A thin titania sol-gel layer was prepared on thermal oxide silica-on-silicon and borosilicate surfaces with spin-coating techniques under cold (room temperature) processing conditions. The physical structure and chemical uniformity of these layers were examined by a series of spectroscopic (FTIR, UV-VIS spectroscopy and ellipsometry) and microscopic (light microscopy, SEM and EDS) techniques. Selective binding with hydroxy-containing (-OH) organic compounds is explored.
©2012 Optical Society of America
Silica (SiO2) and titania (TiO2) sol-gels are increasingly used to fabricate films of potential use in optical devices. The sol-gel process in principle allows many organic compounds to be integrated into the structure . Sol-gel fabrication procedures follow relatively simple methodologies, and can be used to build sub-micron layers under the influence of a template structure . The process for fabricating films usually involves chemically preparing the silica or titania gel, then allowing for sedimentation over a substrate. The structure would be affected during the drying process, where the material undergoes densification, but more so by subsequent heat treatment (annealing) at temperatures up to 1000 °C to sinter the material. It is usually during annealing that the gel structure is assumed to convert into a specific form . Much work on sol-gel construction requires precise control over preparation conditions [4–6]; obtaining the intended structure requires planning the conditions of the process. At these high temperatures, it is not possible to integrate beforehand organic species or other volatiles so these are often introduced post sol-gel fabrication.
Some recent studies on the binding mechanism of silane layers to silica substrate would indicate that there are 3 Si-O bonds attached to the surface Si atoms . The role of terminating Si attaching directly to the substrate is also proposed . This is in contrast to the classically accepted model of only a single sol-gel Si-O bond, and raises significant questions on interface induced strain between the substrate and subsequent lattice. Perhaps this may be the principle reason why sol-gel is prepared in sub-micron layers when fabricating useful micron-sized integrated optics [9,10]. It is this networking structure that account for the porous nature of the glassy silica system. It seems reasonable to expect that titania-based layers would follow similar formation pathways and have a similar structure. Hence, using a cold-preparation method would directly affect the TiO2 , often assumed to be crystalline without direct supporting evidence, generally leading to a much more porous, glassy structure [12,13]. In a previous example, we reported depositing titania sol-gel layers within specially micro-structured fibres at room temperature  in which we combine fibre-based sensor technology with acid-base chemistry of a selected porphyrin, 5,10,15-tris(di-tert-butylphenyl)-20-(p-hydroxyphenyl) porphyrin (henceforth abbreviated as hydroxyporphyrin) . Titanates have the affinity to bind to a large variety of functional groups, including –OH [16,17], and have greater chemical (catalytic) and biocompatibility than silicates for specific compounds [18,19]. TiO2 also has higher refractive index which can serve to increase the evanescent field overlap within the holes, both through optical impedance matching and potentially through extended resonator enhancements , ultimately leading to higher sensitivity. Cold-preparation was suitable for our applications because the higher porosity allows for greater uptake of selected compounds, and also because the high temperatures used to complete gel-to-oxide conversion  and decompose organic compounds are avoided. Cold-processing has other advantages, including the potential for integrating materials during fabrication that would not be possible in the presence of thermal annealing. A key problem was that the TiO2 appeared to break up after repeated cycling of the sensor; this appears to be consistent with tetragonal rutile formation, which seems less inclined to form uniform films.
In this work, we characterise the methodology for cold-processed, TiO2 sol-gel based surface fabrication on silica and silicate substrates for various applications, particularly the integration of organic materials of potential optoelectronic value. In place of the flow of sol-gel under pressure through the structured optical fibres , the TiO2 sol-gel layers are deposited on pathology grade borosilicate (B2O3/SiO2) slides and on similar silica-based surfaces, particularly Si wafers coated with thermal SiO2 layer arising from oxide termination, using the method of spin-coating which is known to give sub-micron scale films . In principle, these silica layers ought to be similar to the inner surface of the structured optical fibre previously reported, whilst the general properties of wafers are compared using the borosilicate pathology grade slides. The key aim in this work is to determine whether or not sufficiently uniform and compacted and stable TiO2 layers can be achieved at room temperature, and we show by spin coating at very high speeds this appears possible. By characterizing the TiO2 layer properties, we can further consider their effectiveness for selective binding when applied to other optical components in planar or optical fibre form.
2. Experimental method
The surfaces used were borosilicate glass slides (pathology grade) and silica-terminated silicon, (thermal oxide thickness ~20 µm). The surfaces had rectangular dimensions (slide: 76.2 mm x 25.4 mm, wafer: ~10 mm x 10 mm). These were washed with NH3/H2O2/H2O (1:1:5) for surface activation (-OH generation), followed by washing with de-ionized water and drying under N2 flow. The TiO2 gel was formed with a 5% titanium isopropoxide (tetra-isopropyl titanate) in isopropanol and left overnight for aging. The slides were spin-coated at 1000 rpm for 60 seconds with the TiO2 sol-gel, which produced a thin, faintly iridescent TiO2 coating. The surface was then placed in organic solutions (hydroxyporphyrin, rhodamine B, thymol blue and bromothymol blue) overnight for post-binding. The surfaces were finally washed with isopropanol to remove unbound organic material and dried at room temperature with N2 flow. Variations of the preparation conditions to control layer thickness included reduction of TiO2 concentration to 1%, increase spin-rate to 6000 rpm, and heating to 300 °C for 3 hours. The layer was analyzed by FTIR, UV-VIS, ellipsometry, light microscopy, scanning electron microscopy (SEM) and energy dispersive X-ray scattering (EDS).
3. Results and discussion
3.1 Titania layer characterisation by FTIR
A cross-sectional view of the surface by light microscopy of a layer deposited on the borosilicate slide indicates the presence of a colorless layer of micron-scale thickness, formed from the centripetal force causing the titania layer to spread (Fig. 1 ) and assemble to a matrix of titanium oxide. However, a top-down view of the surface indicates some aggregation of titania particles in various regions consistent with rutile formation , and groove-like features on the outer regions are most likely caused by the liquid motion during spin-coating on the rectangular surface (76.2 mm x 25.4 mm) leading to uneven non-isotropic force on different regions, a common phenomena observed for the sol-gel process. The chemical integrity of the titania layers were characterised with FTIR measurements at different regions of the TiO2 surface on both the silicon wafers and borosilicate microscope slide [Fig. 2(a) ]. Measurements on both surfaces showed the titania (Ti-O) bend peak at 1639 cm−1 [Fig. 2(b)], comparable to the primary FTIR peak of the reference titania powder in a solid matrix [Fig. 2(c)] observed at 1637 cm−1. These results were similar to that of the Lopez group, who used FTIR to determine the specific form of TiO2 at different annealing temperatures, different pH conditions, and interactions with bound compounds [24–26]. Other primary peaks observed are the broad O-H peak (3500 to 3000 cm−1 range) and the Si-O peak (1000 cm−1). The same observed Ti-O peak in both surfaces justifies the previous assumption – the films prepared on borosilicate slides and thermal oxide silica have similar chemical environment despite the presence of borate in the pathology slides. We may therefore make some generalizations to all silica and silicate surfaces, such the structured optical fibres we reported previously . The FTIR spectra of different regions of TiO2 surface appear to be the same with no shifts of the Ti-O band observed on different areas, indicating uniform chemical coverage over the surface. Therefore, the process was confirmed to be reproducible on different SiO2-based surfaces.
3.2 Refractive index measurement
The refractive index of TiO2 sol-gel layers can vary depending on the porosity introduced . It is nearly always higher than fused silica (n = 1.458) but much lower than that of the rutile polymorph (n = 2.609). The thickness of the TiO2 sol-gel on the SiO2 thermal oxide layer on Si wafer was measured by scanning electron microscopy (SEM) [Fig. 3(a) ]. The composition can be partially extracted from energy dispersive spectroscopy (EDS) [Fig. 3(b)]. The structure of these TiO2 layers is a thick, gel-like covering of thickness t ~3 µm over a thinner regular layer of t ~300 nm. EDS measurements indicate the presence of titanium in both layers. From these thicknesses, ellipsometry was used to determine the refractive index of the titania layer to lie over a range between n ~(1.9 – 2.1) [Fig. 3(c)]. These measurements are an average of the thicker top layer and the more uniform thin layer which is likely to have higher index. This refractive index is lower than literature values of TiO2 layers prepared with thermal annealing, which gave thinner and less porous TiO2 with a greater extent of sol-gel polymerisation , consistent with the lower layer being denser and having a higher index.
3.3 Binding of organic compounds
Titania-based sol-gels are well known to have a greater affinity for binding to a variety of organic compounds containing hydroxyl (-OH) or carboxyl (-COOH) groups . To consider chemoselectivity, the compounds hydroxyporphryin, rhodamine B, thymol blue and bromothymol blue were considered. These compounds were selected due to their functional groups, high extinction coefficient and chemical changes under different acidic/basic conditions  making them commonly used acid/base indicators (with the exception of rhodamine B which was selected due to its characteristic fluorescence). The TiO2 surfaces with these compounds bound were analysed by UV-VIS spectroscopy, since these compounds have strong visible absorbance. TiO2 itself shows absorbance mostly in the UV-region , and therefore does not overlap with the spectra of the bound compound. Complete washing of the surface removes excess substrate, leaving only the compounds with strong absorption to the layer. The hydroxyporphyrin, rhodamine B and thymol blue showed strong affinity for binding to the TiO2 surface, given that their characteristic spectra were observed for all samples [Figs. 4(a) , 4(e) and 4(f)]. The bromothymol blue showed some absorbance, which can be explained by noting its structure with a sterically hindered –OH; it is likely that the –SO3- can also bond to the titania. This is evidence supporting the assertion that several attachment mechanisms other than –OH can be present. The substances bound onto TiO2 show interaction with acidic hydrogen chloride (HCl) and basic ammonia gas (NH3). These gaseous species affect the visible spectra of the solid compounds, justifying the possibility of using them with TiO2 as acid sensors. One drawback to this method is that time-based degradation, simulated by repeated heating of the layer to 80°C followed by rinsing in isopropanol, would lead to partial breakdown of the TiO2. This is observed, for example, as the release of porphyrin material [Fig. 4(b)], indicating that the TiO2 is not sufficiently stable. Titania that has been heat sintered to 300°C does not have strong porphyrin uptake, but porphyrin can be incorporated before sintering, which gives a layer without time-based degradation and release of porphyrin [Fig. 4(c)]. Although the porphyrin withstood the sintering, many other organic compounds are known to decompose at such high temperatures.
3.4 Modifications to the sol-gel processing technique
As observed by SEM [Fig. 5(a) ], much of the TiO2 was a gel-like layer over a more uniform-thickness layer. Although the thicker gel-like layer would have the desired porosity, methods to reduce the layer thickness were explored. The sol-gel preparation was modified to have lower TiO2 concentration, which lead to a less-uniform coverage of the surface [Fig. 5(b)]. Heat treatment to 300 °C [Fig. 5(c)] seems to have compacted the structure. Alternatively, control of the flow rate can be used on the slide surface; this can be done by increasing the spin-coating velocity to generate much higher effective flow rates. The relative centrifugal force F (in units of gravity) of the layer should be proportional to the rotor radius R and the square of the spin rate (S), following the equation F = (1.118 x 10−5) RS2. Therefore, a six-fold increase of spin-rate would lead to an equivalent of a thirty-six times greater force. In addition to spreading and thinning the film out, the resistance of the gel to flow as it condenses, through Poissons relations, can generate an effective compression force that produces a much more compact and thinner layer [Fig. 5(d)]. Using a greater force and flow speed would control layer thickness with reasonable precision using this process.
The fabrication of TiO2 sol-gel layers on borosilicate and thermal oxide silica surfaces at room temperature has been studied. Little distinction is observed between borosilicate and silica glasses. Cold processing is simpler than conventional fabrication, with a greater range of tuning of its porous properties. This makes it more suitable for various applications such as optoelectronic sensors and devices, provided that high temperature operation is not critical. Spin-coating at high speeds allows circumvention of conventional thermal annealing to produce compact and stable films, though more work is required to fully assess annealing properties. Nevertheless, we also find that brief, rapid sintering at 300° C of both titania and porphyrin together can be tolerated by the porphyrin. This sintering process does improve the stability by comparison with pre-sintering of the titania which reduces the chemisorption of porphyrin considerably, a disadvantage. Sintering at lower temperatures in combination with rapid spinning may also help further stabilise films further, allowing other organic species to be integrated. The structure of these layers was observed to be uniform and reproducible. The layer thickness could be physically modified and chemical functionalisation during or after fabrication can enable selective binding to the titania. Although this process can be reproduced on different silica surfaces to give similar structures, using this process for a specific optical device may involve additional mechanical and chemical manipulation to suit the substrate used. This will be the emphasis for future application-specific studies.
References and links
1. D. Avnir, V. R. Kaufman, and R. Reisfeld, “Organic fluorescent dyes trapped in silica and silica-titania thin films by the sol-gel method. Photophysical, film and cage properties,” J. Non-Cryst. Solids 74(2-3), 395–406 (1985). [CrossRef]
2. R. A. Caruso and M. Antonietti, “Sol-gel nanocoating: an approach to the preparation of structured Materials,” Chem. Mater. 13(10), 3272–3282 (2001). [CrossRef]
3. C. C. Wang and J. Y. Ying, “Sol-gel synthesis and hydrothermal processing of anatase and rutile titania nanocrystals,” Chem. Mater. 11(11), 3113–3120 (1999). [CrossRef]
4. P. C. A. Alberius, K. L. Frindell, R. C. Hayward, E. J. Kramer, G. D. Stucky, and B. F. Chmelka, “General predictive syntheses of cubic, hexagonal, and lamellar silica and titania mesostructured thin films,” Chem. Mater. 14(8), 3284–3294 (2002). [CrossRef]
5. S. J. Park, Y. C. Kang, J. Y. Park, E. A. Evans, R. D. Ramsier, and G. G. Chase, “Physical characteristics of titania nanofibers synthesized by sol-gel and electrospinning techniques,” J. Eng. Fiber Fabr. 5(1), 50–56 (2010).
6. Z. Miao, D. Xu, J. Ouyang, G. Guo, X. Zhao, and Y. Tang, “Electrochemically induced sol-gel preparation of single-crystalline TiO2 nanowires,” Nano Lett. 2(7), 717–720 (2002). [CrossRef]
7. C. A. Schlecht and J. A. Maurer, “Functionalization of glass substrates: mechanistic insights into the surface reaction of trialkoxysilanes,” RSC Adv. 1(8), 1446–1448 (2011). [CrossRef]
8. A. Depla, D. Lesthaeghe, T. S. van Erp, A. Aerts, K. Houthoofd, F. Fan, C. Li, V. Van Speybroeck, M. Waroquier, C. E. A. Kirschhock, and J. A. Martens, “29-Si NMR and UV-Raman Investigation of Initial Oligomerization Reaction Pathways in Acid-Catalyzed Silica Sol-Gel Chemistry,” J. Phys. Chem. C 115(9), 3562–3571 (2011). [CrossRef]
9. L. Yang and S. S. Saavedra, “Chemical Sensing Using Sol-Gel Derived Planar Waveguides and indicator Phases,” Anal. Chem. 67(8), 1307–1314 (1995). [CrossRef]
10. R. M. Almeida, “Sol–gel planar waveguides for integrated optics,” J. Non-Cryst. Solids 259(1-3), 176–181 (1999). [CrossRef]
11. L. X. Zhang, P. Liu, and Z. X. Su, “Low temperature preparation of titania coated PAN fiber and its photocatalytical property,” J. Chin. Chem. Soc. 53(2), 319–324 (2006).
12. K. Farhadi, F. Kheiria, and M. M. Golzan, “Cold deposition as a novel procedure for the preparation of titania sol-gel: a development of a high sensitive electrochemical method for determination of Cu(II) in the presence of arsenic(III),” J. Chin. Chem. Soc. 55(5), 1113–1118 (2008).
14. G. Huyang, J. Canning, M. L. Åslund, M. Naqshbandi, B. Gibson, I. Petermann, D. Stocks, and M. J. Crossley, “Porphyrin-doped solgel-lined structured optical fibers for local and remote sensing,” Opt. Lett. 36(11), 1975–1977 (2011). [CrossRef] [PubMed]
15. G. Huyang, J. Canning, M. L. Åslund, D. Stocks, T. Khoury, and M. J. Crossley, “Evaluation of optical fiber microcell reactor for use in remote acid sensing,” Opt. Lett. 35(6), 817–819 (2010). [CrossRef] [PubMed]
16. J. Y. Wen and G. L. Wilkes, “Organic/inorganic hybrid network materials by the sol−gel approach,” Chem. Mater. 8(8), 1667–1681 (1996). [CrossRef]
17. U. Diebold, “The surface science of titanium dioxide,” Surf. Sci. Rep. 48(5-8), 53–229 (2003). [CrossRef]
18. J. Yu and H. Ju, “Preparation of porous titania sol-gel matrix for immobilization of horseradish peroxidase by a vapor deposition method,” Anal. Chem. 74(14), 3579–3583 (2002). [CrossRef] [PubMed]
19. Q. Li, G. Luo, J. Feng, Q. Zhou, L. Zhang, and Y. Zhu, “Amperometric Detection of Glucose with Glucose Oxidase Absorbed on Porous Nanocrystalline TiO2 Film,” Electroanalysis 13(5), 413–416 (2001). [CrossRef]
20. J. Canning, W. Padden, D. Boskovic, M. Naqshbandi, H. de Bruyn, and M. J. Crossley, “Manipulating and controlling the evanescent field within optical waveguides using high index nanolayers,” Opt. Mater. Express 1(2), 192–200 (2011). [CrossRef]
21. R. Deb Roy, D. Sil, S. Jana, P. K. Biswas, and S. K. Bhadra, “Experimental Study of Perfectly Patterned Silica-Titania Optical Waveguide,” Photon. Sens. 2(1), 81–91 (2012). [CrossRef]
22. U. Selvaraj, A. V. Prasadarao, S. Komarneni, and R. Roy, “Sol-gel fabrication of epitaxial and oriented TiO2 thin films,” J. Am. Ceram. Soc. 75(5), 1167–1170 (1992). [CrossRef]
23. P. I. Gouma and M. J. Mills, “Anatase-to-Rutile Transformation in Titania Powders,” J. Am. Ceram. Soc. 84(3), 619–622 (2001). [CrossRef]
24. T. Lopez, E. Sanchez, P. Bosch, Y. Meas, and R. Gomez, “FTIR and UV-Vis (diffuse reflectance) spectroscopic characterization of TiO2 sol-gel,” Mater. Chem. Phys. 32(2), 141–152 (1992). [CrossRef]
25. T. Lopez, E. Ortiz, R. Gomez, and M. Picquart, “Amorphous sol-gel titania modified with heteropolyacids,” J. Sol-Gel Sci. Technol. 37(3), 189–193 (2006). [CrossRef]
26. T. Lopez, E. Ortiz-Islas, E. Vinogradova, J. Manjarrez, J. A. Azamar, J. J. Alvarado-Gil, and P. Quintana, “Structural, optical and vibrational properties of sol–gel titania/valproic acid reservoirs,” Opt. Mater. 29(1), 82–87 (2006). [CrossRef]
27. L. H. Lee and W. C. Chen, “High refractive-index thin films prepared from trialkoxysilane-capped poly(methylmethacrylate)-titania materials,” Chem. Mater. 13(3), 1137–1142 (2001). [CrossRef]