NANO LETTERS
Mesoporous Bragg Stack Color Tunable Sensors
2006 Vol. 6, No. 11 2456-2461
Sung Yeun Choi,†,| Marc Mamak,†,⊥ Georg von Freymann,‡ Naveen Chopra,§ and Geoffrey A. Ozin*,† Materials Chemistry Research Group, Department of Chemistry, UniVersity of Toronto, 80 St. George Street, Toronto, Ontario, Canada M5S 3H6, Institut fu¨r Nanotechnologie, Forschungszentrum Karlsruhe in der Helmholtz-Gemeinschaft, D-76021 Karlsruhe, Germany, and Xerox Research Centre of Canada, 2660 Speakman DriVe, Mississauga, Ontario, L5K 2L1, Canada Received July 10, 2006; Revised Manuscript Received September 14, 2006
ABSTRACT Herein we report a novel self-assembly synthesis, structural and optical characterization of mesoporous Bragg stacks (MBS) composed of spin-coated multilayer stacks of mesoporous TiO2 and mesoporous SiO2. Investigation of the optical response of MBS to the infiltration of alcohols and alkanes into its pores reveals better sensitivity and selectivity than conventional Bragg reflectors. Furthermore, we demonstrate that the chemical sensing ability can be tuned via layer thickness, composition and surface properties.
Structural color in biology (e.g., butterfly wings, peacock feathers, and sea mouse whiskers) and geology (opal gem stones) originates from light that has undergone Bragg diffraction from organic, inorganic, or composite materials, forming a periodic dielectric lattice at the scale of light.1,2 Such constructs are known as photonic crystals. Their synthetic counterparts are becoming quite abundant in the materials chemistry literature, most examples coming from the field of opal or inverse opal 3-D photonic crystals made of self-assembled microsphere building blocks or replicas thereof.3 Through physical4-9 and chemical10-13 induced alterations of materials properties of the photonic lattice, this kind of color from structure is widely and continuously wavelength tunable. While 1-D photonic crystals are inherently structurally simpler than 2-D or 3-D analogues, much less has been reported on color adjustable 1-D multilayer stacks of materials, known as Bragg stacks. In this context, there are various ways one can envision building tunable structural color into Bragg stacks. First, one can employ stacks of dense layers whose thickness-refractive index variations can be reversibly altered, for example, through metal-nonmetal transitions, nonlinear optical or swelling effects.14-16 Second, a porous layer can be judiciously integrated at a chosen * Corresponding author. E-mail:
[email protected] † University of Toronto. ‡ Forschungszentrum Karlsruhe. § Xerox Research Centre of Canada. | Current address: Material Science Division, Argonne National Laboratory, 9700 S. Cass Ave., Argonne, IL 60439. ⊥ Current address: Plastic Additives Research & Technology, Ciba Specialty Chemicals, 540 White Plains Rd., Tarrytown, NY 10591. 10.1021/nl061580m CCC: $33.50 Published on Web 10/04/2006
© 2006 American Chemical Society
location into a stack of dense layers to take advantage of reversible adsorption-desorption induced changes in the refractive index of the Bragg stack.17,18 Third, and the focus of the work reported herein, one can make the entire Bragg stack out of porous layers. This has been realized only in the case of porous silicon whose periodic modulation of refractive index contrast has been designed into the Bragg stack through current or voltage modulated anodic oxidation of the porosity of a heavily p-doped single-crystal silicon wafer.19 Variations in refractive index between constituent layers is achieved through intelligent design of porosity and tunable color controlled by introducing guests into the porous silicon layers. Although there are numerous prospective applications of the color tunable Bragg stack, chemical/biochemical sensing applications have been the most actively studied, since techniques for the simple, rapid, and continuous in situ monitoring of various chemical/biological species are of great interest in medical, pharmaceutical, environmental, and food technologies.17,18 Structural color of a Bragg stack stems from the periodic variation in the effective refractive index, which causes high reflectivity for a certain wavelength of light (λB) determined by the Bragg condition λB ) 2neff Λ
(1)
with Λ being the period of a pair of layers and neff its effective refractive index. It is clear from eq 1 that a change in neff will result in a change of λB for a porous Bragg stack with a given Λ. Beyond this fundamental factor, a Bragg
stack composed of nanoscale pores offers additional unique advantages. For example, nanoscale pores can work as a molecular sieve, which excludes molecules larger than its diameter from the sensing event. It can also provide high sensitivity to the low concentration of volatile molecules due to capillary condensation within nanoscale pores. Earlier mentioned porous silicon represents the best example of a nanoporous optical sensor, which successfully integrated nanoscale phenomena into a molecular sensing platform. This system has been shown to be incredibly versatile for optically sensing a range of chemical and biochemical analytes.19 Along these lines a rather attractive way to expand the immense inherent capability of the porous Bragg stack for sensing applications would be to make “all of the layers” out of surfactant and block copolymer templated mesoporous materials because of their broadly tunable composition, pore size, hexagonal, cubic and lamellar mesostructures, and functionalizable surfaces.20,21 If such kinds of mesoporous Bragg stacks (MBS) with layers filled with designed nanopores could be reduced to practice, this would introduce structural color plus molecular selectivity into the world of templated mesoporous materials, which has only previously been possible by building chromophores and fluorophores such as dyes, quantum dots, transition metal ions and complexes into the voids or walls of an inherently colorless open framework.22-26 In this paper we describe the synthesis, structural and optical characterization of MBS, whose reversible variations of hue originate from adsorption-desorption of organic analytes chosen from a series of alcohol and alkane homologues. The superior molecule sensitivity of the MBS compared to a conventional Bragg stack control device bodes well for the future development of a new generation of chemical sensors. As a proof-of-concept, we prepare MBS by alternately spin casting meso-TiO2 and meso-SiO2 into a multilayer stack. This system has been chosen because of the high refractive index contrast and well-established synthesis methods for making structurally well-defined meso-TiO2 and meso-SiO2 materials. The preparation of each layer follows methods described in previous reports.27,28 The novelty of our synthetic approach to MBS lies in achieving an ABAB multilayer mesoporous film through a simple yet effective method of alternating layer deposition by spin casting and subsequent calcination. The key to achieving a robust crack-free, high optical quality multilayer film in which each layer adheres well to the previous layer lies in the choice of calcination conditions. This must be done without causing differential thermally induced stress, originating from removal of organic templates, condensation of the inorganic network, and crystallization of pore walls, which could result in delamination and cracking in and between layers and the substrate. The detailed procedure for achieving this goal is provided in the Supporting Information. Briefly, we fabricate MBS by multiple casting of meso-TiO2 and meso-SiO2 precursor solutions by spin-coating. Calcination at 300 °C for 1 h between each layer deposition is critical to obtain a crackfree optically transparent MBS, since residual organic Nano Lett., Vol. 6, No. 11, 2006
Figure 1. High-resolution SEM images of a mesoporous Bragg stack composed of six alternating composition layers of meso-TiO2 and meso-SiO2 (a and b), STEM images displaying the pore architecture of meso-TiO2 (c) and meso-SiO2 (d) and its smallangle XRD pattern compared to those of single layer films of mesoTiO2 and meso-SiO2.
moieties on the film surface often hinders good adhesion with the next deposited layer, resulting in cracking of the film deposited thereupon and degrading the optical quality of the MBS. A film aging step for obtaining an ordered mesostructure can be included between deposition and calcination if so desired (see Supporting Information). Although it is widely acknowledged that elaborate procedures are often required to obtain crack-free, high optical quality stacked films using the sol-gel method,29 in this study we have been able to obtain such high quality MBS having up to six layers using the aforementioned procedure. A representative ABAB six-layer MBS displayed in Figure 1 is composed of three layers of meso-TiO2 alternated with three layers of meso-SiO2 on a glass substrate. The thickness of each layer is very uniform over a large cross sectional area (Figure 1a, and b). The adhesion between layers is good enough to prevent delamination between layers while it is scraped off the substrate with a razor blade for sampling. 2457
Figure 2. (a) Transmittance spectra obtained from ABAB stacked MBS(x), where x is the number of layers in MBS. (b) Transmittance spectra obtained from the four-layered MBS before and after postthermal treatment at 450 °C.
The brighter meso-TiO2 layers are clearly distinguished from the darker meso-SiO2 layers in scanning electron microscopy (SEM) images due to its higher electron density. Although the deposition conditions are identical for each layer, the first layer deposited on the substrate is usually thicker than subsequent layers deposited on each other (Figure 1b) due to different surface wetting properties of the substrate and meso-TiO2 and meso-SiO2. Adjusting the casting parameters for the first layer allows us to adjust the thickness of the first layer as will be shown later on. The pores of mesoTiO2 and meso-SiO2 layers can be observed in scanning transmission electron microscopy (STEM) images as seen in Figure 1c and d, respectively. MBS templated by P123 for meso-TiO2 and CTAC for meso-SiO2 exhibit very uniform pore diameters estimated around 9 nm for mesoTiO2 and about 3 nm for meso-SiO2, in good correspondence 2458
with the pore sizes reported in previous papers.27,28 The smallangle X-ray diffraction (XRD) pattern obtained from this MBS confirmed the presence of two distinct sizes of mesopores in the film (Figure 1e). The two XRD reflections of the MBS at 0.91° 2θ and 1.81° 2θ are well matched with XRD reflections for a single layer of meso-TiO2 and mesoSiO2. The observed small shift to higher angle and slight broadening can be attributed to the multiple calcinations during the preparation of the MBS. These results (clean interface between adjacent layers in SEM, well-defined pore architectures in STEM, the distinct reflections in small-angle XRD pattern, and response of MBS to the different analytes, see below) suggest that the interdiffusion of layer species upon deposition is negligible. The optical properties of MBS were investigated by UVvis spectroscopy. The transmittance spectra obtained from MBS constructed on a glass substrate are seen in Figure 2. An MBS composed of just one bilayer of meso-TiO2/mesoSiO2 already exhibits a transmittance dip at 580 nm. With increasing number of layers, this dip becomes narrower in peak width and grows in intensity. The slight shifting to shorter wavelength with increasing number of meso-TiO2/ meso-SiO2 layers can be attributed to the multiple calcination steps, which induce certain shrinkage of the mesoporous metal oxide layers each time. It is noteworthy that the transmittance of MBS reached 44% at 544 nm with only six layers in the stack. The post-thermal treatment leads to a blue shift of the transmittance dip without much change of its intensity or shape (Figure 2b). The similar background of the spectra suggests that there is negligible degradation of optical quality of the MBS at the higher temperature thermal treatment. Let us now discuss the suitability and sensitivity of MBS for optical chemical sensing. The color can be reversibly altered by introducing/removing an analyte into/out of its pores. As seen in Figure 3c and d, the MBS soaked in ethanol exhibits longer wavelength color compared to when the MBS was exposed only to air (Figure 3a and b). Furthermore, adjusting the viewing angle allows for tuning the center wavelength to, e.g., adjust the MBS optical response to a suitable detector. In order to investigate the response of MBS to different analytes more precisely, 8 mm × 8 mm size MBS built on a Si wafer are set in a 1 mm-thick quartz cuvette as seen in Figure 3 and their reflectance spectra are collected using a fiberoptic spectrometer coupled to an optical microscope (see Supporting Information). Figure 4a displays representative spectral changes of MBS(x) brought about by different analytes. The number x indicates the number of mesoporous metal oxide layers in the MBS sample. As expected from the Bragg condition (eq 1), this optical change is induced by changing the effective refractive index of MBS due to the infiltrating analyte. The reflectance maxima of each sample are displayed as a function of the refractive index of analyte in Figure 4b. The analytes used in this investigation are air (1, n ) 1.000), deionized water (2, nD20 ) 1.330), anhydrous ethanol (3, nD20 ) 1.360), anhydrous 2-propanol (4, nD20 ) 1.377), and anhydrous 1-butanol (5, nD20 ) 1.399). A Bragg stack consisting of a single layer Nano Lett., Vol. 6, No. 11, 2006
Figure 3. Four-layer MBS4 in air (a and b) and in ethanol (c and d) observed from different viewing angles.
of meso-TiO2 and dense, non-templated SiO2/TiO2 stacks (denoted MBS1) is prepared as a control sample for this investigation since it would be representative of a conventional chemooptical sensor assembly as mentioned earlier. The reflectance maximum of each MBS composed of two, four, and six layers exhibits a linear correlation with the refractive indices of the analyte. The calculated sensitivities (∆λ/∆n) of each sample are 74 for MBS2, 78 for MBS4, and 113 for MBS6, whereas that of the reference MBS1 is only 30. This result demonstrates, by more than a factor of 2, superior sensitivity of MBS to the conventional Bragg reflector having only a single porous layer as the chemical sensing element. In addition to this enhanced sensitivity, MBS exhibit a tunable selectivity in the detection of analytes. We observe that the response of MBS depends not only on the refractive index of an analyte but also on other physical properties, such as hydrophilicity. A comparison of the response of MBS to the series of alcohols including anhydrous ethanol (e, nD20 ) 1.360), anhydrous 2-propanol (ip, nD20 ) 1.377), anhydrous 1-butanol (b, nD20 ) 1.399), and anhydrous 1-hexanol (h, nD20 ) 1.418) and to the series of alkanes including anhydrous n-hexane (6, nD20 ) 1.375), anhydrous n-octane (8, nD20 ) 1.398) and anhydrous n-decane (10, nD20 ) 1.411), clearly demonstrates this effect. Even in the case that an alcohol and an alkane have similar refractive indices like n-hexane and 2-propanol, we observe that the MBS response is not the same. As seen in Figure 5, a linear correlation between the reflectance maximum and the refractive index of analyte exists only within the same family of analytes. The sensitivity of MBS also varies between families. It is noteworthy that the selectivity and sensitivity are altered by the composition of MBS, suggesting that MBS devices can Nano Lett., Vol. 6, No. 11, 2006
Figure 4. (a) Reflectance spectra of two-layer, four-layer, and sixlayer MBS(x) deposited on a Si wafer in air and ethanol. (b) Reflectance maxima obtained from two-layer MBS(2), four-layer MBS(4), and six-layer MBS(6) in air (1), water (2), ethanol (3), 2-propanol (4), and 1-butanol (5) as a function of refractive index of analyte as compared to the response of a single meso-TiO2 layer deposited on the top of a Bragg reflector composed of dense SiO2/ TiO2 layer stacks synthesized without a structure directing template (MBS1).
be tailored for certain applications by selecting the number and composition of the mesostructured layers. In the case of four-layer MBS consisting of a cross-sectional area of 1/2 meso-TiO2 and 1/2 meso-SiO2 (Figure 5a), the ∆λ in the series of alcohols is larger than that of the series of alkanes, though the sensitivity (∆λ/∆n) of MBS for the two series is similar. On the other hand, MBS consisting of a cross-sectional area of 1/3 meso-TiO2 and 2/3 meso-SiO2 (Figure 5b) exhibits larger ∆λ in the series of alkanes and higher sensitivity in the series of alcohols. This interesting interconversion in the sensitivity and the selectivity of MBS can be attributed to the change of the composition ratio of 2459
Figure 5. ∆λ obtained from two different four-layer MBS in a series of alcohols (black squares, ethanol(e), isopropanol(ip), 1-butanol(b) and 1-hexanol(h)) and a series of n-alkanes (red circles, hexane(6), octane(8), and decane(10)) as a function of refractive index of the analyte. SEM images under the graphs are the crosssection of the MBS providing those responses. The scale bars seen in SEM images indicate 200 nm.
meso-TiO2/meso-SiO2. By increasing the relative portion of meso-SiO2, which is relatively more hydrophobic than mesoTiO2 in MBS, the affinity to the hydrophobic medium should increase accordingly. The response shown in Figure 5b is a larger spectral shift for the alkane series, whereas the response to the series of alcohols decreased. Furthermore, the increased affinity to hydrophobic analytes results in an increased sensitivity for distinguishing among members of the alcohol series, which is twice as large as that of MBS having similar cross-sectional areas of meso-TiO2/meso-SiO2. This implies that minute difference in hydrophilicity for the members of the alcohol series can be discriminated since the hydrophilicity of the analyte acts as an additional factor for recognition in the MBS having a larger portion of mesoSiO2. These results demonstrate that the selectivity and sensitivity of MBS are highly dependent on the internal surface properties of the mesoporous metal oxide layers constituting MBS and their adsorption affinity to the analyte of interest. This distinctive performance can be attributed to its large surface-to-volume ratio and nanosized pore architecture. Together with the aforementioned unique advantages enabled by designed nanopores for the molecular sensing, the enhanced molecular discrimination performance is an additional benefit of the MBS platform for advanced chemooptical sensors. Because of recent exciting advances in the synthesis and structures of templated mesoporous metal oxides, one can envision the development of cheap MBS sensors for liquids and vapors tailored for specific applications by tuning the composition, pore size, surface properties, and layer thickness of the device. In summary, mesoporous Bragg stacks (MBS) have been assembled by alternate multiple coatings of meso-TiO2 and 2460
meso-SiO2. Their optical response is reversibly altered by the analyte in their pores. Mesoporous metal oxide diffractive optical elements of this genre exhibit enhanced sensitivity to a change of refractive index of the analyte when compared to conventional Bragg stacks having a single porous adsorption layer integrated into a stack of dense layers. Investigation of the response of MBS to a series of alcohols and alkanes revealed that their sensitivity and selectivity are highly dependent on the properties of the mesoporous metal oxide layers constituting the MBS. It is worth noting that the majority of mesoporous metal oxides are wide band gap semiconductors/insulators devoid of color. Fashioning them as Bragg stacks endows them for the first time with “structural color” that depends on the composition, pore architecture, surface properties, and layer thickness of the layers as well as the analyte contained therein. This all bodes well for the development of a new generation of chemooptical sensors based on MBS. Further investigations of the interdiffusion of precursor species between adjacent layers during self-assembly of MBS are in progress, together with studies aimed at optimizing the synthesis conditions for MBS sensors for detecting and discriminating specific chemical/ biological species. Acknowledgment. G.A.O. is Government of Canada Research Chair in Materials Chemistry. The authors thank the Xerox Research Centre of Canada, NSERC, and the University of Toronto for sustained financial support. SEM/ STEM imaging was performed at the CFI funded “Centre of Nanostructure Imaging” at the Chemistry Department, University of Toronto. The research of G.v.F. is supported by the Deutsche Forschungsgemeinschaft under project FR 1671/2-3 and FR 1671/4-3 (Emmy-Noether program). Supporting Information Available: Experimental details. This material is available free of charge via the Internet at http://pubs.acs.org. References (1) John, S. Phys. ReV. Lett. 1987, 58, 2486. (2) Joannopoulos, J. D.; Villeneuve, P. R.; Fan, S. Nature 1997, 386, 143. (3) Arsenault, A.; Fournier-Bidoz, S.; Hatton, B.; Miguez, H.; Tetreault, N.; Vekris, E.; Wong, S.; Yang, S. M.; Kitaev, V.; Ozin, G. A. J. Mater. Chem. 2004, 14, 781. (4) Busch, K.; John, S. Phys. ReV. E 1998, 58, 3896. (5) Kim, S.; Gopalan, V. Appl. Phys. Lett. 2001, 78, 3015. (6) McPhail, D.; Straub, M.; Gu, M. Appl. Phys. Lett. 2005, 87, 091117. (7) Golosovsky, A.; Neve-Oz, Y.; Davidov, D. Phys. ReV. 2005, 71, 195105. (8) Schmidt, M.; Eich, M.; Huebner, U.; Boucher, R. Appl. Phys. Lett. 2005, 87, 121110. (9) Arsenault, A. C.; Clark, T. J.; von Freymann, G.; Cademartiri, L.; Sapienza, R.; Bertolotti, J.; Vekris, E.; Wong, S.; Kitaev, V.; Manners, I.; Wang, R. Z.; John, S.; Wiersma, D.; Ozin, G. A. Nature Mater. 2006, 5, 179. (10) Arsenault, A.; Miguez, H.; Kitaev, V.; Manners, I.; Ozin, G. A. AdV. Mater. 2003, 15, 503. (11) Te´treault, N.; Arsenault, A. C.; Mihi, A.; Wong, S.; Kitaev, V.; Manners, I.; Miguez, H.; Ozin, G. A. AdV. Mater. 2005, 17, 1912. (12) Fleischaker, F.; Arsenault, A. C.; Kitaev, V.; Peiris, F. C.; von Freymann, G.; Manners, I.; Zentel, R.; Ozin, G. A. J. Am. Chem. Soc. 2005, 127, 9318. (13) Fleischaker, F.; Arsenault, A. C.; Manners, I.; Ozin, G. A. AdV. Mater. 2005, 17, 2455.
Nano Lett., Vol. 6, No. 11, 2006
(14) Ozaki, R.; Matsuhisa, Y.; Ozaki, M.; Yoshino, K. Appl. Phys. Lett. 2004, 84, 1844. (15) Nemec, H.; Duvillaret, L.; Laret, F.; Kuzel, P.; Xavier, P.; Richard, J.; Rauly, D. J. Appl. Phys. 2004, 96, 4072. (16) Khartsev, S. I.; Grishin, A. M. Appl. Phys. Lett. 2005, 87, 122504. (17) Marazuela, M. D.; Moreno-Bondi, M. C. Ana;. Bioanal. Chem. 2002, 372, 664. (18) Lee, B. Optical Fiber Technol. 2003, 9, 57. (19) Sailor, M. J.; Link, J. R. Chem. Commun. 2005, 1375. (20) Soler-Illia, G. J. A. A.; Sanchez, C.; Lebeau, B.; Patarin, J. Chem. ReV. 2002, 102, 4093. (21) Antonietti, M.; Ozin, G. A. Chem. Eur. J. 2004, 10, 28 (22) Yang, P. D.; Wirnsberger, G.; Huang, H. C.; Cordero, S. R.; McGehee, M. D.; Scott, B.; Deng, T.; Whitesides, G. M.; Chmelka, B. F.; Buratto, S. K.; Stucky, G. D. Science 2000, 287, 465. (23) Frindell, K. L.; Bartl, M. H.; Robinson, M. R.; Bazan, G. C.; Popitsch, A.; Stucky, G. D. J. Solid State Chem. 2003, 172, 81.
Nano Lett., Vol. 6, No. 11, 2006
(24) Vogel, R.; Meredith, P.; Kartini, I.; Harvey, M.; Riches, J. D.; Bishop, A.; Heckenberg, N.; Trau, M.; Rubinsztein-Dunlop, H. Chem. Phys. Chem. 2003, 4, 595. (25) Bartl, M. H.; Puls, S. P.; Tang, J.; Lichtenegger, H. C.; Stucky, G. D. Angew. Chem., Int. Ed. 2004, 43, 3037. (26) Bartl, M. H.; Boettcher, S. W.; Hu, E. L.; Stucky, G. D. J. Am. Chem. Soc. 2004, 126, 10826. (27) Choi, S. Y.; Mamak, M.; Coombs, N.; Chopra, N.; Ozin, G. A. AdV. Funct. Mater. 2004, 14, 335. (28) Hatton, B. D.; Landskron, K.; Whitnall, W.; Perovic, D. D.; Ozin, G. A. AdV. Funct. Mater. 2005, 15, 823. (29) Rabaste, S.; Bellessa, J. Brioude, A.; Bovier, C.; Planet, J. C.; Brenier, R.; Marty, O., Mugnief, J.; Dumas, J. Thin Solid Films 2002, 416, 242.
NL061580M
2461