Nanorough silica coatings by chemical vapor deposition

Abstract

Control over the growth of nanoscale roughness on surfaces and interfaces is critical for applications ranging from tunable wettability to anti-reflective coatings. Herein, we investigate the growth of nanoscale roughness on silicon substrates. Nanoscale roughness is introduced by chemical vapor deposition (CVD) of tetraethoxysilane under mild conditions and quantified by scanning force microscopy. We study the dependence of nanoroughness on the surface properties of the substrate before CVD, in particular on annealing and plasma cleaning, while taking an untreated substrate as reference. Nanoroughness is highest on annealed samples and always increases over longer CVD time. Furthermore, we report the time-dependent growth of the thickness of the silica layer. These results contribute to understanding of the role of surface pretreatment in the context of growing silica on a substrate.

---

Introduction

Nanoscale roughness plays a vital role in controlling wetting,^1,2 and optical³ and mechanical properties⁴ of particles and substrates.^1,5 Depending on the application, low⁶ or high roughness⁷ is required. Smooth surfaces are used to induce surface order of polymers, lipids or liquid crystals.^6,8 Increasing the nanoscale roughness decreases adhesion^4,9 and allows the alignment or aggregation of particles by tuning depletion attraction.¹⁰ By conferring roughness in the micro- and nanoscale, the wetting of surfaces can be tuned to prepare water-repellent (superhydrophobic),^2b superwetting,¹¹ or antifogging¹² materials. In order to tune the contact area of a liquid with a surface and to control the pinning strength, the role of roughness over different characteristic length scales has been studied. The role of spherical nanoasperities or nanofibers on wetting and pinning of flat and rough surfaces has been investigated.^1,5,13 Depending on whether the liquid penetrates the rough substrate or not, adhesion may increase or decrease, respectively. It was suggested that when water sits on top of the hydrophobic nanoasperities, there is a suppression of penetration of water into the pores of the fibrous structure. Thus nanoroughness aids in superhydrophobicity.^5b

However, apart from roughness the chemical properties of the surface layer are important to tune its wettability and mechanical stability. Silica is an especially favorable coating material, as a variety of substrates can be coated by silica making use of a Stöber reaction,¹⁴ sputtering, or chemical vapor deposition (CVD).^5b The thickness of the coating can be tuned by the process conditions.¹⁵ Silica can easily be chemically modified and different methods are established to chemically bind or adsorb molecules to its surfaces.^15a,16 Silica can also be used to coat particles or fibers with a shell of thickness of a few tens of nm to improve their mechanical properties or to chemically bind neighboring particles together or to the substrate.¹⁷

CVD is one of the most common ways to apply silica coatings. In general CVD can be performed under acidic or basic conditions, using, for example, hydrochloric acid or ammonia as catalyst.¹⁸ Depending on the conditions, largely different morphologies can be achieved. Special attention received the formation of one-dimensional silicon nanofibers that formed under acidic conditions on a wide number of surfaces.^13,19 CVD of tetraethoxysilane catalyzed by aqueous ammonia is a mild method to introduce nanoroughness,^5b to increase the mechanical strength of a coating²⁰ or to introduce reactive binding sites at the surface of silica. It can even be used to coat polymer-based surfaces with a nanorough silica shell.^5b

Despite widespread application of CVD for silica coatings, the growth behavior of silica formed by CVD at ambient temperature is poorly characterized.^13a,b In this work, we aim for a better understanding of the evolution of this nanoscale roughness during the CVD of tetraethoxysilane (TES), striving for a precise control of the thickness and roughness of the silica layer. CVD of TES depends on the existence of reactive binding sites at the substrate, i.e. silanol groups and two-membered silicon rings.^14b,21 To vary the density of binding sites on the silicon surface, we adopted different standard pretreatment procedures before growing silica through CVD. We show that the root mean square (RMS) roughness and thickness of these thin silica films can be tailored based on the specific pretreatment procedure and duration of CVD.

Experimental

Materials

Tetraethoxysilane (TES) (98%) and tetrahydrofuran (THF) (99.9%) were purchased from Sigma Aldrich, ammonia solution (28% in water) from VWR International. Polystyrene (PS) (average molecular weight 66 000 g mol⁻¹, polydispersity index = 1.05) was synthesized by anionic polymerization. Si (100) wafers purchased from Si-Mat, Germany, were used as substrates. Millipore system operated at 18.2 MΩ cm was used for providing Milli-Q water. The wafers were cut into 2.5 × 1.5 cm² substrates and were systematically cleaned by sonicating twice using a 2% solution of Hellmanex (Hellma GmbH), rinsing with Milli-Q water and drying under nitrogen for a minimum of 30 minutes.

Substrate pre-treatment

The cleaned Si wafers were pre-treated in three different ways to vary the surface density of binding sites: they were annealed, treated with argon plasma or used as such. Pre-annealing of the wafers was carried out at 500 °C for 3 hours in a furnace under ambient air (Thermo Scientific, Germany). Plasma cleaning was applied to achieve a hydrophilic surface by removal of organic residues of the wafer. The plasma chamber was evacuated and flushed with argon three times before administering the plasma in order to ensure a high percentage of argon in the ambient. The plasma cleaner (Harrick Plasma, USA) was operated at a power of 300 Watt under argon atmosphere (20 mbar, 5 minutes). In a third set of experiments the untreated Si wafers were used immediately after cleaning. We verified that the plasma pretreatments did not change the roughness of the substrates by imaging the substrates before and after plasma treatment using scanning force microscopy. Indeed, we observed no differences in the RMS roughness after the plasma treatments.

Chemical vapor deposition (CVD) of TES

The cleaned pre- or untreated silicon wafers were placed in a desiccator together with two open vials, one containing 1 ml of TES and the second 1 ml of aqueous NH₃. CVD of TES was carried out at ambient temperature for time durations between 0 and 48 h. Silica was formed by hydrolysis and condensation of TES vapor in the presence of aqueous ammonia as a catalyst in analogy to the Stöber procedure.²²

Height profile and root mean square (RMS) roughness measurements

The flow chart (Fig. 1) illustrates the method we used to prepare surfaces coated with a silica film of varied thickness and roughness. The Si wafers were prepared as described above and were then half-dipped into a PS–THF solution (5 g of PS dissolved in 20 ml of THF) for about 10 seconds to coat part of the wafer with a PS “shield” layer. After letting THF evaporate for one minute, part of the wafer was coated with a thin layer of PS. Thereafter, the wafers were placed in a desiccator, together with an open glass filled with 1 ml TES and a second open glass filled with 1 ml of ammonia for various time periods between 0 and 48 h. Silica grows only on the uncoated part of the wafer.^5b Finally, the PS shield layer was removed by combustion at 500 °C. The silica film thickness was determined by scanning force microscopy (SFM) over a 30 × 30 μm² area in the vicinity of the step created by the PS shield layer. Line profiles spanning this area were used to quantify the film thickness. Root mean square (RMS) roughness of the silica films was measured after different time periods of CVD of TES by SFM over a 1 × 1 μm² area.

Fig. 1 Fabrication of nanorough silica-coated Si wafers via chemical vapor deposition, allowing for measurement of thickness and root mean square roughness of the silica film.

RMS roughness measurements were also performed on silicon wafers that were not half-coated with a PS shield layer. Hereby, we exclude possible contributions from the precedent PS–THF coating step. In this case, the Si wafers prepared under different pretreatment conditions were directly used for CVD of TES.

Characterization

SFM measurements of film thickness and topography were obtained using a Dimension 3100 CL scanning probe microscope from Digital Instruments (Veeco Metrology Group, Plainview, NY). Imaging under tapping-mode in air was conducted at 24 °C ± 2 °C using silicon SFM tips (OMCL-AC TS, Olympus, Japan) with a resonance frequency around 300 kHz and a nominal tip radius less than 10 nm, working in closed loop mode. Larger 30 × 30 μm² images at a scan rate of 0.2 Hz were recorded to estimate the film thicknesses while 1 × 1 μm² images were taken at 0.7 Hz to investigate the surface morphology. Line profiles over the 30 × 30 μm² scan area were used to quantify the film thickness. Each data point is the arithmetic mean of 5 thickness values and the error bar is the standard deviation of 3 individual measurements on different regions on one sample. Root mean square (RMS) roughness values were calculated over a 1 × 1 μm² scan area. Reported RMS roughness values are the average over 3 different regions on a given sample and the error is the standard deviation.

Results and discussion

First, we studied the role of the pretreatment on the thickness of the silica coating, as it affects the number of silanol groups which react with TES.^14b,21,23 To this end, the cleaned wafers were annealed, treated with argon plasma or used as such. We determined the thickness of the silica film by measuring the height of a step created by growing silica on one half of a silicon wafer while shielding the second half with a layer of polystyrene. Ammonia catalyzes the reaction of tetraethoxysilane to silica in analogy to the Stöber¹⁵ procedure, resulting in the formation of a homogeneous and rough silica layer devoid of cracks and fissures (Fig. 2a). Finally, the PS shield layer was removed by combustion at 500 °C.

Fig. 2 (a) SFM image of a Si wafer where silica was grown on one part of the wafer. The color scale corresponds to height. (b) The height profile across a step as outlined by the white line in (a). The profile was averaged over 15 adjacent lines and resulted here in a height of 24.8 nm.

The SFM images show a step profile, where the height of the step increases with duration of CVD of TES. Fig. 2a shows a representative image, used to determine the film thickness created after 24 h of CVD of TES on an untreated silicon wafer. The brighter side (light brown, left) and the darker side (brown, right) denote the silica coating and bare silicon wafer respectively. Notably, the use of a polystyrene shield layer resulted in films with a clean and sharp edge without any delamination. The polystyrene layer ideally prevents silica growth.^5b The height profile across a step resulted here in a height of 24.8 nm.

Independently of the pretreatment the thickness of the silica layer increases with the duration of CVD of TES (Fig. 3). We could not measure the thickness of the silica films for times shorter than 6 hours, as the thickness was only a few nm, i.e., comparable to the roughness caused by CVD of TES. During this initial incubation period, single molecules adsorb to the wafer. Neighboring molecules may grow together and molecules will condense from the vapor phase, likely resulting in the formation of terraces,²⁴ small particles or fibers.¹ With increasing duration of CVD more molecules adsorb to the terraces, nanoparticles or nanofibers resulting in a steady increase of the thickness (Fig. 3).^5b,25 Si wafers pretreated with argon plasma show the fastest growth rate (2.1 ± 0.3 nm h⁻¹) and reach a film thickness of 102.8 ± 4.8 nm after 42 h of CVD. In contrast, the annealed and untreated wafers show slightly lower growth rates (1.4 ± 0.3 nm h⁻¹ and 1.1 ± 0.3 nm h⁻¹, respectively) and reach final thickness values of 49.2 ± 2.4 nm and 67.9 ± 6.9 nm, respectively after 42 h. The higher film thickness for argon plasma pretreated samples can be explained through the higher density of silanol¹⁴ groups leading to a quick monolayer formation and subsequent film growth. The thickness of the annealed amorphous silica samples is almost identical to those of the untreated wafers. It has been shown that on normal silica surfaces exposed to air, silanol groups are the main adsorption sites for the adsorption of water and other chemicals.²⁶ Annealing causes condensation of neighboring silanols (Si–OH), leading to the formation of two-membered (2M) silicon rings. However, already the presence of a thin layer of condensed water or ammonia is sufficient to induce hydroxylation and thereby formation of new silanol sites. Indeed, ab initio molecular dynamic simulations show that water and ammonia molecules physisorb on the acidic silicon atoms of the 2M silicon ring.^21,23 Successive chemisorption of ammonia or water proceeds via an electrophilic attack of the oxygen atom in the ring, inducing hydroxylation and formation of new silanol groups.

Fig. 3 Dependence of film thickness on the duration of chemical vapor deposition of tetraethoxysilane for different pretreated samples. The dotted lines are guides to the eye.

Aiming to relate the thickness of the step to the roughness of the surface we monitored the evolution of nanoroughness with increasing CVD times (0–48 h) by imaging the surface using scanning force microscopy. Fig. 4 shows images of silicon wafers with different pretreatment procedures after 26 h of CVD of TES. Silica nanoasperities with a rounded morphology and a rough topography (bright spots), irregularly distributed on the film, are evident for all the samples. The average diameter of the almost spherical nanoasperities is slightly larger (∼20 nm) for the argon plasma pretreated and pre-annealed samples (Fig. 4a and b) as compared to the untreated samples (∼15 nm, Fig. 4c). We measured a height difference of about 20 nm between the tip of the nanoasperities and the valleys (underlying silica film) for the pre-annealed and untreated samples while the Ar plasma pretreated samples showed a slightly lower value of 13 nm. Note however, this height difference might be affected by the finite size of the cantilever and should therefore be considered as a lower limit of the maximal height differences.

Fig. 4 SFM images taken on a Si wafer subjected to (a) pre-annealing (b) Ar plasma and (c) no pretreatment followed by 26 h of CVD of TES (independent of PS–THF coating). The RMS roughness values were (a) 3.61 nm, (b) 2.13 nm and (c) 3.22 nm.

To gain information on the growth mechanism we investigated the time evolution of nanoroughness for different periods of CVD of TES (Fig. 5). To quantify the nanoroughness, we determined the root mean square roughness over a 1 × 1 μm² area. Independently of the type of pretreatment, the RMS roughness increases with duration of CVD of TES. The pre-annealed wafers showed the highest roughness with a RMS roughness of 1.3 nm after 24 h and 2.8 nm after 42 h. The untreated and argon plasma pre-treated samples also showed increased roughness over longer CVD time. Note, that all samples remain comparatively smooth; even after 42 h of reaction time the RMS roughness was much lower than the lateral dimensions of the nanoasperities (Fig. 4).

Fig. 5 Dependence of RMS roughness on the duration of CVD of TES. The samples were exposed to THF vapor.

Surfaces that were not half-coated with PS and therefore did not come in contact with THF vapor show slightly higher roughness (2.2 nm after 24 h of CVD of TES).^5b Therefore, we investigated whether remaining residuals of THF can be used to modify the roughness. We repeated the previously described experiments without using the PS–THF shield layer i.e., we directly performed CVD of TES on the differently pretreated Si wafers. Indeed, the roughness of the samples increases (Fig. 6). For instance, a RMS roughness of 4.3 nm (pre-annealed), 3.1 nm (untreated) and 1.7 nm (Ar plasma) were observed after 26 h of CVD (Fig. 6). This implies that THF gives rise to a more homogeneous silica nucleation. We expect that THF activates the surface twofold: THF as a Lewis base may coordinate to the Si-atoms of remaining 2M silicon rings inducing a ring opening followed by an electrophilic attack of a TES molecule to the oxygen atom of the opened 2M ring. Secondly, THF forms hydrogen bonds with the silanol groups activating the oxygen atom of the silanol group for an electrophilic attack of a TES molecule. As soon as the substrate is completely coated with silica, the nucleation rate and density changes, giving rise to a different growth behaviour of the silica coating. For comparison we investigated the surface roughness in excess catalyst, i.e. aqueous ammonia (10 ml instead of 1 ml). The RMS roughness continuously increases (Fig. 6) and is much higher than for samples that were in contact with THF. The difference is especially pronounced for the annealed samples showing the lowest density of silanol groups if not in contact with THF. Contrary, the RMS roughness only weakly increases for the Ar plasma treated samples, as Ar activates the surfaces resulting in the highest density of silanol groups. This results in a more homogeneous adsorption, nucleation and growth of tetraethoxysilane. Notably, after 24 h the RMS roughness of the untreated samples that were not in contact with THF is within experimental accuracy identical to those measured for samples with less ammonia (1 ml aqueous ammonia).^5b This might be due to more silanol groups on the surface of the silicon wafer.

Fig. 6 Dependence of the RMS roughness after different periods of CVD of TES on the surface pretreatment. The samples were not exposed to THF vapor (vapor pressure of THF: 173 mbar at 20 °C). The samples were placed in a desiccator together with two open bottles, containing 1 ml of TES and 10 ml ammonia, respectively for varied periods of time. Dashed lines: least square fits through the data points.

This data also shed light on the formation of fibers under acidic conditions. So far the formation of fibers has only been reported under acidic conditions.¹ It was speculated that the reaction speed may cause the difference of the growth mechanisms, which might be too fast under basic conditions. Here, even the growth rate is below those observed under acidic conditions. The surfaces remain comparatively smooth and no fiber formation is observed, independent on the density of silanol groups at the surface. This hints that fiber formation is not caused by the slow reaction speed but that the growth mechanism of silica under basic and acidic conditions follows different reaction paths. This is also supported by the observation that polymerization of silicic acid at acidic pH values leads to chaining of nanometer sized particles, whereas no chaining has been observed for a pH value below 7.^18a

Conclusion

We studied the growth and roughness of amorphous silica thin films prepared by chemical vapor deposition of tetraethoxysilane at ambient temperature, a mild procedure to coat a surface with silica. The thickness of the homogeneous films increases linearly with the duration of CVD of TES. The growth rate depends on the pretreatment, with the fastest rate for Ar plasma pretreated samples. Likely, this is due to a greater density of binding sites i.e. silanol groups compared to untreated and annealed ones. We examined the growth of nanoscale roughness with CVD time and wafer pretreatments. Nanoscale roughness increases with reaction time irrespective of the kind of pretreatment. Annealing produced a surface with highest roughness possibly due to an inhomogeneous nucleation and growth process owing to fewer binding sites. Presence of THF residuals gives rise to smoother films. We expect that THF as a Lewis base activates nucleation. Our results demonstrate how strongly the growth behavior is determined by both the surface properties of the substrate and the reaction conditions. To identify and control the growth behavior and therefore the morphology of the coating a deeper understanding of underlying molecular processes is required.

Acknowledgements

We are grateful to G. Glasser, U. Rietzler, G. Schäfer, D. Donadio, and R. Berger for technical support and stimulating discussions. D.V. acknowledge financial support from SPP 1273, L.M. from SPP 1486.

References

1. J. Zimmermann, F. A. Reifler, U. Schrade, G. R. J. Artus and S. Seeger, Long term environmental durability of a superhydrophobic silicone nanofilament coating, Colloids Surf., A, 2007, 302(1–3), 234–240.
2. (a) A. Y. Fadeev and T. J. McCarthy, Trialkylsilane monolayers covalently attached to silicon surfaces: Wettability studies indicating that molecular topography contributes to contact angle hysteresis, Langmuir, 1999, 15(11), 3759–3766; (b) D. Quere, Wetting and roughness, Annu. Rev. Mater. Res., 2008, 38, 71–99; (c) R. Shuttleworth and G. L. J. Bailey, The Spreading of a Liquid over a Rough Solid, Trans. Faraday Soc., 1948, 44(3), 16–22.
3. (a) E. Yu, H. J. Lee, T. J. Ko, S. J. Kim, K. R. Lee, K. H. Oh and M. W. Moon, Hierarchical structures of AlOOH nanoflakes nested on Si nanopillars with anti-reflectance and superhydrophobicity, Nanoscale, 2013, 5(20), 10014–10021; (b) G. Zhou and J. H. He, Antireflective Coatings on Fresnel Lenses by Spin-Coating of Solid Silica Nanoparticles, J. Nanosci. Nanotechnol., 2013, 13(8), 5534–5541.
4. M. Gotzinger and W. Peukert, Particle adhesion force distributions on rough surfaces, Langmuir, 2004, 20(13), 5298–5303.
5. (a) T. Verho, J. T. Korhonen, L. Sainiemi, V. Jokinen, C. Bower, K. Franze, S. Franssila, P. Andrew, O. Ikkala and R. H. A. Ras, Reversible switching between superhydrophobic states on a hierarchically structured surface, Proc. Natl. Acad. Sci. U. S. A., 2012, 109(26), 10210–10213; (b) L. Mammen, X. Deng, M. Untch, D. Vijayshankar, P. Papadopoulos, R. Berger, E. Riccardi, F. Leroy and D. Vollmer, Effect of nanoroughness on highly hydrophobic and superhydrophobic coatings, Langmuir, 2012, 28(42), 15005–15014.
6. S. P. Pujari, L. Scheres, T. Weidner, J. E. Baio, M. A. C. Stuart, C. J. M. van Rijn and H. Zuilhof, Covalently Attached Organic Mono Layers onto Silicon Carbide from 1-Alkynes: Molecular Structure and Tribological Properties, Langmuir, 2013, 29(12), 4019–4031.
7. M. Ohmori and E. Matijevic, Preparation And Properties Of Uniform Coated Colloidal Particles. 7. Silica On Hematite, J. Colloid Interface Sci., 1992, 150(2), 594–598.
8. L. Zhang, M. Kappl, G. K. Auernhammer, B. Ullrich, H.-J. Butt and D. Vollmer, Surface-Induced Ordering of Liquid Crystal on Modified Surfaces, in Surface and Interfacial Forces – from Fundamentals to Applications, ed. G. K. Auernhammer, H. J. Butt and D. Vollmer, 2008, vol. 134, pp. 39–47.
9. Y. I. Rabinovich, J. J. Adler, M. S. Esayanur, A. Ata, R. K. Singh and B. M. Moudgil, Capillary forces between surfaces with nanoscale roughness, Adv. Colloid Interface Sci., 2002, 96(1–3), 213–230.
10. (a) K. Zhao and T. G. Mason, Suppressing and enhancing depletion attractions between surfaces roughened by asperities, Phys. Rev. Lett., 2008, 101(14), 148301; (b) K. Zhao and T. G. Mason, Directing colloidal self-assembly through roughness-controlled depletion attractions, Phys. Rev. Lett., 2007, 99(26), 268301.
11. X. Feng and L. Jiang, Design and creation of superwetting/antiwetting surfaces, Adv. Mater., 2006, 18(23), 3063–3078.
12. X. F. Gao, X. Yan, X. Yao, L. Xu, K. Zhang, J. H. Zhang, B. Yang and L. Jiang, The dry-style antifogging properties of mosquito compound eyes and artificial analogues prepared by soft lithography, Adv. Mater., 2007, 19(17), 2213–2217.
13. (a) L. Gao and T. J. McCarthy, (CH₃)(3)SiCl/SiCl₄ azeotrope grows superhydrophobic nanofilaments, Langmuir, 2008, 24(2), 362–364; (b) D.-a. E. Rollings, S. Tsoi, J. C. Sit and J. G. C. Veinot, Formation and aqueous surface wettability of polysiloxane nanofibers prepared via surface initiated, vapor-phase polymerization of organotrichlorosilanes, Langmuir, 2007, 23(10), 5275–5278; (c) G. R. J. Artus, S. Jung, J. Zimmermann, H.-P. Gautschi, K. Marquardt and S. Seeger, Silicone nanofilaments and their application as superhydrophobic coating, Adv. Mater., 2006, 18(20), 2758–2762.
14. (a) M. D'Acunzi, L. Mammen, M. Singh, X. Deng, M. Roth, G. K. Auernhammer, H.-J. Butt and D. Vollmer, Superhydrophobic surfaces by hybrid raspberry-like particles, Faraday Discuss., 2010, 146, 35–48; (b) L. Zhang, M. D'Acunzi, M. Kappl, A. Imhof, A. van Blaaderen, H. J. Butt, R. Graf and D. Vollmer, Tuning the mechanical properties of silica microcapsules, Phys. Chem. Chem. Phys., 2010, 12(47), 15392–15398; (c) C. Graf, D. L. J. Vossen, A. Imhof and A. van Blaaderen, A general method to coat colloidal particles with silica, Langmuir, 2003, 19, 6693–6700.
15. (a) A. Van Blaaderen and A. P. M. Kentgens, Particle morphology and chemical microstructure of colloidal silica spheres made from alkoxysilanes, J. Non-Cryst. Solids, 1992, 149(3), 161–178; (b) A. Padermshoke, S. Konishi, M. Ara, H. Tada and T.-A. Ishibashi, Novel SiO₂-Deposited CaF₂ Substrate for Vibrational Sum-Frequency Generation (SFG) Measurements of Chemisorbed Mono Layers in an Aqueous Environment, Appl. Spectrosc., 2012, 66(6), 711–718.
16. S. P. Pujari, E. Spruijt, M. A. C. Stuart, C. J. M. van Rijn, J. M. J. Paulusse and H. Zuilhof, Ultralow Adhesion and Friction of Fluoro-Hydro Alkyne-Derived Self-Assembled Monolayers on H-Terminated Si(111), Langmuir, 2012, 28(51), 17690–17700.
17. X. Deng, L. Mammen, Y. Zhao, P. Lellig, K. Muellen, C. Li, H.-J. Butt and D. Vollmer, Transparent, Thermally Stable and Mechanically Robust Superhydrophobic Surfaces Made from Porous Silica Capsules, Adv. Mater., 2011, 23(26), 2962–2965.
18. (a) R. K. Iler, The chemistry of silica, Wiley, 1979; (b) E. P. Plueddemann, Silylated surfaces, Gordon and Breach, 1980.
19. J. Zimmermann, F. A. Reifler, G. Fortunato, L.-C. Gerhardt and S. Seeger, A Simple, One-Step Approach to Durable and Robust Superhydrophobic Textiles, Adv. Funct. Mater., 2008, 18(22), 3662–3669.
20. (a) X. Deng, L. Mammen, H. J. Butt and D. Vollmer, Candle Soot as a Template for a Transparent Robust Superamphiphobic Coating, Science, 2012, 335(6064), 67–70; (b) X. Deng, L. Mammen, Y. F. Zhao, P. Lellig, K. Mullen, C. Li, H. J. Butt and D. Vollmer, Transparent, Thermally Stable and Mechanically Robust Superhydrophobic Surfaces Made from Porous Silica Capsules, Adv. Mater., 2011, 23(26), 2962–2965; (c) L. Mammen, P. Papadopoulos, K. Friedemann, S. Wanka, D. Crespy, D. Vollmer and H. J. Butt, Transparent and airtight silica nano- and microchannels with uniform tubular cross-section, Soft Matter, 2013, 9, 9824–9832.
21. F. Zipoli, D. Donadio and M. Bernasconi, Simulation of the grafting of organosilanes at the surface of dry amorphous silica, J. Phys.: Condens. Matter, 2008, 20, 224011.
22. W. Stöber, A. Fink and E. Bohn, Controlled Growth of Monodisperse Silica Spheres in Micron Size Range, J. Colloid Interface Sci., 1968, 26(1), 62–69.
23. P. Masini and M. Bernasconi, Ab initio simulations of hydroxylation and dehydroxylation reactions at surfaces: amorphous silica and brucite, J. Phys.: Condens. Matter, 2002, 14(16), 4133–4144.
24. E. Carlon and H. vanBeijeren, Networks of steps on crystal surfaces, Phys. Rev. Lett., 1996, 76(22), 4191–4194.
25. K. Oura, V. G. Lifshits, A. A. Saranin and A. V. Zotov, Surface Science: An Introduction, Springer, Berlin, 2003.
26. A. P. Legrand, The surface properties of silica, Wiley, 1998.

---