Periodic Si nanopillar arrays by anodic aluminum ... - OSA Publishing

Periodic Si nanopillar arrays by anodic aluminum oxide template and catalytic etching for broadband and omnidirectional light harvesting Hsin-Ping Wang,1 Kun-Tong Tsai,1,2 Kun-Yu Lai,1 Tzu-Chiao Wei,1 Yuh-Lin Wang,2 and Jr-Hau He1,* 1

Institute of Photonics and Optoelectronics, & Department of Electrical Engineering, National Taiwan University, Taipei10617, Taiwan 2 Institute of Atomic and Molecular Sciences, Academia Sinica, Taipei, Taiwan * [email protected]

Abstract: Large-area, periodic Si nanopillar arrays (NPAs) with the periodicity of 100 nm and the diameter of 60 nm were fabricated by metalassisted chemical etching with anodic aluminum oxide as a patterning mask. The 100-nm-periodicity NPAs serve an antireflection function especially at the wavelengths of 200~400 nm, where the reflectance is decreased to be almost tenth of the value of the polished Si (from 62.9% to 7.9%). These NPAs show very low reflectance for broadband wavelengths and omnidirectional light incidence, attributed to the small periodicity and the stepped refractive index of NPA layers. The experimental results are confirmed by theoretical calculations. Raman scattering intensity was also found to be significantly increased with Si NPAs. The introduction of this industrial-scale self-assembly methodology for light harvesting greatly advances the development of Si-based optical devices. ©2011 Optical Society of America OCIS codes: (000.0000) General; (000.2700) General science.

References and links J. M. Choi and S. Im, “Ultraviolet enhanced Si-photodetector using p-NiO films,” Appl. Surf. Sci. 244(1-4), 435–438 (2005). 2. B. S. Richards, “Enhancing the performance of silicon solar cells via the application of passive luminescence conversion layers,” Sol. Energy Mater. Sol. Cells 90(15), 2329–2337 (2006). 3. S. Chhajed, M. F. Schubert, J. K. Kim, and E. F. Schubert, “Nanostructured multilayer graded-index antireflection coating for Si solar cells with broadband and omnidirectional characteristics,” Appl. Phys. Lett. 93(25), 251108–2511083 (2008). 4. J. Ullmann, M. Mertin, H. Lauth, H. Bernitzki, K. R. Mann, D. Ristau, W. Arens, R. Thielsch, and N. Kaiser, “Coated optics for DUV excimer laser application,” Proc. SPIE 2000(3902), 514–527 (2000). 5. M. D. Kelzenberg, S. W. Boettcher, J. A. Petykiewicz, D. B. Turner-Evans, M. C. Putnam, E. L. Warren, J. M. Spurgeon, R. M. Briggs, N. S. Lewis, and H. A. Atwater, “Enhanced absorption and carrier collection in Si wire arrays for photovoltaic applications,” Nat. Mater. 9(3), 239–244 (2010). 6. J. S. Li, H. Y. Yu, S. M. Wong, X. C. Li, G. Zhang, P. G. Q. Lo, and D. L. Kwong, “Design guidelines of periodic Si nanowire arrays for solar cell application,” Appl. Phys. Lett. 95(24), 243113–2431133 (2009). 7. L. Li, T. Y. Zhai, H. B. Zeng, X. S. Fang, Y. Bando, and D. Golberg, “Polystyrene sphere-assisted onedimensional nanostructure arrays: synthesis and applications,” J. Mater. Chem. 21(1), 40–56 (2010). 8. J. Zhu, Z. F. Yu, G. F. Burkhard, C. M. Hsu, S. T. Connor, Y. Q. Xu, Q. Wang, M. McGehee, S. H. Fan, and Y. Cui, “Optical absorption enhancement in amorphous silicon nanowire and nanocone arrays,” Nano Lett. 9(1), 279–282 (2009). 9. Y. R. Lin, K. Y. Lai, H. P. Wang, and J. H. He, “Slope-tunable Si nanorod arrays with enhanced antireflection and self-cleaning properties,” Nanoscale 2(12), 2765–2768 (2010). 10. C. X. Lin and M. L. Povinelli, “Optical absorption enhancement in silicon nanowire arrays with a large lattice constant for photovoltaic applications,” Opt. Express 17(22), 19371–19381 (2009). 11. H. C. Chang, K. Y. Lai, Y. A. Dai, H. H. Wang, C. A. Lin, and J. H. He, “Nanowire arrays with controlled structure profiles for maximizing optical collection efficiency,” Energy Environ. Sci. 4(8), 2863–2869 (2011). 12. O. L. Muskens, J. G. Rivas, R. E. Algra, E. P. Bakkers, and A. Lagendijk, “Design of light scattering in nanowire materials for photovoltaic applications,” Nano Lett. 8(9), 2638–2642 (2008). 1.

#158504 - $15.00 USD (C) 2011 OSA

Received 22 Nov 2011; revised 11 Dec 2011; accepted 12 Dec 2011; published 21 Dec 2011 2 January 2012 / Vol. 20, No. S1 / OPTICS EXPRESS A94

13. S. L. Diedenhofen, G. Vecchi, R. E. Algra, A. Hartsuiker, O. L. Muskens, G. Immink, E. Bakkers, W. L. Vos, and J. G. Rivas, “Broad-band and omnidirectional antireflection coatings based on semiconductor nanorods,” Adv. Mater. (Deerfield Beach Fla.) 21(9), 973–978 (2009). 14. Y. C. Chao, C. Y. Chen, C. A. Lin, and J. H. He, “Light scattering by nanostructured anti-reflection coatings,” Energy Environ. Sci. 4(9), 3436–3441 (2011). 15. P. B. Clapham and M. C. Hutley, “Hutley, Reduction of lens reflection by moth eye principle,” Nature 244(5414), 281–282 (1973). 16. H. Sai, H. Fujii, K. Arafune, Y. Ohshita, M. Yamaguchi, Y. Kanamori, and H. Yugami, “Antireflective subwavelength structures on crystalline Si fabricated using directly formed anodic porous alumina masks,” Appl. Phys. Lett. 88(20), 201116–201116-3 (2006). 17. W. Chern, K. Hsu, I. S. Chun, B. P. Azeredo, N. Ahmed, K. H. Kim, J. M. Zuo, N. Fang, P. Ferreira, and X. L. Li, “Nonlithographic patterning and metal-assisted chemical etching for manufacturing of tunable light-emitting silicon nanowire arrays,” Nano Lett. 10(5), 1582–1588 (2010). 18. H. Sai, Y. Kanamori, K. Arafune, Y. Ohshita, and M. Yamaguchi, “Light trapping effect of submicron surface textures in crystalline Si solar cells,” Prog. Photovolt. Res. Appl. 15(5), 415–423 (2007). 19. Y. R. Lin, H. P. Wang, C. A. Lin, and J. H. He, “Surface profile-controlled close-packed Si nanorod arrays for self-cleaning antireflection coatings,” J. Appl. Phys. 106(11), 114310 (2009). 20. H. P. Wang, K. Y. Lai, Y. R. Lin, C. A. Lin, and J. H. He, “Periodic si nanopillar arrays fabricated by colloidal lithography and catalytic etching for broadband and omnidirectional elimination of Fresnel reflection,” Langmuir 26(15), 12855–12858 (2010). 21. E. C. Garnett and P. Yang, “Silicon nanowire radial p-n junction solar cells,” J. Am. Chem. Soc. 130(29), 9224– 9225 (2008). 22. W. A. Nositschka, C. Beneking, O. Voigt, and H. Kurz, “Texturisation of multicrystalline silicon wafers for solar cells by reactive ion etching through colloidal masks,” Sol. Energy Mater. Sol. Cells 76(2), 155–166 (2003). 23. Z. Fan, R. Kapadia, P. W. Leu, X. Zhang, Y. L. Chueh, K. Takei, K. Yu, A. Jamshidi, A. A. Rathore, D. J. Ruebusch, M. Wu, and A. Javey, “Ordered arrays of dual-diameter nanopillars for maximized optical absorption,” Nano Lett. 10(10), 3823–3827 (2010). 24. Y. A. Dai, H. C. Chang, K. Y. Lai, C. A. Lin, R. J. Chung, G. R. Lin, and J. H. He, “Subwavelength Si nanowire arrays for self-cleaning antireflection coatings,” J. Mater. Chem. 20(48), 10924–10930 (2010). 25. H. Masuda and K. Fukuda, “Ordered metal nanohole arrays made by a two-step replication of honeycomb structures of anodic alumina,” Science 268(5216), 1466–1468 (1995). 26. C. H. Liu, J. A. Zapien, Y. Yao, X. M. Meng, C. S. Lee, S. S. Fan, Y. Lifshitz, and S. T. Lee, “High-density, ordered ultraviolet light-emitting ZnO nanowire arrays,” Adv. Mater. (Deerfield Beach Fla.) 15(10), 838–841 (2003). 27. A. P. Li, F. Muller, A. Birner, K. Nielsch, and U. Gösele, “Hexagonal pore arrays with a 50-420 nm interpore distance formed by self-organization in anodic alumina,” J. Appl. Phys. 84(11), 6023–6026 (1998). 28. Z. P. Huang, X. X. Zhang, M. Reiche, L. F. Liu, W. Lee, T. Shimizu, S. Senz, and U. Gösele, “Extended arrays of vertically aligned sub-10 nm diameter [100] Si nanowires by metal-assisted chemical etching,” Nano Lett. 8(9), 3046–3051 (2008). 29. X. Li and P. W. Bohn, “Metal-assisted chemical etching in HF/H2O2 produces porous silicon,” Appl. Phys. Lett. 77(16), 2572–2574 (2000). 30. K. Q. Peng, J. J. Hu, Y. J. Yan, Y. Wu, H. Fang, Y. Xu, S. T. Lee, and J. Zhu, “Fabrication of single-crystalline silicon nanowires by scratching a silicon surface with catalytic metal particles,” Adv. Funct. Mater. 16(3), 387– 394 (2006). 31. K. Bhatt, S. Tan, S. Karumuri, and A. K. Kalkan, “Charge-selective Raman scattering and fluorescence quenching by “nanometal on semiconductor” substrates,” Nano Lett. 10(10), 3880–3887 (2010). 32. S. J. Wilson and M. C. Hutley, “The optical-properties of moth eye antireflection surfaces,” Opt. Acta (Lond.) 29(7), 993–1009 (1982). 33. P. K. H. Ho, D. S. Thomas, R. H. Friend, and N. Tessler, “All-polymer optoelectronic devices,” Science 285(5425), 233–236 (1999). 34. M. Erman, J. B. Theeten, P. Chambon, S. M. Kelso, and D. E. Aspnes, “Optical properties and damage analysis of GaAs single crystals partly amorphized by ion implantation,” J. Appl. Phys. 56(10), 2664–2671 (1984). 35. K. Hadobás, S. Kirsch, A. Carl, M. Acet, and E. F. Wassermann, “Reflection properties of nanostructure-arrayed silicon surfaces,” Nanotechnology 11(3), 161–164 (2000). 36. Y. C. Lee, C. F. Huang, J. Y. Chang, and M. L. Wu, “Enhanced light trapping based on guided mode resonance effect for thin-film silicon solar cells with two filling-factor gratings,” Opt. Express 16(11), 7969–7975 (2008). 37. M. Born and E. Wolf, “Principles of optics,” (Cambridge University Press, 1999), sec. 8.6.1, Eq. (8). 38. H. A. Haus, “Waves and fields in optoelectronics,” (Prentice-Hall, Englewood Cliffs, NJ, 1984) 39. L. Cao, P. Fan, A. P. Vasudev, J. S. White, Z. Yu, W. Cai, J. A. Schuller, S. Fan, and M. L. Brongersma, “Semiconductor nanowire optical antenna solar absorbers,” Nano Lett. 10(2), 439–445 (2010). 40. L. Cao, B. Nabet, and J. E. Spanier, “Enhanced Raman scattering from individual semiconductor nanocones and nanowires,” Phys. Rev. Lett. 96(15), 157402 (2006). 41. W. S. Shi, H. Y. Peng, Y. F. Zheng, N. Wang, N. G. Shang, Z. W. Pan, C. S. Lee, and S. T. Lee, “Synthesis of large areas of highly oriented, very long silicon nanowires,” Adv. Mater. (Deerfield Beach Fla.) 12(18), 1343– 1345 (2000).

#158504 - $15.00 USD (C) 2011 OSA


1. Introduction Si is a popular material for optical devices, such as solar cells (SCs) and photodetectors (PDs) owing to its monolithic integratibility with low-cost complementary metal-oxidesemiconductor technology. However, the low quantum efficiency particularly in UV region (500 nm) of the nanostructured surface is less effective in suppressing undesired Fresnel reflection at the wavelengths below 500 nm [9, 19, 22]. To overcome this problem, thick nanostructures (i.e., large deff) could be used to obtain broadband AR performance, but would increase parasitic resistance and disturb carrier collection decreasing the device efficiency [23, 24]. It is a key challenge to fabricate a well-controlled thin nanostructure to suppress broadband reflection but still facilitate effective charge-carrier transport. Therefore, to produce short but high-aspect-ratio Si NPAs with excellent AR characteristics, a small Λ of Si NPAs is required. In this study, we fabricated close-packed Si NPAs with 100 nm in Λ and 60 nm in diameter using an anodic aluminum oxide (AAO) template for surface structuring combined with metal-assisted chemical etching. The 100-nm-periodicity Si NPAs broadbandly eliminate the Fresnel reflection at the angles of incidence (AOI’s) up to 60°. The interaction between the incident light and the Si NPAs is realized through the simulation based on the finitedifference time domains (FDTD) analysis. Through rigorous coupled-wave analysis (RCWA), we demonstrated that as the Λ is small, the light absorption can be significantly improved in short wavelength regions due to the grating effect, confirming our experimental results. Enhanced Raman scattering also demonstrated the AR ability of Si NPAs. An AAO template combined with metal-assisted chemical etching might be a promising surface structuring method for efficient light harvesting for next-generation Si optical devices.

#158504 - $15.00 USD (C) 2011 OSA


2. Experimental section Annealed high-purity (99.99%) aluminum foil was electropolished in a mixture of HClO4 and C2H5OH (volume ratio = 1:5) until the root-mean-square surface roughness of a typical 10 µm × 10 µm area was ca. 1 nm. The foil was then anodized in 0.3 M oxalic acid at 1 °C at a constant voltage of 40 V for 3 min using two-step anodization process to obtain AAO substrates with nanochannel arrays of self-organized honeycomb structure [25]. After anodization, the nanochannels were pre-opened using a 6 wt% solution of H3PO4 at 36 °C to obtain AAO substrates with arrays of nanopores. Subsequently, a thick protecting layer of nail polish was coated on the top of AAO substrate for maintaining mechanically robust during the removal of aluminum and the barrier layer. Then, the underlying aluminum was removed in a mixture of CuCl2 and HCl, and the remaining barrier layer in the bottom of AAO was dissolved in H3PO4. After removing the nail polish by acetone, the AAO film as a patterned mask was transferred to the single crystalline p-type (001) Si substrate with ρ = 8-12 Ω-cm. SF6/O2 plasma was applied to etch Si substrate through the AAO pores using SAMCO RIE10NR to form small etched holes on the surface of the substrate. The AAO mask was removed in H3PO4 before a 30-nm-thick Ag layer was deposited on the patterned-Si substrate using RF sputtering at a power of 50 W and chamber pressure lower than 2 × 10−6 Torr. The metal-assisted wet etching was carried out by immersing the Ag-patterned Si substrate in the solution of HF/H2O2 for 160 seconds to obtain the hexagonal NPA structures. The reflectance measurement of the Si NPAs over the wavelength regions from 200 to 850 nm was performed by a JASCO V-670 UV-VIS-IR spectrometer with an integrating sphere. The integrating sphere collects all light reflected by the samples and measures the Rtotal. The coherent reflectance of a collimated incident light beam (Rspec) was determined by collecting the specularly reflected cone of light within an acceptance angle of 5°. The omnidirectional property of the antireflective Si NPAs was characterized by measuring the reflectance at the AOI from 5° to 80° with the fixed incident wavelength of 250 nm. The theoretical calculations based on RCWA and FDTD were employed to simulate the reflectance spectra and |E| distribution of polished Si and periodic Si NPAs, respectively. The Raman spectroscopy was obtained by a micro-Raman Jobin Yvon T64000 system equipped with a coherent VerdiV10 532 nm laser as the excitation source. The Raman signals were detected with the back illuminated UV enhanced CCD detector. 3. Results and discussion Figure 1(a) illustrates the flowchart of the experimental process. First, the AAO membrane was placed on a Si substrate, which was cleaned by standard RCA process. The selfassembled AAO membranes used as the templates can be easily formed with a large area of controlled pore diameters and Λ’s [26–28]. The fabrication of AAO membranes is described in Experimental Section. After the RIE treatment to etch the unmasked Si (through the pores of AAO membrane), the close-packed hexagonal pattern was transferred to the Si substrate. Subsequently, the AAO membrane was removed by H3PO4, and then a 30-nm-thick Ag layer was deposited by sputtering for the following chemical etching process. Finally, the Si NPAs were obtained by immersing the Ag-patterned Si substrate in the etching solution of HF/H2O2 for 160 seconds. Figures 1(b)-1(e) are SEM images of the corresponding experimental process. Figure 1(b) shows AAO membrane/Si substrate with uniform holes after the RIE treatment, indicating that the positions and diameters of the periodic holes match those of AAO membrane pores. The SEM image of the patterned substrate with the 30-nm-thick Ag layer is shown in Fig. 1(c). In the sputtering process, Ag was not only deposited on the substrate surface (Ag network), but also aggregated in the holes etched by the RIE treatment (Ag nanoparticles), as shown in the inset of Fig. 1(c). When the Ag-coated substrate was immersed

#158504 - $15.00 USD (C) 2011 OSA


Fig. 1. (a) The flowchart of experimental process for fabricating periodic Si NPAs. (b)-(e) SEM images of the corresponding experimental procedures.

in the HF/H2O2 solution, the chemical etching took place. Upon the adhesion to Si, Ag attracts the electrons from Si due to its strong electronegativity. The electron transfer from Si to Ag induces the etching effect, which involves Ag-induced local oxidation and subsequent dissolution of the oxidized Si by HF [20, 29, 30]. Such electron transfer is greatly related to the Schottky-barrier height (SBH) at the Ag/Si interface, and the SBH is found to be increased with reduced Ag/Si contact areas [31]. Since the contact areas of the Ag nanoparticles deposited in the Si holes are smaller than those of the film-like Ag network deposited on the substrate surface, the Ag nanoparticles (compared with the film-like Ag) are of high SBH with Si. The high SBH prevents the electrons from transporting from Si to Ag nanoparticles, and thus slows down the oxidation rate of the Si underneath Ag nanoparticles, giving rise to low etching rate in the holes. These different etching rates by HF/H2O2 on the substrate surface and in the holes eventually lead to the formation of Si NPAs; i.e., when the etching depth of the Ag network overtook that of the Ag nanoparticles in the holes, the Si underneath the Ag nanoparticles appears with the wire-like shape, as shown in Fig. 1(d). After 160second etching and removal of the Ag by HNO3, Si NPAs were obtained, as shown in Fig. 1(e). From the inset of Fig. 1(e), the Λ and the diameter of Si NPAs are 100 and 60 nm, respectively. The NPAs duplicate geometric features of the AAO template; i.e., the Λ and the diameter of NPAs can be controlled by AAO templates. The self-assembled AAO templates can be fabricated with controlled pore diameters ranging from 10 to 200 nm and Λ’s ranging from 50 to 420 nm by varying the electrochemical parameters, such as the voltage and the electrolyte [27]. However, the formation of NPAs is due to the different etching rates by HF/H2O2 on the substrate surface (with Ag network) and in the holes (with Ag particles). The large pore diameters of AAO membranes cause the difficulty in nanopillar formation because of little difference in the etching rate between substrate surface (with Ag network) and holes (with Ag particles). The optimized spatial filling ratio of NPAs is 0.33, as shown in Fig. 1(e). The average length of nanopillars is 344 nm, which only depends on the etching time. However, long nanopillars would increase parasitic resistance and disturb carrier collection, which decreases the device efficiency [23, 24]. Accordingly, it is the main reason that we emphasize the importance of small Λ of NPAs.

#158504 - $15.00 USD (C) 2011 OSA


Fig. 2. (a) Total reflectance (b) specular reflectance (c) diffuse reflectance and (d) diffusion order ratio of polished Si and Si NPAs over the wavelength regions of 200~850 nm.

The dependence of reflectance on the deff of nanostructured layer and the wavelength (λ) has been studied by several groups [15, 32]. Clapham et al. found that the AR effect would be pronounced as the ratio of deff/λ is comparable to or larger than 0.4 [15]. In our case, deff is given by deff = neff*d, where d is the structural thickness of the NWA layer and neff is the effective refractive index. For the wavelength region from 200 to 850 nm, neff is among 1.01~2.23 determined by the spatial filling ratio ( = 0.33) of NWAs, estimated from the inset of Fig. 1(e), using the Bruggeman effective medium approximations [33–35]. The deff/λ ratios are among 0.41~3.84. These ratios are expected to result in the broadband AR ability. In order to confirm the AR performance, the total reflectance (Rtotal) and specular reflectance (Rspec) spectra of the Si NWAs were measured as compared with polished Si over the wavelength region from 200 to 850 nm [Fig. 2(a) and 2(b)]. As shown in Fig. 2(a), the Si NWAs exhibit much lower Rtotal than that of the polished Si, and particularly at the wavelength region from 200 to 400 nm the Rtotal of NWAs is about a tenth of the value of the polished Si (from 62.9% to 7.9%). In Fig. 2(b), the Rspec of NWAs noticeably decreases from the long wavelength to the short wavelength region. The diffuse reflectance (Rdiff), defined by Rtotal-Rspec, decreases as the wavelengths increase, as shown in Fig. 2(c). The different tendencies of Rspec and Rdiff of the NWAs can be explained using Fig. 2(d), in which the ratios of Rspec/Rtotal and Rdiff/Rtotal for Si NWAs and the polished Si are plotted as the function of wavelengths. These two samples present distinct behaviors. On the polished Si, Rtotal is dominated by Rspec (average Rspec/Rtotal = 90.4%), and the ratios remain nearly unchanged at the entire wavelength range. This is because the reflection from a mirror-like Si surface is governed by the ordinary theorem of geometrical optics. In contrast, Rdiff/Rtotal on the NPAs is more than 85.2% for the wavelengths below 500 nm, and gradually decreases with wavelengths. This phenomenon manifests the fact that light scattering occurs significantly only when the wavelength is comparable with the Λ of NPAs [12]. The periodic NPAs can be regarded as a diffraction grating [36]. The light impinging the NPAs proceeds with three steps: coupling with the grated surface, diffracting to several beams with different diffraction angles into the NPAs,

#158504 - $15.00 USD (C) 2011 OSA


and re-bouncing between NPAs until being absorbed. At the wavelengths comparable with the Λ, the grating reduces the zero-order reflectance (i.e., Rspec), but the light beams are redistributed to the diffracted orders [32], leading to the high ratios of Rdiff/Rtotal. The high diffracted orders caused by surface grating can increase optical path lengths [12, 18]. These multiple scattering light paths inside the NPAs are folded up, which effectively suppresses the reflection and enhances the absorption [12]. Rspec/Rtotal of NPAs increases with the incident wavelengths, suggesting that as the incident wavelength increases, the scattering on the periodic structure diminishes [35], and Rspec dominates Rtotal, being close to the situation on polished Si, as shown in Fig. 2(d). Because the wavelengths are much longer than Λ ( = 100 nm) of NPAs, the light interacts with the whole Si NPA layer rather than each nanopillars. It is found that Rspec of the NPAs gradually increases with the wavelengths, indicating that the periodic NPAs become less resolved by the light with long incident wavelengths and therefore the grating effect is suppressed. Overall, the decrease in Rtotal at the long wavelength regions by NPA surfaces, shown in Fig. 2(a), should be explained by the effective medium theory (EMT) [35]. In EMT, light strikes on the subwavelength structures as if it encounters an AR thin layer with an neff between refractive indices of air and Si, avoiding the abrupt transition of refractive index from air (n = 1) to Si (n = 3.7), and therefore effectively suppresses the reflection.

Fig. 3. The time-averaged, normalized TE electric field distribution (|E|) of polished Si and Si NWAs simulated by FDTD analysis with the wavelength of 250 nm.

To confirm the experimental results, we simulated the optical behavior of light within the near-field regime propagating on the polished and periodic NPA surfaces with Λ = 100 nm using FDTD analysis. A plane wave was launched from z = 1 µm to the Si surface with/without the NPA structures. The grid sizes are ∆x × ∆y × ∆z = 2 × 1 × 5 nm3 in space domain, and the time step for every calculation is 0.0029 fs. In the figures, the distribution of time-averaged electric field (|E|) in polished Si [Fig. 3(a)] and Si NPAs [Fig. 3(b)] is calculated using λ = 250 nm. The AR abilities of the two structures can be compared by the |E| distribution above z = 1 µm, which indicates the reflection from the structure surface without the interference by the incident waves. The improved AR ability of the NPAs is clearly demonstrated by the reduced intensities at z = 1~1.5 µm in Fig. 3(b), which are significantly lower than those in the same region in Fig. 3(a). The result agrees with that presented in Fig. 2(a), and confirms that the NPAs with Λ = 100 nm efficiently eliminate the reflection even in UV region.

#158504 - $15.00 USD (C) 2011 OSA


Fig. 4. Optical properties of Si NWAs with 100 and 500 nm in Λ simulated by RCWA analysis with TE-polarized waves.

It is noticed that our NPAs achieve a significantly low reflectance at the short wavelengths (i.e., UV region) as compared with the results by Lin et al. [9], whose neff profile is similar to that in the present study due to almost identical geometric features (height = 360 nm, filling ratio = 0.35) of NPAs except the Λ ( = 100 and 500 nm). In order to investigate this discrepancy, we employed RCWA to simulate the reflectance spectra with two types of Si NPA structures, i.e., Λ = 100 and 500 nm, to gain the insight of Λ-dependent optical behavior. In the simulation, the reflectance is calculated with a fixed height = 350 ± 10 nm and filling ratio = 0.33, as shown in Fig. 4(a), at the wavelengths from 200 to 850 nm. Figure 4(b) presents simulated Rtotal of the polished Si and the Si NPAs with Λ = 100 and 500 nm. The polished Si exhibits Rtotal>30% while Rtotal is effectively decreased with Si NPA layer over broadband regions, consistent with experimental results [Fig. 2(a)]. Moreover, Si NPAs with Λ = 100 nm reduce Rtotal more effectively than those with Λ = 500 nm in short wavelength regions, exhibiting a superior AR performance. Li et al. shows that for the fixed length but varying the Λ’s of Si nanowire arrays, the lowest reflectance over the whole photon energy spanning 1−4 eV can be reached at the nanowire array with the Λ of 100 nm, consistent with our simulation results [6]. The impact of Λ on the AR properties of NPAs can be further differentiated from Fig. 4(c) and 4(d); i.e., Rspec of NPAs with Λ = 100 nm is much lower than that with Λ = 500 nm for the wavelengths below 400 nm. Figure 4(e) displays the ratios of #158504 - $15.00 USD (C) 2011 OSA


Rspec/Rtotal and Rdiff/Rtotal for Si NPAs with Λ = 100 nm whose tendencies agree well with the experimental results [Fig. 2(d)]. On the other hand, the diffraction order ratios of Si NPAs with Λ = 500 nm [Fig. 4(f)] present distinct tendencies from those with Λ = 100 nm. Because the filling ratio is fixed, i.e., the neff’s of two structures are the same, the discrepancy between two Λ’s structures can be ascribed to the grating effect when the NPAs are regarded as a diffraction grating. The diffraction behavior of NPAs can be described using the grating equation [37, 38],

nt sin θ m − ni sin θ i =

mλ Λ

(1)

where nt is the refractive index of the transmitting medium, ni is the refractive index of the incident medium, θi andθm are respectively AOI and the angle of the mth order diffraction, and λ is the incident wavelength. When the condition of the incident light satisfies the grating equation, the resonance wave couples with the NPAs and diffracts to several orders travelling in different directions. The directions of these beams depend on the Λ of the NPAs and the λ. From Fig. 4(e) and 4(f), for the wavelengths below 400 nm, Rdiff/Rtotal ratios of Si NPAs with Λ = 100 nm maintain high values, but the Si NPAs with Λ = 500 nm show that the Rdiff/Rtotal ratios are lower than Rspec/Rtotal ratios and gradually increase with wavelengths. It means that the effect of light diffraction by Si NPAs with Λ = 100 nm is more pronounced than that with Λ = 500 nm at UV region, and thus Rspec of Si NPAs with Λ = 100 nm is much lower than that with Λ = 500 nm at this region. From Eq. (1), for a constant λ, the small Λ leads to the diffracted beams with large θm for the same diffraction order (m) [38]. The beams transmitting with the larger θm result in the elongation of optical paths and enhance internal bounces within the NPAs, increasing the probability of absorption. On the other hand, at UV wavelength region, the improvement in the light trapping can be realized by reducing the Λ of NPAs (increasing the density of NPAs), which increases the number of reflection between adjacent nanopillars. For the NPAs with large Λ, the incident light virtually strikes on the bottom and reflects to the air, so the light cannot be effectively trapped by NPAs, causing the reflectance to be increased at UV region.

Fig. 5. Specular reflectance as a function of AOI for unpolarized light with the wavelength of 250 nm.

A desirable AR coating should give consideration to angle-dependent effects to suppress Fresnel reflection over a wide range of AOI’s, which is so-called omnidirectionality [39]. In order to investigate the omnidirectional characteristics of the NPAs in the UV region, where the reflectance can be greatly reduced by our structure, the Rspec was measured with the AOI’s ranging from 5° to 80° with the wavelength fixed at 250 nm, as shown in Fig. 5. The reflectance on the NPAs remains below 1.2% for the AOI up to 60°, exhibiting significantly improved omnidirectionality in comparison with the polished surface. The reflectance on the two samples gradually increases after AOI reaches 60°. Intuitively, when the light reaches the

#158504 - $15.00 USD (C) 2011 OSA


NPA layer’s surface at a large AOI, the portion of the light entering the NPA layer is decreased, suggesting that the probability of trapping light within the NPA layer is reduced, and therefore the reflectance is increased.

Fig. 6. Raman spectra of the Si NPAs and the polished Si.

Raman scattering describes the inelastic scattering of lattice vibration, and has been widely used as an analytic spectroscopic technique [40]. Figure 6 presents the Raman spectra of the polished Si and the Si NPAs. Two samples exhibit a peak at 520 cm−1, which caused by the first-order optical phonon mode of single-crystal Si [41]. No frequency shift is observed with the Si characteristic Raman peak, indicating that the Si NPAs were not extensively damaged by the chemical etching process. The scattering intensity on Si NPAs is enhanced by a factor of 8 as compared with the case on polished Si. This can be ascribed to an AR effect. The Si NPAs lead to the major portion of incident laser light entering the structure, and therefore increase the absorption of light. In addition, Raman backscatter traveling toward the surfaces will also encounter the NPA layer, suggesting an additional enhancement factor for the light extraction due to a mediate neff of NPA layer. Therefore, enhanced insertion and extraction of light lead to a substantial Raman signal improvement. 4. Conclusion In summary, periodic Si NPAs with 100 nm in Λ were fabricated by an AAO templating method combined with metal-assisted chemical etching. Their broadband AR performance eliminates the Fresnel reflection at the AOI up to 60° especially at the wavelengths below 400 nm, indicating the importance of small Λ and stepped refractive index of NPA layer, which was confirmed by RCWA and FDTD simulations. Substantial Raman signal improvement also demonstrated the AR ability of Si NPAs. The nanofabrication for broadband omnidirectional light-harvesting demonstrated here will greatly benefit the design of optical devices. Acknowledgment The research was supported by the National Science Council Grant No. NSC 99-2120-M-007012, NSC 99-2112-M-002-024-MY3 and NSC 99-2622-E-002-019-CC3.

#158504 - $15.00 USD (C) 2011 OSA
