Encapsulation of grating structures facilitates an improvement of the optical functionality and/or adds mechanical stability to the fragile structure. Here, we introduce novel encapsulation process of nanoscale patterns based on atomic layer deposition and micro structuring. The overall size of the encapsulated structured surface area is only restricted by the size of the available microstructuring and coating devices; thus, overcoming inherent limitations of existing bonding processes concerning cleanliness, roughness, and curvature of the components. Finally, the process is demonstrated for a transmission grating. The encapsulated grating has 97.5% transmission efficiency in the −1st diffraction order for TM-polarized light, and is being limited by the experimental grating parameters as confirmed by rigorous coupled wave analysis.
© 2015 Optical Society of America
Diffractive optical elements are essential to manipulate and tailor the propagation of light. Based on ongoing technological progress and miniaturization, components with very small periodic and aperiodic structures can be generated in dielectric substrates. A wide range of optical elements are available, e.g. high efficiency transmission gratings, wire grid polarizer gratings, resonant waveguides, or computer generated holograms (CGH) [1–6]. A fundamental problem especially for optical elements designed for transmission applications are losses due to scattering and Fresnel reflection. Improved structuring processes can minimize scattering losses; however, Fresnel reflections are inherent to the optical element and depend on the incident wavelength and angle.
One method to minimize the Fresnel reflections is the usage of gratings with triangle or sinusoidal shapes instead of a binary shape with perpendicular sidewalls [7, 8]. So the refractive index continuously increases with the height, similar to an antireflective coating based on subwavelength nanostructures. Another method was demonstrated by Clausnitzer et al. . They designed highly efficient transmission grating based on index matching at the grating interfaces. Therefore, an encapsulation of the grating in the same material as the substrate is required. In such an element, the incident beam will propagate in the same material as the diffracted beam and it is possible to adjust the grating parameters to eliminate reflection orders of the diffractive grating. Besides enhanced optical properties, the encapsulation of structured surfaces conveys mechanically stability to the pattern and protection against harmful external contaminants such as corrosive materials, fatty components, or dust particles.
Two methods for grating encapsulation were previously demonstrated. Nishii et al.  proposed coating fused silica gratings with periods between 1.5 µm to 7.3 µm with SiO2 by plasma enhanced chemical vapor deposition (PECVD) at 400°C substrate temperature. Depending on the grating parameters, PEVCD will generate detrimental features within the grooves and will diminish the optical performance of the grating . A lower limit of 0.4 for the duty cycle (ridge width/ period) to maintain the groove shape was determined . Another approach is based on the direct bonding of a fused silica cover substrate to the grating [10, 11]. Clausnitzer et al.  bonded a circular standard substrate (25 mm diameter) on 20mm x 20mm substrate with a patterned area of 10 mm x 10 mm. They achieved diffraction efficiencies of 97% for TE-polarized light in the −1st transmitted diffraction order for a grating with a period of 600 nm in the Littrow configuration . Larger encapsulated gratings with patterns up to 60 mm x 20 mm and high transmission efficiencies (98.8%) for TE-polarized light were realized by Kalkowski et al. . However, encapsulation through direct bonding is a challenging process. The bonding must be very accurate over large surface areas because even tiny air pockets or contamination will strongly disturb the optical function . Hence the surfaces, to be bonded, must be very smooth and clean. The surface roughness should not exceed 1 nm . Bonding a massive bulk material on top of a structured surface also requires a constant height profile of the nano- and microstructures. Otherwise the linkage between the grating and cover element only takes place at the highest part of the structure.
This work reports on a bonding-free encapsulation process, where a flat triple layer stack is realized directly on top of the grating. The triple layer stack functions additionally as antireflection coating. In this process, the grating is first embedded in a sacrificial material (Al2O3), the excess material on top of the grating is removed, a cover layer (SiO2) is applied, and finally the sacrificial material is removed by selective wet-chemical etching. Complete removal of the sacrificial material has been confirmed by energy dispersive X-ray spectroscopy (EDX) and scanning electron microscope (SEM) measurements. A transmission grating with 97.5% diffraction efficiency in the −1st order for TM-polarized light has been achieved. The measured efficiency is limited by the non-optimal grating parameters as proven by rigorous coupled wave analysis and can be further increased by improving the grating structure.
2. Experimental methods
2.1. Encapsulation process
The encapsulation process is schematically presented in Fig. 1. First a binary grating was prepared by structuring a fused silica substrate [Fig. 1(a)]. Therefore we structured the sample with standard electron beam lithography following by standard etching processes. The structured sample was carefully cleaned to remove all residuals. Then, a sacrificial layer (Al2O3) was deposited to fill the grooves of the grating. Atomic layer deposition (ALD) was applied to ensure high uniformity of the filling material [Fig. 1(b)]. The deposition was carried out in a commercial available plasma enhanced ALD tool, OpAL Oxford Plasma Technologies. The sacrificial layer was deposited by thermal ALD and plasma enhanced atomic layer deposition (PEALD). Trimethylaluminium (TMA) was used as metallic precursors and water as oxidizer at a substrate temperature of 120°C and a working pressure of 200 mTorr. The pulse and purge times for TMA and water were 20 ms; 6000 ms and 30 ms; 12000 ms, respectively. The chamber was continuously purged with Argon (purity: 99.999%) and the base pressure equaled 16 mTorr. A total layer thickness of 230 nm was deposited corresponding to 2300 cycles. The Al2O3 deposited by PEALD was carried out by TMA as metallic precursors and oxygen plasma at a substrate temperature of 200°C. The pulse and purge times for TMA and oxygen plasma were 20 ms; 6000 ms; 5000 ms and 4000 ms, respectively.
The next step in the encapsulation process contains in the removal of the excess material on the top of the grating [Fig. 1(c)]. Therefor ion beam etching (IBE) was used. If only the excess material is etched by IBE, the intentions of the excess layer are replicated in the groove filling material. To prevent the replica of the intentions in the groove filling material an AZ resist was spun onto the excess material. The AZ resist generates a plan surface. Argon ions (Ekin: 400 eV; angle of incidence: 60°) etch the AZ resist and the excess Al2O3 layer nearly with the same rate, and so a planar filled grating is obtained [cf. Figure 1(c)]. Accordingly, the intentions from the excess layer are not formed in the groove filling material. Then, the cover layer (SiO2) was deposited by plasma enhanced atomic layer deposition (PEALD) [Fig. 1(d)]. Based on this deposition technique, a chemically bonded cover layer will grow on the top of the grating ridges and a strong adhesion between the cover layer and the grating is ensured. Here, we used tris[dimethylamino]silane (3DMAS) as a silicon precursor and oxygen plasma as an oxidizer at 200°C substrate temperature. The pulse and purge times for the 3DMAS and the oxygen plasma correspond to 350 ms; 8000 ms and 3000 ms; 2000 ms, respectively. The thickness of the SiO2 cover layer was 40 nm and 200 nm, respectively.
Thereafter, the cover was patterned by a similar structuring process as we treat the sample before [Fig. 1(e)]. Two microstructure patterns were chosen. First a linear pattern with a period (P) of 5 µm and 300 nm slit, and second a 300 µm period (P) pattern also with 300 nm slit. The upper pattern is tilted 90 °C towards the optical grating. Finally, the element was placed in phosphoric acid aqueous solution (H3PO4 85%) at 80°C. Complete removal of the sacrificial material was observed after 3 h for the 5 µm pattern and after ca. 15 h for the larger 300 µm pattern [Fig. 1(f)]. Finally, the element was rinsed with distilled deionized water and dried under nitrogen flow. In a following step, the slit in the cover can be closed using PVD, as it has been done for the demonstrator.
2.2. Characterization methods
The SiO2 and Al2O3 layers were analyzed by spectroscopic ellipsometry (Woollam Inc. M2000 ellipsometer). The ellipsometry data were fitted by a Cauchy model. The reflectivity and transmission measurements were done by optical spectroscopy in a custom made system operating at 5° angle of incidence in reflectance. Additionally the density of the Al2O3 layer was determined by Rutherford Backscattering Spectrometry (RBS).
The SEM Hitachi S-4800 and Neon 60 CrossBeam® were used to investigate the grating geometry. The grating parameters were determined using the SEM Data Manager software, from the electron micrographs. The energy dispersive X-ray spectroscopy (EDX) was performed at the SEM Hitachi S-4800. The electrons are accelerated to 15 keV to ensure broad element detection.
The transmission diffraction efficiency measurement was performed in a custom made setup with a motorized goniometer, which was equipped with an integrating sphere and a tunable laser diode as light source. The measurement setup has an estimated error of ± 0.5%. The custom-made tunable laser is operating in the spectral range from 1000 nm to 1070 nm.
3. Encapsulated gratings
Etching precisely defined structures into a substrate material is crucial in the fabrication of diffractive optical elements. The pattern which is generated by an exposure tool e.g. electron beam lithography in a resist must be transferred to a mask, and further into the substrate material. High etch selectivity and an anisotropic etch process is necessary to generate small period, high aspect ratio structures in the substrate material. In recent years, reactive ion etching turned out to be the method of choice for patterning nanostructured optics [5, 6]. Reactive ion etching is a well-controlled process concerning the uniformity, profile and quality of the achieved nanostructures. An anisotropic behavior guaranties high aspect ratio structures without undercutting the mask. In contrast, wet chemical etching received less attention in the fabrication of optical elements, although it is a standard technique in semiconductor industry for nanostructuring. Generally, wet chemical etching has an isotropic behavior and the etching is not directionally defined in an isotropic substrate such as fused silica. This hinders wet chemical etching from generating defined high aspect ratio nanostructures by transferring them from a mask. Nevertheless, wet chemical etching has other important advantages, e.g. very high etching selectivity, and low contamination or damage behavior to the substrate material [12, 13]. The semiconductor industry takes advantage of these benefits, e.g. selectively removing of sacrificial layers. Often SiO2 is used as a sacrificial layer and etched by fluoric acid, whereby the Si substrate remains unaffected [12, 13]. Furthermore, anodic oxidation can generate porous layers by wet etching of nano-sized long tubes [14, 15].
Here, we apply wet chemical etching in the fabrication of diffractive optical elements. Particularly, we take advantage of the isotropic behavior inherent to wet etching to develop a novel encapsulation process for high aspect ratio gratings. Therefore, high selectivity must be ensured by properly choosing the etching reactant, the substrate and sacrificial materials. Selective etching in high aspect ratio gratings has previously been demonstrated in a damage-free replication process of gratings . Optical elements for transmission applications commonly consist of structured fused silica substrates. Fused silica is chemically inert against phosphoric acid (H3PO4), whereas other oxides such as alumina (Al2O3) can be efficiently removed by this etching material [17–19]. Hence, alumina can serve as an adequate sacrificial layer in the advanced structuring of fused silica optics.
Figure 2(a) shows electron micrograph in cross section of the encapsulated grating before H3PO4 etching. The grooves are filled with alumina, and a 40 nm thin SiO2 layer closes the grating. Platinum was locally deposited as metallic layer prior to the inspection. The same grating is shown in Fig. 2(b) after the H3PO4 treatment. The H3PO4 has completely cleared the grooves without damaging the fused silica grating or the ultrathin SiO2 layer [Fig. 2(b)], whereas the shape of the ridges remains unaffected by the wet etching process. Detailed size measurements demonstrate a very good agreement between the grating height, ridge width and SiO2 layer thickness in both SEM images. Hence, the grating remains completely unaffected by the H3PO4. Additionally, no residues in the grooves are discernible. This is also confirmed by an EDX-measurement [Fig. 2(c)]. While a strong Kα peak from aluminum at 1.48 keV was detected before H3PO4 etching, the peak disappears after H3PO4 etching. The triangular shape at the bottom of the grating groove results from the reactive ion etching of the fused silica substrate and remains also unaffected by the H3PO4 treatment. Figure 2(d) shows an overview image of the encapsulated grating. The two SEM images in the right upper corner show magnified section of Fig. 2(d). Through the very thin SiO2 cover layer, the grating and the empty grooves are visualized. The authors emphasize the very thin SiO2 cover (ca. 40 nm) and highlight the uniform covering of the grating on a large scale.
A large period, small slits and appropriate thickness of the cover layer diminishes the effect of the upper pattern on the optical function of the grating. In order to enlarge the period of the upper microstructure, the etching process was extensively optimized. The first pattern we applied for the cover layer has a period (P) of 5 µm and the duty cycle is 0.94, corresponding to a slit of 300 nm. The goal is to minimize the area covered by slits. In the ideal case, the removal of the sacrificial layer could be realized at the edges of the substrate. The removal of the Al2O3 sacrificial layer by H3PO4 from the nanochannels is considered as reaction controlled based on the reaction mechanism study reported by Zhou et al. . In contrast to the diffusion-controlled reaction, the etch rate is independent of the diffusion of reactants and products, and remains nearly unaffected when the surface to be etched is structured, e.g. small trenches or pores. Zhou et al.  investigated the reaction between sputtered Al2O3 and H3PO4 in terms of reaction temperature, acid concentration and reaction time. They reported an etch rate of 22.5 nm/min at 50°C in 85% H3PO4.
To optimize the etch rate of the sacrificial layer, two Al2O3 coatings deposited by thermal ALD and PEALD are compared. Generally, PEALD is able to coat denser films than thermal ALD. The Al2O3 deposited by thermal ALD at 120°C substrate temperature had a density of 2.74 g/cm3, whereas Al2O3 prepared by PEALD at 200°C substrate temperature had a density of 2.96 g/cm3, as determined by RBS. A lower density results in less material to be etched in the same volume which effects the etch rate. Figure 3(a) displays the amount of Al2O3 etched as a function of the reaction time for both coating processes. The analysis was performed at 60°C reaction temperature and the acid concentration of 30% was chosen, in order to get reasonable results, because the reported etch rates are too high at higher reaction temperature and acid concentration [17, 18]. The alumina coating deposited by thermal ALD at 120°C substrate temperature achieved a nearly two times higher etch rate, 1.7 µm/h, compared to the alumina deposited by PEALD at 200°C substrate temperature. Furthermore, the etch rate will probably increase several orders of magnitude, if the reaction temperature and the acid concentration are higher.
Second, the effect of the phosphoric acid on the fused silica gratings and the SiO2 cover layer must be minimized. An etch rate close to 0 nm per min is desirable. Figure 3(b) illustrates etch rate as a function of the reaction temperatures in aqueous 85% H3PO4. The reaction time was kept constant at 2.5h for all SiO2 samples. The SiO2 was deposited by PEALD at two different substrate temperatures, 120°C and 200°C. At 80°C reaction temperature, the etch rate equals ca. 0.16 nm/h for the SiO2 that was deposited by PEALD at 200°C. Hence, the following demonstrator is etched at 80°C reaction temperature in aqueous 85% H3PO4.
A demonstrator was developed according to the encapsulation process previously discussed. The Al2O3 sacrificial layer was deposited at 120°C in a thermal ALD process, whereas PEALD deposited the cover SiO2 layer at 200°C. The refractive index of the SiO2 layer deposited by PEALD at 200°C was determined by a Cauchy model from ellipsometry data to 1.45 at a wavelength of 632 nm. Reflectivity and transmittance measurements via optical spectroscopy verified that the SiO2 coating has no optical losses in the visible and near infrared spectral range.
4. High efficiency transmission grating
4.1. Theoretical design of the transmission grating
A highly demanding transmission grating has been designed. The grating is optimized for TM-polarized light, to achieve a high transmission in the −1st diffraction order at a wavelength of 1064 nm for a grating with a period of 800 nm in the Littrow configuration at 41.7° angle of incidence. The efficiencies have been calculated based on rigorous coupled wave analysis (RCWA) . Figure 4 displays the design study for the encapsulated grating as function of the duty cycle and the grating depth coded through different efficiencies. All red shaded areas indicate grating parameter combination where the theoretical transmission efficiency exceeds 99%. A technically feasible solution exists at a duty cycle of 0.7 and 3.1 µm grating depth. Nevertheless, such high aspect ratio gratings pose extremely high demands on the fabrication process.
Figure 4 shows the transmission efficiency in dependence on the duty cycle and grating depth for a binary grating structure. However, the grating ridges have rather trapezoidal than perfect perpendicular sidewalls and extensive optimization of the reactive ion etching to control the tilt angle of the ridges and the final profile is necessary. Here, an optimized tilt angle of 2.28° has been applied in the design study of the gratings because the tilt angle is obtained by a stabilized etching process. Table 1 summarizes the ideal and the real grating parameter for a trapezoidal grating structure.
4.2. Measured transmission efficiency
A fused silica grating with the parameters provided by the theoretical design was targeted. A cross section of the final encapsulated grating is shown in Fig. 5(b). Figure 5(a) presents the measured transmission efficiency (black squares) in the wavelength range from 1030 nm to 1070 nm. The results show transmission efficiencies above 95% in the wavelength range from 1040 nm to 1070 nm with a peak efficiency of 97.5% at 1050 nm. The black curve illustrates the efficiency of the ideal grating, whereas the dashed curve is a simulation of the real grating. A very good agreement between the theoretical values of the real grating and the experimental data is obtained. The grating model used for the simulation is schematically shown in Fig. 5(c). All dimensions of the model grating were implemented according to the measured parameters of the experimentally encapsulated grating by SEM analysis. These grating parameters are also summarized in Table 1. The SEM image and the EDX spectrum (not shown) indicate alumina free grooves. Consequently, the measured efficiency is limited by the non-optimal grating parameters and can be further increased by improving the grating structure.
In contrast to Fig. 2(a), the cover layer is flat in Fig. 5(b). A flat cover layer can be generated, when the initial intentions in the excess material are smoothen out. So a flat surface on the filled grating emerges. The cover was structured with a period of 300 µm and a slit of 300 nm. Remarkably, the channels, to be etched considering diffusion of the etching reactant from both sides, have a length of 150 µm and a width of ca. 150 nm at the narrowest point in the groove (measured in cross section of the channel). The encapsulated grating was placed in the phosphoric acid bath for 15.5 h, resulting in an etch rate of 10 µm per hour. Furthermore, small structures at the grating bottom, which appeared during the fused silica reactive ion etching, do not hamper the clearance of the groove by the phosphoric acid. The slit for the removal of the sacrificial layer introduces a defect area in the grating element. In this case, the defect area is ca. 0.1% of the total surface since the ratio of the slit to the upper structure period is 10−3. Transmission efficiency losses due to the effect of the structured cover are negligibly small (≤ 1ppm), as verified by comprehensive RCWA study.
The encapsulated grating was finally refined by adding antireflection coatings (ARC) on the top and backside of the element. The ARC consists of a SiO2/Ta2O5/SiO2 system deposited by sputtering. On top of the grating, a triple layer stack with a total thickness of ca. 700 nm consisting of 400 nm SiO2, 30 nm Ta2O5, and 284 nm SiO2 was applied. The first SiO2 layer corresponds to both PEALD and PVD deposited SiO2; however, no boundary between these two materials can be identified in the SEM image [see Fig. 5(b)]. The efficiency of the transmission grating prior to the ARC deposition was 95% at 1064 nm wavelength, and improved to 97,5% [see Fig. 5(a)] after the refinement. The ARC coating closes additionally the slit completely.
The adhesion of the cover to the grating element was evaluated. The adhesive strength of the cover was controlled by cross cut and tape test according to the ISO:9211-4:2007-03 norm. Therefore, a priori, a damage structure was generated on the encapsulated grating with an authorized cutting tool. Then, cello tape was carefully stuck on the damaged encapsulated grating and abruptly pulled away. Figure 6(a) displays a light microscope image of the encapsulated grating after the tape test. No visible peel off at edges in the cross cutting can be observed, and the cover stays bonded. According to the ISO appraisal system, the result classes with 0 or non-damaged. Hence, a high adhesive strength between the cover layer and the grating was confirmed.
Figure 6(b) shows a FIBSEM image where small debris was encapsulated in the cover layer. The impurity has a height of ca. 50 nm and was completely embedded in the coating. The surrounding region is not affected by the impurity. Such debris would significantly hinder directly bonding a cover substrate to the grating due to mechanical stiffness of the elements, but it does not affect the proposed process.
The proposed encapsulation process is applicable to a variety of conceivable structures, e.g. also periodic and aperiodic 2D and 3D patterns. The requirements on the substrate cleanness, surface roughness, substrate flatness and material selection are much lower than in commonly used direct bonding processes. Contamination on the patterned sample affects the encapsulation by the investigated process only locally. Particles on the surface will be coated as well as their surroundings. Therefore the encapsulation takes place at the surrounding of the particle, while the structure directly beneath the particle is excluded from the encapsulation. In contrast, the same particle would have a large effect on the encapsulation by direct bonding and form a large-scale non-bound environment strongly depending on their size. As a consequence, direct bonding also inhibits adequate encapsulation of structures with different height levels which is possible with the proposed encapsulation process. So far, the encapsulation of curved patterned substrates seems impossible to realize with the technologies currently available. The presented strategy is virtually independent of the sample flatness as long as the thickness uniformity and conformity of the ALD deposition remains in their tolerances. Thus, it is conceivable to encapsulate curved substrates. This would open up a wide range of new applications and optical devices.
The achieved transmission efficiency is comparable with the highest reported efficiency in [6, 11] which are between 96% and 98.5% for transmission gratings. Efficient transmission gratings are highly demanded in standard four-pass chirped pulse amplification (CPA) systems. In a four-pass CPA of a high power laser, diffraction gratings in transmittance or reflectance are used for pulse manipulation in the Littrow configuration. The final laser power scales with the fourth power of the efficiency of the diffraction grating (η4). A diffraction efficiency of 93% will lead to more than 25% losses in the CPA system. Hence, transmission gratings with nearly 100% are highly desired. Želudevičius et al.  used transmission gratings as compressor which reach >94% efficiency. The total compressor achieves an efficiency of 84%. Besides to the losses caused by the gratings, misalignment and depolarization losses influence the efficiency of the CPA. Nevertheless, high transmission efficiencies are essential to receive high total CPA efficiency.
In this study, we present the design, fabrication and characterization of fully encapsulated dielectric gratings. The process is divided into five major steps, based on atomic layer deposition and micro structuring. Firstly, the grating will be embedded in a sacrificial layer, alumina. The excess material on top of the grating is removed and a cover layer (SiO2) is applied. Finally the sacrificial material is etched by phosphoric acid through openings in the cover layer which has been structured prior to the etching. It is demonstrated, that the phosphoric acid is able to clear nano-scaled encapsulated grooves with a length of at least 300 µm within a reasonable reaction time. The proposed process overcomes tight specifications imposed on the substrates inherent to the direct bonding processes such as very low surface roughness (<1nm) or excellent substrate flatness. It opens up a wide range of new optical devices, which are prevented by direct bonding, e.g. encapsulation of curved substrates. The proposed encapsulation process is successfully demonstrated for a diffractive grating optimized for TM-polarized light. The demonstrator reaches a transmission efficiency of 97.5% in the −1st diffraction order for TM-polarized light. The experimentally measured efficiency is in excellent agreement with the theoretical value obtained by rigorous wave coupled analysis. The transmission efficiency can be further increased by improving the grating structure.
We highly acknowledge the financial support within the Emmy Noether program of the German Science Foundation (SZ 253/1-1). The authors are indebted to Frank Fuchs (Fraunhofer-Institut für Angewandte Optik und Feinmechanik) for providing the RCWA software and to Peter Munzert for the physical vapor deposition.
References and links
1. A. Szeghalmi, M. Helgert, R. Brunner, F. Heyroth, U. Gosele, and M. Knez, “Tunable guided-mode resonance grating filter,” Adv. Funct. Mater. 20(13), 2053–2062 (2010). [CrossRef]
3. T. Weber, T. Kasebier, A. Szeghalmi, M. Knez, E. B. Kley, and A. Tünnermann, “High aspect ratio deep UV wire grid polarizer fabricated by double patterning,” Microelectron. Eng. 98, 433–435 (2012). [CrossRef]
5. T. Clausnitzer, J. Limpert, K. Zöllner, H. Zellmer, H. J. Fuchs, E. B. Kley, A. Tünnermann, M. Jupé, and D. Ristau, “Highly efficient transmission gratings in fused silica for chirped-pulse amplification systems,” Appl. Opt. 42(34), 6934–6938 (2003). [CrossRef] [PubMed]
8. H. C. Cao, C. H. Zhou, J. J. Feng, P. Lv, and J. Y. Ma, “Polarization-independent triangular-groove fused-silica gratings with high efficiency at a wavelength of 1550 nm,” Opt. Commun. 283(21), 4271–4273 (2010). [CrossRef]
10. T. Clausnitzer, T. Kämpfe, E. B. Kley, A. Tünnermann, A. V. Tishchenko, and O. Parriaux, “Highly-dispersive dielectric transmission gratings with 100% diffraction efficiency,” Opt. Express 16(8), 5577–5584 (2008). [CrossRef] [PubMed]
11. G. Kalkowski, U. Zeitner, T. Benkenstein, J. Fuchs, C. Rothhardt, and R. Eberhardt, “Direct wafer bonding for encapsulation of fused silica optical gratings,” Microelectron. Eng. 97, 177–180 (2012). [CrossRef]
12. K. R. Williams, K. Gupta, and M. Wasilik, “Etch rates for micromachining processing - part II,” J. Microelectromech. Syst. 12(6), 761–778 (2003). [CrossRef]
13. G. Spierings, “Wet chemical etching of silicate-glasses in hydrofluoric-acid based solutions,” J. Mater. Sci. 28(23), 6261–6273 (1993). [CrossRef]
15. D. Gong, C. A. Grimes, O. K. Varghese, W. C. Hu, R. S. Singh, Z. Chen, and E. C. Dickey, “Titanium oxide nanotube arrays prepared by anodic oxidation,” J. Mater. Res. 16(12), 3331–3334 (2001). [CrossRef]
16. A. Szeghalmi, K. Sklarek, M. Helgert, R. Brunner, W. Erfurth, U. Gosele, and M. Knez, “Flexible replication technique for high-aspect-ratio nanostructures,” Small 6(23), 2701–2707 (2010). [CrossRef] [PubMed]
17. J. Oh, J. Myoung, J. S. Bae, and S. Lim, “Etch behavior of ALD Al2O3 on HfSiO and HfSiON stacks in acidic and basic etchants,” J. Electrochem. Soc. 158(4), D217–D222 (2011). [CrossRef]
18. B. Zhou and W. F. Ramirez, “Kinetics and modeling of wet etching of aluminum oxide by warm phosphoric acid,” J. Electrochem. Soc. 143(2), 619–623 (1996). [CrossRef]
19. J. A. Aboaf, “Deposition and properties of aluminum oxide obtained by pyrolytic decomposition of an aluminum alkoxide,” J. Electrochem. Soc. 114(9), 948–952 (1967). [CrossRef]
20. M. G. Moharam and T. K. Gaylord, “Diffraction analysis of dielectric surface-relief gratings,” J. Opt. Soc. Am. 72(10), 1385–1392 (1982). [CrossRef]
21. J. Želudevičius, R. Danilevičius, K. Viskontas, N. Rusteika, and K. Regelskis, “Femtosecond fiber CPA system based on picosecond master oscillator and power amplifier with CCC fiber,” Opt. Express 21(5), 5338–5345 (2013). [CrossRef] [PubMed]