Accepted Manuscript Confined mesoporous silica membranes for albumin zero-order release Nicola Gargiulo, Ilaria De Santo, Filippo Causa, Domenico Caputo, Paolo Antonio Netti PII: DOI: Reference:
S1387-1811(12)00216-8 10.1016/j.micromeso.2012.04.003 MICMAT 5470
To appear in:
Microporous and Mesoporous Materials
Please cite this article as: N. Gargiulo, I. De Santo, F. Causa, D. Caputo, P.A. Netti, Confined mesoporous silica membranes for albumin zero-order release, Microporous and Mesoporous Materials (2012), doi: 10.1016/ j.micromeso.2012.04.003
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Confined mesoporous silica membranes for albumin zero-order release Nicola Gargiuloa,*, Ilaria De Santoa,b,*, Filippo Causaa,c, Domenico Caputoc, Paolo Antonio Nettia,b,c a
Centro di Ricerca Interdipartimentale sui Biomateriali, Università Federico II, P.le
Tecchio 80, 80125 Napoli, Italy b
Centre for Advanced Biomaterials for Health Care, Italian Institute of Technology,
Largo Barsanti, 80125 Napoli, Italy c
Dipartimento di Ingegneria dei Materiali e della Produzione, Università Federico II,
P.le Tecchio 80, 80125 Napoli, Italy *
These two authors contributed equally to this work.
Corresponding author: Domenico Caputo (e-mail:
[email protected]; address: Dipartimento di Ingegneria dei Materiali e della Produzione, Università Federico II, P.le Tecchio 80, 80125 Napoli, Italy; telephone: +397682396; fax: +397682394) Abstract In this work, the transport capabilities of anodic alumina membranes confining SBA-15-like nanochannels in their pores were investigated. The mean pore size of the confined mesophase was set around 7.0 nm, thus to achieve single passage of the selected Bovine Serum Albumin (BSA) protein, having comparable hydrodynamic diameter, through the nanochannels. Shape, size and orientation of the manufactured mesophase were characterized by means of small-angle X-ray scattering, transmission electron microscopy and N2 adsorption / desorption at 77 K. The usage of
mesostructure-containing samples having pore tuned over the size of the released BSA allowed achieving a long-term zero-order release profile by single-file diffusion up to several weeks, which delivers a constant amount of drug by unit time independently from drug concentration and drug accessible area, while control anodic alumina membranes showed a classic Fickian diffusion release. The attained BSA constant release rate was about 2 μg/day. The possibility of rate tuning could be further exploited by varying the length of the siliceous nanochannels inside the alumina pores, i.e. through different amounts of mesostructure filling the host support. Keywords: Mesoporous materials; Drug delivery; Confined diffusion; Zero-order release 1. Introduction The morphological control of porous systems on the nanoscale is a field of research that recently turned out to be very active. In fact, the development of structures consisting in ordered arrays of oriented nanochannels would allow to efficiently perform operations such as the inclusion synthesis of a vast range of nanowire systems through hard templating, the separation of large biomolecules, or the long-term and controlled delivery of drugs. A shining example of such structures that is subject of many recent research activities is represented by hierarchical systems made of mesoporous siliceous materials confined in the channels of anodic alumina membranes (AAMs) [1, 2]. In these systems the silica-surfactant nanocomposite is assembled at the alumina pore walls to form the surfactant-templated silica-nanochannels which are typical of such mesostructures. These materials can be synthesized by different methods, the most investigated of which relies on the so-called evaporation-induced self-assembly (EISA) [3]. In one of the first studies about mesoporous silica confined in AAMs by EISA,
different porous alumina membranes having distinct channel diameters (i.e., from 18 to 80 nm) were used as substrates and dip-coated with a synthesis solution containing Pluronic P123 as templating agent. The development of different mesostructures (from single chains of spherical mesopores to concentric or chiral helical mesopores) was found to depend from the different diameters of the hosting alumina nanochannels [4]. More recently, Bein and coworkers conducted thorough experiments on the formation of mesophases inside AAMs by means of two-dimensional small-angle X-ray scattering (2D SAXS) and transmission electron microscopy (TEM) [5]. These studies were then deepened by performing grazing incidence SAXS (GISAXS) measurements [6], by formulating a possible formation mechanism for the confined mesostructure [7] and by investigating the influence of the addition of inorganic salts on the orientation of the channels in confined mesoporous silica templated by non-ionic triblock copolymer surfactants [8]. Later on, Bein and coworkers also considered possible applications of confined silica mesostructures as drug delivery systems (DDSs). These studies reported the adsorption and the in vitro release kinetics of mesostructure-containing AAMs previously loaded with vancomycin [9] or ibuprofen [10], showing that so-called hexagonal columnar systems (i.e., systems in which the cylindrical pores of the confined phase are straight and parallel to those of the hosting AAM [1]) attain a slower release over an extended time period, and pointing that major parameters affecting drug release are pore shape, accessibility, length, and diameter compared to that of the eluting molecule. Indeed, the release or adsorption of bioactive agents from mesoporous materials depends in a complex way from the physicochemical properties of both the biomolecule and the
device. Furthermore, the dynamic behavior of the transported molecules depends in turn on the surface chemistry and size of the pores. Although the reported works already tested mesostructure-containing AAMs as drug carriers, none were devoted to the assessment of their transport properties as drug delivery membranes. The most appropriate configuration of the confined mesophase for this kind of application is represented by the hexagonal columnar orientation, which can also be tailored in pore size on the nanometric scale by accurately choosing the templating non-ionic tri-block copolymer surfactant (e.g., Brij 56 instead of Pluronic P123 [8]). Indeed, the exploitation of mesostructure-containing AAMs as drug delivery membranes could be advantageous for the long-term delivery of biomolecules, proteins in particular. Among the diverse release mechanisms, the most fruitful for pharmaceuticals applications consists in the zero order release mechanism, which delivers a constant amount of drug by unit time independently from drug concentration and drug accessible area [11]. Zero order release can be achieved through a single file diffusion mechanism of molecules trafficking in a pore having similar size. Indeed, the single-file diffusion of the released molecule is made possible when the diameter of the constraining 2D geometry in the membrane is smaller than twice the size of the released molecule, and, as a result, the effective release rate becomes constant with time due to the single molecule passage [12]. Recent studies have already demonstrated single-file diffusion of protein drugs through nanoprous membranes. In particular, membranes consisting of cylindrical block copolymers nanochannels were developed and obtained single-file diffusion of proteins through the fine-tuning of the pore membrane over the size of the bioactive agent to release [13]. The pore size was indeed precisely modulated to that of the released agents
through Au deposition. Other groups showed the constant release rate of proteins through nanoporous silicon membranes having a slit shaped pore nanometric in height, although obtained through multiple steps of precision silicon fabrication techniques on silicon-on-insulator substrates [14]. Here, we investigated the transport capabilities of AAMs confining SBA-15-like columnar nanochannels in their pores (SBA-15-AAMs). The mean pore size of the manufactured SBA-15-AAMs was set around 7.0 nm, and showed the long-term capability of zero-order release of the fluorescent BSA proteins (molecular size 7.2 nm) up to several weeks. 2. Experimental 2.1 Synthesis of AAM-confined mesoporous silica The confinement of mesoporous silica in AAMs was performed following the procedure reported in [8]. First, 2.08 g of tetraethyl orthosilicate (Aldrich) were mixed with 3.00 g of a 0.2 M HCl (J. T. Baker) aqueous solution, 1.80 g of H2O and 4.00 g of ethanol (Fluka). This mixture was then heated at 60 °C for 1 h to achieve acid-catalyzed hydrolysis-condensation of the silica precursor. The pre-hydrolyzed silica was then mixed with 0.75 g of the non-ionic triblock copolymer surfactant Pluronic P123 (Aldrich) that was preliminarily dissolved in 11.85 g of ethanol. In order to achieve the hexagonal columnar orientation of the mesoporous siliceous channels, 0.085 g of LiCl (Aldrich) were added to the resulting mixture [8]. The AAMs used (Anodisc, Whatman) have a diameter of 13 mm and an average pore size of 200 nm. Two drops of the precursor mixture were homogeneously spread on the membrane surface and let to evaporate at 30 °C at about 50% relative humidity. Mesostructure-containing membranes were then calcined using a heating rate of 0.5 °C/min with annealing
periods of 10 h at 120 °C, and 5 h at 200, 300 and 500 °C, respectively. Because the confined phase may be thought to resemble SBA-15 mesoporous silica in both its structure and textural properties [15], the final product was labeled as SBA-15-AAM. 2.2 Characterization of SBA-15-AAM SBA-15-AAM samples were characterized with SAXS by means of an Anton Paar SAXSess instrument operating in point-collimation mode with a Genix Microsource X-ray generator (CuKα radiation) at 50 kV and 1 mA. The experimental setup is similar to that reported in [8]: in detail, SBA-15-AAM samples were stuck onto a Variostage X-Rotation cell using a small amount of vacuum grease; the tilting angle was set to 10° and the data collection was performed in 5 exposures of 120 s each that were successively averaged into one image. Moreover, TEM images of SBA-15-AAM samples were collected by means of a Philips EM 208 instrument equipped with a MegaView camera. Prior to the observation, membrane samples were embedded within resin and then ultramicrotomied. In order to analyze the pore structure of the synthesized hierarchical structure, partially ground samples of SBA-15-AAM were submitted to nitrogen adsorption/desorption at 77 K by means of a Micromeritics ASAP 2020 apparatus. The density functional theory (DFT) pore size distribution was determined modeling experimental adsorption data with a kernel function for oxide materials having cylindrical pore geometry. 2.3 Release experiments In order to assess protein release through nanochannels, SBA-15-AAMs were glued by a cyanoacrylate solution onto Transwell support, in place of the original membrane, and between two reservoirs. Control Anodisc membranes, having 200 nm pore size, were attached to a Transwell support in the same way. The proposed experimental
system envisages the uniaxial flow of biomolecules through the mesoporous membrane. The membrane separates an initially filled biomolecules-containing reservoir from a phosphate buffered saline (PBS) solution environment where biomolecules are depleted. Before loading, membranes were preconditioned in 30% ethanol solution for thirty minutes and then put in water for about ten days. After preconditioning, the volume above the membranes was filled with 0.5 ml of 2.5 M fluorescent BSA-TetraMethylRhodamine protein solution (Molecular Probes, MW 67 kDa), whereas the lower volume contained 2 ml of PBS plus 0.03% sodium azide. The release system was kept at 23 °C on a 50 rpm shaking plate. Protein concentration in the lower reservoir was measured by means of a Perkin-Elmer Spectrofluorimeter (485-535 nm). The BSA concentration was determined from a 100 L solution aliquot removed from the sink reservoir and replaced with an equal amount of fresh PBS plus 0.03% sodium azide at intervals of approximately 48 h and followed up to two months. 3. Results and discussion 3.1. Characterization of SBA-15-AAM Fig. 1 shows the SAXS pattern of a SBA-15-AAM sample: as reported in the literature [8], SAXS measurements on siliceous mesostructures confined in AAMs usually result in diffraction patterns with two or four visible first-order reflections that can be correlated to the orientation of the mesoporous channels with respect to the macroporous ones of the hosting AAM. The reflections in the horizontal plane of the primary beam are called in-plane (ip) reflections, and are detectable when the hexagonal columnar orientation occurs, but also when the mesophase assumes a closed, donut-shaped circular configuration [1]. The reflections out of the horizontal plane are called oop reflections and occur only in correspondence of the circular orientation.
When ionic surfactants (such as cetyltrimethylammonium bromide) are used as templates for the assembly of the confined phase, the process spontaneously leads to the hexagonal columnar arrangement of the mesostructure [5]. On the contrary, when using non-ionic triblock copolymer surfactants (e.g., Pluronic P123) as templates, the assembly evolves towards the circular phase. The addition of an inorganic salt (LiCl) to the P123-based sol allows to mimic the behavior of a sol based on an ionic surfactant: by this way, the resulting confined mesostructure will be characterized by a high accessible porosity due to its columnar orientation, and also by a high pore size due to the usage of a long-chain, non-ionic surfactant as template. These observations are confirmed by the analysis of Fig.1, in which the scattering pattern of the SBA-15-AAM sample shows distinct ip reflections at a value of the scattering vector q of about 0.51 nm-1 (corresponding to a d-spacing of 12.3 nm), and only weak oop reflections with almost the same d-spacing (~12.7 nm). As a consequence, the mesophase is mainly hexagonal columnar, with only a small fraction of hexagonal circular domains: this result is well comparable with others reported in the literature [8]. [Figure 1] Fig. 2 shows a TEM image of a SBA-15-AAM sample, in which one of the 200 nm-wide pores of the AAM is filled with the siliceous mesostructure. In previous papers dealing with similar topics [8], TEM observations consisted in top views of the membranes, and the samples were prepared by dimple grinding and successive ion milling. On the contrary, in this case, the sample was embedded within resin and then ultramicrotomied: this approach, though basically more destructive than the aforementioned one, allowed to directly observe a cross section of the membrane. In Fig. 2, the hexagonal columnar orientation of the mesoporous siliceous channels is
clearly visible, confirming what already pointed out from SAXS experiments. Moreover, the adherence of the confined phase to the AAM channel surface is plainly demonstrated: as reported in the literature [8], when the confinement of the P123-templated mesostructure is performed without the addition of inorganic salts to the precursor sol, the resulting circularly oriented phase turns out to be significantly detached from part of the wall of the hosting macropores after the assembly process; such phenomenon is further stressed after the calcination treatment. On the contrary, the presence of LiCl in the synthesis environment, apart from promoting the columnar orientation of the siliceous channels, aids to protect the confined phase from such detachment during both the gelation and the template removal steps. [Figure 2] Fig. 3 shows the N2 adsorption/desorption isotherm at 77 K on partially ground samples of SBA-15-AAM: the shape of the isotherm is coherent with what reported in the literature [8]. In particular, it is classifiable as IUPAC type IV, which is typical of mesoporous materials [16]; moreover, a H1-type hysteresis loop is highlighted. The BET specific surface area turned out to be about 40 m2/g, in fair accordance with the literature: such result seems odd if compared with those (one order of magnitude higher) usually associated to bulk mesoporous silica particles, but becomes plausible when considering that the mesophase content inside the composite membrane may be as low as 10 wt% [8]. [Figure 3] Fig. 4 shows the DFT pore size distribution of partially ground samples of SBA-15-AAM obtained from modeling of experimental data reported in Fig. 3. Again, the results are coherent with those reported in the literature [8]: the distribution has a
main peak at about 7.0 nm, that actually corresponds to the pore size of the confined mesostructured phase. The tail that follows such peak in Fig. 4 may be related to defects originated during the grinding process, but also by the occasional merging of silica pores. This phenomenon may be considered as a side effect of the increased interactions of the mesophase with the alumina channel walls due to the addition of an inorganic salt to the synthesis system. In fact, while in the absence of LiCl the mechanical stress caused to the circularly oriented siliceous channels during the calcination process leads to the detachment from the alumina pore walls, the columnar phase originated in the presence of the inorganic salt remains attached to the alumina, and the mechanical stress related to the thermal treatment propagate to the silica structure, causing the rupture of a small fractions of its pore walls [8]. [Figure 4] 3.2. Release experiments Bovine Serum Albumin (BSA) protein was selected for the evaluation of the transport properties of the manufactured membrane. Indeed, BSA is characterized by a molecular weight of 67000 Da and a hydrodynamic diameter of around 7.2 nm [17], which is close to the main peak value of the mesopore size distribution of SBA-15-AAM samples. The quality of BSA release profile through SBA-15-AAM was assessed through the evaluation of the time-dependent transport of proteins across a membrane mounted in a Transwell support. Fig. 5 shows the release profile of BSA through SBA-15-AAM, and through the control alumina membrane. Open circles represent the released BSA amount from the control AAM having nominal pore size of 200 nm, while full stars the quantity released from a SBA-15-AAM with a mean pore size of around 7.0 nm. No
significant time lags are recognized in Fig. 5, which were instead present in absence of membranes preconditioning (data not shown). [Figure 5] The constant release profile of BSA is obtained in SBA-15-AAM, whereas a classic Fickian diffusion release profile is obtained for the AAM. In the case of AAM, the release profile has been adjusted following the mass balance, which applies under the assumption of Fickian diffusion for a quasi-stationary process [18] reported below:
(
)
ln[1 − C (t ) / C(∞)] = − DnA p / L × (1/V1 +1/V2 ) × t
(1)
In Eq. (1), C(t) is the protein concentration in the sink reservoir at a time t, C(∞) is the concentration of the system at infinite time, D is the protein diffusion coefficient in the pores, n and Ap are the number of pores and their section respectively, L represents the membrane thickness, V1 and V2 are the source and sink volumes. The BSA diffusion coefficient in AAM could be roughly esteemed from Eq. (1), attaining 3.5E-9 cm2/s, which is around two orders of magnitude smaller than in bulk, 6E-7 cm2/s. Since the ratio of molecule diameter over pore size is in this case 0.04, such a reduction of diffusion coefficient cannot be uniquely ascribed to constrained diffusion, and could be rather due to an over esteem of the pore number in AAMs, which are declared having a broad pore density range. On the other hand, in the case of SBA-15-AAM the BSA release profile exhibits a clear zero order release, which follows the equation Qt = Q0 + K 0 × t
(2)
where Qt is the protein amount dissolved at time t, Q0 is the initial amount of drug in the solution, and K0 is the zero order release constant independent of the solvent accessible area.
In particular, since during the first two days of observation, the release curve of SBA-15-AAM shows a higher slope compared to the longer times, probably ascribable to low sensitivity of the reading apparatus, the linear fit of the profile in Fig. 5 is evaluated from day three of release attaining a 0.990 r-square, which suggests a zero order release rate of 1.83 μg/day, and the total released quantity is around 70% of the protein loaded in the source. Indeed, BSA proteins have comparable size of membrane pores, of about 7.0 nm. This condition justifies a zero order release rate by the single file passage of molecules through the nanochannel. This size effect has to be coupled to considerations regarding surface interactions, which of course cannot be neglected in the case of high surface to volume ratios accomplished by nanoporous geometries [19]. Indeed, since BSA isoelectric point is 4.7, albumin shows a net negative charge of about -17e at pH 7. This would suggest that also electrostatic repulsion between a negative charged SBA-15 inner surface [20] and BSA, avoiding adsorption [21], could play a determinant role in the zero order release shown. Zero order release is highly desired in pharmaceuticals applications, since guarantees constant release rate independently from reservoirs concentrations. Our formulation attained a constant rate of about 2 μg/day, which is on the same order of magnitude of what obtained in nanofluidic membranes [22]. This rate could be tuned to satisfy the small dosages required for long term therapies such as growth factor delivery, useful in tissue engineering applications for different tissues repair aims, such as vascularization and bone regeneration. Growth factors as vascular endothelial VEGF and bone morphogenetic proteins BMPs, control self-renewal, migration, differentiation as well as other cell fate processes of progenitor cells. In vascularization as well as osteo-
regenerative processes, the release rates are indeed preferably around 100 ng - 10 µg/day/cm3 of injured tissue [23]. However, the obtained rate could be further optimized and adjusted on the specific drug dosage required for the considered application by fine-tuning the exposed membrane surface. Otherwise, the amount of silica precursor solution used in the initial sol quantities, could be lowered thus attaining higher release rates by decreasing the length of the siliceous nanochannels inside the alumina pores, i.e. filling less alumina membrane thickness. Indeed, it would be ideally possible to tune the release slope, thus release velocity, by confining less mesostructure inside the alumina pore in order to change the effective thickness of the mesoporous channel system. Since zero order release induced by single molecule passage can be achieved as soon as the diameter of the constraining 2D geometry in the membrane is smaller than twice the size of the released molecule, the manufactured pore size of SBA-15-AAM could be used to perform zero order release of several other proteins, including hemoglobin and human insulin (hydrodynamic diameters 7.0 and 5.4 nm, respectively). Indeed, the dimensional range explored covers most of large proteins of pharmaceutical interest, which the tested BSA belongs to. In addition, it was shown already the achievement of confined mesoporous silica membranes of about 5.0 nm pore size in the presence of ionic surfactants (i.e. cetyltrimethylammonium bromide [8-10]). These membranes could be tested to obtain single passage of smaller proteins, such as lysozyme and chymotrypsinogen (hydrodynamic diameters 3.8 and 4.8 nm, respectively). Moreover, it would be interesting to explore the possibility to expand the pores of SBA-15-AAM by modifying the starting sol, for applying the manufactured membrane
to even larger proteins such as apoferritin and thyroglobulin (hydrodynamic diameters 16.4 and 20.2 nm, respectively). Different works already demonstrated zero order release for BSA diffusing in nanochannel of nanometric height, though this result was never achieved so far for mesoporous membranes produced by EISA. The major advantage in this configuration is the relatively ease of manufacturing compared to costly and time-consuming production schemes developed for the microprocessor industry.
4. Conclusions Samples of anodic alumina membranes confining SBA-15-like nanochannels in their pores were successfully synthesized and characterized by means of small-angle X-ray scattering, transmission electron microscopy and N2 adsorption / desorption at 77 K. SAXS and TEM measurements confirmed the hexagonal columnar orientation of the mesoporous siliceous channels hosted by the 200 nm-wide alumina pores. The mean pore size of SBA-15-AAMs, as confirmed by microporosimetric analysis, was set around 7.0 nm, which is comparable with the hydrodynamic diameter of the Bovine Serum Albumin protein selected for release testing. Long-term zero order release of BSA through SBA-15-AAM samples was demonstrated up to several weeks, obtaining a release rate of about 2 μg/day. The possibility of tuning the release rate through different amounts of AAM-filling mesostructure could be exploited. Further investigations are needed in order to test protein biological activity intactness after several weeks of protein-mesostructure interaction. These findings claim a deeper insight into transport phenomena through confined silica mesopores in order to better design drug delivery systems.
Acknowledgements The authors acknowledge the help provided by Anton Paar GmbH (Austria).
References [1] B. Platschek, A. Keilbach, T. Bein, Adv. Mater. 23 (2011) 2395-2412. [2] A. Yamaguchi, F. Uejo, T. Yoda, T. Uchida, Y. Tanamura, T. Yamashita, N. Teramae, Nature Mat. 3 (2004) 337-340. [3] C. J. Brinker, Y. Lu, A. Sellinger, H. Fan, Adv. Mater. 11 (1999) 579-585. [4] Y. Wu, G. Cheng, K. Katsov, S. W. Sides, J. Wang, J. Tang, G. H. Fredrickson, M. Moskovits, G. D. Stucky, Nat. Mater. 3 (2004), 816-822. [5] B. Platschek, N. Petkov, T. Bein, Angew. Chem. Int. Ed. 45 (2006) 1134-1138. [6] B. Platschek, R. Köhn, M. Döblinger, T. Bein, Langmuir 24 (2008) 5018-5023. [7] B. Platschek, R. Köhn, M. Döblinger, T. Bein, ChemPhysChem 9 (2008) 2059-2067. [8] B. Platschek, N. Petkov, D. Himsl, S. Zimdars, Z. Li, R. Köhn, T. Bein, J. Am. Chem. Soc. 130 (2008) 17362-17371. [9] V. Cauda, B. Onida, B. Platschek, L. Mühlstein, T. Bein, J. Mater. Chem. 18 (2008) 5888-5899. [10] V. Cauda, L. Mühlstein, B. Onida, T. Bein, Microporous Mesoporous Mater. 118 (2009) 435-442. [11] P. Costa, J. M. Sousa Lobo, European J. Pharm Sciences 13 (2001) 123-133. [12] Q. Wei, C. Bechinger, P. Leiderer. Science 287 (2000) (5453) 625-627. [13] S. Y. Yang, J.-A. Yang, E.-S. Kim, G. Jeon, E. J. Oh, K. Y. Choi, S. K. Hahn, J. K. Kim, ACS Nano 4 (2010) 3817-3822. [14] F. Martin, R. Walczak, A. Boiarski, M. Cohen, T. West, C. Cosentino, M. Ferrari, J. Controlled Release 102 (2005) 123-133.
[15] D. Zhao, Q. Huo, J. Feng, B. F. Chmelka, G. D. Stucky, J. Am. Chem. Soc. 120 (1998) 6024-6036. [16] S. Lowell, J.E. Shields, M.A. Thomas, M. Thommes, Characterization of Porous Solids and Powders: Surface Area, Pore Size and Density, first ed., Kluwer Academic Publishers, Dordrecht, 2004. [17] P. G. Squire, P. Moser, C. T. O'Konski, Biochemistry, 7 (1968) 4261-4272. [18] D. S. Cannell, F. Rondelez, Macromolecules 13 (1980) 1599. [19] I. De Santo, F. Causa, P. A. Netti, Anal. Chem. 82 (2010) 997-1005. [20] E. Papirer, Adsorption on Silica Surfaces, Marcel Dekker, New York, 2000. [21] S.W. Song, K. Hidajat, S. Kaw, Langmuir 21 (2005) 9568-9575. [22] D. Fine, A. Grattoni, S. Hosali, A. Ziemys, E. De Rosa, J. Gil, R. Medema, L. Hudson, M. Kojic, M. Milosevic, L. Brousseau III, R. Goodall, M. Ferrari, X. Liu, Lab Chip 10 (2010) 3074-3083. [23] E. A. Silva, D. J. Mooney, Biomaterials 31 (2010) 1235-1241.
Figure captions Figure 1. SAXS pattern of SBA-15-AAM evidencing in-plane (ip) and out-of-plane (oop) reflections. Figure 2. TEM image of SBA-15-AAM. Figure 3. N2 adsorption (solid symbols) / desorption (open symbols) isotherm at 77 K on partially ground SBA-15-AAM samples. Figure 4. DFT differential pore size distribution of partially ground SBA-15-AAM samples. Figure 5. Release profiles of BSA through SBA-15-AAM (full stars) with a pore size of 7.0 nm, and through the supporting AAM with a nominal pore size of 200 nm (open circles).
Graphical abstract
Highlights > Usage of nanoporous silica-alumina composites as drug delivery membranes. > Investigation about protein transport capabilities of mesophase-confining membranes. > Long-term zero-order release profile of bovine serum albumin up to several weeks.
