J O U R N A L O F C H E M I S T R Y Materials Surface morphology study of corona-poled thin films derived from sol–gel processed organic–inorganic hybrid materials for photonics applications Yoo Hong Min,a Kwang-Sup Lee,*b Choon Sup Yoona and Lee Mi Doc aDepartment of Physics, Korea Advanced Institute of Science and T echnology, Taejon 305-701, Korea bDepartment of Macromolecular Science, Hannam University, T aejon 300-791, Korea cElectronics and T elecommunications Research Institute, P.O.Box 106, Yusong, Taejon 305-600, Korea Dye-doped and a dye-attached sol–gel films using the chromophore (E)-N-butyl-(4-{2-[4-bis(2-hydroxyethyl)amino]phenyl}- ethenyl)pyridinium tetraphenylborate (BPTP) and N,N-bis(2-hydroxyethyl)amino-N¾-methylstibazoliumtoluene-p-sulfonate (BAST) were fabricated by sol–gel processing.Surface morphology of the films was studied using an atomic force microscope. In the dye-doped film, many surface defects were observed. The morphology of the surface defects varied as a function of chromophore concentration. The shapes of the defects became more complicated as the chromophore concentration increased. On the other hand, no evidence of such defects was present in the dye-attached film and the quality of the film was high.Also for the dye-attached film, the surface was optically flat and smooth, and the surface roughness was measured to be less than 4 nm. It was found that corona poling processes also aVected the quality of the dye-attached film significantly. Corona poling with a needle electrode resulted in many irregular shaped defects in the film.However, poling with a tungsten wire stabilized the corona discharge and thus prevented the formation of defects, which led to films of excellent quality. The electro-optic coeYcient, r33, of the BPTP dye-attached film was measured to be 5.0 pm V-1 and this value was maintained even after 54 days at room temperature. Poled side-chain nonlinear optical (NLO) polymers, where The sol–gel processing technique is often utilized for fabricating inorganic polymer thin films.7 Using sol–gel processes, two organic chromophores with a large molecular hyperpolarizdi Verent types of sol–gel composite films were fabricated in ability are covalently bonded to the polymer backbone, have our experiment, namely dye-doped and dye-attached systems.been of great interest for photonics applications.1 Using such In the former the organic chromophores are physically blended a material, electro-optic modulators with a 40 GHz bandwidth with the silica matrix. The dye-attached system was obtained have already been realized at the laboratory level, and a by using the functionalized silicon alkoxide with one alkoxide number of passive and active photonic devices were also functional group substituted by an organic chromophore. successfully fabricated.2 In order to apply these sol–gel films to waveguide devices, From the initial stages of NLO polymeric materials research the optical propagation loss should be lower than 1 dB cm-1.in the mid-1980s, thermal relaxation and low values of NLO The causes of optical loss have been attributed to optical coeYcients of chromophores in these systems were the major absorption, Rayleigh scattering, nonuniform refractive index obstacles for practical applications.Since then there have been profile, defects in the film, surface roughness and surface continuous eVorts toward reducing thermal relaxation and damages generated by corona poling, etc.8 Since these factors increasing the NLO activity of the materials.3 In recent years, are closely related to the method of fabrication, and thus the in addition to enhancing these two fundamental factors, the resulting thin film quality, it is important to assess how research was extended to improve the optical quality, prodi Verent sol–gel processes and corona poling processes aVect cessability and the chemical stability of NLO materials, which the parameters mentioned above.are the properties closely related to those of matrix polymers. In this work, dye-doped and dye-attached sol–gel films with NLO polymer systems which are currently being investigated a stilbazolium salt dye were fabricated. The surface morpho- can be divided into two major categories.These include organic logies of the films obtained from diVerent sol–gel processes, polymer systems with a high glass transition temperature (Tg) curing and poling conditions were investigated using atomic polymer as a matrix,4 and cross-linked inorganic polymer force microscopy (AFM). Surface defects, the electro-optic systems.5 Among inorganic polymers, silicate glass is the most coeYcient and the photobleaching eVect are also described.studied and important. It has been reported by several research groups that the use of high Tg polymers can remarkably reduce the thermal relaxation of the molecular dipoles. Such improved stability is attributed to the restricted free volume and the Experimental chain stiVness of the polymer matrix. However, it has been Materials pointed out that organic polymers generally produce serious surface damage when poled by the corona discharge method.6 Tetraethylorthosilicate (TEOS) (Adrich) and 3-isocyanato- Using an inorganic polymer, such problems of surface damage propyl triethoxysilane (IP-TEOS) (Lancaster) were used can be easily overcome.Silicate glass has a high Tg and also as-received without any further purification.N,Nexhibits transparency in the spectral region of device appli- Dimethylformamide (DMF) (Junsei) solvent was distilled cations, hence it is considered to be an attractive alternative under reduced pressure over anhydrous magnesium sulfate. The NLO chromophores, (E)-N-butyl-(4-{2-[4-bis(2-hydroxy- to organic polymer systems.J. Mater. Chem., 1998, 8(5), 1225–1232 1225ethyl)amino]phenyl}ethenyl)pyridinium tetraphenylborate (0.0462 g, 1 M) were added and stirred for 15 days for the appropriate viscosity for spin casting to be obtained. (BPTP) and N,N-bis(2-hydroxyethyl)amino-N¾-methylstilbazoliumtoluene- p-sulfonate (BAST) were synthesized according to the methods reported elsewhere.9 Film casting and corona poling Sol–gel films were fabricated by the spin casting method.Dye-doped system. TEOS (10.0 g, 48.0 mmol) and ethanol Coating solutions were prepared by the method described in (2.2 g) were mixed in a 40 ml vial with HCl aqueous solution the previous section. The viscous sol was filtered through a (2.9 g, 2 M), which was added dropwise while stirring gently at Teflon syringe filter of 0.45 mm pore size.Thin films were spin 0 °C. Further stirring was performed at room temperature for coated on indium tin oxide (ITO) substrates with a rotation 12 h by using a magnetic spin bar with a rotation speed of speed of 1000–2000 rpm. The film thickness ranged from 400 rpm to complete the sol reaction. Then BPTP (0.302 g, 0.457 mmol) and DMF (15 ml ) were added to the sol and stirred again until the sol became viscous enough for film casting.Another batch of sol with a diVerent BPTP concentration (0.453 g, 0.686 mmol) was also prepared using the same procedure. Dye-attached system. DMF solvent (2 ml), IP-TEOS (0.394 g, 1.59 mmol) and BPTP (0.5 g, 0.76 mmol) were mixed together in a 20 ml vial and stirred at 90 °C for 3 h, during which time the bonding reaction of BPTP molecules with IP-TEOS took place to form urethane linkages.After cooling to room temperature, TEOS (0.039 g, 0.189 mmol) and HCl aqueous solution (0.0618 g, 1 M) were added. This mixture was then stirred for 10 days so that an appropriate viscosity for the spin casting might be obtained. Similarly, IP-TEOS (0.221 g, 0.895 mmol) and BAST (0.200 g, 0.425 mmol) were dissolved in DMF (2 ml) and the solution was stirred at 90 °C for 3 h to form urethane bonding between BAST molecules Fig. 2 UV–VIS absorption spectra for cured (———) and poled (- - - -) and IP-TEOS. After cooling the solution to room temperature, samples. The poling condition was at 160 °C for 2 h with an applied voltage of 13 kV. TEOS (0.0490 g, 0.235 mmol) and HCl aqueous solution Fig. 1 (a) Two diVerent types of sol–gel routes for dye-doped system (SG-I) and dye-attached system (SG-II). (b) Chemical equation shows the sol–gel monomer for SG-II where the BPTP or BAST salt-type chromophore is covalently bonded to the alkoxysilane by the urethane linkage. 1226 J. Mater. Chem., 1998, 8(5), 1225–1232Measurements UV–VIS spectra were obtained by a Shimadzu UV-3010PC spectrophotometer.The surface morphologies of the films were observed by using a Seiko SPA-300 atomic force microscope (AFM) equipped with an SPI-3700 controller. The 100 mm long cantilever for the AFM (Olympus) was microfabricated from pyramidal Si3N4 and the spring constant was 0.09 N m-1. The electro-optic coeYcient was measured at 1.3 mm by the simple reflection method proposed by Teng and Man.10 Results and Discussion Two diVerent sol–gel processes for preparing dye-doped and dye-attached systems are shown in Fig. 1. In sol–gel process I (SG-I), the sol–gel film was a physically blended composite where the NLO chromophore was merely distributed in the silica matrix without any chemical attachment. In sol–gel Fig. 3 Optical photograph of a dye-doped sol–gel film (magnification process II (SG-II), the NLO chromophore was covalently ×100) bonded to the silica matrix by using the functionalized sol–gel monomer described in Fig. 1( b). Considering the disappearance of the stretching vibration mode peak of the isocyanate group around 2300 cm-1 in the FTIR spectrum, the urethane 0.5 mm to 3 mm for the dye-attached system, depending on the viscosity and the rotation rate.However for the dye-doped linkage forming reaction between diol of the chromophore and isocyanate of IP-TEOS was assumed to proceed nearly system, only films of the thickness less than 1 mm were free of cracks. quantitatively. Fig. 2 shows the UV–VIS absorption spectra for the BPTP The alignment of molecular dipoles in the film was established by the corona poling method.The poling system con- dye-attached films prepared by the SG-II process. The spectra were measured in the wavelength range between 250 and sisted of a base and an electrode, and the whole system was covered by a glass bell jar. The base was electrically grounded 800 nm. Two samples prepared from the same film were used to investigate the poling eVect.The absorption maximum for and sat on top of a hot-plate. The atmosphere above the films was not controlled in any way. Either a stainless-steel needle the unpoled film was at 473 nm. For the poled film, the height of the absorption peak was reduced as compared with the with an edge angle of 60° or a tungsten wire of 20 mm diameter was used as an electrode. In the case of needle poling, the unpoled film, but the peak position remained the same. Considering the same curing conditions for both films, the needle electrode was positioned 2 cm above the ground with a poling voltage of 13 kV, at which a stable corona discharge reduction of the peak height can be accounted for by the corona poling eVect, which indicates the alignment of chromo- started to be generated.For the wire poling, the distance between the wire electrode and the ground was set at 1 cm phore dipoles normal to the film surface. For the dye-doped system, fabrication of sol–gel films thicker and a poling voltage of 5 kV was suYcient for generating a stable corona discharge. than 1 mm was very diYcult because of severe cracking prob- Fig. 4 AFM images of a BPTP dye-doped composite film with 0.46 mmol of BPTP, (a) 20×20 mm2, ( b) 2×2 mm2.The film was thermally cured at 150 °C for 2 h. J. Mater. Chem., 1998, 8(5), 1225–1232 1227Fig. 5 AFM images of a BPTP dye-doped composite film with 0.69 mmol of BPTP, (a) 30×30 mm2, ( b) 10×10 mm2. The film was thermally cured at 150 °C for 2 h. lems (Fig. 3). The films were very brittle and the adhesion to the ITO layer was poor.During thermal curing, the color of the dye-doped film changed from bright orange to pale yellow. The melting temperature of the chromophore was 79 °C and decomposition started at around 150 °C, as determined by thermogravimetric analysis (TGA) using a very slow heating run. Therefore, it is very likely that the change in color during the thermal treatment may be caused by decomposition and/or vaporization of chromophore molecules at elevated temperatures.The AFM images of two dye-doped sol–gel films are shown in Fig. 4 and 5. Both films have the same concentration of TEOS, ethanol and H2O, and were prepared using the same thermal curing procedures (150 °C for 2 h). The only diVerence between Fig. 4 and 5 is that the doping concentration of BPTP in the latter was 1.5 times greater than in the former, as was described in the Experimental section (0.46 mmol versus 0.69 mmol).In Fig. 4, crater-like defects with diameters of 100–200 nm and a depth of 20 nm were scattered randomly on the surface. On the other hand, in Fig. 5, canyon-like defects run in zigzag fashion, twisted and mingled together.The depth and width of these canyons were less than 20 nm and several hundred nm, respectively, and the length ranged from submicron to several mm. The present result shows that the film quality deteriorated at higher concentrations of BPTP. This may indicate that the distribution of the dye chromophore molecules becomes less inhomogeneous at the nanometer scale as the BPTP concentration increases.The maximum doping ratio of the NLO chromophore without any phase separation was about 5–6 wt.%.11 Here, no phase separation is defined as there was no observable aggregation of the dye through an optical microscope. Whether or not the distribution of the chromophores in the silica matrix was uniform on the nanometer scale could not be proved. When the concentration of chromophore exceeded a certain critical value, the film was observed to be translucent and surface defects were observed. Fig. 6(a) is the SEM micrograph of the dye-doped film containing BPTP chromophore (20 wt.%). This film was not cured so that the dye aggregations remained on the surface or inside the film. On the other hand, no aggregation was found on the Fig. 6 SEM micrographs for (a) a dye-doped film of 20 wt.% dye concentration and (b) a dye-attached film of 64 wt.% dye concentration surface of the dye-attached film [Fig. 6(b)]. 1228 J. Mater. Chem., 1998, 8(5), 1225–1232Fig. 7(a) and (b) show the AFM images, at diVerent magnifi- concentration of the chromophore increased, the film flexibility approached that of organic polymers and the adhesion to the cations, of a BPTP dye-attached sol–gel film fabricated by the SG-II route without corona poling.In this case, the thin film ITO electrode layer also improved. Fig. 7(c) and (d) shows AFM images at diVerent magnifi- was thermally cured at 160 °C for 80 min. As seen from Fig. 7(a), the surface was relatively flat and defects such as cations of the SG-II sol–gel film which was cured at 160 °C for 80 min while poling at 13 kV with a corona discharge craters or canyons could not be observed.The surface roughness was estimated to be less than 2 nm from AFM images using a needle electrode. Unevenly distributed defects, in the form of irregular-shaped craters characterized by several of higher magnification (2×2 mm2) [Fig. 7(b)]. It is believed that the reasons for the remarkable improvement of the dye- hundred nm diameters and 12–26 nm depths, were generated.These defects might been formed by the bombardment of attached film quality were that the chemical bonding of chromophore molecules to the silica matrix prevented the dye nonuniform and unstable plasma produced from the rough edge of the needle. The depths of most craters were smaller molecules from aggregating and that the chromophore molecules provided elasticity between the stiV SiO2 backbones.than the wavelength of visible light. However, the diameters were much larger than the UV–VIS wavelength. Therefore, the Such elasticity can accommodate severe contractions and hence prevent the formation of cracks [Fig. 6(b)].12 Film thicknesses scattering loss by these defects may not be neglected when used as a waveguide.of 0.5–3 mm could be fabricated in the dye-attached system and no cracks were found after the thermal treatment. As the The dye-attached sol–gel film using the NLO chromophore Fig. 7 AFM images of BPTP dye-attached sol–gel composite films: (a,b) without and (c,d) with corona poling using a needle electrode. The image sizes were 20×20 mm2 for (a,c), 2×2 mm2 for (b) and 2.5×2.5 mm2 for (d).The films were thermally cured at 160 °C for 80 min. J. Mater. Chem., 1998, 8(5), 1225–1232 1229with tosylate anion (BAST) instead of tetraphenylborate anion was also prepared by the SG-II process. AFM images for two diVerent BAST sol–gel films without and with corona poling using a needle electrode are exhibited in Fig. 8(a) and (b), respectively. The overall surface morphologies show features similar to the BPTP dye-attached sol–gel film. The surface of the cured only film is very flat and smooth [Fig. 8(a)] and the surface roughness is less than 4 nm. However, the AFM images of the BAST thin film, which was corona-poled by using a needle electrode, showed many defects of irregular shapes and sizes similar to those shown by the BPTP film [Fig. 7(c), (d)]. This problem could be overcome by using a tungsten wire electrode. Fig. 9 shows the surface morphology of a 20 mm diameter tungsten wire poled sample and nearly no surface damage can be observed. The results imply that if corona discharge is stabilized by the tungsten wire, it is possible to avoid the surface damage generated by the plasma and poled films of excellent quality can be obtained.In order to investigate the influence of corona poling on dye-attached films, cross-sectional AFM images were analyzed. As shown in Fig. 10, the surfaces of unpoled BPTP (a) and BAST (b) films are very smooth and flat. In contrast, BAST film (c) poled by the needle electrode has valleys of 3–8 nm depth and 0.8–1.2 nm width, and peaks of 3–4 nm height and 1.5–3 mm width.The results of roughness measurements are summarized in Table 1. The roughness of the BAST thin film, poled by using a needle electrode, was four times greater than those of unpoled BAST and BPTP dye-attached films. The dye-attached sol–gel film showed a good stability towards stabilized corona discharge.This obviously contrasts with the case of organic polymer systems where severe surface damage, such as cracks, pin holes, chemical reaction, surface deformation, etc., was generated.6 The endurance of the sol–gel film may be due to the strong covalent bonds which forms the SiO2 matrix. Electro-optic properties of the poled dye-attached BPTP films were investigated. Measurements on dye-doped films were not possible because of low dye concentration (lower than 5 wt.%), sublimation and/or decomposition of the dye during the poling process, and rapid thermal relaxation.The Fig. 8 AFM images (20×20 mm2) of BAST dye-attached sol–gel composite films (a) without and (b) with corona poling using a needle electrode Fig. 9 AFM image (20×20 mm2) of a BAST dye-attached sol–gel composite film.The film was poled with an applied voltage of 13 kV using 20 mm tungsten wire. (a) Top view and (b) inclined side view. 1230 J. Mater. Chem., 1998, 8(5), 1225–1232Fig. 11 Temporal stability of the electro-optic coeYcient for the SG-II sample at room temperature large area illuminator (Oriel) as the UV source in the spectral Fig. 10 (a) Cross-section of Fig. 7(a); ( b) cross-section of Fig. 8(a); and range of 320 to 450 nm. The BPTP dye-attached film was (c) cross-section of Fig. 8( b) exposed to 55 mW cm-1 UV intensity. The change of refractive index was measured at 1.3 mm as a function of exposure time Table 1 Roughness measurements of dye-attached films by using a prism coupling method. As shown in Fig. 12, the refractive index decreased exponentially.The refractive index sample Ra a/nm Rb max/nm Rc z/nm was changed by 0.028 over 3 h exposure. Fig. 13 shows the Fig. 7(a) 0.209 3.563 1.163 change of the UV–VIS absorption curve of the photobleached Fig. 8(a) 0.199 1.290 0.866 film. The absorption peak height decreased as the exposure Fig. 8(b) 1.517 9.480 4.177 time increased. However no new peak appeared as a result of photobleaching.The results suggest that the mechanism of aThe mean roughness (Ra), calculated from Ra=1/L | f (x)|dx, with f (x) photobleaching in the BPTP chromophore is not due to the profile curve and L the length of the profile curve. bRmax was cis–trans isomer transformation, as happens with azo or stil- obtained from the diVerence between the highest and lowest points on the profile.cRz was determined, which gives the mean variation between the twenty highest and twenty lowest points on the profile. electro-optic coeYcient, r33, was measured at a wavelength of 1.3 mm for the samples which were poled by a needle electrode at 160 °C for 2 h with an an applied voltage of 13 kV. The electro-optic coeYcient values at diVerent sites of a film remained the same inside a circular boundary of about 10 mm diameter. However the electro-optic coeYcient values of diVerent films varied between 3.1 and 5.0 pm V-1 with the most probable frequency at 4 pm V-1, since the poling eYciency varied greatly from time to time even under the same poling conditions. As shown in Fig. 2 the absorption peak is quite far from the wavelength used for measuring the electro-optic coeYcient.Hence the measured electro-optic coeYcient can be regarded as a non-resonant value. In corona poling, conductivity can hardly be quantified Fig. 12 Change of refractive index of the BPTP dye-attached sol–gel because the leakage current, applied electric field and conduc- film by photobleaching eVects tion area are not well defined. However an estimation of the conductivity of the film can be made for contact poling. A very large conduction was observed at elevated temperatures over 100 °C, and as a consequence the film was damaged and dc contact poling was not possible. The leakage current was about 100 mA over an electrode area of 30 mm2 in which the applied field strength was 50 V mm-1 which corresponds to a value two orders of magnitude larger than that observed in nonionic organic sol–gel films, such as 4-[N,N-di(2-hydroxyethyl) amino]-4¾-nitrostilbene (DANS diol).The large electrical conduction observed in the ionic sol–gel films would certainly reduce the poling field by draining the charges piled up on the film surface, and this could lead to a poorer poling eYciency than it might otherwise have. Therefore it is very likely that the electro-optic coeYcient values measured were underestimated.The film showed good temporal stability. The initial value of the electro-optic coeYcient remained unchanged for 54 days (Fig. 11). Fig. 13 Change of UV–VIS absorption spectrum of the BPTP dyeattached sol–gel film by photobleaching eVects Photobleaching experiments were performed by using a J.Mater. Chem., 1998, 8(5), 1225–1232 1231S. L. Kwiatkowski, G. F. Lipscomb and R. S. Lytel, Appl. Phys. bene dyes.13 However, the results support the possibility that L ett., 1991, 58, 1730. the UV radiation could break the double bond of BPTP 3 J. W. Wu, J. F. Valley, S. Ermer, E. S. Binkley, J. T. Kenny and chromophore molecules, which would then lead to the R.Lytel, Appl. Phys. L ett., 1991, 59, 2213; D. Yu, A. Gharavi and reduction of charge transfer absorption in the photobleached K. Yu, Appl. Phys. L ett., 1995, 66, 1050; M. Chen, L. R. Dalton, film. L. Yu, Y. Q. Shi and W. H. Steier,Macromolecules, 1992, 25, 4032. 4 B. K. Mandal, Y. M. Chen, J. Y. Lee, J. Kumar and S. Tripathy, Appl. Phys. L ett., 1991, 58, 3; C. Xu, B.Wu, L. R. Dalton, Conclusions P. M. Ranon, Y. Shi and W. H. Steier, Macromolecules, 1992, 25, 6716; P. M. Ranon, Y. Shi, W. H. Steier, C. Xu, B. Wu and Using sol–gel processes, a BPTP dye-doped film and BPTP, L. R. Dalton, Appl. Phys. L ett., 1993, 62, 2605; J. A. F. Boogers, BAST dye-attached films were fabricated. Surface morphology P. Th. A. Klaase, J. J. de Vlieger, D. P.W. Alkema and was studied by using AFM and many surface defects were A. H. A. Tinnemans, Macromolecules, 1994, 27, 197; C.-K. Park, observed in the dye-doped film. As the chromophore concen- J. Zieba, C.-F. Zhao, B. Swedek, W. M. K. P. Wijekoon and P. N. Prasad, Macromolecules, 1995, 28, 3713; K. M. White, tration increased, the defect concentration also increased and D. K. Kitipichai and C.V. Francis, Appl. Phys. L ett., 1995, 66, the shapes of defects became more complicated. On the other 3099; Y. Shi, W. H. Steier, L. Yu and L. R. Dalton, Appl. Phys. hand, surface defects were not observed in the dye-attached L ett., 1992, 60, 25; A. K.-Y. Jen, K. J. Drost, Y. Cai, V. P. Rao and sol–gel film. Surface roughness was measured to be less than L. R. Dalton, J.Chem. Soc., Chem. Commun., 1994, 965; T. Verbiest, a few nm. In the film poled by the needle electrode, many D. M. Burland, M. C. Jurich, V. Y. Lee, R. D. Miller and W. Volksen, Macromolecules, 1995, 28, 3005; D. Yu, A. Gharavi defects of irregular shapes were present due to nonuniform and L. Yu, Macromolecules, 1995, 28, 784; S.-K. Ham, S.-H. Choi, distribution of the corona discharge.When the corona dis- B.-H. Lee and K. Song, Polymer (Korea), 1997, 21, 201; H. K. Kim, charge was stabilized by a 20 mm diameter tungsten wire I. K. Moon, M. Y. Jin and K.-Y. Choi, Korea Polym. J., 1997, 5, 57. electrode, such defects were not observed and poled film of an 5 J. Kim, J. L. Plawsky, R. LaPerta and G. M. Korenowsky, Chem. excellent quality was obtained. Mater., 1992, 4, 249; Y.Zhang, P. N. Prasad and R. Burzynski, The electro-optic coeYcient, r33, of the poled BPTP dye- Chem. Mater., 1992, 4, 851; R. J. Jeng, Y. M. Chen, A. K. Jain, J. Kumar and S. K. Tripathy, Chem. Mater., 1992, 4, 1141; attached film was measured to be 5 pm V-1 and the film R. J. Jeng, Y. M. Chen, A. K. Jain, J. Kumar and S. K. Tripathy, exhibited good temporal stability.The initial value of r33 Chem. Mater., 1991, 4, 972; G. H. Hsiue, J. K. Kuo, R. J. Jeng, remained unchanged even after 54 days at room temperature. J. I. Chen, X. L. Tiang, S. Marturunkakul, J. Kumar and A photobleaching experiment was performed with the BPTP S. K. Tripathy, Chem. Mater., 1994, 6, 884; F. Chaput, J.-P. Boilot, dye-attached film and a large refractive index change of 0.028 D.Riehl and Y. Levy, SPIE Proc., 1994, 2288, 286. 6 R. A. Hill, A. Knoesen and M. A. Mortazavi, Appl. Phys. L ett., was induced by UV irradiation of 55 mW cm-2 for 3 h. 1994, 65, 1733. In summary, the dye-attached organic–silica hybrid films 7 C. J. Brinker and G. W. Scherer, Sol–Gel Science, Academic Press, are characterized by excellent temporal stability, controllable Boston, 1990; L.C. Klein, Sol–Gel Optics: Processing and refractive index and defect free processability. Hence by making Applications, Kluwer Academic Press, Boston, 1994; E. J. A. Pope, further improvements in their electro-optic coeYcient, the dye- S. Sakka and L. C. Klein, Sol–Gel Science and T echnology, attached system may be a potential material for photonics American Ceramic Society, Westerville, 1995. 8 T. C. Kowalczyk, T. Kose and K. D. Singer, J. Appl. Phys., 1994, applications. 76, 2505; C. C. Teng, M. A. Mortazavi and G. K. Boudoughian, Appl. Phys. L ett., 1995, 66, 667. This research was supported by the Basic Science Research 9 K. J. Moon, H.-K. Shim, K.-S. Lee, J. Zieba and P. N. Prasad, Macromolecules, 1996, 29, 861; K. J. Moon, Dissertation, KAIST, Institute Program, Ministry of Education and in part by the Taejon, 1996.Korea Science and Engineering Foundation through the Opto- 10 C. C. Teng and H. T. Man, Appl. Phys. L ett., 1990, 56, 1734. Electronics Research Center (KAIST). The authors are grateful 11 H.-H. Huang, B. Orler and G. L. Wilkes, Macromolecules, 1987, to Prof. M. Fujihira of Tokyo Institute of Technology for 20, 1322; M. W. Ellsworth and B. M. Novak, J. Am. Chem. Soc., AFM experiments. 1991, 113, 2756; S. Chakrabarti, J. Sahu, M. Chakraborty and H. M. Acharya, J. Non-Cryst. Solids, 1994, 180, 96; H. Kaji, K. Nakanishi and N. Soga, J. Non-Cryst. Solids, 1995, 185, 18. 12 G. Philipp and H. Schmidt, J. Non-Cryst. Solids, 1984, 63, 283; References B. Litner, N. Arfsten, H. Dislich, H. Schmidt, G. Phillip and 1 K.-S. Lee, M. Samoc and P. N. Prasad, in Comprehensive Polymer B. Seiferling, J. Non-Cryst. Solids, 1988, 100, 378; H. Schmidt, Science, ed. S. L. Aggarwal and S. Russo, Pergamon Press, Oxford, J. Non-Cryst. Solids, 1989, 112, 419. 13 Y. Shi, W. H. Steier, L. Yu, M. Chen and L. R. Dalton, Appl. Phys. 1992, p. 407; P. N. Prasad and D. J. Williams, Nonlinear Optical L ett., 1991, 58, 1131; M. B. J. Diemer, F. M. Suyten, E. S. Trommel, EVects in Molecules and Polymers, Wiley Interscience, New York, A. McDonach, J. M. Copeland, L. W. Jenneskens and 1991; G. A. Lindsay and K. D. Singer, Polymers for Second-order W. H. G. Horsthus, Electron. L ett., 1990, 26, 379. Nonlinear Optics, American Chemical Society, Washington DC, 1995. 2 C. C. Teng, Appl. Phys. L ett., 1992, 60, 1538; D. G. Girton, Paper 8/00917A; Received 3rd February, 1998 1232 J. Mater. Chem., 1998, 8(5), 1225–1232