Effect of sol-age on the surface and optical properties of sol-gel derived mesoporous zirconia thin films

Mesoporous films have recently attracted much attention for a wide range of applications i.e. microelectronics,¹ hosts for nanostructures^2–4 and low dielectric constant materials.⁵ The reason behind this lies in the fact that porous microstructure can enhance the surface activity, can modulate the dielectric constants and can allow the filling of variable amounts of reinforcement. Average size of pores and their distribution influence the diffusion and equilibrium of molecules adsorbed in the porous structure. Materials with pore diameter in mesoporous regime (between 2 and 50 nm) have been proven to be ideal for hosting the noble metal nanostructures for plasmonic applications, since in this size regime metal nanostructures exhibit excellent size dependent properties. Such fine requirements place strong demands on manipulation of microstructure to obtain controlled porosity and average pore size in mesoporous range with narrow size distribution. In this regard, there are tremendous efforts reported for the synthesis of mesoporous silica and titania.^6–9 Periodic ordered pores in the range of 5-30 nm has been obtained with using amphiphilic triblock copolymers.⁶ Lu et al. reported in situ evolution of the mesophase by fluorescence depolarization in silica films, in which the pores are connected in a three-dimensional network.⁷ Kluson et al. prepared mesoporous titania using a surfactant-mediated sol–gel method.⁸ The successful synthesis of mesopores with hexagonal and cubic symmetries through the use of surfactant micellar structures as template or organizing agents made these materials particularly attractive as covalent scaffolding hosts used in catalysis and drug delivery.⁹ 

The synthesis process of the porous oxide films often imposes restrictions on the utilization of films for practical applications.^10–12 For example, supercritical evacuation method provides high porous films but their strength and adhesion to the substrate is poor.¹⁰ The technique of using chemical etching agents for porous films shows very poor control on the pore-size and uniformity of pores.¹¹ The technique using the incorporation and burning out of the sacrificial components, as reported by Borrás et al.¹²may induce undesired thermal effects in the inorganic matrix during the formation of porous films. These limitations have forced researchers to develop porous films by sol-gel method, which is regarded as a relatively promising method to obtain controlled porosity by controlling the processing steps.^8,13

Zirconia (ZrO₂) is an interesting oxide material which has found wide applications such as bio-active material, thermal barrier coating, antireflection coating, optical humidity sensor and for surface with controlled hydrophilicity.^5,14–16 For the synthesis of ZrO₂ thin film, variety of techniques i.e. r.f. magnetron sputtering,¹⁷ ion beam assisted deposition,¹⁸ e-beam deposition,¹⁹ plasma spraying,²⁰ pulsed laser deposition²¹ and sol-gel^5,22,23 are used. Though the reports on sol-gel derived ZrO₂ thin films show porous nature, the effect of sol-age on the porosity and in turn on the surface and optical properties of films has not been studied in details.

In this paper, the role of pH of starting polymeric sol and duration of sol-age on the microstructure, surface and optical properties of films has been presented with the emphasis on controlled porosity.

Polymeric sol of 0.1 M was prepared by mixing zirconium n-propoxide (23-28% free alcohol, STREM chemicals) precursor in 2-propanol solution. Appropriate quantities of HCl are added for varying the pH of starting polymeric sol and controlling the hydrolysis rate. All sols are stirred for 18 h keeping the temperature at 35°C and relative humidity between 40% and 50%. These polymeric sols, referred as ZrO₂ sol throughout this paper, are kept for ageing for 1, 2, 3, 5 and 10 days. Films are deposited by a vibration-free dip coating system on glass, quartz and on KCl substrates using sol of different age. Pulling speed has been kept constant as 15 cm/min. After each deposition the film is kept for 10 min in air and then dried for 10 min at 100 °C in static air. For getting desired thickness, multiple coatings are deposited under identical conditions.

The pH of sols has been measured with Elico Analyser (model: LI 612). Film thickness was measured using a surface profiler AMBIOS-XP II. TEM and HRTEM studies were carried out on films deposited on KCl substrates, which were carefully floated on copper grid and then characterized with PHILIPS TEM CM-12. Surface topography of films was recorded in contact mode with Atomic Force Microscope (Nanoscope 3a Veeco). Optical transmittance (T) and reflectance (R) of films coated on glass/quartz substrates were measured with UV-VIS-NIR spectrophotometer (Perkin Elmer Lambda 900). Diffused reflectance was measured using the specular excluded geometry (8^o/hemispherical) in integrating sphere (150 mm spectralon) attached with the spectrophotometer.

To identify the stability of sols, pH of starting sol is optimized by observing the variation of pH with ageing time. For this, different amount of HCl is added in 5 identical sols of 0.1 M and variation in pH of these sols are measured with sol-age as shown in Fig. 1. The sols of starting pH > 3 (sol 4, and sol 5) show gelation after 2 days and hence their pH is not further monitored after 2 days. The sols prepared with high amount of HCl (>1.5 ml HCl in 100 ml ZrO₂ sol), shows starting pH <1 which remain <1 when monitored up to 6 days. These sols are found to be stable for more than two months. Hence, it is concluded that less amount of HCl leads to the fast gelation of the sol whereas high amount of HCl slows down the gelation process drastically. Sol of starting pH 0.5 is used to deposit films in present experiments.

Since, in the sol-gel process there are many sub-processes i.e. alcolysis, condensation, polymerization, shrinkage of molecular skeleton, and agglomeration, which undergo simultaneously, depending on the chemical (moisture/water) and physical (ambient temperature) environment.^24–26 The variation in pH with the ageing period can be explained on the basis of variation in viscosity of the sol.²⁷ Tomita et al. showed that the gelation time of the SiO₂ sol can be varied from 10 minutes to 3 hours by changing the pH, temperature and concentration of the surfactant added.²⁷ Pramer et al. reported that when the pH of SiO₂ sol is decreased from 6 to 4, there is systematic increase in gelation time from 40 min to 3 days.²⁸ It should be noted that in case of ZrO_2, isoelectric point of pH is more than 4, which shows the longer gelation time at pH <4 as observed in present case.

Fig. 2 shows the typical TEM micrographs of (a) ZrO₂ sol (5 days aged) dispersed on carbon coated Cu grid and of (b) film prepared with this sol. Micrograph (a) shows uniform dispersion of nanosized pores. Majority of pores are observed having diameter less than 20 nm. The pore walls are observed to collapse and leading to interconnecting the pores, when electron beam exposure is extended on selected area for more than 10 seconds. Micrograph (b) shows uniform crack free deposition of film and exhibits mesoporous microstructure. Here pore walls were stable and that is why no collapsing of pore walls or interconnecting of pores has been observed. Selected area diffraction (SAD) pattern is shown in inset of (b), which shows a bright halo which confirms that films are amorphous.

The AFM micrographs of ZrO₂ films; prepared with sol age 1, 2, 3 and 5 days are shown in Fig. 3 as (a), (b), (c) and (d), respectively. The scan area is selected as 1x1 μm and height scale is 10 nm/division. Micrographs show that films are very smooth, free from pinholes and having pores in nanometer range. The root mean square (RMS) roughness in these films is estimated to be 0.42 nm, 0.43 nm, 0.79 nm and 2.80 nm for films prepared with sol age of 1 day, 2 days, 3 days and 5 days, respectively. The variation of RMS roughness with sol age (in days) is shown in Fig. 4. It initially increases slowly but rises sharply when sol age is increased to 5 days. Detailed surface analysis is also carried out to know the asymmetry of height variation about the mean value by skewness calculation which gave us negative values -0.058 (Fig. 3(a)), -0.128 for (Fig. 3(b)), -0.041 for (Fig. 3(c)) and -0.212 for (Fig. 3(d)) in all cases, confirming that the films are porous in nature. This should be mentioned here that skewness measures positive for the flat surface with peaks, and negative for the flat surface with holes.²⁹ 

Spectral variation in T and R of films deposited on quartz substrate using 1, 3, 5 and 10 days aged sol is shown by curves (i), (ii), (iii), and (iv), respectively in Fig. 5(a). Thicknesses of these films are measured to be 88 nm, 117 nm, 160 nm and 254 nm for the sol-age 1, 3, 5 and 10 days, respectively. All the films show high T (62.4%- 90%) in entire visible and NIR region. The complimentary R values for each film exhibit that the absorptance (A=100-T-R) is negligible in visible and NIR region. The film prepared with 1 day sol-age does not show any interference maxima either in T or R, however the films prepared with higher sol-age exhibit interference fringes in T and R whose number of interference maxima increases with increasing the sol-age. This effect is ascribed to be due to the increase in thickness of the films, which was found in agreement with the measured thicknesses of films.

To identify a correlation of surface roughness of the films as observed in AFM measurement with the optical properties, DR of ZrO₂ films deposited with sol of 1, 3 and 5 days age are shown in Fig. 5(b) as curves (i), (ii) and (iii), respectively. DR values of all films are found to be <2% throughout the 300-2500 nm region, indicating that films are smooth and surfaces roughness is very less. The values of DR increase from the curves (i) to (ii) vary slowly and then significantly in curve (iii). Thus DR results suggest that surface roughness increases initially slowly up to 3 days and then at faster rate when sol-age is further increased.

Absorbance of ZrO₂ sol (not shown here) shows a peak centred on 242 nm, confirms that ZrO₂ formation starts in sol itself. The absorption coefficient (α_o) is calculated from the T and R spectra of films deposited on quartz using expression α_o =ln [(100-R)/T]/d. The optical band gap of films has been determined from the analysis of the Tauc's plots near the absorption edge. Tauc's plots of all the films are plotted for different possible transitions and best fit is obtained for direct allowed transition. Fig. 6 shows the dependence of (α_ohυ)² on hυ of films prepared with sol-age (i) 1, (ii) 3 and (iii) 5 days. All curves show an absorption edge and an absorption tail. The linear extrapolating of absorption edge gives band gap value as 5.74 eV, which is found to be same for all films. However, the absorption tail shows enhanced value of area under this region in films prepared with higher sol-age. The absorption tail is attributed to be due to the remnant hydroxyl parts in films, and higher contribution of absorption tail in films of higher sol-age indicates an enhancement in the surface area/porosity of films with increasing the sol-age.

For the porosity estimation, ZrO₂ films are dipped in aqueous solution of methylene blue (MB) and change in A of solution is measured before and after the immersion of films prepared with sol age of 1, 3 and 10 days and plotted in Fig. 7(a) with curves (i), (ii), (iii) and (iv), respectively. A values of solutions after the soaking in films were normalized to their respective thickness. The absorbance plots show two characteristic peaks of MB at 613 nm and at 664 nm. The intensity of these peaks decreases from (i) to (iv), suggesting that soaking of films adsorb the MB molecules at its porous surface. Since the adsorption is supposed to be proportional to available active surface sites, the difference in area under the curves can be associated with the surface area of mesoporous ZrO₂. Area under the peaks in 500 nm-800 nm region and the difference of area under the peaks are plotted in Fig. 7(b). This shows the enhancement in difference of area under the curves from 3.8 to 21.32 for increasing the sol age from 1 day to 10 days. The increase of 5.6 times of ΔA confirms the excellent improvement in surface area/ porosity of the films with the sol age. It should be mentioned here that however, this approach gives a fair estimation of relative porosity; it does not estimate the pore size.

In our earlier work, we have presented the change in chemical properties and composition of ZrO₂ films with sol-age by FTIR and X-ray photo electron spectroscopy.²³ The work presented in this paper provides further information on the synthesis process and role of sol-age. The hydrolysis of zirconium n-propoxide and subsequent condensation is sensitive to the solvent of the system, used catalyst and on the atmospheric conditions. To control the hydrolysis and condensation processes, various molecular organic stabilizers viz polyethylene glycol, acetylacetone, diethanolamine, and catalysts viz acetic acid, HNO₃, NaOH, NH₄OH are reported to be used.^5,12,22 In present case, the use of HCl as catalyst reduces the reactivity of zirconium n-propoxide precursor and provides transparent sol in 2-propanol without any trace of precipitation, which in turn enhances the stability of sol. The rate of gelation process is adjusted with varying HCl amount. The gelation occurs when the pH of sol gets stabilized. Thus, the use of HCl in place of other chemicals used for chelating or stabilization of sol is promising as it provides sol stabilization for months, economic in comparison of other chemicals and feasible to yield controllable porosities in ZrO₂ thin films.

2-propanol, used as solvent in present work, has less than 0.5 % water content. Hence, higher activity of alcohol dominates the alcolysis process in poly-condensation reactions. The alcolysis of primary zirconium alkoxides are polymeric oxide-alkoxides. Initially 2-propanol alters some of the bonds of the propoxide molecules, and these altered species react with each other or with unaltered species to form large molecules. In these reactions, diffusion and statistical interactions play significant roles in determining porosity, average size of particles and surface morphologies. Formed large molecules construct a network embedded in solvent and drying this sol removes the solvent and leaves the porous network as seen in Fig. 2(a). For a fixed starting pH of sol, the variation in ageing time affects the viscosity of sol, which in turn decides the thicknesses of films. Increase in sol-age also affects the branching of dispersed molecules in solvent and yields bigger pore size. Bigger pore size and enhanced thickness combinedly result in the enhancement of volumetric porosity as observed in Fig. 7(b).

For the porosity estimation typical techniques use nitrogen isotherm, Hg porosimetry and permeability measurements based on different versions of adsorption Brunauer–Emmet–Teller porosimetry. When the mass of material is very less as in case of thin films where mass of a few nm thick dielectric film is many orders of magnitude less than the mass of substrate, it is almost impossible to measure the change of the film mass from nitrogen adsorption. Hence, in thin films these methods have the limitation of detection of adsorbed material (gas/liquid) and may lead to huge error in porosity estimation. Also, typical analyte molecules may not be able to penetrate the smallest pores present in sol-gel films. Porosity measurement made by nitrogen gas may grossly overestimate the accessibility of the analyte to the encapsulated indicator species.³⁰ Therefore, in this work, the spectroscopic results through the absorption measurements and porosity/density simulation have been chosen to determine the relative variation in porosity of thin films. Three dimensional molecular size of MB has been reported as 1.43 nm × 0.61 nm × 0.4 nm,³¹ hence MB is taken as the adsorbed specie which is dimensionally less than the expected pore size in the films. Using similar technique, Harris et al. reported the porosity estimation in in sol-gel silica thin films.³² It should be emphasized that porosity estimation in present work is convenient in use because it is non-destructive and does not require low-temperature or controlled pressure condition. The earlier work on porous ZrO₂ thin films reported the pore size distribution centered at 3.6 nm (for two-coatings) and 15.5 nm (for one coating) when films are prepared using zirconium butoxide as precursor and acetic acid as chelating ligand.⁵ Average pore size is reported to be around ∼4 nm, when ZrOCl₂.8H₂O has been used as precursor,.²² In comparison of these reports, films prepared in this work have comparable pore size and have better control on volumetric porosity. Films with desired porosity can be prepared with presented methodology, which could play promising role in sensing and biomedical applications.

A sol-gel based synthesis process is modified by using HCl as catalyst to control the gelation process for deposition of mesoporous zirconia thin films having controlled porosity. Experimental results show that lower values of starting pH of polymeric sol are appropriate for the stability of sol for several days. The thickness of films is varied from 88 nm to 254 nm by increasing the sol age. Pore size of the films in mesoporous regime and nature of the as deposited films as amorphous has been confirmed by TEM study. The enhancement in diffused reflectance as well as in RMS roughness of the films with the increase in sol-age establishes a strong correlation between the surface topography and porosity. A relative enhancement of 5.6 times in porosity was found in films prepared with sol-age of 10 days with respect to that of 1 day. The band gap of films is estimated by Tauc's plot and found to be invariant (5.74 eV) to any change in sol-age.

Author expresses his appreciation to Dr. Deepak Varandani for AFM measurement, and Mr. V.K. Khanna for TEM imaging during this work.