Open Access
Add to BibSonomyAbstract
Silicon antireflection is realized with vertical-aligned SiNWs by using improved metal-induced etching technique. The spectral responses of the transmission, reflection, and absorption characteristics for the SiNWs of different lengths are investigated. In order to realize short SiNWs to provide sufficiently low reflection, a post chemical etching process is developed to make the nanowires have a larger length fluctuation and/or tapered structure. The use of short SiNWs can allow a faster process time and avoid the sub-bandgap absorption that frequently occurs in long nanowires. Short SiNWs can also provide more compatible material structure and fabrication procedures than long ones can for applying to make optoelectronic devices. Taking the applications to solar cells as examples, the SiNWs fabricated by the proposed technique can provide 92% of solar weighted absorption with about 720 nm long wires because of the resultant effective graded index and enhanced multiple optical scattering from the random SiNW lengths and tapered wires after KOH etching.
©2011 Optical Society of America
1. Introduction
Surface antireflection techniques are important for improving the performance of many optical and optoelectronic devices such as solar cells, photodetectors, displays and various sensing devices [1–5]. Although the use of deep surface texturing with spikes and pits having sizes of several micrometers on Si can reduce the surface reflection over a broad spectral range [6], its antireflective effect has strong angle dependence. Fine surface structures, comprising features on the nanometer scale, can provide excellent antireflective performance [2–5]. However, it is currently difficult to achieve large wafer-scale fabrication with low cost and good uniformity. Recently, we have realized tall and tapered two-dimensional silicon photonic crystals over a large sample area by using low cost interference lithography and novel single-step deep reactive ion etching [2]. This material is able to achieve angular-independent antireflection with a reflectance Rs of 1-3% on silicon with good controllability and repeatability. Silicon nanowires (SiNWs) can also serve as a good antireflective structure for silicon [7–11]. Disordered SiNW arrays can be realized by immersing clean silicon wafers in metal-contained aqueous solution, such as HF/AgNO3 [7,8], or by immersing metal-distributed silicon wafers in oxidizing hydrofluoric acid solution, such as HF/H2O2 [9,10]. Silver or gold particles are deposited mostly by electroless plating process as the catalysts in this etching process. Simple, low-cost and mass producible fabrication is the obvious advantages of this method. However, the agglomeration of fabricated SiNWs at the top usually takes place and results in large voids and hence degrades its optical properties. This is partially due to the random electroless deposition of metal particles which have random sizes and locations. Therefore, tall SiNWs are necessary for suppressing the silicon surface reflection [3]. Nevertheless, it is very difficult to obtain low contact resistance on the top of tall SiNWs, which leads to lower conversion efficiency as integrating this material into practical solar cells [5]. Higher sub-bandgap absorption also occurs in taller SiNWs [12].
In this work, we present a simple fabrication process for synthesizing SiNW arrays by depositing thin silver film on Si surface prior to wet chemical etching in oxidizing HF solution, and then followed by another chemical etching to further improve surface antireflection for short and aligned SiNWs. With the proposed scheme, vertical-aligned SiNWs with controllable heights are realized over large area. By characterizing the optical properties, this antireflective material is expected to be feasible for improving the performance of wafer-scale photovoltaic devices and other optoelectronic devices.
2. Sample preparation
Figure 1(a) shows the process comparison between the conventional metal-induced wet etching scheme and our scheme. The main idea of our proposed process is to replace randomly distributed silver particles in conventional scheme [7–10,12] with network-like percolated silver film [13], trying to avoid the agglomeration issue which usually happens in the conventional electroless deposition scheme. Besides, unlike the conventional scheme requiring critical control of the chemical solution, our approach is relatively simple and advantageous for realizing uniform SiNWs over large area because it only requires deposition of a thin silver film. The silver nano-clusters act as catalysis in the metal-induced etching process [13]; and we optimize the entire etching process in this work for obtaining vertically-aligned and uniform SiNW arrays. The proposed process comprises three steps: 1) a thin silver film is deposited onto the cleaned Si surface by electron-beam evaporation; 2) the metal-treated Si wafers are immersed in oxidizing hydrofluoric acid solution for anisotropic silicon wet chemical etching; and 3) the as-prepared samples are treated in nitric acid to completely remove the residual silver particles. The deposition of the thin silver film can also be done by sputtering over a large wafer for mass production.
Fig. 1 (a) Process comparison between conventional metal-induced wet etching scheme and the proposed scheme. (b) Cross-sectional SEM view of an etched Si sample at a distance of about 90 μm from the surface, showing that thin silver film can act as uniformly-distributed silver nano-particles which locate at the bottom of the channels after etching.
The synthesis of aligned SiNW arrays is carried out on front-side polished, boron-doped (100) crystalline silicon wafers with a resistivity of 1-10 Ω-cm. The control of silver thickness is the key to the success of SiNW formation since the morphology of thin silver film changes with the thickness [14]. In this work, the process condition is set as the concentrations of HF, H2O2 and H2O being 6.72 M, 0.79 M and 48 M, respectively, initial silver thickness of 20 nm, and etchant temperature of 25 degree Celsius. The resultant SiNW profiles are characterized by using a field-emission scanning electron microscope (SEM). Figure 1(b) verifies that thin silver film with right film thickness can act as uniformly-distributed silver nano-particles which locate at the bottom of the channels after wet etching.
The SiNW formation rate is very stable (~1.2 μm/min at room temperature), so the etching depth can be easily controlled. Figure 2 shows the top and cross-sectional SEM views of 120-nm, 800-nm and 5250-nm tall SiNWs realized with different etching time. All SiNWs are vertically aligned with high uniformity and good reproducibility. The reduction in agglomeration issue by the proposed process is due to the network-like silver nano-clusters which restrict each silver nanoparticle to move freely, thus leading to collective sinking of silver nano-particles in vertically downward direction. The realized 120-nm tall SiNWs have an initial volume ratio of around 0.276 after 5 seconds of etching. The SiNW volume ratio initially increases with the etching time and reaches its peak at around 0.36 when the SiNWs are 800 nm tall with a density of 3.6 × 107 wires per square millimeter. Further increase of the etching time will decrease the SiNW volume ratio. The volume ratio of 5250-nm tall SiNWs is only about 0.204. From these results, we can conclude that the proposed fabrication process is capable of realizing uniform and vertical-aligned SiNWs with lengths down to 120 nm with good controllability and reproducibility. We demonstrate the fabrication of vertical-aligned SiNWs with good uniformity over 4-inch wafer area, as shown in Fig. 3 . The SiNW sample appears totally black with no mirror image, as compared to grey and polished surface of original silicon substrate with clear mirror image, verifying low surface reflection over whole 4-inch sample area. It is worth mentioning that it is relatively difficult to yield uniform SiNW with a length below 1 μm by means of conventional electroless deposition and wet etching process [12].
Fig. 2 Top and cross-sectional SEM views of (a) 120-nm tall, (b) 800-nm tall and (c) 5250-nm tall SiNWs realized with different etching time.
The use of short SiNWs (less than 1μm in length) as the antireflective material suffers from insufficient light trapping which is the roadblock to further reduce the surface reflection of silicon, especially for longer wavelength region. Previous work has focused on sharpening 10-μm tall SiNWs to obtain lower surface reflection [15]. They attributed the enhanced antireflection of sharpened SiNWs to the separation of agglomerated top-ends of the SiNWs, multiple optical scattering and the graded-index effects. It is also reported that aligned SiNW arrays with structural randomness show absorption enhancement, comparing to the aligned and ordered one [16]. In this work, a post chemical etching process with a diluted KOH solution is developed to produce larger length fluctuation and/or tapered structure for enhancing the antireflection of short SiNWs. Figure 4 show the SEM images of resultant SiNWs after post etching with 10% KOH solution. Schematic illustrations of the resultant SiNW profiles and their corresponding effective refractive index profiles across the air-to-wire axis are also shown in Fig. 4. Since KOH is an anisotropic etchant for silicon, the post-etching process results in random heights of SiNWs due to the originally random diameters after the silver-induced etching. It can be observed from Fig. 4(b) and 4(c) that SiNWs with gradually-changed effective index are realized after later (60 seconds) KOH etching. However, longer KOH etching time would lead to greatly decreased height and density of SiNWs, thus resulting in much mismatched refractive index on SiNW-Si interface, as shown in Fig. 4(c). Therefore, the enhanced surface antireflection in short and sharpened SiNWs is not only due to the graded-index but also because of the random SiNW length after KOH etching which enhances optical scattering inside the material.
Fig. 4 SEM pictures of (a) original SiNWs, (b) SiNWs after initial KOH etching and (c) SiNWs after later KOH etching. Schematic illustrations of the resultant SiNW profiles and their corresponding effective refractive index profiles across the air-to-wire axis are also shown.
To further verify that the height variation in SiNWs benefits the reduction of surface reflection, the total optical reflection of SiNWs is calculated by solving the Maxwell equation with rigorous coupled-wave analysis (RCWA) method. This method can account for both the dispersion and absorption of materials and is originally designed for simulating periodic structures using the unit cell concept. In this case, we introduce position, diameter and height randomness into 13 by 13 two-dimensional periodic rods which originally have a lattice constant of 150 nm and a rod radius of 37.5 nm according to the averaged distance between wires and the averaged radius of our real SiNWs. The whole structure is then regarded as a unit cell in RCWA simulation. Figure 5(a) shows the simulated total reflection spectra of a bare silicon and the SiNWs with and without random heights. In this case, the wires are placed randomly within the unit cell with a diameter fluctuation between 66 and 115 nm for the SiNW structures. The simulation indicates that the use of SiNWs greatly reduces the surface reflection of silicon. The introduction of random heights in SiNWs can further improve its antireflection performance, as indicated in Fig. 5(a). The contribution of SiNW height variation to the reduction of total surface reflection of SiNWs is shown in Fig. 5(b). About +/− 100 nm height variation in SiNWs is able to achieve the lowest averaged reflectance.
Fig. 5 (a) Simulated total optical reflection spectra of bare silicon and SiNWs with and without random heights and (b) the contribution of SiNW height variation to the reduction of total optical reflection of SiNWs.
3. Optical characterization
The optical reflection and transmission of the fabricated SiNW samples are measured over the spectral regions above and below the bandgap of silicon with a spectrophotometer (JASCO UV-Vis/NIR spectrophotometer V-670) equipped with an integrating sphere. The optical specular reflection of SiNW samples is measured with a thin-film measurement system (Filmetrics F20-UVX). Figure 6(a) shows that the optical specular reflectance of 0.59-μm and 1.06-μm tall SiNWs can be below 5% and 2% in the UV-visible region, respectively, as compared to above 30% specular reflectance of a bare silicon. The use of longer SiNWs can result in a lower optical specular reflectance. Since our approach can realize uniform SiNWs with low surface roughness, the short SiNWs can be regarded as a uniform thin film with an effective index, leading to the resonance-like reflection spectrum shown in Fig. 6(a). The effective index of the fabricated SiNWs can thus be derived from the spectrum and is depicted in Fig. 6(b). The fitted effective index of SiNWs increases linearly with the wavelengths and its average value is around 2.2. This value agrees well with the calculation from the volume ratio of realized SiNWs.
Fig. 6 (a) Measured optical specular reflection spectra of bare silicon and SiNWs with 0.59-μm and 1.06-μm in height, respectively. (b) Fitted effective index of 0.59-μm tall SiNWs with the wavelengths.
Figure 7(a) compares the optical reflection between bare silicon and SiNWs of different heights. In the short-wavelength region, the reflection spectrum of original bare silicon shows the high optical reflectance of the front polished surface. The higher optical reflectance below the fundamental absorption bandgap Egap of silicon is due to the additional diffuse reflection from the back silicon surface. The formation of 0.12-μm, 0.59-μm and 6.98-μm tall SiNWs atop the silicon surface can reduce the total surface reflectance from 33~75% down to around 8~40%, 3~7% and 1~4% in the above-bandgap spectral region, respectively. However, in the below-bandgap region where Si is transparent, the decreased reflection of 6.98-μm tall SiNW surface does not lead to an expected increase but decrease in the transmitted light, as indicated in Fig. 7(b). This indicates that there is a strong residual absorption below the Si bandgap for tall SiNWs. We attribute this residual absorption to strong infrared light trapping coupled with the presence of surface states on the nanowires that absorb the below-bandgap light. This sub-bandgap absorption also occurred in micro-structured silicon [17] and can be slightly decreased with post annealing process [12,17]. On the contrary, more transmitted light is observed from the surface of 0.12-μm and 0.59-μm tall SiNWs, indicating that the sub-bandgap absorption in this case is very low such that the reduced reflection contributes directly to the increased transmission.
Fig. 7 Measured total optical (a) reflection and (b) transmission spectra of bare silicon and SiNWs with different heights. (c) Comparison of total and specular-only reflection spectra of 0.59-μm and 6.98-μm tall SiNWs.
Figure 7(c) compares the total and specular-only reflection spectra between short (0.59-μm) and tall (6.98-μm) SiNWs. It can be found that both specular and diffuse reflection play an important role in total surface reflection of short (0.59-μm) SiNWs while diffuse reflection determines the total surface reflectance in tall (6.98-μm) SiNWs. Repeat experiments show that there exists a reflection peak at around 450 nm of wavelength in the total reflection spectrum of tall (6.98-μm) SiNW. Figure 7(c) clearly shows that this reflection peak comes from the diffuse reflection inside the material. We attribute this reflection peak to the optical crossover feature between enhanced optical scattering inside the material (depends on wavelength as λ−3 for rods) and increased optical absorption coefficient of silicon (jumps at around 400 nm of wavelength for crystalline silicon). This reflection peak is not observed in the total optical reflection spectrum of shorter SiNWs.
Figure 8(a) shows the 3D contour plot of the measured total optical reflection spectra of SiNWs after different etching time. Since the proposed fabrication method enables high SiNW formation rate, low surface reflection in above-bandgap spectral region can be obtained with a short process time, comparing to relatively lower reduction rate of the total surface reflection of SiNWs realized with conventional electroless deposition and wet etching scheme [3]. It can be observed from Fig. 8(b) that the formation of SiNWs to achieve the lowest total surface reflectance at 530 nm of wavelength (peak of AM1.5 solar spectrum) only requires about 30 seconds. Longer SiNW etching time would slightly increase the reflectance due to the presence of reflection peak at 450 nm of wavelength in the total reflection spectrum of tall SiNWs, as indicated in Fig. 7(c). Total optical absorption of SiNWs at 1300 nm of wavelength (Si transparent region) under different etching time is also shown in Fig. 8(b) to determine the sub-bandgap absorption for SiNWs with different heights. The total optical absorption A is calculated from the corresponding reflection R and transmission T according to R + T + A = 1. We observe that the sub-bandgap absorption increases with the SiNW heights from 2% (bare silicon), 10% (0.12-μm tall SiNWs) to 25% (6.98-μm tall SiNWs). Therefore, it is better to use shorter SiNWs as an antireflection layer to avoid the presence of a large number of surface states which lead to the higher sub-bandgap absorption.
Fig. 8 (a) 3D contour plot of measured total optical reflection spectra of SiNWs after different etching time. (b) Total optical reflection of SiNWs at 530 nm of wavelength (peak wavelength of AM1.5 solar spectrum), total optical absorption of SiNWs at 1300 nm of wavelength (Si transparent region) and etched SiNW height under different etching time.
Therefore, short SiNWs (less than 1μm in length) are favored over tall SiNWs as the antireflection layer due to its sufficiently low surface reflection, faster process time, lower sub-bandgap absorption, and more compatible structure to photovoltaic devices. The antireflection property of short SiNWs can be further enhanced by using an additional etching process. In Section 2, a post KOH etching process is developed to enhance multiple optical scattering inside the SiNW materials by producing larger length fluctuation and/or tapered structure, as shown in Fig. 4. Figure 9 shows the comparison of total surface reflection and transmission between original and KOH-etched SiNWs. The surface reflectance of KOH-etched SiNWs is lower than that of original short SiNWs, as shown in Fig. 9(a), especially for longer wavelengths. It indicates that KOH-etched SiNW surface indeed provide longer optical path inside the material, thus allowing further suppression of surface reflection in the wavelength regions where silicon is not strongly absorptive. Since reduced reflection is directly contributed to the increased transmission, as shown in Fig. 9(b), no additional residual sub-bandgap absorption is produced after post KOH etching.
Fig. 9 Measured total optical (a) reflection and (b) transmission spectra of bare silicon, original SiNWs and KOH-etched SiNWs.
4. Implications for photovoltaic applications
In order to understand how SiNWs can contribute to the conversion efficiency of a solar cell from an optical perspective, we define the solar weighted absorption which indicates the portion of sun light that SiNWs can absorb by overlap integrating the total optical absorption spectra of SiNWs with the AM1.5 solar spectrum.
Figure 10(a) shows the total optical absorption spectra of bare silicon and SiNWs with different heights covering only above-bandgap spectral region. These curves are calculated from the measured reflection and transmission spectra shown in Fig. 7(a) and 7(b). With the help of vertical-aligned SiNWs as the intermediate layer between air and silicon substrate, the amount of absorbed light inside the material is significantly increased, especially for shorter wavelengths. Above 90% and 95% total optical absorption in the wavelength range between 280 nm and 1000 nm are observed for the case of 0.59-μm and 6.98-μm tall SiNWs, respectively. Figure 10(b) shows that the amount of absorbed sun light is significantly increased from 60% (bare silicon) to 95% (tall SiNWs). It is worth noticing that around 90% solar weighted absorption can be obtained from SiNWs with a height of only 800 nm. To further increase this value up to maximum 95% requires at least 2~3 times taller SiNWs which cause higher sub-bandgap absorption and require much complicated procedures for fabricating solar cells.
Fig. 10 (a) Total optical absorption spectra of bare silicon and SiNWs with different etched heights for mapping AM1.5 solar spectrum. (b) Solar weighted absorption of SiNWs with different etched heights.
The weighted absorption can be enhanced by the post-etching process. Figure 11 shows the resultant height and the solar weighted absorption of resultant SiNWs after KOH etching of the originally 800-nm tall SiNWs. There exists an optimal etching condition for obtaining the largest enhancement in the optical absorption. The antireflection and light trapping can be enhanced by the increasing random and tapered structures from the KOH etching up to a point that the antireflection is worsen by the decrease in the height and density of SiNWs from over-etching. In our experimental case, the best KOH etching time is between 20 and 40 seconds. The solar weighted absorption can be increased from the original 90% to 92% by 40 seconds of KOH etching. According to Fig. 10(b), to achieve 92% solar weighted absorption requires about two times taller SiNWs without performing post chemical etching process.
Fig. 11 SiNW height and solar weighted absorption under different KOH etching time. The volume ratio is indicated with SiNW height curve.
5. Conclusion
Silicon antireflection is realized with vertical-aligned SiNWs by using an improved metal-induced chemical etching technique. The new process replaces the randomly distributed silver particles in the prior approaches with a network-like percolated silver film as the catalyst for SiNW formation, thus avoiding the agglomeration issue which usually happens in the conventional electroless deposition and wet etching scheme. The proposed scheme is capable of realizing vertical-aligned SiNWs with lengths down to 120 nm and up to several tens of micrometers with good controllability and reproducibility. We demonstrate black nonreflecting silicon surface over 4-inch sample area by realizing uniform SiNWs over silicon substrate, verifying the feasibility of this technique for practical wafer-scale applications. Around 3~7% and 1~4% of total surface reflectance are observed over entire Si absorbing region for the case of 0.59-μm and 6.98-μm tall SiNWs, respectively.
The use of short SiNWs as the antireflective material is favored over tall ones due to its faster process time and lower sub-bandgap absorption. Short SiNWs can also provide more compatible material structure and fabrication procedures than long ones can for applying to make optoelectronic devices. In order to further reduce the surface reflectance of short SiNWs, a post chemical etching process with a diluted KOH solution is developed to make the nanowires have a larger length fluctuation and/or tapered structure. Taking the applications to solar cells as examples, the 720-nm long SiNWs fabricated with the proposed two-step etching technique can provide 92% of solar weighted absorption which is equivalent to the performance of at least two times longer SiNWs without performing post chemical etching process. We attribute this enhanced antireflection of SiNWs to the resultant effective graded index and enhanced multiple optical scattering from the random SiNW lengths and tapered wires after KOH etching.
Acknowledgment
This work is supported in part by the National Science Council, Taiwan, under grant NSC97-2221-E-011-077-MY3 and by the Ministry of Education, Taiwan, under the Top University Program.
References and links
1. Z. Fan, D. J. Ruebusch, A. A. Rathore, R. Kapadia, O. Ergen, P. W. Leu, and A. Javey, “Challenges and prospects of nanopillar-based solar cells,” Nano Res. 2(11), 829–843 (2009). [CrossRef]
2. Y.-J. Hung, S.-L. Lee, and L. A. Coldren, “Deep and tapered silicon photonic crystals for achieving anti-reflection and enhanced absorption,” Opt. Express 18(7), 6841–6852 (2010). [CrossRef] [PubMed]
3. S. K. Srivastava, D. Kumar, P. K. Singh, M. Kar, V. Kumar, and M. Husain, “Excellent antireflection properties of vertical silicon nanowire arrays,” Sol. Energy Mater. Sol. Cells 94(9), 1506–1511 (2010). [CrossRef]
4. C. Chen, R. Jia, H. Yue, H. Li, X. Liu, D. Wu, W. Ding, T. Ye, S. Kasai, H. Tamotsu, J. Chu, and S. Wang, “Silicon nanowire-array-textured solar cells for photovoltaic application,” J. Appl. Phys. 108(9), 094318 (2010). [CrossRef]
5. D. Kumar, S. K. Srivastava, P. K. Singh, M. Husain, and V. Kumar, “Fabrication of silicon nanowire arrays based solar cell with improved performance,” Sol. Energy Mater. Sol. Cells 95(1), 215–218 (2011). [CrossRef]
6. L. A. Dobrzanski and A. Drygala, “Surface texturing of multicrystalline silicon solar cells,” J. Achieve. Mater. Manuf. Eng. 31, 77–82 (2008).
7. T. Qiu, X. L. Wu, G. G. Siu, and P. K. Chu, “Intergrowth mechanism of silicon nanowires and silver dendrites,” J. Electron. Mater. 35(10), 1879–1884 (2006). [CrossRef]
8. D. Kumar, S. K. Srivastava, P. K. Singh, K. N. Sood, V. N. Singh, N. Dilawar, and M. Husain, “Room temperature growth of wafer-scale silicon nanowire arrays and their Raman characteristics,” J. Nanopart. Res. 12(6), 2267–2276 (2010). [CrossRef]
9. K. Peng, A. Lu, R. Zhang, and S.-T. Lee, “Motility of metal nanoparticles in silicon and induced anisotropic silicon etching,” Adv. Funct. Mater. 18(19), 3026–3035 (2008). [CrossRef]
10. C. Chartier, S. Bastide, and C. Levy-Clement, “Metal-assisted chemical etching of silicon in HF-H2O2,” Electrochim. Acta 53(17), 5509–5516 (2008). [CrossRef]
11. Y.-J. Hung, K.-C. Wu, S.-L. Lee, and Y.-T. Pan, “Realization and characterization of aligned silicon nanowire array with thin silver film,” IEEE Photon. J. (to be published). [CrossRef]
12. L. Tsakalakos, J. Balch, J. Fronheiser, M.-Y. Shih, S. F. LeBoeuf, M. Pietrzykowski, P. J. Codella, B. A. Korevaar, O. Sulima, J. Rand, A. Davuluru, and U. Rapol, “Strong broadband optical absorption in silicon nanowire films,” J. Nanophotonics 1(1), 013552 (2007). [CrossRef]
13. H. Fang, Y. Wu, J. Zhao, and J. Zhu, “Silver catalysis in the fabrication of silicon nanowire arrays,” Nanotechnology 17(15), 3768–3774 (2006). [CrossRef]
14. R. Koch, “The intrinsic stress of polycrystalline and epitaxial thin metal films,” J. Phys. Condens. Matter 6(45), 9519–9550 (1994). [CrossRef]
15. J.-Y. Jung, Z. Guo, S.-W. Jee, H.-D. Um, K.-T. Park, and J.-H. Lee, “A strong antireflective solar cell prepared by tapering silicon nanowires,” Opt. Express 18(Suppl 3), A286–A292 (2010). [CrossRef] [PubMed]
16. H. Bao and X. Ruan, “Optical absorption enhancement in disordered vertical silicon nanowire arrays for photovoltaic applications,” Opt. Lett. 35(20), 3378–3380 (2010). [CrossRef] [PubMed]
17. C. Wu, C. H. Crouch, L. Zhao, J. E. Carey, R. Younkin, J. A. Levinson, E. Mazur, R. M. Farrell, P. Gothoskar, and A. Karger, “Near-unity below-band-gap absorption by microstructured silicon,” Appl. Phys. Lett. 78(13), 1850–1853 (2001). [CrossRef]
