nmat1107-print
4/8/04
12:08 PM
Page 337
ARTICLES
Self-assembly of a silica–surfactant nanocomposite in a porous alumina membrane AKIRA YAMAGUCHI1,2, FUMIAKI UEJO1, TAKASHI YODA1, TATSUYA UCHIDA1*, YOSHIHIKO TANAMURA1†, TOMOHISA YAMASHITA1 AND NORIO TERAMAE1,2‡ 1
Department of Chemistry,Graduate School of Science,Tohoku University,Aoba-ku,Sendai 980-8578,Japan CREST,Japan Science and Technology Agency (JST),Sendai 980-8578,Japan *Present Address: School of Life Science,Tokyo University of Pharmacy and Life Science,Hachioji 190-0392,Japan † Present Address: Research Institute for Electronic Science,Hokkaido University,Kita-ku,Sapporo 001-0021,Japan. ‡ e-mail:
[email protected] 2
Published online: 11 April 2004; doi:10.1038/nmat1107
A mesoporous membrane composed of nanochannels with a uniform diameter has a potential use for precise sizeexclusive separation of molecules. Here, we report a novel method to form a hybrid membrane composed of silica–surfactant nanocomposite and a porous alumina membrane, by which size-selective transport of molecules across the membrane becomes possible. The nanocomposite formed inside each columnar alumina pore was an assembly of surfactant-templated silicananochannels with a channel diameter of 3.4 nm; the channel direction being predominantly oriented along the wall of the columnar alumina pore. Molecules could be transported across the membrane including the silica–surfactant nanocomposite with a capability of nanometre-order size-exclusive separation. Our proposed membrane system has a potential use not only for separation science, but also catalysis and chip technologies.
A
mesoporous membrane composed of nanochannels with a uniform diameter of molecular dimensions has a potential use for precise size-exclusive separation of molecules1–3.For this purpose, several methods to synthesise the nanochannel-based membrane have been proposed:for example,deposition of a metal layer within the pores of a polycarbonate membrane2, and deposition of silica nanotubes within the pores of the alumina membrane by a sol–gel template synthesis3. Use of thin films of surfactant-templated mesoporous silica materials has also been anticipated in separation science, because these films are assemblies of surfactant-templated silica-nanochannels with a uniform diameter ranging from a few to tens of nanometres.In general, mesoporous films have been synthesised by using a spontaneous organization of silica–surfactant nanocomposite at solid–liquid interfaces4–8. The procedure for this film formation is quite simple and rapid; the films of surfactant-templated silica-nanochannels can be grown at the substrate surface by simply immersing the substrate in a precursor solution4, which contains surfactant and tetraethoxysilane (TEOS) as a silica source, or by a spin-coating method5–7. However, the channel direction of the resulting mesoporous films is oriented parallel to the substrate surface, and transportation of molecules across the film is not possible. From the standpoint of molecular separations, the channel direction of the mesoporous films should be oriented perpendicular to the film surface,making transport of molecules across the mesoporous film possible. Macroscopic structures of surfactant-templated mesophases grown at the interfacial region are known to depend on the shape of the interfaces8–10. For example, aerosol-assisted silica mesophases have a spherical structure9, and a spherical or fibre structure is formed at the oil–water interface10. These results suggest the possibility of controlling the macroscopic structures of the silica–surfactant nanocomposite with mesoporosity grown at substrate surfaces. When the silica–surfactant nanocomposite is grown inside capillary tubes (tens or hundreds of nanometres in the internal diameter), the tube wall is expected to assist
nature materials | VOL 3 | MAY 2004 | www.nature.com/naturematerials
337 ©2004 Nature Publishing Group
©2004 Nature Publishing Group
nmat1107-print
4/8/04
12:08 PM
Page 338
ARTICLES a
b
c
d
e
f ~ 5 µm
Front surface
5–20 µm 60 µm Back surface
200 nm
Figure 1 SEM images of the alumina membrane with columnar structures inside the alumina pores. a,Top view of the alumina membrane before treatment with the precursor solution.b,Top view of the alumina membrane after treatment with the precursor solution.c,Cross-sectional view of the alumina membrane after treatment with the precursor solution. To obtain the clear cross-sectional view,the alumina membrane was etched slightly to broaden the alumina pores.d,Columnar structures formed in the alumina pore.The columnar structures were obtained as a white precipitate by complete etching of the alumina membrane and collected by filtration.e,Low-magnification SEM image of d.Scale bars correspond to 1 µm.f,Schematic illustration of the columnar structures formed inside the columnar alumina pores.
the self-organization of the silica–surfactant nanocomposite, and the resulting mesophase might be an assembly of surfactant-templated silica-nanochannels oriented along the capillary tube. In this work, we demonstrate that a porous anodic alumina membrane can serve as a solid material to form the silica–surfactant nanocomposite with a desirable orientation of nanochannels for sizeseparation of molecules. The porous alumina membrane is fabricated by anodizing an alumina substrate in a strong acidic solution11. The porous alumina membrane has a packed array of columnar pores ranging from tens to hundreds of nanometres in diameter, depending on the anodizing condition, and the channel direction of the columnar pore is perpendicular to the membrane surface. When a precursor solution, containing cetyltrimethylammonium bromide (CTAB) surfactant and TEOS as the silica source, is introduced into the alumina pores, the silica–surfactant nanocomposite is assembled at the pore walls to form the surfactant-templated silica-nanochannels, and the direction of the silica-nanochannels is predominantly oriented along the columnar alumina pores. As a result, the alumina membrane including the silica–surfactant nanocomposite has a capability of nanometre-order size-exclusive separation of molecules. Figure 1 shows scanning electron microscopy (SEM) images of the alumina membranes before and after treatment with the precursor solution. Energy dispersive X-ray spectroscopy (EDS) was used to identify elements on the surface of each sample (see Supplementary Information, Fig. S1). The procedures for sample preparation and the SEM observation are described in detail in the Methods section. There are no significant changes in surface structures of the membrane after introduction of the precursor solution (Fig. 1a and b),and the peak intensity of Si in EDS is very weak (Fig. S1a and b). In contrast,
columnar structures are observed inside the alumina pores after slight etching of the alumina membrane (Fig. 1c) and the intensity of the Si peak becomes large in EDS (Fig. S1c). These columnar structures are located about 5–10 µm from the front surface and there are no columnar structures in the vicinity of the back surface. Fig. 1d and e shows the columnar structures collected after complete etching of the alumina membrane. The column diameter is about 200 nm (Fig. 1d) and this value corresponds to the diameter of the columnar alumina pores (Fig. 1a and b).These columns are a few to 20 µm in length (Fig. 1e) and a strong Si peak observed in EDS (Fig. S1d) reveals that Si is the major element of these columns. A schematic illustration of the columnar structures inside the alumina pores is shown in Fig. 1f. The mesoporosity of the columnar structures was evaluated by N2 adsorption–desorption measurements after calcination (Fig. 2). The isotherm obtained for the alumina membrane without the columnar structures (curve B) can be classified as type II (ref. 12). When the columnar structures are present inside the alumina pores, the amount of adsorbed N2 increases (curve A) and the isotherm shows type-IV mesopore sorption behaviour12. The type-IV isotherm is often observed in surfactant-templated mesoporous silica materials13, and the inflection position in p/p0 is similar to that found13,14 in CTAB-templated silicate molecular sieves (MCM-41). Thus, the columnar structure can be regarded as a silica–surfactant nanocomposite and the CTAB surfactant micelles serve as templates for the formation of the mesoporous silica. The diameter of silica mesopores was estimated as 3.4 ± 0.2 nm based on Barret–Joyner–Halenda (BJH) analysis, assuming that isotherm A in Fig. 2 was ascribed only to the mesoporous silica within the alumina pores. nature materials | VOL 3 | MAY 2004 | www.nature.com/naturematerials
338 ©2004 Nature Publishing Group
©2004 Nature Publishing Group
nmat1107-print
4/8/04
12:08 PM
Page 339
ARTICLES 70
Volume adsorbed ( ml g –1)
60 A
50 40
Adsorption
30
Desorption
20 10 B
0 0
0.2
0.4
0.6
0.8
1
Relative pressure (p / p0)
Figure 2 Typical N2 adsorption–desorption isotherms for the alumina membrane. p = pressure of nitrogen gas; p0 = saturated vapour pressure. Curves show data with (A) and without (B) the silica–surfactant nanocomposite.The BET surface areas of the alumina membranes A and B were estimated to be 139 m2 g–1 and 4.0 m2 g–1, respectively.The average pore diameter of the mesoporous silica within the columnar alumina pores was estimated from the adsorption–desorption isotherm (A) using BJH analysis, and the value was 3.4 ± 0.2 nm.
The local structure of the silica–surfactant nanocomposite was further investigated by transmission electron microscopy (TEM) measurements. As shown in Fig. 3a, all of the alumina pores are filled with the silica–surfactant nanocomposite. At the alumina wall, a few layers of ordered structures of silica mesopores are observed, and the structural order becomes slightly distorted in the interior region (Fig. 3b). In the region with an ordered mesopore structure, the
a
pore-to-pore distance is approximately 5 nm. Considering the wall thickness of about 1 nm in surfactant-templated mesoporous silica materials6,13–15, this value means a pore diameter of approximately 3 nm, which agrees with the values estimated from the N2 adsorption–desorption isotherm (Fig. 2, curve A). The TEM side-view of the porous alumina membrane including the silica–surfactant nanocomposite shows that the mesoporous silica-nanochannels run predominantly parallel to the wall of the columnar alumina pore (Fig.3c).In previous works on the synthesis of thin films of mesoporous silica, it was reported that the self-organization of the silica–surfactant nanocomposite commenced at a substrate–precursor solution interface,and that highly ordered silica-nanochannels were observed in the vicinity of the substrate surface4,7. In contrast, structural orders of nanochannels tended to be distorted in the region far from the substrate surface7. Hence, the ordered layers of silica-nanochannels observed in the present study indicate that the self-organization of the silica–surfactant nanocomposite commences at the wall surfaces of the alumina pores, and the alumina walls assist the growth of the nanocomposites in a one-dimensional direction along them. In the evaporation-induced self-assembly of silica–surfactant nanocomposites at substrate surfaces, it has been proposed that preferential evaporation of ethanol induces the formation of surfactant micelles and further organization of the mesophase8. This preferential evaporation process may contribute to the formation of the silica–surfactant nanocomposite inside the columnar alumina pores, because the precursor solution was introduced into the columnar alumina pores under moderate aspiration. Furthermore, it is considered that adsorption of cationic CTAB would also be responsible for the formation of the nanocomposite. As CTAB is adsorbed easily at the hydrophilic alumina wall, the concentration of CTAB in the vicinity of the alumina wall would increase by adsorption, and CTAB micelles can be formed at the alumina wall,even if the CTAB concentration in the precursor solution is below the critical micelle concentration.As a result, the silica–surfactant nanocomposite would grow at the alumina wall. When the precursor solution is passing through the columnar alumina pore, the concentrations of CTAB and TEOS will drop with increasing penetration depth due to adsorption of CTAB and formation of the nanocomposite. Thus, the decrease in the concentrations of CTAB and
b
c
Figure 3 TEM images of the alumina membrane with the silica–surfactant nanocomposites inside the columnar alumina pores.a,Low-magnification of top view.b,Highmagnification of top view.c,Side view.Scale bars correspond to100 nm (a) and 50 nm (b,c). nature materials | VOL 3 | MAY 2004 | www.nature.com/naturematerials
339 ©2004 Nature Publishing Group
©2004 Nature Publishing Group
nmat1107-print
4/8/04
12:08 PM
Page 340
ARTICLES a
b
Moles transported (nmol)
Moles transported (nmol)
100
200
100
0 0
2
4 Time ( h )
50
6
0 0
2
4
6 8 Time ( h )
10
12
Rhodamine B
a way not only to obtain the desired mesoporous materials for separation of molecules,but also to fabricate array chips by patterning of the silica–surfactant nanocomposite in the alumina membrane. For example, the simple fabrication procedure, just spotting the precursor solution onto the alumina membrane, allowed formation of the silica–surfactant nanocomposite at a desired region in the alumina membrane (see Supplementary Information, Fig. S2). This simple patterning method enables us to develop chip devices, particularly sensor-array systems, based on mesoporous materials. The present membrane system composed of the alumina membrane and surfactant-templated silica-nanochannels can be easily expanded to the use of another porous membrane with columnar pores, such as porous polycarbonate and porous silicon membranes. The method proposed here will contribute to the development of surfactant-templated mesoporous materials in separation, catalytic and chip technologies.
Vitamin B12 Myoglobin
METHODS
Bovine serum albumin PREPARATIONS
Figure 4 Time-dependent transport of molecules through the alumina membrane. a,b,The amount of moles transported through the alumina membrane without (a) and with (b) the silica–surfactant nanocomposite.
TEOS in the vicinity of the back surface is considered to be responsible for the formation of the nanocomposite only near the front surface as illustrated in Fig. 1f. The molecular transport across the membrane was examined using a series of molecules with different molecular sizes (Fig. 4). These molecules are usually used to evaluate the molecular-weight cutoff of ultrafiltration membranes and to estimate the pore size of such membranes16. In the absence of the silica–surfactant nanocomposite in the alumina membrane, all molecules are transported through the membrane (Fig. 4a). In contrast, the alumina membrane including the silica–surfactant nanocomposite shows selective transport of molecules depending on the molecular size; relatively small molecules, such as rhodamine B (molecular size,∼1.0 nm) and vitamin B12 (∼2.4 nm),are transported, but not large molecules, such as myoglobin (∼4.0 nm) and bovine serum albumin (∼7.2 nm) (Fig. 4b). The transport through the membrane is completely rejected for molecules whose size is larger than the channel diameter (3.4 ± 0.2 nm). In the transport experiments, myoglobin and bovine serum albumin were not detected at all in the receiver phase after 20 h and 24 h, respectively. Although no significant time lags are recognized in Fig. 4a,the time lags for the break-through of the smaller molecules are observed in Fig. 4b and they are approximately 4 h and 7 h for rhodamine B and vitamin B12, respectively. These time lags might be ascribed to the difference in diffusivity within the surfactant-templated silica-nanochannels,and/or distribution rate of these molecules to the nanocomposite from the feed solution. These experimental results suggest that the large molecules cannot permeate into the surfactant-templated silica-nanochannels inside the columnar alumina pores, and the assembly of surfactanttemplated silica-nanochannels inside the columnar alumina pores can work to separate molecules with a capability of a nanometre-order sizeexclusive characteristic due to the channel diameter. The silica–surfactant nanocomposite composed of an assembly of the surfactant-templated silica-nanochannels described here was easily formed inside the columnar pores of the alumina membrane,and size-selective separation of molecules was demonstrated using the membrane including the nanocomposite.The present method provides
To form the silica–surfactant nanocomposite, a precursor solution was prepared according to the method reported earlier4,7, with some modifications. A mixture of ethanol (7.68 g), TEOS (11.57 g) and 1 ml of HCl aqueous solution (2.8 mM) was refluxed at 60 °C for 90 min. Then, ethanol (15 g), 4 ml of HCl solution (55 mM) and CTAB (1.52 g) were added to the refluxed solution, which was stirred for 30 min to give the precursor solution. Porous alumina membranes (pore diameter = ∼200 nm, thickness = ∼60 µm, diameter of membrane = 25 or 47 mm) were obtained from Whattman (Anodisk). The alumina membrane was set in an ordinary membrane filtration apparatus, and the precursor solution was dropped onto the alumina membrane. Moderate aspiration was applied so that the precursor solution penetrated into the columnar alumina pores. The alumina membrane including the precursor solution was then dried in air at room temperature. As shown in Fig. 1f, the upper surface of the membrane in the filtration apparatus is designated as the ‘front surface’. The experiments on SEM and TEM observation, N2 adsorption–desorption isotherms, and molecular transport across the membrane, were carried out using a hybrid membrane system where all the columnar pores of the alumina membrane were filled with the silica–surfactant nanocomposite.
MEASUREMENTS SEM images were measured on a JSN-6320F (JEOL). Sample membranes were fixed on a SEM stage using carbon tape, and SEM images were measured after deposition of a thin Pt layer (thickness ∼2 nm) using a JFC 1300 Auto Fine Coater (JEOL). To obtain a clear cross-sectional view (Fig. 1c) of the alumina membrane with the silica–surfactant nanocomposite, the alumina membrane was etched slightly by immersing it into a 10-wt% phosphoric acid solution for 10 min and then was cleaved. Complete etching of the alumina membrane was also carried out using a 10-wt% phosphoric acid solution to obtain the silica–surfactant nanocomposite formed inside the pore of the alumina membrane. The nanocomposite was obtained as a white precipitate and it was collected by filtration and used for observation of SEM images (Fig. 1d and e). N2 adsorption and desorption isotherms were measured on a Micrometrics ASAP2010 instrument. The alumina membrane including the silica–surfactant nanocomposite inside the columnar alumina pore was cut into small pieces of about 3-mm square for placement in the measurement cell. These membrane pieces were calcined at 400 °C for 3 h before the measurement. TEM images were measured on a JEM-200EX (JEOL) equipped with a field-emission gun and operated at 200 kV. The TEM top-views were obtained after mechanical polishing, dimpling and argonion milling. To obtain the TEM side-view, the membrane was embedded in an epoxy resin and cut orthogonally to the membrane surface by mechanical polishing, dimpling and argon-ion milling. For the experiment on the molecular transport across the membrane, the alumina membrane (diameter = 25 mm) including the silica–surfactant nanocomposite was used without calcinations, and was mounted in a U-tube permeation cell. The feed solution contained 0.2 mM of each sample molecule in a 100 mM 2-(N-morpholino)ethanesulphonic acid (MES) buffer (pH 7.0), and the permeate solution contained the same buffer. The amount of moles transported was analysed by measuring ultraviolet–visible absorption spectra of each molecule in the permeate solutions. The absorption measurements were carried out using a V-570 Spectrophotometer (JASCO).
Received 5 January 2004; accepted 23 February 2004; published 11 April 2004. References 1. Rouhi, A. M. From membrane to nanotubes. Sci. Technol. 79, 29–33(2001). 2. Jirage, K. B., Hulteen, J. C. & Martin, C. R. Nanotube-based molecular-filtration membranes. Science 278, 655–658 (1997). 3. Lee, S. B. et al. Antibody-based bio-nanotube membranes for enantiomeric drug separations. Science 296, 2198–2200 (2002). 4. Yang, H., Kuperman, A., Coombs, N., M.-Afara, S. & Ozin, G. A. Synthesis of oriented films of mesoporous silica on mica. Nature 379, 703–705 (1996). 5. Ogawa, M. Formation of novel oriented transparent films of layered silica–surfactant nanocomposites. J. Am. Chem. Soc. 116, 7941–7942 (1994). 6. Ogawa, M. A simple sol-gel route for the preparation of silica–surfactant mesostructured materials. Chem. Commun. 1149–1150 (1996). 7. Lu, Y. et al. Continuous formation of supported cubic and hexagonal mesoporous films by sol-gel dipcoating. Nature 389, 364–368 (1997).
nature materials | VOL 3 | MAY 2004 | www.nature.com/naturematerials
340 ©2004 Nature Publishing Group
©2004 Nature Publishing Group
nmat1107-print
4/8/04
12:08 PM
Page 341
ARTICLES 8. Brinker, C. J., Lu, Y., Sellinger, A. & Fan, H. Evaporation induced self-assembly: nanostructures made easy. Adv. Mater. 11, 579–585 (1999). 9. Lu, Y. et al. Aerosol-assisted self-assembly of mesostructured spherical nanoparticles. Nature 398, 223–226 (1999). 10. Schacht, S., Huo, Q., V.-Martin, I. G., Stucky, G. D. & Schüth, F. Oil-water interface templating of mesoporous macroscale structures. Science 273, 768–711 (1996). 11. Diggle, J. W., Downie, T. C. & Goulding, C. W. Anodic oxide films on aluminum. Chem. Rev. 69, 365–405 (1969). 12. Sing, K. S., Everett, D. H., W. Haul, R. A., Moscou, L. & Pierotti, R. A. Reporting physisorption data for gas/solid systems with special reference to the determination of surface area and porosity. Pure Appl. Chem. 57, 603–619 (1985). 13. Kresge, C. T., Leonowicz, M. E., Roth, W. J., Vartuli, J. C. & Beck, J. S. Ordered mesoporous molecular sieves synthesized by a liquid-crystal template mechanism. Nature 359, 710–712 (1992). 14. Beck, J. S. et al. A new family of mesoporous molecular sieves prepared with liquid crystal template. J. Am. Chem. Soc. 114, 10834–10843 (1992). 15. Tanamura, Y. et al. Ship-in-a-bottle synthesis of copper phthalocyanine molecules within mesoporous channels of MCM-41 by a chemical vapor deposition method. Nano Lett. 1, 387–390 (2001).
16. Sarbolouki, M. N. A general diagram for estimating pore size of ultrafiltration and reverse osmosis membrane. Sep. Sci. Technol. 17, 381–386 (1982).
Acknowledgements We thank Eiji Aoyanagi and Yuichiro Hayasaka, High-Voltage Electron Microscope Laboratory, Tohoku University, and Shun Ito, Analytical Research Core for Advanced Materials, Institute for Material Research, Tohoku University, for SEM and TEM measurements. We also thank Hiroaki Misawa, Hokkaido University, for useful discussions about alumina membranes. This work was supported in part by a Grant in Aid for Scientific Research (No. 14204074, No. 13129201, No. 15750062) from the Ministry of Education, Culture, Sports, Science and Technology, Japan, JSPS-RFTF (Research for the Future Program from the Japan Society for the Promotion of Science), and the Asahi Glass Foundation. Correspondence and requests for materials should be addressed to N. T. Supplementary Information accompanies the paper on www.nature.com/naturematerials
Competing financial interests The authors declare that they have no competing financial interests.
nature materials | VOL 3 | MAY 2004 | www.nature.com/naturematerials
341 ©2004 Nature Publishing Group
©2004 Nature Publishing Group
