pubs.acs.org/Langmuir © 2009 American Chemical Society
Proton Absorption in As-Synthesized Mesoporous Silica Nanoparticles as a Structure-Function Relationship Probing Mechanism :: Maria Strømme,*,† Rambabu Atluri,† Ulrika Brohede,† Goran Frenning,‡ † and Alfonso E. Garcia-Bennett †
:: Nanotechnology and Functional Materials, Department of Engineering Sciences, The A˚ngstrom Laboratory, Uppsala University, Box 534, SE-75121 Uppsala, Sweden and ‡Department of Pharmacy, Uppsala Biomedical Centre, Uppsala University, Box 580, 751 23 Uppsala, Sweden Received January 9, 2009. Revised Manuscript Received February 23, 2009
A new method to investigate the effect of pore geometry on diffusion processes in mesoporous silica nanoparticles and other types of micro- and mesoporous structures is put forward. The method is based on the study of proton diffusion from a liquid surrounding the mesoporous particles into the particle pore system. The proton diffusion properties are assessed for a variety of as-synthesized mesoporous nano- and microparticles with two-dimensional and three-dimensional connectivity. Results show that the diffusion coefficients are higher for the proton absorption process than for the release of surfactant template molecules, and that they overall follow the same trend with the more complex three-dimensional mesocaged particles showing the highest diffusion coefficients. The pore geometry (cylindrical pores versus cage-type pores) and structure connectivity are found to play a key role for the effects observed. The results put forward in the present work should offer a valuable tool in the development of porous nanomaterials in a range of applications including the use as catalysis and separation enhancers in the petrochemical industry, as scaffolds for hydrogen storage, and as drug delivery vehicles for sustained release and gene transfection.
Research on ordered mesoporous silica nano- and microparticles is currently being carried out worldwide owing to their large internal surface areas and pore volumes, their tunable pore sizes, and the controllability of both the chemical surface properties and particle size. These properties open up the possibility to tailor-make the specific functionalities of the particles in a number of technological applications ranging from catalysis and separation enhancers in the petrochemical industry to scaffolds for hydrogen storage in energy efficient applications as well as drug delivery vehicles for sustained release and gene transfection in the pharmaceutical world. Recently, AMS-n mesoporous silicas have been prepared using anionic surfactants and alkoxysilane as costructure directing agents (CSDAs).1-4 The AMS-n mesoporous materials constitute an important subgroup to the ordered mesoporous silica because of their wide structural diversity, ease of functionalization, and economic and nontoxic surfactant pore forming agents. Unlike conventional mesoporous preparations, the synthesis mechanism of AMS-n mesoporous materials is governed by the interaction of the CSDA with the selfassembling (micellar) surfactant. Using well established solgel techniques, a silicate wall is then condensed surrounding the micellar-CSDA complex. A porous solid is eventually achieved through the preferential calcination (or solvent extraction) of the organic surfactant from the condensed silica particle. The micellar “liquid crystal”-like phase of the anionic surfactant is hence reproduced or templated into silica form. *Corresponding author. E-mail:
[email protected]. (1) Che, S.; Garcia-Bennett, A. E.; Yokoi, T.; Sakamoto, K.; Kunieda, H.; Terasaki, O.; Tatsumi, T. Nat. Mater. 2003, 2, 801. (2) Garcia-Bennett, A. E.; Che, S.; Terasaki, O.; Tatsumi, T. Chem. Mater. 2004, 16, 813. (3) Garcia-Bennett, A. E.; Kupferschmidt, N.; Sakamoto, Y.; Che, S.; Terasaki, O. Angew. Chem., Int. Ed. 2005, 44, 2. (4) Garcia-Bennett, A. E.; Brohede, U.; Hodgkins, R. P.; Hedin, N.; Strømme, M. Langmuir 2007, 23, 9875.
4306
DOI: 10.1021/la900105u
The use of a CSDA has a double advantage as it not only, as the name indicates, acts as a structural directing agent, but if the surfactant is removed through solvent extraction, it leads to a homogeneous coating of organic functional groups on the internal surface of the pores. A schematic diagram of the synthesis mechanism and the final material composition is shown in Figure 1 where N-lauroyl glutamic acid is used as the templating anionic surfactant, 3-aminopropyl triethoxysilane (APES) as the costructure directing agent, and tetraethyl orthosilicate (TEOS) as the silica source. Because of the high potential of ordered mesoporous silica particles in applications where entities such as drugs, peptides, proteins, growth factors, or hydrogen containing substances are to be protected and released in a predetermined manner, methods to evaluate how the geometry of the internal pore system in the particles affects diffusive transport are greatly needed. One way to probe how different pore structures affect diffusion of entities hosted in the particles is to study the release process of the surfactant template molecules present in the uncalcined material.5 Another hitherto scarcely explored method is to introduce the molecular probe to be used as a cotemplate.6 The drawback with both of these approaches is that the anionic amphiphiles employed in the preparation of AMS-n type materials are very sensitive to a chemical (ionic) environment, and only specific types of surfactant template or cotemplate molecules afford ordered mesoporous structures. To compare a broad range of structures, different molecules will have to be used. This introduces additional parameters other than the geometrical ones affecting the diffusion process, thus preventing one to isolate information about how the (5) Brohede, U.; Atluri, R.; Garcia-Bennett, A. E.; Strømme, M. Curr. Drug Delivery 2008, 5, 177. (6) Ruthstein, S.; Schmidt, J.; Kesselman, E.; Popovitz-Biro, R.; Omer, L.; Frydman, V.; Talmon, Y.; Goldfarb, D. Chem. Mater. 2008, 20, 2779.
Published on Web 3/12/2009
Langmuir 2009, 25(8), 4306–4310
Letter
Figure 1. Schematic representation of the self-assembly mechanism of AMS-n mesoporous nanoparticles. pore geometry affects the diffusion process. Another approach that allows one to use the same type of molecule to probe the structural effects on diffusion processes in a number of structures is to postsynthesis load the materials with a probe molecule and study the subsequent release process. The drawback with this method is that the loading capacity may differ substantially between different structures,7 and hence, when studying the release process, the initial conditions will differ considerably between the different experiments. In this Letter, we propose a new method to investigate the effect of pore geometry on diffusion processes in mesoporous silica nanoparticles and other types of micro- and mesoporous structures. The method is based on the study of proton diffusion from a liquid surrounding the mesoporous particles into the particle pore system. The method originates from our findings during several years of mesoporous silica nanoparticle research that uncalcined mesoporous silica particles absorb protons when introduced in a low pH solution before the template or cotemplate molecule release process starts. The driving force behind the proton diffusion into the mesoporous silica particles is the large concentration gradient introduced at the interface between the high proton concentration solution outside the particles and the water monolayer present in the uncalcined particle pores.8 The proton absorption can be easily monitored by an increased pH and a decreased conductivity in the solution surrounding the particles. By assuming that diffusion is the rate-limiting process, it can be shown that the proton concentration C(t) in the well-mixed solution of volume V surrounding approximately spherical particles of radius R and porosity ε changes according to
when the protons diffuse into the particles (see Supporting Information I) by an effective diffusion coefficient Deff. Here, C0 is the initial proton concentration in the solution and S is the proton binding-site concentration, defined as the ratio between the maximum amount of protons that may be bound inside the particles to the total volume VP occupied by the particles in the solution. The general synthesis of AMS-n mesoporous materials has been reported previously.2 All reagents were obtained from Sigma-Aldrich and used as received. N-Lauroyl amino acid
derived surfactants were obtained from Nanologica AB (Uppsala, Sweden). In a typical experiment, the surfactant is dissolved at temperatures between 60 and 80 °C in distilled water. After complete dissolution of the surfactant, the CSDA 3-aminopropyl triethoxysilane (APES) is added under stirring at the synthesis temperature followed by the addition of the silica source tetraethyl orthosilicate (TEOS) after a period of time x (where x may vary according to the desired AMSn mesostructure). For the synthesis of cubic mesocage structure AMS-8, N-lauroyl glutamic acid (C12-Glut) was employed where the final molar composition in the synthesis gel was C12-Glut/H2O/APES/TEOS = 1:1830:2.67:11.8; the dissolution and synthesis temperature was 60 °C, and x was 2 min. The synthesis of tetragonal mesocage structure AMS-9 is identical to that of AMS-8 except in the time x of addition of TEOS, which is delayed to 5 min. The cylindrical bicontinous cubic structure AMS-6 and hexagonal cylindrical structure AMS-3 were prepared with the N-lauroyl alanine (C12-Ala) surfactant, and the detailed synthesis procedure can be found in the literature.9 For these structures, the final molar composition in the synthesis gel was C12-Ala/H2O/APES/TEOS = 1:309.1:1.25:6.7, and the dissolution and synthesis temperature was 80 °C. The time of addition of TEOS (drop by drop) was 2 min for AMS-6 and AMS-3 samples. AMS-8 and AMS-9 synthesis gels were maintained at the synthesis temperature for a period of 1 day, before transferring them to an oven at 100 °C for a further 2 days. For AMS-6 and AMS-3, the synthesis gels were kept at RT under stirring for 24 h before transferring them to hydrothermal treatment at 100 °C for 3 and 9 days, respectively. The solid samples were then filtered, washed, and dried overnight at RT, under the flow of air. The conductance changes in the solution surrounding the nano- and microparticles were monitored on 3-5 samples each from the three mesoporous particle types under study in 20 mL of nitric acid of initial pH around 2 by using ac conductivity measurements as implemented by the alternating ionic current (AIC) technique.10 In each measurement, 100 mg of particles was added to the solution. The solution was kept at a temperature of 65 ( 2 °C during the measurement by hosting the measuring cell in an incubator (incucell IC 55, BMT a.s., Brno, Czechnia). The experimental setup used was identical to that described elsewhere4 using a cubic measuring cell equipped with stainless steel electrodes
(7) Vallet-Regı, M. Chem.;A Eur. J. 2006, 12, 5934. (8) Anderson, M. W.; Egger, C. C.; Tiddy, G. J. T.; Casci, J. L.; Brakke, K. A. Angew. Chem., Int. Ed. 2005, 44, 3243.
(9) Atluri, R.; Hedin, N.; Garcia-Bennett, A. E. Chem. Mater. 2008, 20, 3857. (10) Frenning, G.; Ek, R.; Strømme, M. J. Pharm. Sci. 2002, 91, 776.
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 3VP 2C0 Deff t ð2S þ εC0 Þ CðtÞ = C0 SR2 2V
Langmuir 2009, 25(8), 4306–4310
ð1Þ
DOI: 10.1021/la900105u
4307
Letter
Figure 2. SEM images of mesoporous particles of (a) AMS-3, (b) AMS-6, (c) AMS-8, and (d) AMS-9, showing the variation in particle morphology and particle size. covering two opposite surfaces of the cell. The cell was covered by a lid to prevent water evaporation during the measurements. A 12 mm magnet, rotating at approximately 60 rpm, stirred the liquid during the entire measurement. A function generator (HP 3325A) applied an alternating voltage (1 VRMS, 10 kHz) to the electrodes of the cell, and the conductance of the cell was calculated from measurements of the voltage across the cell and the current that passed through it using two digital multimeters (Agilent 34401A). This experimental setup minimizes delays caused by transport times to the measurement site and thus enhances the temporal resolution of the measurement. The pH of the solution was measured at the beginning of each conductance measurement as well as at the end of the initial decay region using a Mettler Toledo pH meter. Scanning electron microscopy (SEM) images of the particles were recorded using a LEO 1550 scanning electron microscope, equipped with a Schottky field emisson gun. The scanning electron microscope was operated at 3 kV and at magnifications of between 20 000 and 50 000. Analysis of SEM images (Figure 2) showed that all synthesized particles were approximately spherical and fairly monodisperse. The AMS-3 and AMS-6 samples prepared in this project, however, contained some rod shaped particles of approximately the same diameter (but with a somewhat larger length) as the spherical ones (Figure 2a). The radii used for the assessment of the diffusion coefficients were obtained by 4308
DOI: 10.1021/la900105u
extracting the particle sizes from more than 10 SEM images of each sample. The radii are presented in Table 1, and, as can be seen, the AMS-8 and AMS-9 samples consisted of microparticles, while for AMS-3 and AMS-6 the particle radii were ∼200 and 160 nm, respectively. Figure 3a shows how the conductance in an aqueous solution of hydrogen nitrate (nitric acid) changes as a function of time when four different types of uncalcined mesoporous silica particles are introduced. The initial pH in the solution for all measurements was in the interval 1.84-2.12. The conductance first decreases rapidly for 10-25 min and then increases very slowly for several hours. The initial conductance decrease is associated with an increase of the solution pH and can be related to protons diffusing into the particles, thus leaving fewer charge carriers in the solution contributing to conductance. Thermogravimetric measurements show that the slow increase in conductance following the initial decrease can be linked to the release of the templating surfactants present in the uncalcined particles. Matching the conductance to the measured pH, the initial contribution to the conductance from ionic species other than protons can be subtracted and the proton concentration as a function of time for the initial decay can be extracted (cf. inset in Figure 3a). In order to extract the proton diffusion coefficient in the mesoporous structures using eq 1, C(t) is plotted versus the square root of time (Figure 3b-e), and a linear fit to the data is made. Langmuir 2009, 25(8), 4306–4310
Letter Table 1. Textural and Diffusion Data of the AMS-n Mesoporous Particles under Study
1 Pores sizes and unit cells are not to scale and shown for comparison only. The exact structure of AMS-9 has yet to be resolved. Through electron crystallography modeling, AMS-9 is thought to be composed of three distinct types of micellar cages, close packing in the P42/mnm space group; however, the exact connectivity is still unknown. 2 Reference 5.
The obtained proton diffusion coefficients are displayed in Table 1 together with diffusion coefficients extracted from the surfactant release process in neutral pH solutions at the same temperature used in the present experiment5 (see Supporting Information II for details). From the table it is clear that one obtains the same ranking of the three-dimensional (3D) AMS-6, 8, and 9 structures in terms of diffusion speed using both surfactant release5 and proton absorption to monitor the structures. One also finds that the proton diffusion coefficient in these three structures is ∼2 orders of magnitude larger than that for the surfactants. This is the same relationship between surfactant and proton diffusion coefficients as in free water. The size of the pore space left by the surfactants and entered by the protons is very different; the surfactants leave a pore structure corresponding to that of a calcined material, albeit with a layer of functional groups on the pore wall surface, while the protons enter an environment in which the mesopores are almost entirely filled by surfactants. This strongly indicates that it is the geometry in terms of pore connectivity and tortuosity rather than pore size and pore wall interaction that determine the diffusion speed both for the surfactant release process and for the proton absorption process for AMS-6, 8, and 9. For the two-dimensional (2D) AMS-3 nanoparticles, it was previously found very difficult for the surfactants to escape the structure in spite of the very low tortuosity of this structure with hexagonal cylindrical 2D pores.5 It was argued that the low diffusion coefficient found for the surfactant release process from this structure could be linked to the lack of Langmuir 2009, 25(8), 4306–4310
connectivity between the 2D cylinders and also the relatively few openings to the external surface. For the proton absorption process, these factors do not appear to pose hindrance for the proton movement, as the diffusion coefficient is found to be almost 4 orders of magnitude higher than that for the surfactants, and of the same order of magnitude as that in the 3D cylindrical cubic AMS-6 structure (Table 1). The fact that the proton movement occurs at similar speed in these two structures of similar pore shape (cylindrical channels) indicates that proton diffusion is a better probe of the structure than assessment of the diffusion coefficient for the much larger surfactants. Moreover, one finds that the 3D material (AMS-6) with cylindrical and uniform pores provides lower diffusion coefficients than the 3D material with cagelike pores (AMS-8) having only small openings between the pores. The explanation to this most likely lies in the description of the two types of structures. AMS-6 is composed of two interwoven networks of cylindrical pores, while the structure of AMS-8 is composed of two types of networks of cages. Hence, although both are 3D-connected, their level of connectivity is not the same owing to their different symmetries. While a channel in the AMS-6 structure branches into two further channels, the cage type structure of AMS-8 is composed of larger and smaller cages which are tetrahedrally connected to each other. Although cage windows (or cage connections) may offer a diffusion barrier, it appears that for smaller probe molecules (ions, protons) the connectivity of the pore structure itself may play a more important role in determining the speed of diffusion. DOI: 10.1021/la900105u
4309
Letter
Figure 3. Conductance changes in nitric acid solution as a function of time when four different types of uncalcined mesoporous AMS-n silica particles are introduced; the solution proton concentrations as a function of time for the initial conductance decay period are displayed in the inset (a). Linear curve fits (solid lines) with corresponding equations are shown for recorded proton concentration data points in solution containing AMS-3 (b), AMS-6 (c), AMS-8 (d), and AMS-9 (e) particles versus square root of time.
The proton absorption properties have been successfully measured for a variety of as-synthesized mesoporous nanoand microparticles with 2D and 3D connectivity. Results show that the diffusion coefficients are higher for proton absorption than for surfactant release, and that they overall follow the same trend with the more complex 3D mesocaged materials showing the highest diffusion coefficients. A variety of factors have been used to explain this phenomena, but clearly pore geometry (cylindrical pores versus cage-type pores) and structure totuosity are responsible for the effects observed. In previous work, the 2D hexagonal structure of AMS-3 nanoparticles showed an extremely low diffusion rate with respect to surfactants, whereas in the current work the proton diffusion coefficient was found to be of the same order of magnitude in both cylindrically shaped pore systems under study. The latter indicates that proton diffusion is a better tool to obtain information about pore geometry in porous nanoparticles than assessment of movement of larger ionic species.
4310
DOI: 10.1021/la900105u
The results put forward in the present work should offer a valuable tool in the development of porous nanomaterials as catalysis and separation enhancers in the petrochemical industry, as scaffolds for hydrogen storage, and as drug delivery vehicles for sustained release and gene transfection. Acknowledgment. The Swedish Research Council and the Knut and Alice Wallenberg Foundation are greatly acknowledged for financial contribution to the present work. Supporting Information Available: Supporting Information I showing a pseudosteady state analysis of proton absorption in porous spherical particles, Supporting Information II giving the details on the calculation of proton concentration in solution, and Supporting Information III describing the proton binding site concentration S in mesoporous particles. This material is available free of charge via the Internet at http://pubs.acs.org.
Langmuir 2009, 25(8), 4306–4310
