dielectric resonators with hexagonal cross section,” Phys. Rev. Lett. 93, 103903 (2004). [CrossRef] [PubMed]
7. R. Chen, B. Ling, X. W. Sun, and H. D. Sun, “Room temperature excitonic whispering gallery mode lasing from high-quality hexagonal ZnO microdisks,” Adv. Mater. 23, 2199–2204 (2011). [CrossRef] [PubMed]
8. T. Kouno, K. Kishino, and M. Sakai, “Lasing action on whispering gallery mode of self-organized GaN hexagonal microdisk crystal fabricated by RF-plasma-assisted molecular beam epitaxy,” IEEE J. Quantum Electron. 47, 1565–1570 (2011). [CrossRef]
9. C. Tessarek and S. Christiansen, “Self-catalyzed, vertically aligned GaN rod-structures by metal-organic vapor phase epitaxy,” Phys. Status Solidi C 9, 596–600 (2012). [CrossRef]
10. N. Grandjean, J. Massies, and M. Leroux, “Nitridation of sapphire. Effect on the optical properties of GaN epitaxial overlayers,” Appl. Phys. Lett. 69, 2071–2073 (1996). [CrossRef]
11. X. J. Chen, G. Perillat-Merceroz, D. Sam-Giao, C. Durand, and J. Eymery, “Homoepitaxial growth of catalyst-free GaN wires on N-polar substrates,” Appl. Phys. Lett. 97, 151909 (2010). [CrossRef]
12. S. F. Li, S. Fuendling, X. Wang, S. Merzsch, M. A. M. Al-Suleiman, J. D. Wei, H.-H. Wehmann, A. Waag, W. Bergbauer, and M. Strassburg, “Polarity and itsiInfluence on growth mechanism during MOVPE growth of GaN sub-micrometer rods,” Cryst. Growth Des. 11, 1573–1577 (2011). [CrossRef]
13. R. Koester, J. S. Hwang, C. Durand, D. Le Si Dang, and J. Eymery, “Self-assembled growth of catalyst-free GaN wires by metal-organic vapor phase epitaxy,” Nanotech. 21, 015602 (2010). [CrossRef]
14. C. Cheze, L. Geelhaar, O. Brandt, W. M. Weber, H. Riechert, S. Münch, R. Rothemund, S. Reitzenstein, A. Forchel, T. Khegias, P. Komninou, G. P. Dimitrakopulos, and T. Karakostas, “Direct comparison of catalyst-free and catalyst-induced GaN nanowires,” Nano Res. 3, 528–536 (2010). [CrossRef]
15. M. A. Reshchikov and H. Morkoc, “Luminescence properties of defects in GaN,” J. Appl. Phys. 97, 061301 (2005). [CrossRef]
16. J. Wiersig, “Hexagonal dielectric resonators and microcrystal lasers,” Phys. Rev. A 67, 023807 (2003). [CrossRef]
17. M. J. Bergmann, Ü. Özgür, H. C. Casey, H. O. Everitt, and J. F. Muth, “Ordinary and extraordinary refractive indices for AlGaN epitaxial layers,” Appl. Phys. Lett. 75, 67–69 (1999). [CrossRef]
18. B. C. Joshi, M. Mathew, B. C Joshi, D. Kumar, and C. Dhanavantri, “Characterization of GaN/AlGaN epitaxial layers grown by metalorganic chemical vapor deposition for high electron mobility transistor applications,” Pranama - J. Phys. 74, 135–141 (2010). [CrossRef]
19. C. Bulutay, C. M. Turgut, and N. A. Zakhleniuk, “Carrier-induced refractive change and optical absorption in wurtzite InN and GaN: Full-band approach,” Phys. Rev. B 81, 155206 (2010). [CrossRef]
20. G. Cywinski, R. Kudrawiec, W. Rzodkiewicz, M. Krysko, E. Litwin-Staszewska, B. Lucznik, J. Misiewicz, and C. Skierbiszewki, “Doping-induced contrast in the refractive index for GaInN/GaN structures at telecommunication wavelengths,” Appl. Phys. Express 2, 111001 (2009). [CrossRef]
21. E. F. Schubert, I. D. Goepfert, W. Grieshaber, and J. M. Redwing, “Optical properties of Si-doped GaN,” Appl. Phys. Lett. 71, 921–923 (1997). [CrossRef]
22. T. Nobis, A. Rahm, C. Czekalla, M. Lorenz, and M. Grundmann, “Optical whispering gallery modes in dodecagonal zinc oxide microcrystals,” Superlatt. Microstr. 42, 333–336 (2007). [CrossRef]
23. T. Nobis and M. Grundmann, “Low-order optical whispering-gallery modes in hexagonal nanocavities,” Phys. Rev. A 72, 063806 (2005). [CrossRef]
24. A. C. Tamboli, M. C. Schmidt, A. Hirai, S. P. DenBaars, and E. L. Hu, “Observation of whispering gallery modes in nonpolar m-plane GaN microdisks,” Appl. Phys. Lett. 94, 251116 (2009). [CrossRef]
25. S. Li and A. Waag, “GaN based nanorods for solid state lighting,” J. Appl. Phys. 111, 071101 (2012). [CrossRef]
26. F. Vollmer, D. Braun, A. Libchaber, M. Khoshsima, I. Teraoka, and S. Arnold, “Protein detection by optical shift of a resonant microcavity,” Appl. Phys. Lett. 80, 4057–4059 (2002). [CrossRef]
27. K. Srinivasan, A. Stintz, S. Krishna, and O. Painter, “Photoluminescence measurements of quantum-dot-containing semiconductor microdisk resonators using optical fiber taper waveguides,” Phys. Rev. B 72, 205318 (2005). [CrossRef]
28. D. J. Gargas, M. C. Moore, A. Ni, S.-W. Chang, Z. Zhang, S.-L. Chuang, and P. Yang, “Whispering gallery mode lasing from zinc oxide hexagonal nanodisks,” ACS Nano 4, 3270–3276 (2010). [CrossRef] [PubMed]
