3748
OPTICS LETTERS / Vol. 39, No. 13 / July 1, 2014
Etching of nanostructures on soda-lime glass Elmer Wang and Yang Zhao* Department of Electrical & Computer Engineering, Wayne State University, Detroit, Michigan 48202, USA *Corresponding author: [email protected] Received May 1, 2014; revised May 25, 2014; accepted May 26, 2014; posted May 27, 2014 (Doc. ID 211261); published June 18, 2014 Nanostructures were created on the surface of optical glass using nanosphere lithography. The substrates were etched with vapor-phase hydrofluoric (HF) acid. The etching rate was studied and compared with existing results of wet and dry HF etching. An empirical etching rate formula is found for etching depth up to 300 nm. The subsequent artificial material layer demonstrated enhanced transmittance in optical wavelengths. © 2014 Optical Society of America OCIS codes: (110.4235) Nanolithography; (220.4241) Nanostructure fabrication; (310.6628) Subwavelength structures, nanostructures; (160.2750) Glass and other amorphous materials. http://dx.doi.org/10.1364/OL.39.003748
Nanostructures fabricated on the surface of an optical material have interesting properties for many unique and useful applications. These sophisticated nanostructures can potentially guide, refract, reflect, absorb, or maneuver incident light. They will serve as a technology platform for future generations of functional surfaces, including light emission from nanostructured LEDs, diffusion filters for video screens, efficient absorption of light for thin film technology and solar cells, and super-hydrophobic surfaces for self-cleaning. Nanostructures for optical applications typically have feature sizes around 100–300 nm for visible and near infrared wavelengths [1,2]. Direct fabrication of nanostructures on glass substrate is of particular interest because glass is one of the most commonly used optical materials. In addition, glass surfaces can have controllable hydrophilic or hydrophobic properties for applications including microfluidic devices and chemical sensors. [3,4]. Since most glass is amorphous and chemically stable, it is a challenging task to find effective lithography techniques and etchants in order to create well-patterned nanoscale features. Current nanoscale fabrication methods include deep UV photolithography, electron beam lithography, sol-gel based nanoimprinting coating, and direct growth of nanorods using electro-chemical processes. Each of these methods is limited by various factors that hinder fabrication of nanostructures, such as time requirements and/or capital investment [1]. There are attempts to etch glass surfaces using standard cleanroom techniques. Most of these methods are for microstructures on pure (and fused silica) silicon dioxide substrates etched by using CF4 , SF6 , and HF (liquid and gas). [3–13]. This research focuses on the etching of nanostructures on glass substrate by using effective and low-cost methods. Specifically, nanosphere lithography (NSL) with vapor-HF as an etchant is applied to nanostructure fabrication on soda-lime glass. Soda-lime glass and crown glass are two very common types of glass available commercially. Soda-lime glass is typically constituted of 73% SiO2 , 14% Na2 O, 9% CaO, 4% MgO, 0.15% Al2 O3 , 0.03% K2 O, 0.02% TiO2 , and 0.1% Fe2 O3 . All these impurities affect how the glass is etched with HF. We report the etching process, etched nanostructure patterns on glass surface, and the etching rate. 0146-9592/14/133748-04$15.00/0
NSL is a cost-effective technology that uses nanospheres to create a mask on a substrate [14]. Figure 1 illustrates the steps used in NSL. Nanoparticles of uniform size are self-assembled into a monolayer of spheres to be used as a mask [Figs. 1(a) and 1(b)]. This allows for the fabrication of arrays with controlled spacing based on the nanosphere diameters and packing efficiency. The holes between the particles create the space for etchants or other patterning technologies to reach the substrate. After the mask is made, different methods can be used to etch away the substrate to create a uniform, honeycomb spike structure [Fig. 1(c)]. Finally the nanostructure is created after the particles are washed away [Fig. 1(d)]. It should be noted that the created pattern in Fig. 1(d) is the replica of the mask. Details of this technology are described in [1,14,15]. Polystyrene nanospheres from Bangs Laboratories with a 10% solid in colloid concentration and mean diameter of 200 nm were used in this study. The colloid was diluted about 10 times using deionized water. A slide coating method was used in the self-assembly procedure for creating a high-quality monolayer of nanoparticles. It should be noted that the space between neighboring nanoparticles in the monolayer is a fraction of the particle diameter. Figure 2 shows the atomic force microscopic (AFM) scan of monolayer nanoparticles (top view) after a
Fig. 1. Side view of NSL process. (a) Unprocessed substrate. (b) Self-assembled monolayer of nanoparticles as a mask. (c) Etched substrate. (d) Final pattern on substrate after mask removal. © 2014 Optical Society of America
July 1, 2014 / Vol. 39, No. 13 / OPTICS LETTERS
SiO2 6HF− > H2 SiF6 2H2 O.
Fig. 2.
Top view of self-assembled nanoparticle layer.
successful self-assembly procedure on glass surface. Figure 3 shows patterns with less than one layer. Both results can be used for various applications that require desired refractive indices—most notably as an antireflection surface tunable for various central wavelengths. These patterns are to be transferred onto a glass substrate via vapor-phase hydrofluoric (HF) acid etching and the etching rate is studied. HF acid is the typical chemical used to etch glass. Using liquid HF acid is a common practice when it comes to silica; however the method is not suitable for NSL as the nanoparticles that are used as the template may be washed off. An HF vapor etching method, which is far less studied and has a slower etch rate [16], is selected in this study for glass etching. The vapor form of HF acid is an effective tool in NSL for etching glass over large areas. The gas interacts with the glass in a way that doesn’t cause the nanoparticles to be moved or otherwise disordered. The vapor phase is just as effective as the liquid form when etching glass. The etching process is characterized by the following two equations: SiO2 4HF− > SiF4 g 2H2 O
(1)
and
3749
(2)
When HF is placed in a reservoir, it passively generates into its vapor phase. The vapor pressure of HF acid [30.7 mbar (23 mmHg) at 20°C (68°F)] allows it to convert to vapor form from liquid at room temperature. This makes it ideal for etching glass without any special equipment. In this study, 20 mL of HF was poured into a petri dish and left for 5 min so that steady vapors would begin forming. Figure 4 shows how all the components were positioned for the etching process. The setup for the etching process involves a computer controlled arm that alternates the position of the sample. On the two ends of the arc of the arm are a hot plate and the petri dish with HF in it. The hot plate is to evaporate the H2 O by-product of the glass etching. If excess H2 O accumulates on the sample, a thin film of H2 O will prevent the HF vapors from reaching the work piece. The substrate is placed 11.5 cm above the hot plate that has been set to 200° C. Since the substrate is in a fume hood and is a considerable distance away from the hot plate, the temperature is elevated to account for the heat lost. The petri dish at the other end of the path is placed so that the substrate is approximately 1.2 cm from the surface of the acid. The procedure for etching involved setting the time spent above the acid, the time spent above the hot plate, and then the number of cycles of hot plate and acid interaction. The amount of time spent above the acid varied while maintaining a constant hot plate exposure of 30 seconds per cycle. When the etching cycles were completed, the substrate would rest over the hot plate and then removed to be examined under the AFM with the nanoparticle mask still on. Then the mask would be washed off and examined by AFM again. Figure 5 demonstrates successful pattern transfer via NSL and the above etching process. It can be seen that the general feature of the mask shown in Fig. 2 is created on the glass substrate. The mean etching depth in Fig. 5 is around 250 nm, and the standard deviation is about 70 nm, which indicates the roughness in the etching and is mainly caused by nonuniform impurities and the amorphous nature of the glass substrate. While not much research has been done with vapor phase glass etching in nanostructure fabrication, the result herein demonstrates that it is a feasible solution. This technology appears to be the most effective and inexpensive way to etch glass for NSL without damaging the nanosphere mask.
HF Vapor
Fig. 3. Top view of less than one layer of nanoparticles.
Fig. 4.
Setup for HF vapor phase etching glass nanostructures.
3750
Fig. 5. area.
OPTICS LETTERS / Vol. 39, No. 13 / July 1, 2014
AFM scan of etched glass substrate over a 7.5 × 10 μm
It is interesting to study the etching rate of vapor phase HF for nanostructures and compare it to that of liquid HF and for bulk material. Table 1 lists the etching time over the HF vapors versus the etching depth. The total time was controlled by the number of cycles used in the etching. The data for etched depth was gathered from AFM pictures via the horizontal line profile between the highest point and the lowest point on particular slides and averaged. The reason for taking multiple data points on a horizontal line profile accounts for the vertical imbalance that has been observed on the AFM. Figure 6 shows the measured etching depth versus time. A power curve fitting is also shown for etching depth up to 300 nm and is given by y 77.817x0.5128 , where y is the etching depth in nm and x is the time in minutes. An asymptote can be observed in Fig. 6. This is because the measurements taken by the AFM are relative depths. At some point, the mask formed by the nanoparticles becomes etched underneath from the sides—undercut. The pillars that are originally formed from the initial etching start disappearing as the undercutting becomes more severe. It is obvious that we can avoid under etching and under cutting by using proper etching time. There are other approaches to etching silicon oxide with wet or dry HF. There is a good study done to compare wet HF etching and vapor HF etching for silicon oxide removal in microelectromechanical systems (MEMS) [16]. The etching rates in that study are compared with the results from this research. The comparison is done using their data for thermal oxide (as deposited). Note that the oxide is a thermal oxide grown at 975°C and 1200 nm thick on a silicon substrate [16]. For comparison with wet etching, we used the data from Fig. 6 and the results from [16]. It should be noted that wet etching is not applicable in this research because it Table 1. Slide No. 89 92 87 88 91
Fig. 6.
Depth of etch versus total time over HF acid vapor.
would damage the nanoparticle mask, and the data is presented here just to compare the etching rate. The wet etching uses an HF:H2 O solution with a ratio 24.5%∶75.5% or 14.2 mol∕L. Figure 7 shows that the etch rate of vapor HF is much slower than wet HF but more controllable. The vapor phase HF etching rate in [16] is also compared with that in this study. Figure 8 shows the two etching rates. It can be seen that the etch rate in [16] is much slower than the findings in this research. The etch rate of HF vapor is affected by the temperature of the substrate. The colder the substrate, the faster the etch rate. The etch rate from [16] was determined with a 10 min wait time over the heater stage and includes a N2 flow for delivery of HF vapor. This dilutes the concentration per cm3 of HF vapor per second compared to a direct delivery over HF that is vaporizing in this study. The above factors can be the main reasons why the etching performed in this research is at a more rapid rate. Our etching rate is consistent with the results from micromachining processing [17] of vapor HF etching of SiO2 . Their results show a 66 nm/min etch rate that is almost the same as the results produced in this research.
Etching Data
Total Time over Acid (min)
Average Etching Depth (nm)
1 2 5 10 15
70.0 123.3 186.7 266.7 282.5
Fig. 7. Comparison of (A) wet etching rates of SiO2 and (B) vapor etching rates of optical glass in this study.
July 1, 2014 / Vol. 39, No. 13 / OPTICS LETTERS
Fig. 8. Vapor phase HF etching rate comparison between (A) optical glass in this study and (B) SiO2 from [16].
Fig. 9. Enhanced transmission of etched glass slide relative to that of a blank glass slide.
As a test of the etched surface layer with reduced refractive index, a Perkin Elmer Lambda900 UV/VIS/ NIR spectrometer was used to measure the reflectivity of the nanostructures within the visible spectrum of light. Figure 9 demonstrates the enhanced transmittance of the surface modification. Here T is the transmittance of the glass slide with nanostructures on one of its surfaces and T B is the transmittance of a blank glass slide. The nanostructure has an average etching depth of 125 nm, which corresponds to a transmittance peak at 760 nm.
3751
In general, transmittance enhancement is more significant at longer wavelengths. This result is consistent with earlier work on antireflection coatings using nanoparticles as a low refractive index layer [1,18]. It is another indication that the nanoparticle patterns are transferred onto the glass substrate after the vapor HF etching. In summary, we have created nanostructures on the surface of optical glass using nanosphere lithography. The substrates were etched with vapor-phase hydrofluoric (HF) acid. It was shown that the nanoparticle pattern was successfully transferred to the surface of the substrate. The etching rate was studied and compared with existing results of wet and dry HF etching. The effective refractive index of the resulting artificial layer is reduced by depth of etching. The subsequent artificial material layer demonstrated enhanced transmittance in optical wavelengths. References 1. Y. Zhao, J. Wang, and G. Mao, Opt. Lett. 30, 1885 (2005). 2. W. Tan, N. Huang, L. Wang, T. Song, C. Lu, L. Wang, and J. Zhang, J. Solid State Chem. 201, 13 (2013). 3. C. H. Ahn, J. Choi, G. Beaucage, J. H. Nevin, J. Lee, A. Puntambekar, and J. Y. Lee, Proc. IEEE 92, 154 (2004). 4. Y. Takabayashi, M. Uemoto, K. Aoki, T. Odake, and T. Korenaga, Analyst 131, 573 (2006). 5. A. Grosse, M. Grewe, and H. Fouckhardt, J. Micromech. Microeng. 11, 257 (2001). 6. E. Metwalli and C. G. Pantano, Nucl. Instrum. Methods Phys. Res., Sect. B 207, 21 (2003). 7. T. Akashi and Y. Yoshimura, J. Micromech. Miroeng. 16, 1051 (2006). 8. L. Li, T. Abe, and M. Esashi, J. Vac. Sci. Technol. B 21, 2545 (2003). 9. E. Thiénot, F. Domingo, E. Cambril, and C. Gosse, Microelectron. Eng. 83, 1155 (2006). 10. X. Li, T. Abe, and M. Esashi, Sens. Actuators A 87, 139 (2001). 11. T. Ichiki, Y. Sugiyama, T. Ujiie, and Y. Horiike, J. Vac. Sci. Technol. B 21, 2188 (2003). 12. J. H. Park, N.-E. Lee, J. Lee, J. S. Park, and H. D. Park, Microelectron. Eng. 82, 119 (2005). 13. P. W. Leech, Vacuum 55, 191 (1999). 14. C. L. Haynes and R. P. Van Duyne, J. Phys. Chem. B 105, 5599 (2001). 15. E. Wang and Y. Zhao, Proc. SPIE 8818, 881805 (2013). 16. A. Witvrouw, B. Du Bois, P. De Moor, A. Verbist, C. Van Hoof, H. Bender, and K. Baert, Proc. SPIE 4174, 130 (2000). 17. K. Williams and R. Muller, J. Microelectromech. Syst. 5, 256 (1996). 18. H. Jiang, K. Yu, and Y. Wang, Opt. Lett. 32, 575 (2007).
OPTICS LETTERS / Vol. 39, No. 13 / July 1, 2014
Etching of nanostructures on soda-lime glass Elmer Wang and Yang Zhao* Department of Electrical & Computer Engineering, Wayne State University, Detroit, Michigan 48202, USA *Corresponding author: [email protected] Received May 1, 2014; revised May 25, 2014; accepted May 26, 2014; posted May 27, 2014 (Doc. ID 211261); published June 18, 2014 Nanostructures were created on the surface of optical glass using nanosphere lithography. The substrates were etched with vapor-phase hydrofluoric (HF) acid. The etching rate was studied and compared with existing results of wet and dry HF etching. An empirical etching rate formula is found for etching depth up to 300 nm. The subsequent artificial material layer demonstrated enhanced transmittance in optical wavelengths. © 2014 Optical Society of America OCIS codes: (110.4235) Nanolithography; (220.4241) Nanostructure fabrication; (310.6628) Subwavelength structures, nanostructures; (160.2750) Glass and other amorphous materials. http://dx.doi.org/10.1364/OL.39.003748
Nanostructures fabricated on the surface of an optical material have interesting properties for many unique and useful applications. These sophisticated nanostructures can potentially guide, refract, reflect, absorb, or maneuver incident light. They will serve as a technology platform for future generations of functional surfaces, including light emission from nanostructured LEDs, diffusion filters for video screens, efficient absorption of light for thin film technology and solar cells, and super-hydrophobic surfaces for self-cleaning. Nanostructures for optical applications typically have feature sizes around 100–300 nm for visible and near infrared wavelengths [1,2]. Direct fabrication of nanostructures on glass substrate is of particular interest because glass is one of the most commonly used optical materials. In addition, glass surfaces can have controllable hydrophilic or hydrophobic properties for applications including microfluidic devices and chemical sensors. [3,4]. Since most glass is amorphous and chemically stable, it is a challenging task to find effective lithography techniques and etchants in order to create well-patterned nanoscale features. Current nanoscale fabrication methods include deep UV photolithography, electron beam lithography, sol-gel based nanoimprinting coating, and direct growth of nanorods using electro-chemical processes. Each of these methods is limited by various factors that hinder fabrication of nanostructures, such as time requirements and/or capital investment [1]. There are attempts to etch glass surfaces using standard cleanroom techniques. Most of these methods are for microstructures on pure (and fused silica) silicon dioxide substrates etched by using CF4 , SF6 , and HF (liquid and gas). [3–13]. This research focuses on the etching of nanostructures on glass substrate by using effective and low-cost methods. Specifically, nanosphere lithography (NSL) with vapor-HF as an etchant is applied to nanostructure fabrication on soda-lime glass. Soda-lime glass and crown glass are two very common types of glass available commercially. Soda-lime glass is typically constituted of 73% SiO2 , 14% Na2 O, 9% CaO, 4% MgO, 0.15% Al2 O3 , 0.03% K2 O, 0.02% TiO2 , and 0.1% Fe2 O3 . All these impurities affect how the glass is etched with HF. We report the etching process, etched nanostructure patterns on glass surface, and the etching rate. 0146-9592/14/133748-04$15.00/0
NSL is a cost-effective technology that uses nanospheres to create a mask on a substrate [14]. Figure 1 illustrates the steps used in NSL. Nanoparticles of uniform size are self-assembled into a monolayer of spheres to be used as a mask [Figs. 1(a) and 1(b)]. This allows for the fabrication of arrays with controlled spacing based on the nanosphere diameters and packing efficiency. The holes between the particles create the space for etchants or other patterning technologies to reach the substrate. After the mask is made, different methods can be used to etch away the substrate to create a uniform, honeycomb spike structure [Fig. 1(c)]. Finally the nanostructure is created after the particles are washed away [Fig. 1(d)]. It should be noted that the created pattern in Fig. 1(d) is the replica of the mask. Details of this technology are described in [1,14,15]. Polystyrene nanospheres from Bangs Laboratories with a 10% solid in colloid concentration and mean diameter of 200 nm were used in this study. The colloid was diluted about 10 times using deionized water. A slide coating method was used in the self-assembly procedure for creating a high-quality monolayer of nanoparticles. It should be noted that the space between neighboring nanoparticles in the monolayer is a fraction of the particle diameter. Figure 2 shows the atomic force microscopic (AFM) scan of monolayer nanoparticles (top view) after a
Fig. 1. Side view of NSL process. (a) Unprocessed substrate. (b) Self-assembled monolayer of nanoparticles as a mask. (c) Etched substrate. (d) Final pattern on substrate after mask removal. © 2014 Optical Society of America
July 1, 2014 / Vol. 39, No. 13 / OPTICS LETTERS
SiO2 6HF− > H2 SiF6 2H2 O.
Fig. 2.
Top view of self-assembled nanoparticle layer.
successful self-assembly procedure on glass surface. Figure 3 shows patterns with less than one layer. Both results can be used for various applications that require desired refractive indices—most notably as an antireflection surface tunable for various central wavelengths. These patterns are to be transferred onto a glass substrate via vapor-phase hydrofluoric (HF) acid etching and the etching rate is studied. HF acid is the typical chemical used to etch glass. Using liquid HF acid is a common practice when it comes to silica; however the method is not suitable for NSL as the nanoparticles that are used as the template may be washed off. An HF vapor etching method, which is far less studied and has a slower etch rate [16], is selected in this study for glass etching. The vapor form of HF acid is an effective tool in NSL for etching glass over large areas. The gas interacts with the glass in a way that doesn’t cause the nanoparticles to be moved or otherwise disordered. The vapor phase is just as effective as the liquid form when etching glass. The etching process is characterized by the following two equations: SiO2 4HF− > SiF4 g 2H2 O
(1)
and
3749
(2)
When HF is placed in a reservoir, it passively generates into its vapor phase. The vapor pressure of HF acid [30.7 mbar (23 mmHg) at 20°C (68°F)] allows it to convert to vapor form from liquid at room temperature. This makes it ideal for etching glass without any special equipment. In this study, 20 mL of HF was poured into a petri dish and left for 5 min so that steady vapors would begin forming. Figure 4 shows how all the components were positioned for the etching process. The setup for the etching process involves a computer controlled arm that alternates the position of the sample. On the two ends of the arc of the arm are a hot plate and the petri dish with HF in it. The hot plate is to evaporate the H2 O by-product of the glass etching. If excess H2 O accumulates on the sample, a thin film of H2 O will prevent the HF vapors from reaching the work piece. The substrate is placed 11.5 cm above the hot plate that has been set to 200° C. Since the substrate is in a fume hood and is a considerable distance away from the hot plate, the temperature is elevated to account for the heat lost. The petri dish at the other end of the path is placed so that the substrate is approximately 1.2 cm from the surface of the acid. The procedure for etching involved setting the time spent above the acid, the time spent above the hot plate, and then the number of cycles of hot plate and acid interaction. The amount of time spent above the acid varied while maintaining a constant hot plate exposure of 30 seconds per cycle. When the etching cycles were completed, the substrate would rest over the hot plate and then removed to be examined under the AFM with the nanoparticle mask still on. Then the mask would be washed off and examined by AFM again. Figure 5 demonstrates successful pattern transfer via NSL and the above etching process. It can be seen that the general feature of the mask shown in Fig. 2 is created on the glass substrate. The mean etching depth in Fig. 5 is around 250 nm, and the standard deviation is about 70 nm, which indicates the roughness in the etching and is mainly caused by nonuniform impurities and the amorphous nature of the glass substrate. While not much research has been done with vapor phase glass etching in nanostructure fabrication, the result herein demonstrates that it is a feasible solution. This technology appears to be the most effective and inexpensive way to etch glass for NSL without damaging the nanosphere mask.
HF Vapor
Fig. 3. Top view of less than one layer of nanoparticles.
Fig. 4.
Setup for HF vapor phase etching glass nanostructures.
3750
Fig. 5. area.
OPTICS LETTERS / Vol. 39, No. 13 / July 1, 2014
AFM scan of etched glass substrate over a 7.5 × 10 μm
It is interesting to study the etching rate of vapor phase HF for nanostructures and compare it to that of liquid HF and for bulk material. Table 1 lists the etching time over the HF vapors versus the etching depth. The total time was controlled by the number of cycles used in the etching. The data for etched depth was gathered from AFM pictures via the horizontal line profile between the highest point and the lowest point on particular slides and averaged. The reason for taking multiple data points on a horizontal line profile accounts for the vertical imbalance that has been observed on the AFM. Figure 6 shows the measured etching depth versus time. A power curve fitting is also shown for etching depth up to 300 nm and is given by y 77.817x0.5128 , where y is the etching depth in nm and x is the time in minutes. An asymptote can be observed in Fig. 6. This is because the measurements taken by the AFM are relative depths. At some point, the mask formed by the nanoparticles becomes etched underneath from the sides—undercut. The pillars that are originally formed from the initial etching start disappearing as the undercutting becomes more severe. It is obvious that we can avoid under etching and under cutting by using proper etching time. There are other approaches to etching silicon oxide with wet or dry HF. There is a good study done to compare wet HF etching and vapor HF etching for silicon oxide removal in microelectromechanical systems (MEMS) [16]. The etching rates in that study are compared with the results from this research. The comparison is done using their data for thermal oxide (as deposited). Note that the oxide is a thermal oxide grown at 975°C and 1200 nm thick on a silicon substrate [16]. For comparison with wet etching, we used the data from Fig. 6 and the results from [16]. It should be noted that wet etching is not applicable in this research because it Table 1. Slide No. 89 92 87 88 91
Fig. 6.
Depth of etch versus total time over HF acid vapor.
would damage the nanoparticle mask, and the data is presented here just to compare the etching rate. The wet etching uses an HF:H2 O solution with a ratio 24.5%∶75.5% or 14.2 mol∕L. Figure 7 shows that the etch rate of vapor HF is much slower than wet HF but more controllable. The vapor phase HF etching rate in [16] is also compared with that in this study. Figure 8 shows the two etching rates. It can be seen that the etch rate in [16] is much slower than the findings in this research. The etch rate of HF vapor is affected by the temperature of the substrate. The colder the substrate, the faster the etch rate. The etch rate from [16] was determined with a 10 min wait time over the heater stage and includes a N2 flow for delivery of HF vapor. This dilutes the concentration per cm3 of HF vapor per second compared to a direct delivery over HF that is vaporizing in this study. The above factors can be the main reasons why the etching performed in this research is at a more rapid rate. Our etching rate is consistent with the results from micromachining processing [17] of vapor HF etching of SiO2 . Their results show a 66 nm/min etch rate that is almost the same as the results produced in this research.
Etching Data
Total Time over Acid (min)
Average Etching Depth (nm)
1 2 5 10 15
70.0 123.3 186.7 266.7 282.5
Fig. 7. Comparison of (A) wet etching rates of SiO2 and (B) vapor etching rates of optical glass in this study.
July 1, 2014 / Vol. 39, No. 13 / OPTICS LETTERS
Fig. 8. Vapor phase HF etching rate comparison between (A) optical glass in this study and (B) SiO2 from [16].
Fig. 9. Enhanced transmission of etched glass slide relative to that of a blank glass slide.
As a test of the etched surface layer with reduced refractive index, a Perkin Elmer Lambda900 UV/VIS/ NIR spectrometer was used to measure the reflectivity of the nanostructures within the visible spectrum of light. Figure 9 demonstrates the enhanced transmittance of the surface modification. Here T is the transmittance of the glass slide with nanostructures on one of its surfaces and T B is the transmittance of a blank glass slide. The nanostructure has an average etching depth of 125 nm, which corresponds to a transmittance peak at 760 nm.
3751
In general, transmittance enhancement is more significant at longer wavelengths. This result is consistent with earlier work on antireflection coatings using nanoparticles as a low refractive index layer [1,18]. It is another indication that the nanoparticle patterns are transferred onto the glass substrate after the vapor HF etching. In summary, we have created nanostructures on the surface of optical glass using nanosphere lithography. The substrates were etched with vapor-phase hydrofluoric (HF) acid. It was shown that the nanoparticle pattern was successfully transferred to the surface of the substrate. The etching rate was studied and compared with existing results of wet and dry HF etching. The effective refractive index of the resulting artificial layer is reduced by depth of etching. The subsequent artificial material layer demonstrated enhanced transmittance in optical wavelengths. References 1. Y. Zhao, J. Wang, and G. Mao, Opt. Lett. 30, 1885 (2005). 2. W. Tan, N. Huang, L. Wang, T. Song, C. Lu, L. Wang, and J. Zhang, J. Solid State Chem. 201, 13 (2013). 3. C. H. Ahn, J. Choi, G. Beaucage, J. H. Nevin, J. Lee, A. Puntambekar, and J. Y. Lee, Proc. IEEE 92, 154 (2004). 4. Y. Takabayashi, M. Uemoto, K. Aoki, T. Odake, and T. Korenaga, Analyst 131, 573 (2006). 5. A. Grosse, M. Grewe, and H. Fouckhardt, J. Micromech. Microeng. 11, 257 (2001). 6. E. Metwalli and C. G. Pantano, Nucl. Instrum. Methods Phys. Res., Sect. B 207, 21 (2003). 7. T. Akashi and Y. Yoshimura, J. Micromech. Miroeng. 16, 1051 (2006). 8. L. Li, T. Abe, and M. Esashi, J. Vac. Sci. Technol. B 21, 2545 (2003). 9. E. Thiénot, F. Domingo, E. Cambril, and C. Gosse, Microelectron. Eng. 83, 1155 (2006). 10. X. Li, T. Abe, and M. Esashi, Sens. Actuators A 87, 139 (2001). 11. T. Ichiki, Y. Sugiyama, T. Ujiie, and Y. Horiike, J. Vac. Sci. Technol. B 21, 2188 (2003). 12. J. H. Park, N.-E. Lee, J. Lee, J. S. Park, and H. D. Park, Microelectron. Eng. 82, 119 (2005). 13. P. W. Leech, Vacuum 55, 191 (1999). 14. C. L. Haynes and R. P. Van Duyne, J. Phys. Chem. B 105, 5599 (2001). 15. E. Wang and Y. Zhao, Proc. SPIE 8818, 881805 (2013). 16. A. Witvrouw, B. Du Bois, P. De Moor, A. Verbist, C. Van Hoof, H. Bender, and K. Baert, Proc. SPIE 4174, 130 (2000). 17. K. Williams and R. Muller, J. Microelectromech. Syst. 5, 256 (1996). 18. H. Jiang, K. Yu, and Y. Wang, Opt. Lett. 32, 575 (2007).
