Incorporation and Thermal Evolution of ... - ACS Publications

Jun 3, 2009 - (26) Hawkeye, M. M.; Brett, M. J. J. Vac. Sci. Technol. A 2007, 25, 1317. (27) Kosmulski, M. Chemical Properties of Material Surfaces; M...
0 downloads 0 Views 3MB Size
pubs.acs.org/Langmuir © 2009 American Chemical Society

Incorporation and Thermal Evolution of Rhodamine 6G Dye Molecules Adsorbed in Porous Columnar Optical SiO2 Thin Films Juan R. Sanchez-Valencia, Iwona Blaszczyk-Lezak, Juan P. Espinos, Said Hamad, Agustı´ n R. Gonzalez-Elipe, and Angel Barranco* Instituto de Ciencia de Materiales de Sevilla (CSIC-Univ. Sevilla), c/Am erico Vespucio 49, 41092 Sevilla, Spain Received February 26, 2009. Revised Manuscript Received May 15, 2009 Rhodamine 6G (Rh6G) dye molecules have been incorporated into transparent and porous SiO2 thin films prepared by evaporation at glancing angles. The porosity of these films has been assessed by analyzing their water adsorption isotherms measured for the films deposited on a quartz crystal monitor. Composite Rh6G/SiO2 thin films were prepared by immersion of a SiO2 thin film into a solution of the dye at a given pH. It is found that the amount of Rh6G molecules incorporated into the film is directly dependent on the pH of the solution and can be accounted for by a model based on the point of zero charge (PZC) concepts originally developed for colloidal oxides. At low pHs, the dye molecules incorporate in the form of monomers, while dimers or higher aggregates are formed if the pH increases. Depending on the actual preparation and treatment conditions, they also exhibit high relative fluorescence efficiency. The thermal stability of the composite films has been also investigated by characterizing their optical behavior after heating in an Ar atmosphere at increasing temperatures up to 275 °C. Heating induces a progressive loss of active dye molecules, a change in their agglomeration state, and an increment in their relative fluorescence efficiency. The obtained Rh6G/SiO2 composite thin films did not disperse the light and therefore can be used for integration into optical and photonic devices.

Introduction Rhodamine 6G is a well-characterized fluorescent dye molecule widely used for a large variety of applications including laser, probe molecule of surface states, solar cells, or surface coatings.1-5 For the majority of these uses, the dye molecule has to be incorporated into a solid, a feature that has fostered the development of new synthesis and processing methods. Thus, for example, it has been shown that Rh6G molecules can be easily incorporated within the interlayer space of silicates and other layered materials and polymers.6-10 The incorporation of this dye into sol-gel silica11-14 or into mesoporous TiO2 thin films15-17 has been also intended for the fabrication of optical components and devices. *Corresponding author. E-mail: [email protected]. (1) Yang, P.; Wirnsberger, G.; Huanng, H. C.; Cordero, S. R.; McGehee, M. D.; Scott, B.; Deng, T.; Whitesides, G. M.; Cmelka, G. F.; Buratto, S. K.; Stucky, G. D. Science 2000, 287, 465. (2) Sasai, R.; Iyi, N.; Fujita, T.; Arbeloa, F. L.; Martı´ nez, V. M.; Takagi, K.; Itoh, H. Langmuir 2004, 20, 4715. (3) Marlow, F.; McGehee, M. D.; Zhao, D.; Chmelka, B. F.; Stucky, G. D. Adv. Mater. 1999, 11, 632. (4) Tleugabulova, D.; Zhang, Z.; Chen, Y.; Brook, M. A.; Brennan, J. D. Langmuir 2004, 20, 848. (5) Bahadur, L.; Srivastava, P. Sol. Energy Mater. Sol. Cells 2003, 79, 235–248. (6) Martı´ nez-Martı´ nez, V.; Lopez-Arbeloa, F.; Ba~nuelos-Prieto, J.; Arbeloa-Lopez, T.; Lopez-Arbeloa, I. Langmuir 2004, 20, 5709. (7) Gemeay, A. H. J. Colloid Interface Sci. 2002, 251, 235. (8) Barranco, A.; Groening, P. Langmuir 2006, 22, 6719. (9) Bujdak, J.; Iyi, N. J. Phys. Chem. B 2005, 109, 4608. (10) Martı´ nez, V.; Salleres, S.; Ba~nuelos, J.; Lopez-Arbeloa, F. J. Fluoresc. 2006, 16, 233. (11) del Monte, F.; Mackenzie, J. D.; Levy, D. Langmuir 2000, 16, 7377. (12) Loerke, J.; Marlow, F. Adv. Mater. 2002, 14, 1745. (13) Wirnsberger, G.; Yang, P.; Huang, H. C.; Scott, B.; Deng, T.; Whitesides, G. M.; Chmelka, B. F.; Stucky, G. D. J. Phys. Chem. B 2001, 105, 6307. (14) Rao, A. P.; Rao, A. V. Mater. Lett. 2003, 57, 3741–3747. (15) Vogel, R.; Meredith, P.; Kartini, I.; Harvey, M.; Riches, J. D.; Bishop, A.; Heckenberg, N.; Trau, M.; Rubinsztein-Dunlop, H. Chem. Phys. Chem. 2003, 4, 595. (16) Vogel, R.; Meredith, P.; Harvey, M. D.; Rubinsztein-Dunlop, H. Spectrochim. Acta, Part A 2004, 60, 245–249. (17) Tomas, S. A.; Stolik, S.; Palomino, R.; Lozada, R.; Persson, C.; Pepe, I.; Ferreira da Silva, F. J. Appl. Phys. 2005, 98, 073516.

9140 DOI: 10.1021/la900695t

The main interest of these materials relies on the high fluorescence emission of the dye molecule;a feature that, associated with the agglomeration state of the molecule, can be modified by the interaction with the solid matrix. Thus, previous studies on the fluorescence behavior of Rh6G molecules incorporated into host matrices show that they can appear in the form of isolated molecules or higher aggregates such as the so-called H or J dimers, the first one not presenting fluorescent emission.18-20 The different forms are characterized by defined absorption frequencies and fluorescence yields.21 Recently, the UV-vis absorption and fluorescence spectra of hosted Rh6G molecules have been interpreted in terms of the existence of quenching effects induced by adjacent molecules. The quenching efficiency would depend on the distances and angles between the molecules, whereby the formation of dimers or higher aggregates with no or little relative fluorescence efficiency would be a particular case of intermolecular conformation.22 According to these new interpretation schemes, it is more accurate to consider that both the H and J absorption bands pertain to a single intermolecular geometry rather than to assume the formation of dimers with two different configurations. These bands would describe the collective absorption behavior of neighboring molecules, and depending on the actual intermolecular configuration, they are characterized by slightly different wavelengths for their maxima. Another consequence of this way of describing the optical behavior of hosted dye molecules is a continuous variation in their fluorescence frequencies and yields. The present work deals primarily with the incorporation of Rh6G molecules within new type nondispersive thin films of SiO2 (18) Sasai, R.; Fujita, T.; Iyi, N.; Itoh, H.; Takagi, K. Langmuir 2002, 18, 6578. (19) Bujdak, J.; Iyi, N. J. Phys. Chem. B 2006, 110, 2180. (20) Shinozaki, R.; Nakato, T. Langmuir 2004, 20, 7583. (21) Martinez, V. M.; Arbeloa, F. L.; Prieto, J. B.; Lopez, T. A.; Arbeloa, I. L. J. Phys. Chem. B 2004, 108, 20030–20037. (22) Arbeloa, F. L.; Martinez, V. M.; Lopez, T. A.; Arbeloa, I. L. J. Photochem. Photobiol., C 2007, 8, 85.

Published on Web 06/03/2009

Langmuir 2009, 25(16), 9140–9148

S anchez-Valencia et al.

formed by a columnar microstructure with wide open voids. The films have been prepared by physical vapor deposition at glancing angles (GAPVD), a technique known to yield very open and porous microstructures formed by columns and, if controlled, other geometrical forms.23-26 From the point of view of their morphology, these thin films present some similarities with clays and other layered materials. In fact, the two systems have open and extended voids where molecules like Rh6G can be located. By contrast, the GAPVD films do not have charge-compensating cations in the void space. From a practical point of view another important difference is that, while the GAPVD thin films are transparent and can be directly integrated into photonic devices, the clays usually appear in the form of powders or polycrystalline aggregates that scatter the light and, therefore, cannot be used for optical or photonic applications. In the present work we have tried to develop a methodology enabling the controlled infiltration of Rh6G dye molecules into transparent GAPVD thin films to obtain composite materials usable for optical applications. The method presents significant advantages with respect to the current sol-gel and other wet chemical processes where both the synthesis of the host material and the incorporation of the dye proceed simultaneously. In the method developed here, the host films are prepared by GAPVD, whereby critical properties like thickness, refractive index, composition, etc., can be controlled independently. Then, in a second step, the dye is incorporated into the film by infiltration from a solution without altering significantly the optical constants of the film. The infiltration methodology is based on the classical concepts of the point of zero charge (PZC) proposed in the 1960s to account for the surface chemistry of colloidal oxides immersed in solutions at different pHs.27,28 Besides studying the infiltration process, we have also analyzed the UV absorption and fluorescent emission properties of the incorporated dye molecules and the variation of these properties as a function of temperature. To our knowledge, this is the first time that the thermal stability and optical properties of Rh6G molecules incorporated into a solid host have been addressed in the literature. We think that the proposed methodology of preparation has a general character and could be applied to other oxides and/or dye molecules. In this line, we have recently carried out a preliminary study dealing with the incorporation of Rh6G and Rh800 molecules into GAPVD TiO2 and Ta2O5 thin films.29 The present investigation deals primarily with this new kind of composite Rh6G/SiO2 thin films, and the aim is focused on studying the feasibility and main variables of the preparation procedure, including the infiltration process. The thermal stability and the optical behavior of the incorporated molecules as a function of temperature are the other aspect of this research work. The possibilities of this new hybrid methodology based on the use of thin films prepared by dry methods (i.e., PVD) as host materials and wet infiltration process to prepare stable and optically controllable composite materials are stressed. (23) Robbie, K.; Brett, M. J. J. Vac. Sci. Technol. A 1997, 15, 1460. (24) Wang, S.; Xia, G.; He, H.; Yi, K.; Shao, J.; Fan, Z. J. Alloys Compd. 2007, 431, 287. (25) Brett, M. J.; Hawkeye, M. M. Science 2008, 319, 1192. (26) Hawkeye, M. M.; Brett, M. J. J. Vac. Sci. Technol. A 2007, 25, 1317. (27) Kosmulski, M. Chemical Properties of Material Surfaces; Marcel Dekker: New York, 2001. (28) Hiemenz, P. C.; Rajagopalan, R. Principles of Colloids and Surface Chemistry; Marcel Dekker: New York, 1997. (29) Sanchez-Valencia, J. R.; Borras, A.; Barranco, A.; Rico, V. J.; Espinos, J. P.; Gonzalez-Elipe, A. R. Langmuir 2008, 24, 9460–9469.

Langmuir 2009, 25(16), 9140–9148

Article

Experimental Section SiO2 Thin Films. Composite Rh6G/SiO2 thin films were prepared by using porous SiO2 thin films as a host. For this purpose, transparent and amorphous films of this material were prepared by GAPVD at room temperature on quartz and silicon substrates or on a quartz crystal monitor (QCM), the latter for determination of their pore structure. Evaporation was carried out in an electron bombardment evaporator by using SiO2 pellets as a target. Stoichiometric and columnar thin films of SiO2 were obtained by performing the evaporation in 10-4 Torr of O2 by placing the substrates at a glancing angle of 70° with respect to the evaporator source. This geometry produces films with a tilted columnar microstructure.23-26 Films with a thickness around 350 nm were prepared by this method. A characteristic of these films is that they are very porous and, therefore, are characterized by low refractive index values (n smaller than that of the substrate). Owing to the extraordinary small amount of active material available in the prepared samples, determination of porosity of this kind of thin films is not possible by the classical BET methods based on the adsorption of gases (N2, Kr, etc.) at their condensation temperature. To overcome this problem, we have developed a new method based on the use of a QCM and the measurement of water adsorption isotherms at room temperature.30 Although a full account of the experimental method and the procedure used to extract pore size distributions can be found in this previous paper, it is important to mention here that adsorption/desorption isotherms were measured at room temperature by dosing increasing amounts of water vapor in a chamber where a QCM with the SiO2 film deposited on its surface was placed. Prior to the adsorption, the QCM plates were heated under vacuum at ∼393 K by irradiation with a halogen lamp placed in its vicinity. The possibility of heating the samples prior to the adsorption experiment permits to remove vapors (e.g., water vapor from the atmosphere) irreversibly adsorbed in the samples even after they are kept in vacuum for long periods of time. Once the samples have been heated at 393 K, they were cooled to room temperature, and an adsorption isotherm was measured by following the signal of the QCM as a function of the water pressure in the chamber for both the adsorption and desorption branches. To get quantitative results, the measured amount of adsorbed water was corrected by the mass thickness of the thin film as determined by both Rutherford backscattering (RBS) and X-ray fluorescence analysis. The total pore volume was estimated under the assumption that, once the water vapor saturation pressure was reached, all pores were filled with water. Note that this assumption would not be correct if nonaccessible pores exist in the films. The microstructure of the SiO2 thin films deposited on a silicon wafer was examined by field emission scanning electron microscopy (FESEM) in a Hitachi S5200 microscope. Cross-sectional views were obtained by cleaving the silicon substrates. UV-vis absorption spectra in transmission mode were recorded with a Perkin-Elmer Lambda 12 spectrometer for the thin films deposited on quartz plates. Although the absorbance is the typical magnitude used for presenting UV-vis absorption spectra of dyes in liquid solutions, we present here some spectra as the percentage of transmitted light. We have made this choice because this is the usual way of presenting these data when using optical thin films.31 Refractive indices (n) were determined by UV-vis absorption spectroscopy. A detailed description of these experiments can be found in previous works,32,33 and additional data are gathered in (30) Borras, A.; Sanchez-Valencia, J. R.; Garrido-Molinero, J.; Barranco, A.; Gonzalez-Elipe, A. R. Microporous Mesoporous Mater. 2009, 118, 314–324. (31) Gracia, F.; Holgado, J. P.; Caballero, A.; Gonzalez-Elipe, A. R. J. Phys. Chem. B 2004, 108, 17466. (32) Borras, A.; Cotrino, J.; Gonzalez-Elipe, A. R. J. Electrochem. Soc. 2007, 154, 152. (33) Borras, A.; Barranco, A.; Gonzalez-Elipe, A. R. J. Mater. Sci. 2006, 41, 5220.

DOI: 10.1021/la900695t

9141

Article

Figure 1. (a) Cross-section SEM micrograph of a typical SiO2 thin film prepared by GAPVD. (b) UV-vis transmission spectra recorded and simulated for SiO2 thin film deposited on quartz. The simulated spectrum corresponds to a sample of 350 nm and a refractive index (n measured at 550 nm) of 1.317. The inset (c) shows the calculated refractive index n as a function of the wavelength. Figure 1. The method to extract the refractive index from the UV-vis transmission spectra consists of reproducing the interference pattern of the wave fronts of the incoming beam and the partially reflected beams at the film-air and film-substrate interfaces. This interference pattern depends on the difference between the refractive indexes of the substrate and the film. Fluorescence spectra were recorded in a Jobin-Yvon Fluorolog3 spectrofluorometer using grids of 2 nm for the excitation and emission. Depending on the samples, the fluorescence spectra were excited with radiation of 500 or 515 nm and recorded in the front-face configuration. A term generally used to describe the optical behavior of dyes is the fluorescence capacity.34 This magnitude represents the relative fluorescence efficiency of the system and is defined as the ratio between the intensity at the maximum of the fluorescence spectrum and the value of absorbance at the excitation wavelength, in our case the absorbance at 500/515 nm. From now on we will use in the text this term when referring to the relative fluorescence efficiency defined as in the previous paragraph. Infiltration of dye molecules and preparation of Rh6G-SiO2 composite films. A chloride salt of the cationic form of the Rh6G dye molecule has been used for the experiments. The salt was supplied by Aldrich and used without further purification. In solution, the cationic Rh6G bears a positive charge, a feature that is essential for the incorporation of the dye into the SiO2 thin films as explained in the next sections. A 10-4 M solution of the dye in water at controlled values of pH was used for the infiltration experiments. The pH of the solution was controlled between 1.9 and 10.8 by adding giving amounts of HCl or NaOH. The SiO2 films were immersed in these solutions where they were maintained for 10 min. The samples were then removed from the solution and washed with water at the same pH as that of the dye solutions. The washing final point was taken when no more dye was removed from the films. With this washing treatment any dye molecule that is not irreversibly incorporated within the thin film was removed from its surface. The films were (34) Martı´ nez Martı´ nez, V.; Lopez Arbeloa, F.; Ba~nuleos Prieto, J.; Lopez Arbeloa, I. J. Phys. Chem. B 2005, 109, 7443.

9142 DOI: 10.1021/la900695t

S anchez-Valencia et al. then dried at room temperature by blowing dry nitrogen onto their surface. At certain values of the pH, the thin film samples presented the characteristic pink color of the dye solutions used for infiltration. The intensity of the color changed with the pH of the solution, a feature that pointed out that this parameter and the ability to incorporate the dye into the film are connected. To calculate the concentration of dye molecules incorporated into the films at each pH, the composite films were immersed in a given amount of ethanol where they were kept for 20 min. After this time, it was observed that the films became fully transparent while the solutions presented the pink color characteristic of the solutions of Rh6G. By comparison of the absorption intensity of these solutions with that of a series of reference solutions defining a calibration line, we have been able to quantitatively assess the amount of dye molecules incorporated into the films at each pH. This method was not accurate enough for the composite thin films with the lowest amount of incorporated dye. For these films and for the samples obtained by the heating experiments, the evaluation of the dye concentration was made by comparing directly the area of their absorption bands with that of thin films with a high concentration of Rh6G molecules. This permits to compare the data for specimen with slightly different thickness and/or areas. Heating Experiments. The as-prepared Rh6G/SiO2 composite thin films were quite stable and maintained their optical properties even 1 year after their preparation. We have also investigated the thermal stability of the films. For this purpose we heated the films in an Ar atmosphere up to 275 °C. This treatment was carried out in a tubular furnace by placing the films in a quartz tube in Ar flow during the heating and the subsequent cooling treatments. A ramp of 5 °C min-1 was used to reach the final selected temperature at which the sample was kept for 12 h. The samples used for these experiments were prepared by immersing the host thin film in a 10-3 M solution of Rh6G at pH = 10.8 for 12 h. This procedure produces samples with a high concentration of dye molecules and a high agglomeration degree as will be discussed in the next section.

Results Microstructure of the SiO2 Thin Films Used as Hosts. Figure 1a shows a cross-sectional FESEM micrograph corresponding to a SiO2 thin film prepared by GAPVD at an angle of 70° with respect to the evaporation source. To make easier its observation, the film selected for the FESEM analysis was slightly thicker (i.e., 400 nm) than those used for the infiltrations. It is apparent in this figure that the thin films consist of tilted columns and possess a very porous microstructure. An outstanding characteristic of this porous structure is that open pore channels extend from the surface up to the interface with the substrate. This makes likely that molecules or small particles can be homogeneously incorporated within the whole structure of the film. Despite this columnar structure thin films with a thickness smaller than 500 nm did not disperse the light and were optically transparent as deduced from their UV-vis transmission spectra reported in Figure 1b. In this figure we also plot the evolution of the refractive index (n) of the examined films. It is apparent from this figure that the films have a very low n value estimated in 1.317 at 550 nm. This value, smaller than the typical value of quartz (i.e., 1.44635), is attributed to the high porosity of this material. A first estimation of the pore volume of these thin films based on the assessment of n (i.e., by assuming that the actual n value is the result of the contribution of both the SiO2 and the air/condensed water filling the pores) gives an approximate value of 35%. Another interesting optical property of this type of thin films is that when they are supported on quartz plates their (35) Palik, E. D. Handbook of Optical Constants of Solids; Academic Press: New York, 1985.

Langmuir 2009, 25(16), 9140–9148

S anchez-Valencia et al.

Article

Figure 3. (a) UV-vis transmission spectra recorded for the Rh6G/SiO2 thin films prepared at different pHs as indicated. The inset shows an image taken from actual films prepared at the indicated pHs. (b) Normalized UV-vis absorption spectra recorded for the Rh6G/SiO2 thin films prepared at different pHs as indicated. Figure 2. (a) Water adsorption and desorption isotherms of the films measured with a QCM. (b) Pore size distribution curves derived from the adsorption isotherms.

average transparency is slightly higher than that of the bare substrate, thus acting as antireflective coatings (see Supporting Information S1). For the present work, it is critical to have a direct assessment of the porosity of the films. Figure 2a shows the water adsorption/ desorption isotherms measured with a QCM according to the procedure described in the Experimental Section. It is worth nothing in this plot that some water incorporated in the pores of the films during the first adsorption cycle remained irreversibly adsorbed after desorption. This water is mainly filling micropores, and its removal requires to heat the film at moderate temperatures (i.e., t > 110 °C). This result points to that in the Rh6G/SiO2 composites the pores of the film must be partially filled with condensed water from the atmosphere and/or residual water from the solutions where the immersion experiments were carried out. From the analysis of these isotherms it is also possible to extract the corresponding pore size distribution curves.30 Figure 2b shows the corresponding curves in the range 2r > 2 nm where the Kelvin equation for capillary condensation is applicable.30 From them, it is possible to conclude that in these GAPVD thin films there is a continuous variation in the pore sizes in the whole range from micropores (pores diameter smaller than 2 nm) to mesopores (pore diameter larger than 2 nm). Thus, although a quantitative evaluation of micropores is not possible with this methodology, the tendency indicated by the curves in Figure 2b clearly indicates that in these films there is a high concentration of micropores. This conclusion can be also reached by looking to the shape of the isotherms, particularly to the sharp and irreversible increase of adsorption observed at low vapor pressures.30 The pore size distribution curves also indicate the existence of a relatively high concentration of mesopores with pore diameters comprised between 5 and 10 nm. Pores as large as 20 nm (range not comprise in the figure) are also present in the film. It must be Langmuir 2009, 25(16), 9140–9148

also noted that although the concept of surface area is widely used when dealing with powder materials analyzed by the BET method of adsorption, an equivalent parameter is not attainable with the water adsorption isotherms in Figure 2 since the first jump in the adsorption curve cannot be assimilated to the formation of a water monolayer but to the filling of the micropores. Infiltration of Rh6G Molecules into the SiO2 Thin Films. The infiltration of Rh6G molecules in the GAPVD thin films has been analyzed by UV-vis absorption spectroscopy. Figure 3a shows a series of transmission spectra recorded for composite thin films prepared by immersion in dye solutions at increasing pHs between 1.9 and 10.8. This figure clearly shows that the overall intensity of the absorption bands at around 500 nm is directly dependent on the pH of the solution. It also shows that the relative intensity of the two features detected at around 499 and 523 nm, attributed to the Rh6G molecules,21,22 also varies with this parameter. A photograph of some selected films is also included in the figure to illustrate their different color intensity. It is important to note that UV-vis absorption spectra of water solutions of Rh6G at different pHs show a quite similar shape (see Supporting Information S2), thus stressing that the changes in the spectra of Figure 3 must be associated with the incorporation of the dye molecules into the films. To get a more detailed assessment of the spectra intensity and shape, Figure 3b presents, in a normalized scale, the absorbance spectra corresponding to the transmission curves plotted in Figure 3a. This figure clearly shows that the shape of the absorption spectra changes with the pH. Thus, at low pHs, the band presents a main peak at 523 nm and a very small shoulder at 499 nm. According to the literature,18-22 this spectral shape is typical of Rh6G monomers. Since the relative intensity of the shoulder at around 499 nm increases with the pH, particularly for pH > 8.8, it can be also concluded that the degree of interaction between molecules increases with this parameter.18-22 Additional evidence in this sense is provided by the increase of the energy separation between the two bands due to the displacement of the main peak from 523 to 530 nm.22 DOI: 10.1021/la900695t

9143

Article

S anchez-Valencia et al.

Figure 4. Representation of the intensity of the absorption band of the spectra in Figure 3 and of the actual concentration of dye molecules incorporated into the films plotted as a function of the pH of the dye solution used for preparation of the films. The lines are drawn to guide the eyes.

Calculation of the actual area of the spectra in Figure 3b (i.e., before height normalization) enables an evaluation of the concentration of dye molecules incorporated in the films at the different pHs. Determination of the actual concentration of dye molecules into the films has been also done by extracting the dye molecules as described in the Experimental Section. Figure 4 shows the intensity of these bands as well as the actual concentration of Rh6G molecules in the films, both magnitudes plotted as a function of the solution pH. The two plots have a similar shape characterized by a steady increase with the pH for values of this parameter higher than 2, close to the PZC of SiO2.27,28 For a pH < 2, the amount of incorporated Rh6G molecules was practically negligible. The similar shape of the two curves supports that the intensity of the absorption band can be taken as an approximate measurement of the concentration of dye molecules within the studied films. Figure 5a shows, plotted in a normalized scale, the fluorescence spectra of the films prepared at different pHs. This figure evidence that, as the pH increases, there is a gradual shift in the position of the maxima from 544 to 572 nm and a gradual increase in the width of the curve. Figure 5b shows a plot of the fluorescence capacity of the films as a function of the pH. To visually illustrate this different fluorescence capacity, we have inserted in the figure the images of three films illuminated with a fluorescent lamp (λ = 365 nm). Considering first the shape of the spectra, it appears that up to a pH = 4.5 they are characterized by a main peak at 548 nm and a small shoulder at around 600 nm. At pH = 7.7, the main peak shifts to 555 nm, while the relative intensity of the shoulder increases. At pH g 8.8, the main peak is further displaced to 570 nm, and the width of the spectra is almost twice that of the composite thin films prepared at pH < 4.5. These changes in the shape and position of the bands as a function of the pH are accompanied by a drastic variation of the fluorescence capacity as reported in Figure 4b. The values of this magnitude define a sharply decreasing curve that indicate that, even if the concentration of dye molecules incorporated into the films increases constantly with the pH of the solution (cf. Figure 4), their fluorescence efficiency dramatically decreases with this parameter. It is likely that quenching effects between molecules and/ or formation of nonfluorescent dimers are responsible for this behavior.22 At this point, it is important to note that the shape of the excitation spectra (see Supporting Information S3) for the series of investigated thin films is very similar to that of their absorption spectra presented in Figure 3b. This equivalence in spectral shapes supports that in each film there is only one type of 9144 DOI: 10.1021/la900695t

Figure 5. (a) Normalized fluorescence spectra recorded for Rh6G/ SiO2 composite films prepared at increasing pHs as indicated. (b) Representation of the fluorescence capacity as a function of the pH used for preparing the films. The inset shows images taken for actual films prepared at the indicated pHs that are being illuminated with a low energy fluorescent lamp (λ = 365 nm).

molecules/aggregates and discards that, as traditionally considered, the absorption features at shorter and longer wavelengths are due to respectively monomers and H and/or J dimers. This is so because the H dimers are deemed as having no fluorescence,18-20 and the absorption spectra resulting from the contribution of fluorescent monomers and these nonfluorescent dimers would never yield excitation spectra with similar shapes. Therefore, we consider that the molecules/aggregates formed at each pH are characterized by two bands, the so-called H and J bands, which appear at shorter and longer wavelengths with respect to the maximum of the absorption of the molecule in monomeric form.21 Thermal Behavior of the Rh6G/SiO2 Composite Thin Films. For many practical applications requiring the integration of fluorescent thin films into photonic devices, it is very important to know the thermal stability of the components. To check whether the absorption and/or fluorescence spectra of our composite thin films can be affected by the temperature, we have heated a composite film at increasing higher temperatures in an Ar atmosphere. For this experiment we have chosen a sample that presents a maximum concentration of Rh6G molecules and an absorption spectrum that depicts a very intense shoulder at 500 nm. Figure 6a shows a series of UV-vis transmission spectra recorded for this sample after heating at increasing temperatures between 50 and 275 °C. The spectra of the sample heated at 50 and 100 °C are quite similar to the spectrum of the original thin film and are not represented in the figure. At 125 °C there is already a small change in the position of the maximum that changes from 527 to 524 nm (spectrum not shown). At T > 225 °C, the intensity of the absorption bands decreases as the temperature increases. This change is not accompanied by the appearance of any absorption feature in the spectral region at λ < 400 nm, thus discarding that any molecular moiety resulting from the partial Langmuir 2009, 25(16), 9140–9148

S anchez-Valencia et al.

Article

Figure 7. Representation of the actual concentration of dye molecules incorporated into the films plotted as a function of the temperature at which they have been heated. The line is plotted to guide the eyes.

Figure 6. (a) UV-vis transmission spectra recorded for Rh6G/ SiO2 composite thin films prepared by immersion in a dye solution at a pH = 10, 8 and then heated at increasing temperatures as indicated. (b) Normalized absorption spectra of the same films. Note that the intensity of the high-energy shoulder decreases as the temperature increases.

degradation of Rh6G molecules remains in the film at the investigated temperatures.36 To get a more precise evaluation of the changes induced by heating, we have represented in Figure 6b the normalized absorption spectra of this series of samples. As a reference, the spectrum of the sample prepared at pH = 3.0 is also included in the figure. From Figure 6a,b, it is apparent that at 150 °C there is a little decrease in the intensity of the feature at 500 nm and a relative increase in the intensity of the main peak at 524 nm. Between 150 and 225 °C, there is a further decrease in the intensity of the shoulder at 500 nm, while the intensity of the main peak at 524 nm only starts to decrease at 225 °C. On the other hand, heating also affects the position of the main peak that shifts from about 526 nm for the original film to ∼520 nm for the film heated at 225 °C. Above this temperature, there is a sharp decrease in the overall intensity of the spectra, so that at T > 275 °C no absorption features attributed to the Rh6G can be seen in the plots. It is important to remark here that the optical behavior of the composite samples heated at T > 200 °C did not vary substantially after their storage for periods longer than 3 months. Figure 7 shows the evolution of the actual concentration of Rh6G molecules that remain in the heated films at each temperature. This evaluation has been carried out according to the procedure described in the Experimental Section. The resulting plot shows that the molecule concentration in the film remains practically constant up to a heating temperature of 175 °C. At higher temperatures, the Rh6G concentration decreases to disappear completely at T > 275 °C. To complete the optical characterization of the heated films, we have also analyzed their fluorescent behavior. Figure 8 shows, represented in a normalized scale, the recorded fluorescence spectra as well as the value of their fluorescence capacity plotted (36) Mialocq, J. C.; Hebert, P.; Armand, X.; Bonneau, R.; Morand, J. P. J. Photochem. Photobiol., A 1991, 56, 323.

Langmuir 2009, 25(16), 9140–9148

Figure 8. Normalized fluorescence spectra recorded for Rh6G/ SiO2 composite thin films heated at increasing temperatures as indicated (a) and fluorescence capacity of these films represented against the heating temperature (b).

against the heating temperature. The shape and position of the spectral features in Figure 8a change with the heating temperature. The spectrum of the original sample is very wide, and it is characterized by a very low fluorescence capacity (cf. Figure 8b). At 125 °C there is a shift in the position of the maximum from 570 to 590 nm. Then, further shifts occur between 175 and 275 °C to reach a value of 550 nm at the maximum temperature. These changes in the position of the maximum are accompanied by a decrease in the spectral width, so that at 225 °C and higher temperatures the spectral shape is similar to that of the reference sample consisting of monomers (cf. Figure 8a).18-22 Heating at increasing temperatures has also a strong influence on the fluorescence capacity of the samples. Figure 8b shows that the fluorescence capacity is small and almost constant up to a heating temperature of 225 °C when it sharply increases to reach a maximum at 275 °C. DOI: 10.1021/la900695t

9145

Article

S anchez-Valencia et al.

Figure 9. Schematic representation of the state of the open pore surface of SiO2 for pHs smaller than, equal to, and higher than the PZC of this oxide and description of the way how the Rh6G molecular ions are adsorbed under these three conditions. “A” and “C” in the scheme refer to anions and cations in the double layer of charge compensating the negative or positive charge generated on the surface.

Discussion Thin Film Porosity, Dye Adsorption, and pH. Porosity characterization of the SiO2 thin films used as host has shown that they have large open channels extending from the surface of the films up to the interface with the substrate. In addition, in these films there is a considerable volume of smaller mesopores and micropores. This porous structure, characteristic of the columnar thin films prepared by electron evaporation at glancing angles, ensures that the Rh6G molecules can diffuse through all the columnar structure provided that a certain driving force is favoring the process. Our experiments with dye solutions at different pHs clearly shows a direct dependence between this parameter and the amount of Rh6G molecules incorporated within the film (cf. Figures 3 and 4). This behavior proves that pH is a critical parameter for the control of the infiltration of the Rh6G molecular cations. An effective infiltration only occurs when this parameter has a value higher than 2, close to the PZC of SiO2.27,28 In this regard, it is important to stress that different oxides have different PZC values and that the infiltration dependence on pH from will vary from an oxide to the other. Thus, in a previous publication,29 we have shown that for a pH = 4.0 the incorporation of Rh6G in a columnar film of TiO2 (PZC around 5) is negligible while, according to the results in Figures 3 and 4, a significant amount of this dye molecule becomes incorporated into the SiO2 columnar films. A possible way of explaining this behavior is by considering that, according to the classical theories of formation of a double layer of charge in colloidal suspensions of oxides, the surface of the SiO2 is positively charged when this oxide is immersed in a solution with a pH < PZC (for SiO2 colloidal suspensions the PZC has a value around 2.5;27,28 we assume a similar value for the columnar thin films studied here) and negatively charged for a pH > PZC. If we admit that similar charging effects occur on the pore surface of the GAPVD SiO2 thin films, the incorporation of Rh6G molecules can be described according to the model presented in Figure 9. This model shows three different situations of the SiO2 surface for solution pHs equal to, lower than, and higher than the PZC of this material. At low and high pHs the surface of silica is respectively positively and negatively charged due to the formation of -OH2+ and -Osurface groups whose charge is compensated by counterions in the double layer of charge. Our results show that incorporation of charged Rh6G molecules only takes place on the negatively charge surface and that the adsorption degree increases with the amount of negative charge on the surface. This mechanism of adsorption of the Rh6G molecules on the internal surface of the SiO2 thin film presents some similarities, 9146 DOI: 10.1021/la900695t

but is not equivalent, with that proposed to account for the infiltration of this molecule into the interlaminar space of clay and related laminar materials.6-10 In clay minerals free cations in the interlaminar space are compensating for the permanent negative charge of the layers and can be exchanged by the Rh6G molecular ions during the infiltration process. In our system the incorporation is driven by the negative charge developed on the internal surface of the SiO2 thin films only when the pH is higher than its PZC. The adsorption states of the dye molecules and optical behavior of the composite Rh6G/SiO2 thin films. According to Figure 3, the shape of the absorption spectra changes with the amount of Rh6G molecules that become incorporated into the films. Thus, for low concentrations of Rh6G (samples prepared by immersion at pH e 4.5) the obtained spectra are quite similar to those reported for very diluted solutions of this molecule.37-39 Under certain conditions, similar spectra have been also recorded by several authors when this molecular cation adsorbs as a monomer on different surfaces or becomes intercalated into the interlayer space of clays.6-10 By comparison with these previous studies, we can conclude that at low pHs most Rh6G molecules are incorporated into the film in the form of monomers. As the concentration of Rh6G molecules increases for pH > 4.5, the absorption spectra develop a shoulder at around 499 nm (cf. Figure 3b). According to previous works, the development of this shoulder indicates a progressive interaction between nearby molecules.21,22 Additional information about the state of the interacting molecules can be obtained by comparing their UV-vis absorption spectra after area normalization. The representation of the spectra in this way (see Supporting Information S4) does not show any clear isosbestic point, a feature supporting that for the series of investigated samples there is a smooth and continuous change in the geometrical arrangement of Rh6G molecules rather than a variable contribution of two different molecular states.21 According to the porosity analysis previously shown in Figure 2, there are large pores, mesopores, and also micropores in the films. The Rh6G molecules will be adsorbed in any of these pores provided that the entrance to them does not impose an unavoidable steric hindrance. Formation of dimers may in fact occur preferentially in the small pores. However, this is only an hypothesis as we do not have a clear view of the distribution of Rh6G molecules according to the pore sizes. Under the assumption that the molecules are interacting as dimers, the energy separation between the shoulder and the main absorption feature in the spectra, usually named as “exciton splitting”,40 may provide some hints about the geometry of the molecular ensembles. Thus, using the original exciton model of Kasha et al.40 and other related theoretical assumptions linking the shape of the UV-vis spectra and the agglomeration state of the molecules, it is possible to conclude that at high pH the Rh6G molecules are in the form of “oblique head-to-tail dimers”.21,34,41 Table 1 shows the evolution with pH of physical and geometrical parameters of the dimers as deduced from the exciton model theory.40 The corresponding equations are presented in Supporting Information S5. Table 1 gathers the spectroscopic parameters obtained from deconvolution of the experimental absorption bands into two peaks due to the J and H components. These (37) Vogel, R.; Harvey, M.; Edwards, G.; Meredith, P.; Heckenberg, N.; Trau, M.; Rubinsztein-Dunlop, H. Macromolecules 2002, 25, 2063. (38) Arbeloa, F. L.; Ojeda, P. R.; Arbeloa, I. L. J. Chem. Soc., Faraday Trans. 2 1988, 84, 1903. (39) Lopez Arbeloa, F.; LLona Gonzalez, I.; Ruiz Ojeda, P.; Lopez Arbeloa, I. J. Chem. Soc., Faraday Trans. 2 1982, 78, 989. (40) Kasha, M. Pure Appl. Chem. 1965, 11, 371. (41) Fujii, T.; Nishikiori, H. Chem. Phys. Lett. 1995, 233, 424.

Langmuir 2009, 25(16), 9140–9148

S anchez-Valencia et al.

Article

Table 1. Spectroscopic (Wavelengths of the J and H Components, λJ and λH, and the Ratio of the Oscillator Strength of These Bands, f1/f2) and Geometric and Physical (Interaction Energy, U, Torsional Angle, r, and Intermonomeric Distance, R) Magnitude of the Rh6G/SiO2 Thin Films Prepared at Different pHs spectroscopic parameters geometrical and physical magnitudes pH λJ [nm] 3.0 4.5 5.7 7.7 8.8 10.8

522.4 522.4 524.5 525.0 534.6 534.5

λH [nm]

fJ/fH

2U [cm-1]

R [deg]

R [A˚]

489.5 491.7 492.1 501.5 500.1

5.13 2.86 2.66 1.16 1.19

1287.6 1272.6 1273.1 1234.8 1284.7

131.0 117.3 115.3 92.5 93.1

10.4 10.3 10.2 9.9 9.8

parameters are the wavelength of each component (λJ and λH) and the ratio (fJ/fH) between their areas taken as equivalent to their oscillator strengths. The geometric and physical magnitudes deduced from these values are the interaction energy between the units integrating the dimer (U) and the distance and angle between them (for further information Supporting Information S5). On the basis of this type of analysis, it is possible to suggest that the distance between molecules in the aggregated state decreases as the pH increases, with a substantial change in the geometrical parameters for pH g 8.8. The study of the fluorescence spectra of the infiltrated samples confirms the previous conclusions. At low pHs, the shape of the fluorescence spectra is similar to that of monomers of Rh6G. The high fluorescence capacity found for these samples (cf. Figure 5b) indicates that no significant quenching exists by the interaction between these Rh6G molecules. Then, the sharp decrease in fluorescence capacity found for pH > 3.0 clearly indicates a progressive interaction between molecules that, according to the literature, produces a significant quenching of the fluorescence emission.22,34 This situation has been traditionally attributed to the formation of dimers or higher aggregates.34 The increase in the width of the fluorescence spectra and the shift to longer wavelengths in the position of their maxima as the pH increases (cf. Figure 5a) agree with the previous view. In fact, similar changes have been associated in other systems with a progressive interaction among molecules and the formation of dimers with a certain dispersion of their actual geometries.40 Thermal Behavior of the Rh6G/SiO2 Composite Thin Films. Our heating experiments have shown that the optical behavior of the Rh6G/SiO2 composite films does not change substantially up to 150 °C. This first observation sustains the potential use of our original composite films for integration into photonic devices working below this temperature. In addition, our heating experiments have shown that integer Rh6G molecules remain into the film up to 275 °C and that the optical behavior of samples heated at t > 200 °C does not evolve with time after the heating treatment. The high fluorescence capacity (cf. Figure 8b) of these samples sustains that heated Rh6G-SiO2 composite thin films are good candidates for integration into photonic components working at moderate temperatures. The evolution with temperature of the optical characteristics of the composite films provides interesting clues to analyze the state of the Rh6G molecules during these heating treatments. If we assume that the ratio between the intensities of the H and J bands in the UV-vis spectrum of the initial sample is close to 1 (cf. Figure 6b), an assessment of its spectral shape according to the principles proposed by Fujii et al.41 permits to conclude that the adsorbed molecules are forming dimers with oblique transition dipoles. This is an intermediate situation between “twisted sandwich dimers” and “oblique head-to-tail dimers” which, Langmuir 2009, 25(16), 9140–9148

according to the literature,40,41 are characterized by intensity ratios of the H and J bands larger than 1.3 and smaller than 0.7, respectively. From this initial situation there is a continuous change in the shape and intensity of the spectra as the sample is heated. The representation of the normalized areas of the UV-vis absorption spectra of the heated samples does not reveal the existence of any isosbestic point (see Supporting Information S6), a feature that together with the detection of several components in the fluorescence emission spectra for the series of heated samples (cf. Figure 8, see below for further discussion on this point) indicates that the dimers may present different configurations. Although this behavior precludes a very precise analysis of the geometrical parameters of the dimers, additional information can be gathered about the state of the Rh6G molecules when looking in detail to the shape of spectra. Thus, the very small change observed when the samples are heated at 125 °C (i.e., a shift in the position of the maximum from 527 to 524 nm) can be attributed to changes in the polarity of the matrix caused by desorption of water filling partially the pores of the films (cf. Figure 2) and/or by some slight modification in the environment around the Rh6G molecules. For temperatures between 150 and 200 °C only small changes are observed in both the UV-vis absorption and fluorescence spectra. This finding supports that the intramolecular conformation of the Rh6G molecules is little affected by heating within this interval of temperatures. At t > 200 °C, the overall intensity of the UV-vis absorption (cf. Figure 6a) and the actual concentration of active molecules in the film (cf. Figure 7) start to decrease to become negligible at t > 275 °C. This result indicates that some dye molecules are irreversibly removed from the films when they are heated in this interval of temperatures. Besides, the changes in the shape of UV-vis and fluorescence spectra and the fact that the fluorescence capacity sharply increases in this interval of temperatures (cf. Figure 8b) support a considerable modification in the configuration of the remaining molecules that, according to the considerations in the previous sections, must now appear in the form of monomers.

Conclusions The previous results and discussion have shown that a high concentration of Rh6G dye molecules can be incorporated into SiO2 columnar thin films provided that they present a high porosity. Oxide thin films prepared by GAPVD are ideal materials for this purpose as they present a very open columnar structure and are transparent. This second characteristic makes that the final Rh6G/SiO2 composite thin films can be considered as good systems for their integration into photonic devices. In this work, we have described the porosity of the films acting as hosts, studied their optical properties, and developed a simple method that permits the controlled incorporation of Rh6G molecules into this kind of optically transparent thin film materials. An outstanding characteristic of this type of thin film is that the incorporation of Rh6G molecular ions into the films can be easily controlled by adjusting the pH of the solution. Thus, it has been found that the SiO2 thin films prepared by GAPVD behave in a quite similar way as colloidal suspensions of this oxide in water solutions and that the PZC concept, developed for these latter in the 1950s and 1960s,27,28 can be used to account for the dependence on pH of the adsorption capacity of the films. These concepts can be of general use for the preparation of other type of composite films where optically (or magnetically, etc.) active molecules have to be incorporated into a highly porous oxide thin film acting as host. DOI: 10.1021/la900695t

9147

Article

The analysis of the optical behavior of the composite Rh6G/ SiO2 thin films has revealed that for relatively low concentrations of Rh6G molecular ions they appear as monomers and presents a high fluorescence capacity. Meanwhile, at higher concentrations, the molecules start to interact between them, likely forming dimers whose fluorescence capacity decreases by quenching processes between the neighboring molecules. It is thus concluded that maximizing the fluorescence capacity of the films requires a very precise control of the pH during the infiltration process. The thermal stability and the optical behavior of Rh6G or other similar molecules incorporated into host matrices have been scarcely studied in the literature, even though any practical use of the corresponding composite systems requires of a deep knowledge of this question. Our films are stable over more than 1 year since their preparation and present a reasonable good thermal stability. We have found that the dimers or molecular aggregates are stable up to t e 150 °C. At higher temperatures they become unstable, release a certain amount dye molecules, and/or reorganize on the surface in the form of monomers that are stable up to t = 275 °C. These monomers are characterized by a high fluorescence capacity and stability, remaining unmodified after storage for long periods of time. This sustains the use of this type of thin films for practical applications where the optical elements

9148 DOI: 10.1021/la900695t

S anchez-Valencia et al.

are subjected to moderate changes in temperature and/or other environmental conditions. Moreover, owing to the open pore structure of these films, their use as photonic sensors and related applications implying the interaction of the infiltrated molecules with the environment is a clear possibility that is being explored in our laboratory. Acknowledgment. We thank the European Union (STREP Project PHODYE, Contract No. 033793), the Spanish Ministry of Research and Innovation (Projects NAN2004-09317-C04-01, MAT2007-65764, and CONSOLIDER INGENIO 2010CSD2008-00023), and the Junta de Andalucı´ a (Project TEP2275) for financial support. The help of the “Domingo Martı´ nez” Foundation is also recognized. Supporting Information Available: (S1) UV-vis transmission spectrum showing the antireflective character of the porous oxide thin films; (S2) UV-vis absorption and emission spectra of Rh6G solutions prepared at different pHs; (S3, S4) UV-vis absorption and excitation spectra of dye doped films vs pH; (S5) equations used for the calculations of Table 1; (S6) UV-vis spectra of dye-doped thin films heated at increasing temperatures. This material is available free of charge via the Internet at http://pubs.acs.org.

Langmuir 2009, 25(16), 9140–9148