High optical performance coatings prepared by a liquid deposition process have been studied with focus on the parameters playing a role on the layer stacking ability. During the development of multilayer optical coatings, defects such as cracks, scattering and a refractive index gradient could appear. In order to understand the origins of these limitations, the investigation was performed on colloidal stacks of single and multi-materials. This study has rendered it possible to define the main process parameters as well as the physical and chemical parameters of the suspensions influencing the stacking capacity. This work is a first step to obtaining evidence of a relationship between the thin film microstructure induced by deposition conditions and the ability to achieve sol-gel thick films with good optical (homogeneous) and mechanical (crack-free) properties.
© 2011 OSA
Besides the well-known physical vapor deposition (PVD) processes used for optical coating preparation, a considerable effort has been and is still being directed to the production of optical layers using a liquid deposition route . The so-called sol-gel liquid technique is a chemical process widely used for oxide material preparation. Based on room temperature chemistry, the sol-gel technique allows the synthesis of nano-sized materials dispersed in an appropriate liquid medium. The process investigated at the CEA (French Commission for Atomic Energy and Alternative Energies) has been developed to prepare coatings onto mineral or metallic substrates using colloidal oxide-based and/or inorganic-organic hybrid materials. Such a chemical process is sufficiently adjustable to develop custom-designed materials and coatings for optical components, such as laser lenses, windows or mirrors.
The LMJ megajoule-class pulsed laser facility of the CEA requires 7,000-m2 of coated area onto 10,000 large-sized optical components and, to date, several optical coating procedures have been developed to fulfil the need of antireflective (AR) coating or highly-reflective (HR) components made of quaterwave stack of colloidal-based low index and hybrid-based high index thin films . Nevertheless, a difficulty when using solution chemistry and a liquid deposition process to prepare functional coatings is to achieve thick films that maintain a high optical quality. A way to do so is to stack multilayers, thereby obtaining the required thickness for optical purposes including dielectric mirrors and polarizers.
The aim of this work has been to study physical, chemical and process parameters in order to identify those that have the largest influence on the stacking ability. First, single material stacked films were studied in order to determine which parameters could improve the optical and mechanical properties. Subsequently, the obtained results were verified on colloidal multimaterial stacks consisting of alternating layers of quarterwave-thick high and low refractive index materials for preparing highly reflective (HR) coatings with decent optical and mechanical performances. The impact of chemical and process parameters was highlighted using a criterion known as the critical thickness, i.e., the thickness value of the stack just before the deterioration of the optical properties of the coating and the occurrence of cracks.
2. Sample preparation
Coating solutions were prepared by the sol-gel technique, a chemical preparation process consisting of the hydrolysis and condensation of metallic precursors (salts or alkoxides). The inorganic polymerization reaction (nucleation and growth) was controlled by varying the chemical conditions (pH, hydrolysis ratio…) .
Colloidal silica was used as the low refractive index material. The silica suspension was a sol of amorphous silica prepared by the base-catalyzed (NH3) hydrolysis of distilled tetraorthosilicate in pure ethanol according to the Stoëber method . The silica sol consisted of monodispersed roughly spherical particles with ca. 10-nm diameter. After removing the dissolved ammonia through a boiling-off step, the sol was kept at a neutral pH, around 6. The sol had a concentration equal to 4% by weight. The low refractive index coating was obtained by spin-coating or dip-coating with the as-prepared colloidal suspension. A layer refractive index equal to 1.22 was found.
For the high refractive layer, the sol was a colloidal zirconia suspension. It was prepared by urea neutralization of zirconium oxychloride in an aqueous medium followed by hydrothermal crystallization through a method similar to the one described by Somiya et al. . The crystalline powder was a mixture of monoclinic and tetragonal phases. The aqueous suspension was concentrated under vacuum until it reached 40% oxide by weight, after which, it was diluted by the addition of pure methanol to obtain a mixture with 20% water. To increase the percentage of methanol in the sol, it was necessary to carry out dialysis in methanol. The percentage of methanol was controlled by a tensiometry measurement. These sols were then deposited by spin-coating or dip-coating. A layer refractive index equal to 1.63 was found.
3. Experimental procedure
Preparing multilayer optical films is known to generate stressed coatings that can lead to cause crazing, cracks or peeling effect. Indeed, reported limits of the PVD and the sol-gel deposition processes mainly involve the difficulty of achieving thick coatings [6–9] and to build multilayer stacks [6, 10,11]. For this reason, the conditions to enable the thickness increase of deposited sol-gel films were studied here. These conditions were related to the process as well as to physical and chemical parameters. The determination of these parameters was then correlated to the critical thickness, i.e., the stack thickness before the occurrence of cracks. To determine the critical thickness and the refractive index of the coating, UV-vis spectroscopy was used and the occurrence of cracks was verified by optical microscopy  (see Fig. 1 ).
The transmission measurements were performed using a Perkin Elmer Lambda 900 spectrophotometer at normal incidence equipped with an integrating sphere that collects transmitted and scattered light. The method to determine the film thickness and refractive indexes was based on Fresnel’s law [13-14]. For a system with a single layer deposited on the substrate, the amplitude of the Fresnel oscillations of the transmission spectrum with the wavelength depends on the difference in refractive index between the substrate ns and coating nc. In further section, the wavelength extreme of the transmission spectrum is named λ and is defined as a maximum for low index coating (nc <ns) and as a minimum for high index coating (nc>ns).The refractive index of the non-absorbing substrate ns is calculated using the reflection values of bare substrate Rs(λ) at extreme position λ in Eq. (1).Eq. (2).Eq. (5) for both-sided coated substrate or by Eq. (6) for one-sided coated substrate:
Therefore, the physical thickness of the deposited film is calculated using Eq. (7):
Thus, the physical layer thicknesses e of deposited stacks is determined using simple transmission measurements of bare and coated substrate.
4. Results and discussion
For optical thin-film deposition, the methods usually used with liquids are spin-coating and dip-coating. These techniques differ with regard to drying conditions and film formation. Therefore, two laws exist for fitting the thickness of the deposited film: the law of Meyerhöfer  for spin-coating and that of Landau-levitch for dip-coating.
Here, η is the viscosity, ω is the rotation velocity and m and m0 are respectively the total mass of the sol and the mass of the solvent. These means correspond to the sol concentration and the rotation velocity.
For the single material stacks, one can plot the critical thickness and the sol viscosity for different sol concentrations (Fig. 2 ). Two regimes are observed. On the one hand, for the diluted sols, the viscosity remained constant, but the critical thickness varied significantly. On the other hand, for the concentrated sols, the viscosity was then increased, but the critical thickness showed only a slight variation. Consequently, the critical thickness was mainly impacted by the sol concentration, rather than the viscosity of the sol.
Moreover, we observe, in Fig. 3 , a similar behavior when varying the rotation velocity. For thick films with a thickness value above 150nm for the each deposited single layer, regardless of the means that were used to vary the deposited thickness (concentration or velocity), the single layer thickness had only a slight impact on the critical thickness. On the other hand, for thin films (<150nm), the deposited thickness had a significant impact on the critical thickness. The most important parameter for controlling the critical thickness was found to be the deposited thickness, showing a threshold effect around 150nm in thickness.
It was furthermore determined that the drying conditions played a role with regard to the occurrence of cracks. The process, in turn, had an impact on the drying conditions. For example, the drying time for the spin-coating was approximately ten seconds whereas for the dip-coating it was more than one minute for the same deposited solution. The critical thickness was also investigated as a function of the process used to prepare the coatings. Figure 4 presents the influence of the process on the critical thickness.
For dip-coating, the critical thickness was systematically smaller than that obtained with spin-coating. In the same manner, when varying the nature of the solvent with a different stack material (zirconia), it was observed that a short drying time was better for increasing the critical thickness. Table 1 gives results of the variation of the critical thickness with different water content in colloidal zirconia suspensions.
For instance, for a stack consisting of colloidal zirconia with a hydroalcoholic solvent containing 17% water, the critical thickness reached 3 µm whereas for the same stack prepared with only a 2% water-containing solvent, the critical thickness increased up to 9 µm. In summary, a short drying time was favorable to prepare thick coatings.
To fabricate high-reflective mirrors, it is necessary to alternately pile up materials with a low refractive index (silica film) and those with a high refractive index (zirconia film). For this reason, the influence of a short drying time and the use of a small thickness single layer on a (SiO2/ZrO2) stack was investigated. These results are summarized in Table 2 . Thin layers were found to have a slight impact on the critical thickness of multimaterials stacks. The critical thickness was increased from 2.8 µm to 3.2 µm. In addition, a short drying time (related to the nature of the solvent) rendered it possible to enhance the critical thickness even further.
For a mirrored stack comprised of a solvent with 20% water containing colloidal zirconia, the critical thickness was 2.8 µm whereas for a stack made with a 2% water concentration sol, the critical thickness was increased up to 3.8 µm. Using this operating procedure, a highly reflective mirror has been obtained and Fig. 5 shows that this multimaterials stack has a transmission coefficient below 2% at a wavelength of 1053 nm, corresponding to a reflectivity value above 98% neglecting absorption of as-prepared colloidal materials . Thus, results obtained for single material stacks were also valid for their multimaterial counterparts.
Concerning the optical quality of multilayer single material stacks, it was observe for the silica stacks that the transmission significantly decreased at low wavelengths with concentrated sols (Fig. 6 ). This decrease in transmission of the coating’s optical function in comparison with that of the substrate was attributed to light scattering. The higher the concentration, the higher was the scattering. Moreover, the greater the number of interfaces, the lower was the scattering. Scattering was thus not related to the number of interfaces but to the deposited single layer thickness.
From AFM measurements, it was noticed that the thicker the deposited layer, the lower was the roughness value (Fig. 7 and Table 3 ). Thus, one can expect that thicker deposited layer lead to smaller surface scattering effect.
Consequently, one can conclude that observed scattering for thick sol-gel deposited stacks is mainly due to volume part scattering (related to bulk layer structure) rather than surface and interface scattering. The parameters required to obtain the highest critical thickness coating and a good optical performance include higher drying kinetics and the smallest possible deposited thickness, thus leading to a film with a more homogeneous microstructure.
Numerous sol-gel preparation and deposition parameters were studied to determine the main conditions leading to thick films with good optical and mechanical properties. It was noticed that the thickness of deposited single layers and film drying kinetics were the most important parameters to control in order to obtain decent optical performances and crack-free coatings. The control of the microstructure (through drying and deposited thickness) was essential to obtain high stacking films. Very good optical performances (R>98%) were found with a colloidal stack. However, further optical and mechanical characterizations such as M-lines, ultra-nanoindentation, LSAW  are in progress to complete these results and describe the mechanism leading to crack occurence. These results should then be taken advantage of in order to fabricate multilayer optical stacks enabling the preparation of high-performance mirrors and polarizers.
References and links
1. C. Sanchez, P. Belleville, M. Popall, and L. Nicole, “Applications of advanced hybrid organic-inorganic nanomaterials: from laboratory to market,” Chem. Soc. Rev. 40(2), 696–753 (2011). [CrossRef] [PubMed]
2. P. Belleville, P. Prené, C. Bonnin, and M. Montouillout, “Use of sol-gel hybrids for laser optical thin films,” MRS Proceedings 726, 365–380 (2002).
3. C. J. Brinker and G. W. Scherer, Sol-gel science (Academic Press, San Diego, CA 1990).
4. W. Stöber, A. Fink, and E. Bohn, “Controlled growth of monodisperse silica spheres in the micron size range,” J. Colloid Interface Sci. 26(1), 62–69 (1968). [CrossRef]
5. S. Somaya, “Hydrothermal processing of ultrafine single-crystal zirconia and hafnia powders with homogeneous dopants,” in Advances in Ceramics G.L. Messing and a. others, eds. (Wersterville, Ohio, 1987).
7. H. Kozuka, S. Takenaka, H. Tokita, T. Hirano, Y. Higashi, and T. Hamatani, “Stress and Cracks in Gel-Derived Ceramic Coatings and Thick Film Formation,” J. Sol-Gel Sci. Technol. 26(1/3), 681–686 (2003). [CrossRef]
8. H. Kozuka, M. Kajimura, T. Hirano, and K. Katayama, “Crack-Free, Thick Ceramic Coating Films via Non-Repetitive Dip-Coating Using Polyvinylpyrrolidone as Stress-Relaxing Agent,” J. Sol-Gel Sci. Technol. 19(1/3), 205–209 (2000). [CrossRef]
9. R. Brenier, C. Urlacher, J. Mugnier, and M. Brunel, “Stress development in amorphous zirconium oxide films prepared by sol-gel processing,” Thin Solid Films 338(1-2), 136–141 (1999). [CrossRef]
11. A. Mehner, W. Datchary, N. Bleil, H. W. Zoch, M. Klopfstein, and D. Lucca, “The Influence of Processing on Crack Formation, Microstructure, Density and Hardness of Sol-Gel Derived Zirconia Films,” J. Sol-Gel Sci. Technol. 36(1), 25–32 (2005). [CrossRef]
12. M. A. Villegas, M. Aparicio, and A. Durán, “Thick sol-gel coatings based on the B2O3-SiO2 system,” J. Non-Cryst. Solids 218, 146–150 (1997). [CrossRef]
13. H. A. Mc Leod, Thin film optical filters, third edition ed. (Institute of physics publishing, 2001).
14. J. L. Rood, “Some Properties of Thin Evaporated Films on Glass,” J. Opt. Soc. Am. 39(10), 854–859 (1949). [CrossRef]
15. D. Meyerhofer, “Characteristics of resist films produced by spinning,” J. Appl. Phys. 49(7), 3993–3997 (1978). [CrossRef]
16. L. D. Landau and V. G. Levich, “Dragging of a liquid film by a moving plate,” Acta Physicochim. URSS 17, 41 (1942).
17. P. M. Pegon, C. V. Germain, Y. R. Rorato, P. F. Belleville, and E. Lavastre, “Large-area sol-gel optical coatings for the Megajoule Laser prototype,” SPIE 5250, 2004, 170–181.
18. C. M. Flannery, C. Murray, I. Streiter, and S. E. Schulz, “Characterization of thin-film aerogel porosity and stiffness with laser-generated surface acoustic waves,” Thin Solid Films 388(1-2), 1–4 (2001). [CrossRef]