Abstract
Epitaxial Laterally overgrown (ELOG) InGaN materials are investigated using a polarization modulated scanning near-field optical microscope. The authors found that luminescence has spatial inhomogeneities and it is partially polarized. Near-field photoluminescence shows polarization phase fluctuation up to 45 degrees over adjacent domains. These results point toward the existence of asymmetries in carrier confinement due to structural anisotropic strain within the framework of the ELOG structure.
©2008 Optical Society of America
Wide-band-gap group III nitride materials were used to make electronic and optical devices, such as LED and solid state Lasers[1]. One problem in the fabrication of these devices is the high density (109–1011 cm-2) of threading dislocations (TDs) produced when the nitride materials are grown on lattice-mismatch substrates. One method to reduce TDs is to use a technique called Expitaxial Lateral Overgrowth (ELOG), a commonly used method that alters the growth orientation along striped regions of the sample[2, 3, 4].
This growth method is inherently anisotropic and induces the formation of stress in the crystal structure which may produce irregular carrier confinement features that are invisible to standard microscopy and are generally difficult to study. Since models point out that anisotropic strain induces asymmetry in the carrier localization and this results in polarized PL emission[5], we propose here a simple optical polarization modulation near-field technique to investigate on carrier confinement processes by studying the local polarization properties in the proximity of the sample surface.
The ELOG device sample was grown by Nichia Corporation (Japan) with MOCVD methodology: exact growth condition are undisclosed, crystal structure is shown in Fig. 1(a). A base layer of 2µm thick GaN is grown on a sapphire substrate, then a 200nm thin striped pattern of SiO2 is deposited. Each strip is about 16µm wide and the separation between each of them is 4µm. This structure acts as a mask for a subsequent 15µm GaN deposition.
The growth is initially possible only on the GaN surface available between strips. From this region, named “seed”, the growth expands laterally into “wings” that invade the SiO2 mask region and eventually contact the adjacent fronts[6, 7]. The growth extends vertically for 15 µm then a final 3nm Indium doped quantum well (QW) is grown with its 5nm capping layer as in standard (non-ELOG) GaN/InGaN samples. The 4µm wide “seed” region between the SiO2 mask is thought to have a reduced number of TDs because of the different crystal orientation and direction of growth[8].
In Fig. 1(b) we show two photoluminescence (PL) microscope images of the sample taken through polars (a Polaroid film) oriented at 0 and 90 degrees. The images are focused slightly below surface to show the mask structure clearly. No particular optical differences are appreciable. Focusing on the sample surface we imaged at 5 different polars angles, intensity measured in relevant areas are shown in table 1 and Fig. 2, confirming weak polarization effects in this far-field test.
There could be stronger confined fluctuations, local effects that are not detectable with a conventional far field microscope. We setup a home-made polarization-modulation system (PM-SNOM[9, 10, 11, 12]) capable of evaluating the polarization changing local properties of the sample. The apparatus is described in Fig. 3, a typical SNOM setup where a multiple-steps chemically etched 100nm apertured optical probe[13, 14, 15] is collecting the optical signal in the proximity of the surface of the sample. The incoming light that excites the photolumines-
cence is a linearly polarized beam with variable polarization angle; this angle is continuously varied by filtering with a revolving polars film (frequency of rotation 63Hz). In this configuration, if the sample is not sensitive to polarization, fluorescence will be induced with same intensity on all angles; however, if there are birefringence or dichroic properties, the luminescence should produce intensity variations in tune with the modulation frequency. The lock-in amplifier will be able to generate a map of the sample dichroism and this map can be compared with the simultaneously taken photoluminescence map[12]. The natural residual birefringence of the probe will introduce a slight ellipticity of the light; a compensation of this disturbance is not required since the effect is the same everywhere on the sample (it is not position dependent). The effect appears simply as an offset value in the final maps.
The near-field luminescence map is shown in Fig. 4(a), the striped structure of the sample is noticeable with the “seed” and “wing” regions indicated. In (b) each pixel value represents the angular phase of the optical signal. The two maps (a) and (b) are taken simultaneously.
Domains and structures are different, indicating that we are free from cross-talk features[12]. Images are contrast-enhanced for visibility, however the strength of the dichroism can be calculated by treating the absolute signal values that are preserved in the actual data files and used with instrument calibration curves. The lock-in amplifier outputs the signal accordingly to its phase-voltage curve and subsequently an analog to digital (A/D) card converts this value to a unit labeled “counts” by the instrument. Lock-in response is linear with an inclination of about 16.9 deg per Volts, zero degree corresponds to zero volts. Our A/D converter responds also linearly with a sensitivity σs=562Vl+2050 where σs is in counts/Volts and Vl is the lock-in output in volts; with these parameters we can estimate easily the actual dichroism effect measured on each local feature mapped. The result of this is shown in the cross section of Fig. 5(b) along the lines r 1 and r 2 of Fig. 4 which are plotted with the corresponding phase shift values in deg.
We demonstrated strong polarization changing properties in the surface proximity of ELOG GaN/InGaN/GaN devices and we were able to obtain quantitative maps of the PL angular orientation. The investigation points toward the existence of peculiar asymmetries in carrier confinement; this suggests structural anisotropic strain within the polcrystalline framework of the ELOG structure. The technique used is applicable to organic conjugated materials as well as any optical absorbing or optical emitting films. It has an impact on all measurements where localized polarization properties of the optical emission are important. We would like to thank Dr. Akio Kaneta of Kyoto University for his help with sample description and preparation, we also aknowledge the Kyoto Nanotechnology cluster for support, the Italy-Japan bilateral Projects “2A2” and 31 “Near-field polarization contrast for nano-optics applications” for mobility support. We are indebted with Tatsuo Nakagawa of Unisoku Ltd, Osaka, for his assistance with the SNOM system and with A. La Barbera Translations, UK for support with English issues.
References and links
1. S. Nakamura, “The roles of structural imperfections in InGaN-based blue light emittind diodes and Lasers Diodes,” Science 281, 956–961 (1998). [CrossRef]
2. A. Usui, H. Sunakawa, A. Sakai, and A. A. Yamaguchi, “Thick GaN epitaxial growth with low dislocation density by hydride vapor phase epitaxy,” Jpn. J. Appl. Phys. 36, L899–L902 (1997). [CrossRef]
3. T. S. Zheleva, O. H. Nam, M. D. Bremser, and R. F. Davis, “Dislocation density reduction via lateral epitaxy in selectively grown GaN structures,” Appl. Phys. Lett. 71, 2472–2474 (1997). [CrossRef]
4. M. D. Craven, S. H. Lim, F. Wu, J. S. Speck, and S. P. DenBaars, “Threading dislocation reduction via laterally overgrown nonpolar (11(2)over-bar0) a-plane GaN,” Appl. Phys. Lett. 81, 1201–1203 (2002). [CrossRef]
5. O. Stier, M. Grundmann, and D. Bimberg, “Electronic and optical properties of strained quantum dots modeled by 8-band k center dot p theory,” Phys. Rev. B 59, 5688–5701 (1999). [CrossRef]
6. M. Matthews, J. Hsu, S. Gu, and T. Kuech, “Carrier density imaging of lateral epitaxially overgrown GaN using scanning confocal Raman microscopy,” Appl. Phys. Lett. 79, 3086–3088 (2001). [CrossRef]
7. J. Hsu, M. Matthews, D. Abusch-Magder, R. Kleiman, D. Lang, S. Richter, S. Gu, and T. Kuech, “Spatial variation of electrical properties in lateral epitaxially overgrown GaN,” Appl. Phys. Lett. 79, 761–763 (2001). [CrossRef]
8. A. Kaneta, M. Funato, Y. Narukawa, T. Mukai, and Y. Kawakami, “Direct correlation between nonradiative recombination centers and threading dislocations in InGaN quantum wells by near-field photoluminescence spectroscopy,” Phys. Status Solidi C 3, 1897–1901 (2006). [CrossRef]
9. R. Micheletto, Y. Kawakami, C. Manfredotti, Y. Garino, and M. Allegrini, “Dichroism of diamond grains by a polarization modulated near field optical setup,” Appl. Phys. Lett. 89, 121125 (2006). [CrossRef]
10. P. G. Gucciardi, M. Allegrini, R. Micheletto, T. Kotani, T. Hatada, and Y. Kawakami, “Waveguide behavior of Distributed Bragg Reflectors probed by polarization-modulated near-field optical microscopy,” J. Korean Phys. Soc. 47, S101–S108 (2005).
11. B. Bhushan and H. Fuchs, eds., Applied Scanning Probe Methods II: Scanning Probe Microscopy Techniques (Springer, 2005).
12. R. Micheletto, M. Allegrini, and Y. Kawakami, “Artifacts in Polarization Modulation Scanning Near-field Optical Microscopes,” J. Opt. A 9, 431–434 (2007). [CrossRef]
13. S. Mononobe and M. Ohtsu, “Development of a fiber used for fabricating application oriented near-field optical probes,” IEEE Photon. Technol. Lett. 10, 99–101 (1998). [CrossRef]
14. S. Mononobe, M. Naya, T. Saiki, and M. Ohtsu, “Reproducible fabrication of a fiber probe with a nanometric protrusion for near-field optics,” Appl. Opt. 36, 1496–1500 (1997). [CrossRef] [PubMed]
15. T. Matsumoto, T. Ichimura, T. Yatsui, M. Kourogi, T. Saiki, and M. Ohtsu, “Fabrication of a near-field optical fiber probe with a nanometric metallized protrusion,” Opt. Rev. 5, 369–373 (1998). [CrossRef]