Open Access
Add to BibSonomyAbstract
Ta2O5/SiO2 dielectric multi-layer micro-mirror array (MMA) with 3μm mirror size and 6μm array period was fabricated on c-plane sapphire substrate. The MMA was subjected to 1200°C high temperature annealing and remained intact with high reflectance in contrast to the continuous multi-layer for which the layers have undergone severe damage by 1200°C annealing. Epitaxial lateral overgrowth (ELO) of gallium nitride (GaN) was applied to the MMA that was deposited on both sapphire and sapphire with 2.56μm GaN template. The MMA was fully embedded in the ELO GaN and remained intact. The result implies that our MMA is compatible to the high temperature growth environment of GaN and the MMA could be incorporated into the structure of the micro-LED array as a one to one micro backlight reflector, or as the patterned structure on the large area LED for controlling the output light.
©2009 Optical Society of America
1. Introduction
Light emitting diode (LED) and micro-LED array [1],[2],[3],[4] are drawing much attention for purpose of illumination and display. Total internal reflection limits the light extraction efficiency of the LED and efforts for enhancing the light extraction efficiency, either from the front side or from the back side of the LED, are intensively investigated in the field. For the back side enhancement, particularly for GaN-based LED, it is essential that the enhancement structure should be able to sustain the high temperature growth environment of GaN, usually as high as ~ 1100°C. Patterned sapphire substrate together with epitaxial lateral overgrowth (ELO) technique of GaN was demonstrated to serve the purpose [5],[6]. Micro-mirror array (MMA) was proposed [7], but high temperature implementation was not found.
In this paper, we report our results on the fabrication and characteristics for a dielectric multilayer Ta2O5 / SiO2 MMA that can sustain 1200°C high temperature annealing. We also report the results on the fabrication of the GaN with ELO technique to fully embed the MMA structure in the GaN.
2. Micro-mirror array fabrication and annealing
The fabrication process for the MMA is shown schematically in Fig. 1. Continuous Ta2O5/SiO2 multi-layers, 5 pairs in total, were deposited on a c-plane sapphire substrate by ion beam sputtering. The thickness of the Ta2O5 layer and the SiO2 layer was 52nm and 68nm, and the refractive index was 2.32 and 1.46 at 450nm wavelength, respectively. An 8.5μm thick SU-8 photo-resist (PR) was applied on the multi-layers for photo-lithography. The array pattern was hexagonal with 3μm separation between each element. Each element was also hexagonal in shape with 3μm width between the opposite sides. Referring to Fig. 2, the hexagon of the array and the element were aligned with that of the c-plane sapphire. Figure 1(a) shows the top view and the side view of the scanning electron microscope (SEM) image of the PR pattern after development. The sample was then subjected to inductive-coupled-plasma (ICP) dry etching in the CF 4/Ar mixture to remove the multi-layers. Figure 1(b) shows the SEM image of the sample after the dry-etching. The remaining PR was in a cup-shape on top of the element. The sample was then dipped in 3:1 H2SO4/H2O2 for 30 minutes at 160°C, the remaining PR was completely removed and the multi-layer was intact. Figure 1(c) shows the SEM image of the sample after the complete fabrication process. The sample together with another sample that had an identical but continuous multi-layer to be served as standard for comparison were subjected to high temperature annealing in nitrogen gas. The annealing condition was that the temperature was raised from room temperature to 1200°C in 30 minutes and held for 30 minutes then cooled down to room temperature in one hour. Figure 3(a)-(c) show the SEM pictures, at different magnifications, of the top and the side views for the continuous Ta2O5/SiO2 multilayer sample before annealing, and Figs. 3(d)-(f) show that after the annealing. Severe damage, peeling-off and bubbling, for the continuous multi-layer structure by the annealing can be seen. Fig. 4(a) and (b), (c) and (d) show the SEM pictures for the MMA before and after annealing, respectively. The pictures clearly indicate that the layer-structure remained intact after the annealing, the severe damage in the continuous layer structure did not appear in the discrete multi-layer array. We believe that the internal stresses in the layers and between the layers and the substrate are proportional to the area of the layers such that the discrete micro-mirror has much lower stress than the continuous layers to sustain the high temperature annealing.
Fig. 1. SEM pictures of the top and the side views of the sample at different stages of the MMA fabrication process.
Fig. 2. Relative orientation of the pattern. The hexagon of the array and the element are aligned with that of the c-plane sapphire.
Our multi-layer was designed to have a high reflection band at the blue wavelength region in the air. Figure 5(a) shows the reflectance spectrum of the continuous layers measured before annealing. The spectrum was taken by using a spectrophotometer with an integrating sphere. In order to investigate the annealing effect on the reflectance of the MMA, we used a He-Cd laser that has output wavelength at 441.6nm to probe the reflectance of the micro-mirror before and after annealing. The laser beam was focused to a spot size of 0.58μm on top of the micro-mirror, and the surface of the sample was scanned by moving the sample with a micrometer with 0.25μm resolution. A 99% reflectance wide band aluminum mirror was used as the standard for reflectance calibration. The result of the reflectance scanning is shown in Fig. 5(b) for two consecutive micro-mirrors elements. The top line in Fig. 5(b) is the reflectance of the continuous mirror at 441.6nm wavelength as obtained from Fig. 5(a). The reflectance of the micro-mirror before annealing is ~ 2% lower than that of the continuous mirror, and we believe this difference is due to the lithography process induced surface modification. The reflectance of the micro-mirror dropped from 83% to 70% after annealing. Comparing to the reflectance of the interface between the GaN and the sapphire which is about ~ 3% (refractive index of GaN and sapphire is 2.45 and 1.7836 respectively), or comparing to the reflectance of the interface between air and sapphire which is about 6.7%, the micro-mirror with 70% reflectance after annealing still can be considered to be a high reflectance mirror for the sapphire substrate. We assert that the following reason is responsible for the reflectance drop after annealing: referring to the inset of Fig. 4(b) that shows the edge of the micro-mirror after the dry-etching process. Due to the higher lateral etching rate for the SiO2 film than the Ta2O5 film, the SiO2 layers were etched deeper inward than the Ta2O5 films to form the undulated side wall for the micro-mirror, but the undulation disappeared after the annealing as can be seen from the inset of Fig. 4(d).
Fig. 3. SEM pictures of the continuous multi-layers. (a)(b) top view and (c) side view before annealing. (d)(e) top view and (f) side view after 1200°C annealing.
Fig. 4. SEM pictures of the MMA. (a) top view and (b) side view before annealing. (c) top view and (d) side view after 1200°C annealing.
This phenomenon implies that there is material diffusion at the side wall of the micro-mirror during the annealing process, and the diffusion could also be occurred in the interface region of the layers such that the refractive index of the layers were modified to be gradient and the reflectance was therefore changed.
Fig. 5. (a) Reflectance spectrum of the continuous multi-layers before annealing. (b) Reflectance scanning at 441.6 nm wavelength of the MMA before and after 1200°C annealing.
3. Embedding the micro-mirror array in the ELO GaN
Since there is open space between the elements of the MMA, it is therefore possible to grow GaN in between the elements from the substrate vertically, and ELO technique could then be applied to cover the MMA such that the light emitting structure could be grown thereafter. Table.1 shows the process parameters for all the relevant steps of the MOCVD process for growing the undoped GaN. The MOCVD unit was a AIXTRON 2000HT. We have prepared MMA on two types of substrate for the ELO, type I is c-plane bare sapphire for which steps 1 ~ 5 were applied consecutively to cover the MMA; and type II is c-plane sapphire with a 2.56μm thick GaN template for which steps 1 ~ 4 were applied for growing the template, MMA was then deposited on the template and step 4 ~ 5 were followed for ELO. Figures 6(a) and (b) show the SEM pictures of the MMA on type I substrate covered by the ELO GaN but the ELO process was terminated before complete coalescence, and Fig. 6(c) shows that for complete coalescent GaN. Figure 6(d)~(f) show the same but on type II substrate. It is clearly indicated by these pictures that the MMA was fully embedded in the ELO GaN with the multilayer stack being kept intact. Figure 6 also show that, with the process parameters provided in Table.1, the coalescence process of the ELO GaN is more uniform on the template than on the sapphire.
The heat resistive MMA embedded in GaN could be applied as a one-to-one micro backlight reflector for micro-LED array, it could also be applied as a patterned array underneath a large area LED serving as a light extraction enhancement structure similar to the patterned sapphire substrate (PSS) [5],[6] but with high reflectance elements. For those applications, the growth rate of the ELO GaN needs to be optimized to yield a coalescent GaN layer with desired thickness such that the growing of the light emitting structure could be followed. Figure 6(d) implies that the ELO GaN on the MMA, with the ELO process terminated before complete coalescence, could be used as a 2-D photonics crystal. The periodicity is determined by the array pattern, the duty cycle of the period can be determined by terminating the ELO process at different stage. The multi-layers could be etched off to yield a 2-D GaN photonics crystal. The MMA could also be replaced by any arrayed pattern that is made of plateau-like element, for example SiO2, and the ELO could produce a 2-D GaN pattern. The photonics crystal fabricated in such way could be either on the front or on the back surface of a LED to serve the purposes as light extraction enhancement[8], collimation[9], and polarization control[10],[11].
Table 1. MOCVD process parameters
Fig. 6. SEM pictures of the MMA on sapphire (a)-(c) and on sapphire with 2.56μm GaN template (d)-(f). (a)(b)(d)(e): the ELO process terminated half-way. (c)(f): the ELO process was completed.
4. Conclusion
A Ta2O5/SiO2 dielectric multi-layer MMA was fabricated on sapphire and sapphire with GaN template. Due to small area of the array element in contrast to the large area of the continuous multi-layer, the MMA can sustain 1200°C high temperature annealing without damage and peeling-off, and the reflectance of the MMA after the annealing still maintain high to serve as micro mirror. Properly controlling the ELO process, we demonstrated that the MMA can be fully embedded in the GaN by applying the ELO technique and survived the high temperature GaN fabrication process. Our results implied that the MMA could be applied as a one-to-one micro backlight reflector in the micro-LED array, or as the light controlling structure on the front or back surface of LED for light extraction enhancement, collimation, and polarization control. If the ELO process is terminated before the GaN is completely coalescent on the MMA, a 2-D GaN photonics crystal could be fabricated.
References and links
1. S. X. Jin, J. Li, J. Y. Lin, and H. X. Jiang, “InGaN/GaN quantum well interconnected microdisk light emitting diodes,” Appl. Phys. Lett. 77, 3236–3238, (2000). [CrossRef]
2. H. X. Jiang, S. X. Jin, J. Li, J. Shakya, and J. Y. Lin, “III-nitride blue microdisplays,” Appl. Phys. Lett. 78, 1303–1305, (2001). [CrossRef]
3. H. W. Choi, M. D. Dawson, P. R. Edwards, and R. W. Martin, “High extraction efficiency InGaN micro-ring light-emitting diodes,” Appl. Phys. Lett. 83, 4483–4485, (2003). [CrossRef]
4. H. W. Choi, C. Liu, E. Gu, G. McConnell, J. M. Girkin, I. M. Watson, and M. D. Dawson, “GaN micro-light-emitting diode arrays with monolithically integrated sapphire microlenses,” Appl. Phys. Lett. 84, 2253–2255, (2004). [CrossRef]
5. S. Nakamura, M. Senoh, S. I. Nagahama, N. Iwasa, T. Yamada, T. Matsushita, H. Kiyoku, Y. Sugimoto, T. Kozaki, H. Umemoto, M. Sano, and K. Chocho, “InGaN/GaN/AlGaN-based laser diodes with modulation-doped strained-layer superlattices grown on an epitaxially laterally overgrown GaN substrate,” Appl. Phys. Lett. 72, 211–213, (1998). [CrossRef]
6. S. Nakamura, M. Senoh, S. I. Nagahama, N. Iwasa, T. Yamada, T. Matsushita, H. Kiyoku, Y. Sugimoto, T. Kozaki, H. Umemoto, M. Sano, and K. Chocho, “Present status of InGaN/GaN/AlGaN-based laser diodes,” J. of Crystal Growth 189/190, 820–825, (1998). [CrossRef]
7. M. D. Dawson and R. W. Martin, US patent 6563141B1, (2003).
8. K. McGroddy, A. David, E. Matioli, M. Iza, S. Nakamura, S. DenBaars, J. S. Speck, C. Weisbuch, and E. L. Hu, “Directional emission control and increased light extraction in GaN photonic crystal light emitting diodes,” Appl. Phys. Lett. 93, 103502, (2008). [CrossRef]
9. J. J. Wierer, M. R. Krames, J. E. Epler, N. F. Gardner, M. G. Craford, J. R. Wendt, J. A. Simmons, and M. M. Sigalas, “InGaN/GaN quantum-well heterostructure light-emitting diodes employing photonic crystal structures,” Appl. Phys. Lett. 84, 3885–3887, (2004). [CrossRef]
10. S. Noda, M. Yokoyama, M. Imada, A. Chutinan, and M. Mochizuki, “Polarization mode control of two-dimensional photonic crystal laser by unit cell structure design,” Science 293, 1123–1125, (2001). [CrossRef] [PubMed]
11. H. Altug and J. Vuckovic, “Polarization control and sensing with two dimensional coupled photonic crystal microcavity arrays,” Optics Lett. 30, 982–984, (2005). [CrossRef]
