Release and Molecular Transport of Cationic and Anionic Fluorescent

Sep 4, 2008 - ... Technology, AlbaNova University Center, SE.10691 Stockholm, Sweden ... For a more comprehensive list of citations to this article, u...
0 downloads 0 Views 1014KB Size
Download PDF
11096

Langmuir 2008, 24, 11096-11102

Release and Molecular Transport of Cationic and Anionic Fluorescent Molecules in Mesoporous Silica Spheres Jovice B. S. Ng,† Padideh Kamali-Zare,‡ Hjalmar Brismar,‡ and Lennart Bergstro¨m*,† Materials Chemistry Research Group, Department of Physical, Inorganic and Structural Chemistry, Arrhenius Laboratory, Stockholm UniVersity, SE-106 91, Stockholm, Sweden, and DiVision of Cell Physics, Department of Applied Physics, Royal Institute of Technology, AlbaNoVa UniVersity Center, SE.10691 Stockholm, Sweden ReceiVed April 15, 2008. ReVised Manuscript ReceiVed June 18, 2008 We describe here a method for study of bulk release and local molecular transport within mesoporous silica spheres. We have analyzed the loading and release of charged fluorescent dyes from monodisperse mesoporous silica (MMS) spheres with an average pore size of 2.7 nm. Two different fluorescent dyes, one cationic and one anionic, have been loaded into the negatively charged porous material and both the bulk release and the local molecular transport within the MMS spheres have been quantified by confocal laser scanning microscopy. Analysis of the time-dependent release and the concentration profiles of the anionic dye within the spheres show that the spheres are homogeneous and that the release of this nonadsorbing dye follows a simple diffusion-driven process. The concentration of the cationic dye varies radially within the MMS spheres after loading; there is a significantly higher concentration of the dye close to the surface of the spheres (forming a “skin”) compared to that at the core. The release of the cationic dye is controlled by diffusion after an initial period of rapid release. The transport of the cationic dye within the MMS spheres of the dye from the core to near the surface is significantly faster compared to the transport within the surface “skin”. A significant fraction of the cationic dye remains permanently attached to the negatively charged walls of the MMS spheres, preferentially near the surface of the spheres. Relating bulk release to the local molecular transport within the porous materials provides an important step toward the design of new concepts in controlled drug delivery and chromatography.

1. Introduction Inorganic porous materials are widely used as molecular sieves,1 as adsorbents,2 for wastewater and soil remediation,3,4 and as catalysts.5-7 Each application has specific demands on the features and properties of the porous material. Materials where the porosity and pore size distribution can be tailored are interesting for applications where the uptake and/or release of active components need to be controlled, particularly in drug release and gene delivery.8-10 Surfactant-templated mesoporous materials, characterized by a narrow pore size distribution, tunable pore connectivity, and morphology,11-19 have attracted significant interest as host material in controlled release applications. The effect of confinement and tortuosity on the release kinetics has been evaluated by relatively simple bulk release studies.14,20 Recent studies have also demonstrated that * Corresponding author. E-mail: [email protected]. Tel.: +46 8 674 7023. Fax: +46 8 15 21 87. † Stockholm University. ‡ AlbaNova. (1) Breck, D. W. Zeolite Molecular SieVes: Structure, Chemistry and Use; Wiley: New York, 1974; p 245 (2) Takahashi, A.; Yang, R. T. AIChE J. 2002, 48, 1457–1468. (3) Widiastuti, N.; Wu, H.; Ang, M.; Zhang, D.-k. Desalination 2008, 218, 271–280. (4) Kesraoui-Ouki, S.; Cheeseman, C. R.; Perry, R. J. Chem. Technol. Biotechnol. 1994, 59, 121–126. (5) Nefedov, B. K. Chem. Technol. Fuel Oils 1992, 28, 65–67. (6) Mole, T.; Whiteside, J. A.; Seddon, D. J. Catal. 1983, 82, 261–266. (7) Ho¨lderich, W.; Hesse, M.; Na¨umann, F. Angew. Chem., Int. Ed. Engl. 1988, 27, 226–246. (8) Lai, C. Y.; Trewyn, B. G.; Jeftinija, D. M.; Jeftinija, K.; Xu, S.; Jeftinija, S.; Lin, V. S. Y. J. Am. Chem. Soc. 2003, 125, 4451–4459. (9) Horcajada, P.; Ramila, A.; Perez-Pariente, J.; Vallet-Regi, M. Microporous Mesoporous Mater. 2004, 68, 105–109. (10) Silva, G. A.; Coutinho, O. P.; Ducheyne, P.; Reis, R. L. J. Tissue Eng. Regen. Med. 2007, 1, 97–109. (11) Yanagisawa, T.; Shimizu, T.; Kuroda, K.; Kato, C. Bull. Chem. Soc. Jpn. 1990, 63, 988–992.

the affinity between the loaded guest molecules and the surfaces of the mesoporous material can have a significant effect on both the loading and release kinetics.21,22 In contrast to the numerous investigations on bulk uptake and release, there are only a limited number of studies where the molecular transport within the material has been analyzed. The poorly defined morphology and wide particle size distribution of many of the investigated mesoporous materials also limits the possibility to analyze the data quantitatively. It is clear that a better understanding of the mass transport and diffusion of guest molecules in porous materials with a well-defined morphology (e.g., monodisperse spheres) is important when developing new materials for controlled release as well as chromatography and bioseparation applications. Various experimental techniques can be used to study molecular transport dynamics within a mesoscopic space, e.g., nuclear (12) Beck, J. S.; Vartuli, J. C.; Roth, W. J.; Leonowicz, M. E.; Kresge, C. T.; Schmitt, K. D.; Chu, C. T. W.; Olson, D. H.; Sheppard, E. W.; McCullen, S. B.; Higgins, J. B.; Schlenker, J. L. J. Am. Chem. Soc. 1992, 114, 10834–10843. (13) Kresge, C. T.; Leonowicz, M. E.; Roth, W. J.; Vartuli, J. C.; Beck, J. S. Nature 1992, 359, 710–712. (14) Qu, F.; Zhu, G.; Huang, S.; Li, S.; Sun, J.; Zhang, D.; Qiu, S. Microporous Mesoporous Mater. 2006, 92, 1–9. (15) Lu, Y.; Fan, H.; Stump, A.; Ward, T. L.; Rieker, T.; Brinker, C. J. Nature 1999, 398, 223–226. (16) Rao, G. V. R.; Lo´pez, G. P.; Bravo, J.; Pham, H.; Datye, A. K.; Xu, H. F.; Ward, T. L. AdV. Mater. 2002, 14, 1301–1304. (17) Yang, H.; Coombs, N.; Ozin, G. A. Nature 1997, 386, 692–695. (18) Rathousky, J.; Zukalova, M.; Zukala, A.; Had, J. Collect. Czech. Chem. Commun. 1998, 63, 1893–1906. (19) Che, S.; Li, H.; Lim, S.; Sakamoto, Y.; Terasaki, O.; Tatsumi, T. Chem. Mater. 2005, 17, 4103–4113. (20) Serra, E.; Mayoral, A.; Sakamoto, Y.; Blanco, R. M.; Diaz, I. Microporous Mesoporous Mater. 2008, 114, 201–213. (21) Salonen, J.; Laitinen, L.; Kaukonen, A. M.; Tuura, J.; Bjorkqvist, M.; Heikkila, T.; Vaha-Heikkila, K.; Hirvonen, J.; Lehto, V. P. J. Controlled Release 2005, 108, 362–374. (22) Manzano, M.; Aina, V.; Arean, C. O.; Balas, F.; Cauda, V.; Colilla, M.; Delgado, M. R.; Vallet-Regi, M. Chem. Eng. J. 2007, 137, 30–37.

10.1021/la801179v CCC: $40.75  2008 American Chemical Society Published on Web 09/04/2008

Fluorescent Molecules in Mesoporous Silica Spheres

magnetic resonance,23,24 neutron scattering,25 holographic laser interferometery,26 and single-molecule fluorescence microscopy.27 The ability of confocal laser scanning microscopy (CLSM) to optically section the studied object28 and to directly visualize and quantify the spatial (both 2-D and 3-D) fluorescence molecules distribution within the porous system in real time makes this technique highly suited for studies of the uptake and relatively slow molecular transport in mesoporous materials29-31 and chromatographic media.32 Cheng and Landry31 showed, e.g., in a recent study how CLSM and an amine-binding dye can be used to probe the rate and the extent of amine functionalization of the mesoporous silica spheres. In this study we present how CLSM can be used to directly follow the time-dependent transport inside monodisperse silica (MMS) spheres with a well-defined pore size. The spatial evolution of two fluorescent dyes, one cationic and one anionic, within the negatively charged MMS spheres immersed in a phosphate buffer was followed over an extended period of time. The data was analyzed within the framework of a spherical diffusion model. The molecular transport and release of the anionic dye could be described with a single diffusion constant. A schematic model that describes how the adsorption of the cationic dye onto the negatively charged surfaces of the mesoporous silica materials infers a “skin” with a relatively low permeability that regulates the bulk release and the molecular transport within the mesoporous sphere was presented and the diffusion constant in the skin was determined and compared to the diffusion constant in the core.

2. Experimental Section 2.1. Synthesis of Mesoporous Spheres. Polyethylene oxide hexadecyl ether, Brij 56 (Sigma-Aldrich), a nonionic surfactant, was used as the templating amphiphilic molecule, and tetraethoxysiloxane, TEOS (Purum >98%), was used as the silica source. Hydrochloric acid (HCl) (Normapur, VWR) and ethanol (99.7%, Solveco AB) were used as received. Millipore-grade water was used in all experiments. The precursor solution was prepared by mixing a solution consisting of 3.17 g of Brij 56 in 90 g of ethanol with a solution of 5.4 g of HCl (1 M) in 12 g of ethanol. TEOS (10.4 g) was added to the precursor solution and allowed to hydrolyze under vigorous stirring at room temperature for 20 min. The mesostructured silica spheres have been produced by an aerosol-assisted technique using a vibrating orifice aerosol generator (VOAG Model 3450, TSI Inc., USA), essentially following previously established procedures.16,33 In short, the operation of the aerosol generator is based on the breakup of a cylindrical liquid jet by a vibrating orifice (20 µm) into uniform droplets. The precursor solution is forced through the vibrating orifice at constant speed using a syringe attached to a stepper motor. The generated droplets were (23) Gjerdaker, L.; Aksnes, D. W.; Sorland, G. H.; Stocker, M. Microporous Mesoporous Mater. 2001, 42, 89–96. (24) Valiullin, R.; Naumov, S.; Galvosas, P.; Ka¨rger, J.; Woo, H.-J.; Porcheron, F.; Monson, P. A. Nature 2006, 443, 965–968. (25) Benes, N. E.; Jobic, H.; Verweij, H. Microporous Mesoporous Mater. 2001, 43, 147–152. (26) Roger, P.; Mattisson, C.; Axelsson, A.; Zacchi, G. Biotechnol. Bioeng. 2000, 69, 654–663. (27) Kirstein, J.; Platschek, B.; Jung, C.; Brown, R.; Bein, T.; Brauchle, C. Nat. Mater. 2007, 6, 303–310. (28) Sheppard, C. J. R.; Shotton, D. M. Confocal Laser Scanning Microscopy; Bios Scientific Publishers: Singapore, 1997; p 106 (29) Fu, Q.; Rao, G.V. R.; Ista, K.; Wu, Y.; Andrzejewski, B. P.; Sklar, L. A.; Ward, T. L.; Lo´pez, G. P. AdV. Mater. 2003, 15, 1262–1266. (30) Buranda, T.; Huang, J.; Ramarao, G. V.; Ista, L. K.; Larson, R. S.; Ward, T. L.; Sklar, L. A.; Lopez, G. P. Langmuir 2003, 19, 1654–1663. (31) Cheng, K.; Landry, C. C. J. Am. Chem. Soc. 2007, 129, 9674–9685. (32) Schroder, M.; von Lieres, E.; Hubbuch, J. J. Phys. Chem. B 2006, 110, 1429–1436. (33) Vasiliev, P. O.; Faure, B.; Ng, J. B. S.; Bergstrom, L. J. Colloid Interface Sci. 2008, 319, 144–151.

Langmuir, Vol. 24, No. 19, 2008 11097 injected axially from the vibrating orifice with a turbulent dispersion air jet (∼15 cm3/min) to suppress any coalescence of droplets followed by a greater volume of a laminar flow of dilution air (∼40 L/min) into a vertical drying chamber (φ ) 10 cm) where evaporation of solvents (essentially ethanol and water) takes place. The drying chamber is attached by a stainless steel tube (4 cm diameter) to a three-zone furnace held at 250 °C (calibrated with thermocouples) where the silica particles are condensed before the particles are collected on a filter (Pall, A/D Glass fiber filter, with diameter 4.7 mm and pore size 3 µm). The collected mesostructured particles were calcined at 550 °C for 4 h in air to remove the surfactant template. The details of the characterization of the particles are available in the Supporting Information. 2.2. Dye Loading. Two types of fluorescent dyes were used in this study. The positively charged Rhodamine 6G was from Sigma Aldrich and the negatively charged Oregon Green 488 from Molecular Probes Europe BV. All experiments were performed in Dulbecco’s phosphate buffered saline solutions (PBS, 135 mM, pH 7.2) from Sigma. The mesoporous particles were fixed in gelatin (2 × 10-6 mM) on glass-bottom microwell dishes (Mat-Tek Corporation, USA) and 4 mL of the PBS saline solution was added to prevent drying of the gel. Stock solutions of the dyes were prepared at a concentration of 10 mM with Millipore-grade water. The immobilized particles were loaded with dyes by incubation in a 0.025 mM solution in PBS of the respective dye. 2.3. Imaging of Dye Release from Mesoporous Spheres. The release was initiated by exchanging the dye-containing solution with dye-free PBS. The PBS was exchanged at least three times to ensure the media were free of dye molecules. The solution exchange process was performed in such a way to ensure that the gelatin gel was hydrated at all times. The dye-containing mesoporous spheres were allowed to equilibrate for 60 ( 10 min under dye-free conditions prior to the CLSM imaging. This was to ensure that any excess dye molecules not permanently adsorbed on the exterior surfaces of the sphere were removed and did not interfere with the release of dye from the sphere’s interior. An area with a population of at least 15 isolated spheres with the nearest neighboring spheres at least one sphere diameter away was selected for imaging. The dye-containing spheres were optically sectioned into 5-10 equidistant slices separated by approximately 1 µm. By simple spherical geometrical consideration the slice with the largest diameter (nearly identical to the diameter of the sphere) is taken as the equatorial slice. The glass-bottom dish was covered to minimize evaporation. The spheres were imaged every 30 min until a total release time of 7 h. The position of the equatorial slice was verified every time the spheres were imaged by the procedure described above. An inverted Axiovert 100 M microscope with a Zeiss LSM 5 Pascal scanner was used. A 488 nm argon laser (for Oregon Green 488) and a 543 nm HeNe laser (for Rhodamine 6G) were used together with a 40× /1.3 NA oil immersion objective lens to image the dye-containing spheres.

3. Results 3.1. Morphology and Porosity of Mesoporous Silica Spheres. The mesostructured silica spheres have been produced by an aerosol-assisted technique using a vibrating orifice aerosol generator.16,33 This produces a mesostructured, inorganic/organic hybrid material which transformed into mesoporous materials by removal of the templating surfactant, Brij 56, by calcination. Figure 1a shows that the calcined particles are spherical and relatively monodisperse with an average particle diameter of 6.0 ( 0.2 µm (see details in Supporting Information). The nitrogen isotherm, the XRD spectra, and the TEM images of the spheres are available in the Supporting Information (Supporting Figure 1). The isotherms for the treated powders can be classified as type IV isotherms according to the IUPAC nomenclature, which

11098 Langmuir, Vol. 24, No. 19, 2008

Ng et al.

is typically observed for conventional mesoporous materials.34,35 The specific area, SBET, and the pore volume of the mesoporous spheres determined from the nitrogen sorption isotherm are 800 m2/g and 0.44 cm3/g, respectively. The average pore size calculated using the NLDFT method is around 2.7 nm. The TEM image of the crushed spheres shows that the spheres exhibit disordered mesopores. This is consistent with the XRD spectra, which contain a single broad peak at around 2θ ) 1.92° (respective d spacing ∼ 4.5 nm).16 The characteristics of our aerosolgenerated spheres correlate well with previous studies on mesoporous spheres produced by similar method and materials.15,16 3.2. Immobilization of Mesoporous Spheres and Loading of Anionic Dye. The mesoporous spheres were immobilized by adding a small amount of gelatin (5 mg of gelatin in 100 mL of water) and allowing the polymer to form a percolating gel. The use of gelatin to immobilize the particles enabled us to monitor individual spheres over an extended period of time, thus allowing a facile yet well-controlled experiment. The gel used in our experiments contained more than 99.99 wt % water and was kept hydrated at all times. The interaction

between the dyes and the gelatin is weak, indicated by the fluorescent-free background after the saline solution exchange (Figure 2a and Figure 3a). The high molecular weight and large architecture of the gelatin (a helix polymer, Mw ∼300000 and length ∼208 nm, which adopt to a folding configuration when cooled down)36 prohibits penetration into the pores of the mesoporous spheres. The fluorescent dyes were loaded into the immobilized mesoporous spheres from PBS solution containing 0.025 mM dye. The dyes were allowed to penetrate and fill the pores of the mesoporous spheres over a period of 20 h at room temperature. Longer loading times resulted in an insignificant increase in the fluorescence of the dye-containing spheres, indicating that the material has reached its loading capacity in our system. The homogeneity of the dye concentration inside the mesoporous spheres was evaluated by obtaining CLSM images from an optical section at the equator of the dye-containing spheres. Figure 1b shows that the anionic dye, Oregon Green 488, is evenly distributed within the mesoporous spheres. This demonstrates that the mesopores within the entire volume of the spheres are accessible for molecular transport, and there is no evidence of a hollow core or any region that is unavailable or blocked.29 3.3. Time-Dependent Spatial Distribution of Dyes during Release. Fluorescence confocal laser scanning microscopy (CLSM) was used to study the molecular transport kinetics of the two fluorophores within the mesoporous spheres. Due to the depth and spatial resolution of the technique, it is possible to determine the intensity of the fluorescent dyes in different regions, e.g., at the core and near the surface of the spheres. The release experiments were initiated by exchanging the dyecontaining solution with dye-free PBS. PBS was changed three times and the dye-containing mesoporous spheres were allowed to equilibrate for about 60 ( 10 min before commencing the measurements to ensure that freely adsorbed dye molecules on the exterior sphere surfaces were minimized. Optical sectioning of the dye-containing spheres was done to obtain the equator slice of at least 15 dye-containing particles. The time evolution of the fluorescence within the mesoporous spheres filled with the anionic (Oregon Green 488) and cationic (Rhodamine 6G) dyes are shown in Figure 2a and 3a, respectively. Figure 2a shows that the fluorescence intensity decreases with time and is barely observable after 5 h, indicating that most of the anionic dye molecules (Oregon Green 488) have diffused out of the pores of the host material. The fluorescent intensity of the Oregon Green 488-containing spheres (Figure 2b) decreases from the cores toward the surface of the mesoporous material; i.e., there is a higher dye concentration at the center than at the near edge of the slice. The Rhodamine 6G-containing spheres show very different behavior (Figure 3). The region near the edge (Figure 3a) has a higher fluorescent intensity compared to that at the center of the slice (or the sphere). The spatial intensity variation of the Rhodamine 6G-containing spheres in Figure 3a appears to be centrosymmetric, which is supported by the symmetrical intensity curves in Figure 3b. The higher fluorescent intensity close to the surface compared to the core of the spheres is maintained at all time intervals. Even after a release time of 400 min (∼7 h), there is a significant amount of cationic Rhodamine 6G molecules that remains in a region close to the edge of the slice, resulting in a “bright ring” appearance.

(34) Sing, K. S. W.; Everett, D. H.; Haul, R. A. W.; Moscou, L.; Pierotti, R. A.; Rouquerol, J.; Siemieniewska, T. J. Macromol. Sci., Pure Appl. Chem. 1985, 57, 603–619. (35) Thommes, M.; Smarsly, B.; Groenewolt, M.; Ravikovitch, P. I.; Neimark, A. V. Langmuir 2006, 22, 756–764.

(36) Babel, W.; Schulz, D.; Giesen-Wiese, M.; Seybold, U.; Gareis, H.; Dick, E.; Schrieber, R.; Schott, A.; Stein, W. Gelatin. In Ullmann’s Encyclopaedia of Industrial Chemistry; Bohnet, M., Brinker, C. J., Clemens, H., Cornils, B., Evans, T. J., Greim, H., Hegedus, L. L., Heitbaum, J., Herrmann, W. A., Karst, U., Eds.; Wiley-VCH: Weinheim, 2000; Vol. A12, pp 307-317.

Figure 1. (a) Scanning electron microscopy image of calcined mesoporous spheres and (b) CLSM image of mesoporous spheres that have been loaded with Oregon Green 488.

Fluorescent Molecules in Mesoporous Silica Spheres

Langmuir, Vol. 24, No. 19, 2008 11099

Figure 2. Time evolution of the release of Oregon Green 488 from mesoporous spheres. (a) CLSM images of optical sections at the sphere equator at different times. (b) Spatial distribution of the fluorescence intensity as a function of time. Note: The intensity data is uncompensated for photobleaching.

Figure 3. Time evolution of the release of Rhodamine 6G from mesoporous spheres. (a) CLSM images of optical sections at the sphere equator at different times. (b) Spatial distribution of the fluorescence intensity as a function of time. Note: The intensity data is uncompensated for photobleaching.

3.4. Analysis of the Molecular Release. Quantitative fluorescent measurements require that the effects of light extinction and quenching are insignificant and that the effect of photobleaching is quantified. The effect of light extinction of the spheres was performed by analyzing the fluorescent intensity profiles of different optical sections (see Supporting Figures 3-5). We do not find any significant difference between the top and bottom sections, showing that the effect of light extinction is insignificant. The Gaussian distribution of the dye intensity of the Oregon Green 488 (Figure 4) that is preserved as the dye concentration decreases suggests that quenching is insignificant. As for Rhodamine 6G (Figure 3b), significant quenching would result in a different evolution of the fluorescent intensity profiles than we observe. If quenching dominated, the low fluorescent intensity in the center of the sphere should recover to higher values during release. Instead, we find that the fluorescent intensity profile with a low, and decreasing, intensity in the middle of the sphere is retained as the dye diffuses out of the sphere. We have performed a bleaching study where particles loaded with the less photostable dye Oregon Green 488 were repeatedly scanned using the same parameters as those in the release experiments with the exception that there was no delay between the scans. The decrease in signal intensity can thus be attributed to photobleaching and not to release of dye. The study is described in detail in the Supporting Information. The fluorescent intensities for the analysis of diffusion coefficients have been corrected for photobleaching.

Figure 4. Spatial distribution of the fluorescence intensity as a function of time inside an Oregon Green 488-containing sphere fitted with a Gaussian function, (c,t) ) (b/π)1/2 exp(-x2b).

Rhodamine 6G is a highly photostable dye37 that displays minimal photobleaching within the typical number of scans performed in this study. The total amount of dye within the observed slice at the equator of the sphere can be obtained by integration over the entire measurement volume. Performing the integration of slices (37) Wolf, D. E.; Edidin, M.; Dragsten, P. R. Proc. Natl. Acad. Sci. 1980, 77, 2043–2045.

11100 Langmuir, Vol. 24, No. 19, 2008

Ng et al.

Figure 5. Normalized integrated fluorescence intensity (photobleaching compensated) as a function of observation time, τ, for Oregon Green 488-containing mesoporous spheres. Fifteen different particles were followed individually at 11 time steps. The mean values of the normalized intensity at each observation times (represented by filled boxes) were fitted using an exponential curve (black line).

obtained at different release times can then be used to follow the global decrease of the amount of dye within the spheres as a function of time. Figure 5 shows how the integrated fluorescence intensity decreases with time (τ) for 15 different spheres. In our experiments τ ) 0 is the time point where observations started, i.e., 60 ( 10 min after the exchange to a dye-free solution. The data for each sphere has been normalized to the initial maximum value. There is a relatively large spread in the data. This can be attributed to the inevitable spread in the starting time but also to variations in, e.g., the pore volume and pore structure between individual spheres, a known characteristic of mesoporous spheres synthesized using the aerosol-assisted process.29 However, for each sphere, we find that the decay in fluorescence intensity with time is exponential, which suggests that the release mechanism is similar, although there might be a structural variation between the spheres. The release of fluorescent dye from the particles can be treated as a diffusion-controlled process based on Fick’s second law,

∂C(x, t) ∂2C(x, t) )D ∂t ∂x2

(1)

By modifying the spherical model used to describe sorbate uptake in isothermal spherical particles, we can calculate the effective diffusion coefficient, Deff, of the molecular release from our spheres using the following expression,

It 6 ) I0 π2

(

∞

∑ n12 exp 1

-n2π2Deff t R2

)

(

)

After correcting for the photobleaching of the anionic dye, we relate the decrease in the fluorescence intensity of the anionic dye (Oregon Green 488) in the spheres directly to the concentration. The release profile of the anionic dye molecules from the mesoporous spheres (Figure 5) was fitted with the exponential function in eq 3, yielding an effective diffusion coefficient, DeffOR ) 0.62 ( 0.12 × 10-12 cm2 s-1. This is more than 6 orders of magnitude lower than the reported bulk diffusion coefficient of Oregon Green conjugated molecules (in Tris-EDTA buffer),40 DOG ) 1.3 × 10-6 cm2 s-1. Interestingly, the release data for the cationic dye (Rhodamine 6G)-containing spheres (Figure 6) can not be fitted to a simple exponential function. It is possible to identify two different periods during the release: one initial period (Period 1), which is characterized by a relatively rapid release, and a second period (Period 2), commencing after approximately 2.5 h, which is characterized by a relatively slow release. Interestingly, the release rate during Period 1 actually increases with time. The release curve during Period 2 could be fitted using eq 3 with an effective diffusion coefficient, DeffRh6G ) 0.24 ( 0.04 × 10-12 cm2 s-1, which is 7 orders of magnitude lower than the reported bulk diffusion coefficient of the same molecule (in water),41 DOG ) 2.8 × 10-6 cm2 s-1. It is interesting to note that a statistical analysis of the data shows that the standard deviation is similar, around 16 and 19%, of the diffusion coefficient for both the Oregon Green 488 and Rhodamin 6G-filled spheres, respectively.

4. Discussion (2)

The expression can be simplified by analyzing the asymptotes at short or long times only. We used the long time asymptote estimation in our calculations as

It -π2Deff t 6 ) 2 exp I0 π R2

Figure 6. Normalized integrated fluorescence intensity as a function of observation time, τ, for Rhodamine 6G-containing mesoporous spheres. Fifteen different particles were followed individually in 15 time steps. The mean values of the normalized intensity at each observation time (represented by filled boxes) in Period 2 were fitted using an exponential curve (black line).

(3)

where It is the measured intensity and I0 is the initial intensity. (38) Cejka, J.; Bekkum, H. v.; Corma, A.; Schu¨th, F., Introduction to Zeolite Science and Practice; Elsevier: Oxford, 2007.

We were able to describe the time-dependent reduction in fluorescent intensity in the anionic dye (Oregon Green 488)containing spheres (Figure 5) with eq 3 and a single diffusion constant. An analysis of the transport within the Oregon Green-containing spheres ( Supporting Figure 6) was also performed by analyzing the time-dependent concentration profiles in the center and edge of the equatorial slice of the Oregon Green-containing spheres. The release of Oregon Green 488 in the center and in the edge region of the spheres are very similar, which gives further support (39) Crank, J. In The Mathematics of Diffusion, 2nd ed.; Oxford University Press: Oxford, 1979; p 238. (40) Berland, K. M. J. Biotechnol. 2004, 108, 127–136. (41) Magde, D.; Elson, E. L.; Webb, W. W. Biopolymers 1974, 13, 29–61.

Fluorescent Molecules in Mesoporous Silica Spheres

Langmuir, Vol. 24, No. 19, 2008 11101

Figure 8. Schematic model describing the molecular transport of the cationic dye within a negatively charged cylindrical mesopore, at close to the surface (edge) of the sphere, and in the core (center) of the sphere. Dc and De are the effective diffusion coefficients and dc and de are the effective pore diameters in the core and surface zone, respectively.

Figure 7. Normalized fluorescence intensity as a function of observation time, τ, for the Rhodamine 6G-containing mesoporous spheres. Data from two separate zones within the spheres, one zone at the center and one zone near the edge of the optical equatorial slices, is plotted together with the entire slice. The dotted lines are fits of the experimental data during Period 2 to an exponential curve (eq 3). The insert illustrates the two zones.

to the notion that the transport of the anionic dye inside the mesoporous spheres can be described well using only one diffusion constant. Any additional effects related to, e.g., adsorption of the anionic dye to the pore walls can be ignored, which is not surprising considering the negative charge of the silica walls at neutral pH. The time-dependent release of the cationic dye (Rhodamine 6G) with a distinct transition in release rate and change in the shape of the release curve after a few hours (Figure 6), however, cannot be described by a simple diffusion model. Whereas previous works have shown how empirical nonideal factors can be added to Fickian expressions to accommodate guest molecule-host surface interactions,22,29 such approaches cannot describe the observed release behavior of Rhodamine 6G. It is necessary to also consider the initial inhomogeneous distribution of the cationic dye within the mesoporous silica spheres and in particular the significantly higher amount of dye molecules close to the surface compared to that at the center of the slice in the description. It is not clear what mechanism is responsible for the inhomogeneous distribution of the cationic dye molecules within the mesoporous spheres during loading. Allowing the dyes to adsorb for an extended period of time (up to 96 h) had a negligible effect on the initial distribution. Previous work has indicated an inhomogeneous distribution of ionizable silanol groups within the mesoporous spheres.31 We also consider charge-regulation effects, which can result in an inhomogeneous counterion and charge distribution when the cationic dyes are predominantly adsorbed onto the pore walls close to the sphere surface.42 Figure 3 shows that a significant fluorescence remains inside the mesoporous spheres after extended release times. Control experiments at release times up to 168 h showed that a significant fraction of the fluorescence dyes appears to be permanently adsorbed onto the pore walls. These permanently adsorbed dye molecules mainly reside in a zone near the surface of the spheres. The spatial information available from the CLSM images has allowed us to also analyze the molecular transport of the cationic dye within the mesoporous spheres. Figure 7 shows that the time-dependent behavior in the zone near the edge and at the center of the slice mimic the bulk release profile (Figure 6), also included in Figure 7 as a reference. (42) Zhmud, B. V.; Bergstro¨m, L. Charge regulation at the surface of a porous solid. In Surface of Nanoparticles and Porous Materials; Schwarz, J. A., Contescu, C. I., Eds.; Marcel Dekker, Inc.: New York, 1999; pp 567-592.

An initial release burst has been observed and reported in several molecular release studies on mesoporous materials.20-22 However, the observed increase of the release rate toward the later part of Period 1 (Figure 7) suggests that a second process in addition to diffusion of initially nonadsorbed dye molecules also becomes important. A plausible explanation is that the diffusion of the free (nonadsorbed) dye is complemented with a delayed desorption process. It appears that the desorption rate increases with time up to the end of Period 1. At longer time, identified as Period 2, the drastic drop in the release rate suggests that the desorption process becomes less pronounced. Indeed, the good fit of the release rate of the cationic dye during Period 2 to the spherical diffusion model in eq 3 indicates that the release is mainly diffusion-controlled during this period. However, although the molecular transport both in the edge and the center zone of the slice is dominated by diffusion during Period 2, we find that the effective diffusion coefficient, Deff, is almost 3 times slower in the near edge zone (0.20 × 10-12 cm2 s-1) than in the center zone (0.73 × 10-12 cm2 s-1). This suggests that the permanently adsorbed dye molecules predominantly in the edge zone provide an additional diffusion barrier. The average DeffRh6G ) 0.24 ( 0.04 × 10-12 cm2 s-1, estimated from the release of the cationic molecules from the mesoporous spheres, is close to the diffusion constant in the edge zone. This suggests that the permanently adsorbed cationic dye molecules act as a semipermeable “skin”, schematically illustrated in Figure 8. In the edge zone, the effective pore diameter (de) is reduced by the molecules permanently adsorbed onto the pore walls and thus hindered the diffusion of the cationic dye molecules across this zone.

5. Summary and Conclusions Confocal laser scanning microscopy was used to evaluate the release and molecular transport of charged fluorescent dyes in mesoporous silica spheres. The mesoporous spheres were produced with an aerosol-assisted self-assembly method that yields relatively monodisperse spheres with a pore size of 2.7 nm. Immobilization of the mesoporous spheres into the liquid phase by a dilute gelatin gel allowed us to monitor one sphere over extended periods of times. Whereas the time-dependent release of the anionic dye (Oregon Green 488) followed a simple diffusion-driven behavior that could be described with a single diffusion constant, it was not possible to describe the release of the cationic dye (Rhodamine 6G) with any simple model. A significant fraction of the Rhodamine 6G is permanently adsorbed onto the pore walls of the negatively charged mesoporous silica spheres, predominantly near the surface of the spheres. In the oppositely charged guest-host system, the electrostatic attraction between the guest molecules and the host solid resulted in a semipermeable “skin”, which delayed the release of the dye molecules. The self-rate-limiting mechanism imposed by the skin effect should be further exploited for applications such as drug delivery where controlled release is essential.

11102 Langmuir, Vol. 24, No. 19, 2008

Acknowledgment. This work was supported by the Swedish Science Council (VR) and by VINNOVA through a joint program with JST. Special thanks go to Dr. Niklas Hedin for fruitful discussion and to Mr. Jacob Kowalewski for sharing his expertise in MatLab for the data analysis. Supporting Information Available: Detailed information on the characterization of the mesoporous spheres (Supporting Figure 1), details on the photobleaching compensation for the release study of the

Ng et al. Oregon Green 488-containing spheres (Supporting Figure 2), and optical sections of both the Rhodamine-6G containing spheres (Supporting Figures 3 and 4) and the Oregon Green 488-containing spheres (Supporting Figure 5). Supporting Figure 6 is the concentration profiles (with respect to time) at the center and the surface of the Oregon Green 488-containing mesoporous spheres. This material is available free of charge via the Internet at http://pubs.acs.org. LA801179V