Journal of Colloid and Interface Science 279 (2004) 471–478 www.elsevier.com/locate/jcis
Three-dimensional colloidal crystals with a well-defined architecture Stéphane Reculusa, Pascal Massé, Serge Ravaine ∗ Centre de Recherche Paul Pascal, CNRS, 115, avenue du Dr. Schweitzer, 33600 Pessac, France Received 20 April 2004; accepted 27 June 2004 Available online 11 September 2004
Abstract Monodisperse silica spheres with diameters of 220–1100 nm were prepared by hydrolysis of tetraethyl orthosilicate (TEOS) in an alcoholic medium in the presence of water and ammonia. By grafting vinyl or amino groups onto silica surfaces using the coupling agents allyltrimethoxysilane and aminopropyltriethoxysilane, respectively, amphiphilic silica spheres were obtained and could be organized to form a stable Langmuir film at the air–water interface. The controlled transfer of this monolayer of particles onto a solid substrate gave us the ability to build three-dimensional regular crystals with a well-defined thickness and organization. These colloidal crystals diffract light in the UV, the visible, and the near-infrared (NIR) spectral regions, depending on the size of the silica spheres and according to Bragg’s law. The depth of the photonic stop band can be tuned by varying the number of deposited layers of particles. By using successive depositions, we could prepare multilayered films with silica spheres of different sizes. The thickness of each slab in the binary crystals can be tuned at the layer level, while the crystalline order of each layer is well preserved. 2004 Elsevier Inc. All rights reserved. Keywords: Colloidal crystals; Silica; Langmuir–Blodgett technique; Controlled architecture; Binary materials
1. Introduction Intensive efforts have been made over the past decade to fabricate three-dimensional (3D) regular dielectric structures, as these crystals have potential uses in the elaboration of optical filters [1], switches [2], sensors [3], and waveguides [4]. Several techniques have been developed for the creation of such materials, including colloidal selfassembly [5–20], 3D holography using multiple laser beams [21,22] and photolithography [23]. Keeping in mind that a minimum of imperfection should be present in the resulting materials, a key parameter for the elaboration of an optical band stop filter is the control of the thickness of the colloidal crystals, as reflectance is a function of the number of stacked layers. In a procedure developed by Jiang et al. [8], the control of the concentration of silica sphere suspensions which were left to evaporate naturally allowed the elaboration of colloidal multilayers with a controlled thickness. More re* Corresponding author. Fax: +33-556845600.
E-mail address:
[email protected] (S. Ravaine). 0021-9797/$ – see front matter 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2004.06.079
cently, Ozin and co-workers controlled the same parameter to adjust the thickness of colloidal crystals elaborated by isothermal-heating-evaporation-induced self-assembly [24]. Park and co-workers have shown that crystals’ thickness can be tuned by varying the tilted angle of a substrate dipped into a colloidal suspension with a fixed concentration [25]. We have previously reported that colloidal crystals with a thickness controlled at the layer level can be synthesized using the well-known Langmuir–Blodgett technique [26]. Here we report the elaboration of crystals with a perfectly controlled thickness, starting with silica spheres with diameters ranging from 220 to 1100 nm, in order to elaborate SiO2 photonic crystals with diffraction wavelengths from 500 to 2300 nm. In fact, this is the first report of the synthesis of high-optical-quality crystalline spheres arrays with a welldefined thickness which operate in the near-infrared region, which is a challenge to fabricate practical optical devices for telecommunications. Reese and Asher [27], Hamilton and co-workers [28], and Ozin and co-workers [24] have previously reported the elaboration of such materials but without the control of their thickness at the layer level. The capabil-
472
S. Reculusa et al. / Journal of Colloid and Interface Science 279 (2004) 471–478
ity of the Langmuir–Blodgett method to produce a colloidal crystal layer by layer allowed us to build binary materials consisted of a defined number of layers of silica spheres with different sizes. These new types of 3D structures may have potential applications as low-threshold lasers [4] or templates for macroporous materials [29].
2. Materials and methods 2.1. Materials Tetraethoxysilane (TEOS, Fluka), ammonia (29% in water, J.T. Baker), allyltrimethoxysilane (ABCR), and aminopropyltriethoxysilane (Aldrich) were purchased in their reagent grades and used without further purification. Deionized water was obtained with a Milli-Q system (Millipore) whereas ethanol (EtOH), methanol (MeOH), and chloroform (CHCl3 ) were purchased from Prolabo. 2.2. Methods 2.2.1. Synthesis of silica particles The methods employed for the synthesis and the functionalization of silica particles were similar to those described in a previous work where the synthesis of particles with diameters of 460 and 680 nm were detailed [26]. The amounts of reagent solutions employed for the synthesis of spheres of other diameters in the micrometer-size range are given in Table 1. In some experiments, an alcoholic solution of TEOS was prepared separately and introduced continuously in the medium at a precise rate thanks to a single-syringe pump (see below). 2.2.2. Functionalization of silica particles The functionalization of the silica beads was carried out by adding a large amount of allyltrimethoxysilane or aminopropyltriethoxysilane directly into the nanoparticles dispersion. The amount of coupling agent was around 10 times greater than the amount necessary to cover the inorganic surface with a monolayer (the theoretical amount for such a coverage being nominally 2 molecules nm−2 ). After it was left to react overnight, the mixture was held at 80 ◦ C for 1 h to promote covalent bonding of the organosilane to the surface of the silica nanoparticles. The choice of allyltrimethoxysilane and aminopropyltriethoxysilane was driven
by the necessity to avoid the aggregation of the silica particles either in solution before their spreading at the air–water interface or just after this step. 2.2.3. Silica suspensions treatment In order to eliminate the remaining reagents, all the suspensions are dialyzed against water several times (for small particle size) or submitted to several cycles of washing and centrifugation. The final concentration of the suspension is determined by measuring the mass of a dried extract and the measured value is always in agreement with the theoretical one (calculated assuming a complete conversion of TEOS into silica). 2.2.4. Silica particles size measurements Granulometry experiments are performed on a Malvern Mastersizer apparatus. 2.2.5. Formation of a 2D array of particles A diluted suspension of functionalized silica particles in an 80%/20% (v/v) mixture of chloroform and ethanol is prepared according to a previously reported procedure [26]. After spreading on a pure water subphase, a stepwise compression of the 2D particulate film is carried out under continuous dried nitrogen flow, at room temperature (20 ± 1 ◦C), until a surface pressure of ca. 6 mN m−1 , that is the pressure chosen for the transfer. 2.2.6. Colloidal crystal synthesis After compression, the Langmuir film is transferred onto hydrophilic glass slides or silicon wafers. The slides are immersed quickly in the subphase (downstroke speed: 10 cm min−1 ) and then slowly pulled up out of the water (upstroke speed: 0.1 cm min−1 ). In these optimized conditions, for all sizes of particles, the deposition on the substrate only occurs during the upstroke with a transfer ratio close to unity, what allows us to transfer a monolayer of particles at each cycle. 2.2.7. Scanning electron microscopy SEM observations were performed with a JEOL JSM840A scanning electron microscope operating at 10 kV. The specimens were carbon-coated prior to examination.
Table 1 Experimental conditions corresponding to the synthesis of silica spheres with various diameters Reaction medium Volume of alcohol (mL)
Volume of ammonia (mL)
Volume of alcohol (mL)
Solution of TEOS Volume of TEOS (mL)
Rate of addition (mL h−1 )
200 (EtOH) 100 (MeOH) 200 (EtOH)
15 20 20
0 20 (MeOH) 25 (EtOH)
5 20 25
20 8
a TEOS was added at once.
a
Final particle size (nm) 220 360 1100
S. Reculusa et al. / Journal of Colloid and Interface Science 279 (2004) 471–478
473
Table 2 Experimental and calculated values of the area per particle at the collapse of the particulate monolayer for silica spheres with different sizes Diameter of the silica spheres (nm)
Predicted area (µm2 )
Area at the collapse (µm2 )
220 360 460 680 1100
0.042 0.112 0.183 0.400 1.048
0.055 0.161 0.194 0.430 1.280
where D is its diameter. The presence of a small number of defects and holes that can be observed by SEM (see Fig. 2) and a small incertitude on the value of the spheres diameters determined by granulometry can explain the slight difference between experimental and theoretical values. Fig. 1. Π–A isotherms of functionalized silica particles of different sizes (from left to right: 220 nm/amine, 360 nm/amine, 460 nm/vinyl, 680 nm/vinyl, and 1100 nm/amine).
2.2.8. UV–visible/near-IR spectroscopy Near-infrared spectra were recorded on a Magna-IR Spectrometer 750 from Nicolet and UV–visible spectra with a Unicam UV 4 spectrophotometer.
3. Results and discussion 3.1. Π –A isotherms Typical surface pressure/area isotherms corresponding to the compression of the silica particles films are shown in Fig. 1. The slope of the curves is extremely steep, indicating a low compressibility of the films. The areas corresponding at the collapse of the films are listed in Table 2. These values are in good agreement with those predicted by assuming that the area occupied by one silica sphere in a close-packed hexagonal arrangement is equal to √ 3 2 D , 2
3.2. Langmuir–Blodgett films The transfer of the particles films onto a solid substrate was done at a surface pressure of ca. 6 mN m−1 . Transferring a single layer of silica spheres onto a solid substrate is a good way to illustrate their organization state at the gas/liquid interface [30–32]. As shown in Fig. 2, observation of the samples at low magnification shows that the particles are close packed in a hexagonal lattice with some local defects and grain boundaries. Nevertheless, the visual appearance of the LB films (see Supplementary Material) testifies to their high crystalline quality and their uniform thickness at the centimeter scale. The samples exhibit a brilliant color due to Bragg diffraction of visible light. A systematic change of the color can be seen by modifying the orientation of the substrate. Tentative transfers at surface pressure values higher than 6 mN m−1 led to poor quality LB films, due to the high rigidity of the Langmuir films. In Figs. 3 and 4 are presented SEM side views of colloidal crystals obtained through successive depositions of 1100-nm silica spheres and through the transfer of 10 layers of silica particles with various sizes, respectively. The close-packed structure extends uniformly over the samples, whatever their thickness is. Even if typical defects such as sphere vacancies and vertical cracks can
Fig. 2. Top view of a single layer of silica particles (left, 220 nm/amine; right, 680 nm/vinyl) transferred at 6 mN/m onto a glass slide.
474
S. Reculusa et al. / Journal of Colloid and Interface Science 279 (2004) 471–478
(a)
(b)
(c)
(d)
Fig. 3. Side views of colloidal crystals resulting from a multistage deposition of silica particles monolayers (1100 nm/amine). Samples consist of (a) 3, (b) 5, (c) 10, and (d) 20 layers, respectively.
be observed every 100 µm on average, the top surface of the crystals is relatively smooth and the number of deposited layers matches perfectly with the predefined value. The perfect control of the films architecture is also demonstrated by the study of the optical properties of the crystals. Fig. 5 shows the dependence of the near-IR transmission spectra of a colloidal crystal made of 360-nm silica spheres on its thickness. Similar results were obtained with other silica spheres tested in this work. The intensity of the Bragg diffraction peak around 785 nm is found to increase with the number of deposited layers, confirming the regular stacking of the particles layers perpendicularly to the surface of the substrate. The Fabry–Pérot fringes can also be observed around the Bragg diffraction peak. These fringes are mainly caused by the uniform thickness of the colloidal crystals and reflect their high quality. Considering the system as a layer of effective refractive index, ne = φn2s + (1 − φ)n2a , where ns and na are the indices of silica and air, respectively, one can show that the values of the wavelengths of
two Fabry–Pérot fringes, λp and λp+m , are related to the thickness θ of the crystals by mλp λp+m = 2ne θ (λp+m − λp ),
(1)
where m and p are integers indexing the fringes. Fig. 6 illustrates the result of the calculations for the five sizes of silica particles. All the curves exhibit well-defined linear behavior, with a slope equal to θ . Assuming perfect fcc colloidal crystals constituted by N layers, the diameter of the silica spheres, D, can then be calculated according to the following relation: 2 θ = 1+ (2) (N − 1) D. 3 The results given in Table 3 are in good agreement with the values determined by granulometry. A quantitative study of the optical properties of the colloidal crystals based on the scalar wave approximation theory [33,34] has been performed. The solution of the Maxwell equation for an electric field in a periodic medium leads to
S. Reculusa et al. / Journal of Colloid and Interface Science 279 (2004) 471–478
(a)
(b)
(c)
(d)
475
Fig. 4. Side views of colloidal crystals consisting of 10 layers of silica particles with different sizes and surface functions: (a) 220 nm amine; (b) 360 nm amine; (c) 460 nm vinyl; (d) 680 nm vinyl.
Fig. 5. NIR transmission spectra of colloidal crystals consisting of (from bottom to top) 2, 3, 4, 5, 10, 15, and 25 layers of 360-nm silica particles. Inset illustrates the position of the Fabry–Pérot fringes after baseline correction in the range 800–2700 nm.
Fig. 6. Thickness calculation for colloidal crystals consisting of 10 layers of silica particles with various diameters (from bottom to top: 220, 360, 460, 680, and 1100 nm).
476
S. Reculusa et al. / Journal of Colloid and Interface Science 279 (2004) 471–478
Table 3 Comparison of the average values of the silica particles diameters (nm) estimated with several techniques Granulometry Fabry–Pérot fringes SWA model
220 230 230
370 364 360
460 470 480
680 688 690
1100 1195 1070
where k, q, κb , and κg are analytic expressions depending on other parameters not described here (see Ref. [33] for more details). These expressions have been used to fit the experimental UV–visible–NIR transmission spectra. Results of the fitting procedure, carried out with D as the only adjustable parameter, are shown in Fig. 7. A good agreement between experimental and simulated spectra is achieved. The coincidence between the fringes positions is good, taking into account that no baseline correction was done for the experimental spectra. The values of D resulting from the simulation are listed in Table 3 and are in agreement with those calculated by other techniques. 3.3. Binary colloidal crystals
Fig. 7. Experimental (continuous line) and predicted (dotted line) visible–NIR transmission spectra of colloidal crystals consisting of 10 layers of silica particles with various diameters (from bottom to top: 220, 360, 460, 680, and 1100 nm). For better clarity, curves were shifted vertically.
the expressions for the transmission rate T , 1
T=
1 + (κb − 1) sin2 kN 23 D
T=
1
1 + (κg + 1) sinh2 qN 23 D
in band regions,
in gap regions,
We have further taken advantage of the Langmuir– Blodgett technique by engineering binary colloidal crystals with a perfectly controlled structure. In fact, by depositing sequentially a defined number of layers of silica spheres with different sizes, we could prepare different 3D organized materials. As examples, crystals with (ABAABAAAAA) and (AAABBBBBBB) stackings are shown in Fig. 8. Colvin and co-workers were the first to report the elaboration of multilayer crystals using successive deposition of crystals of silica colloids of different sizes [35]. Velikov et al. [36] have developed an original approach based on the layerby-layer deposition of small and large silica particles to fabricate structures with stoichiometry LS, LS2, and LS3. Nevertheless, each layer can only be grown at a low rate, ca. 1 to 2 mm per day. More recently, Ozin and co-workers [24] have elaborated binary colloidal crystals with various structures by multiple depositions. Our approach has the great advantage to be a versatile and fast route to produce colloidal crystals with a predefined architecture, whatever its complexity is. For example, the elaboration of the (ABAABAAAAA) colloidal crystal shown in Fig. 8 was performed in only 6 h.
Fig. 8. Side views of colloidal crystals resulting from consecutive deposition using two different silica sphere sizes: (left) three layers of 360-nm silica particles and seven layers of 680-nm silica particles; (right) one layer of 680-nm/one layer of 220-nm/two layers of 680-nm/one layer of 220-nm/five layers of 680-nm silica particles.
S. Reculusa et al. / Journal of Colloid and Interface Science 279 (2004) 471–478
477
two types of silica spheres with different sizes. These new binary colloidal crystals have potential applications as optical detectors or templates for macroporous materials which may serve as membranes [37]. We are currently extending our strategy to engineer colloidal crystals with a more complex structural organization, based on precursors with different natures and morphologies.
Acknowledgments We thank Béatrice Agricole (CRPP, Pessac) and Elisabeth Sellier (CREMEM, Talence) for Langmuir–Blodgett and SEM experiments, respectively.
Supplementary material
Fig. 9. NIR transmission spectra of colloidal crystals consisting of three layers of 360-nm silica particles (dotted line); seven layers of 680-nm silica particles (dashed line); seven layers of 680-nm silica particles on top of three layers of 360-nm silica particles (continuous line).
Fig. 9 shows that the optical transmission spectrum of the (AAABBBBBBB) colloidal crystal has features which correspond to peaks in each of the spectra of the crystal made of only one size of silica colloids. The intensity of the peak at ca. 760 nm can be explained as it results from the combination of the second-order diffraction of the crystal made of 680 nm particles and the first-order diffraction of the crystal made of 360 nm silica spheres. This result is in agreement with those reported previously [35], and confirms the good crystalline quality of each opal slab. Moreover, the thickness of the binary films can be calculated using Eq. (1) from the positions of the Fabry–Pérot fringes shown at the long wavelength side of the two diffraction peaks in Fig. 9. The value of the thickness is 5.07 µm, in good agreement with the expected value given by (see Eq. (2)) 2 2 θ = θ1 + θ2 = 1 + 2 0.360 + 1 + 6 0.680 3 3 = 4.96 µm.
4. Summary Using the Langmuir–Blodgett technique, we have engineered colloidal crystals with a well-controlled thickness made of silica spheres with diameters of 220–1100 nm. These materials exhibit diffraction properties in UV, visible, or NIR wavelength ranges, depending on the size of the silica particles. Theoretical simulations of the experimental spectra based on the scalar wave approximation gave good results. Binary 3D assemblies with a well-defined architecture have also been fabricated by successive deposition of
Optical pictures showing the quality of colloidal crystals and the evolution of the sample colors upon the orientation of the substrate are available free of charge in the online version of this article. Please visit DOI: 10.1016/j.jcis.2004.06.079.
References [1] J.M. Weissman, H.B. Sunkara, A.S. Tse, S.A. Asher, Science 274 (1996) 959. [2] G. Pan, R. Kesavamoorthy, S.A. Asher, Phys. Rev. Lett. 78 (1997) 3860. [3] J.H. Holtz, S.A. Asher, Nature 389 (1997) 829. [4] W. Lee, S.A. Pruzinsky, P.V. Braun, Adv. Mater. 14 (2002) 271. [5] S.G. Romanov, N.P. Johnson, A.V. Fokin, V.Y. Butko, H.M. Yates, M.E. Pemble, C.M. Sotomayor Torres, Appl. Phys. Lett. 70 (1997) 2091. [6] M. Egen, R. Zentel, Chem. Mater. 14 (2002) 2176. [7] H. Miguez, F. Meseguer, C. Lopez, A. Blanco, J. Moya, J. Requena, A. Mifsud, V. Fornes, Adv. Mater. 10 (1998) 480. [8] P. Jiang, J.F. Bertone, K.S. Hwang, V.L. Colvin, Chem. Mater. 11 (1999) 2132. [9] J. Texter, C. R. Chimie 6 (2003) 1425. [10] Y.-H. Ye, F. LeBlanc, A. Hache, V.V. Truong, Appl. Phys. Lett. 78 (2001) 52. [11] L.M. Goldenberg, J. Wagner, J. Stumpe, B.-R. Paulke, E. Gornitz, Langmuir 18 (2002) 3319. [12] G. Lu, X. Chen, J. Yao, W. Li, G. Zhang, D. Zhao, B. Yang, J. Shen, Adv. Mater. 14 (2002) 1799. [13] W. Wang, B. Gu, L. Liang, W. Hamilton, J. Phys. Chem. B 107 (2003) 3400. [14] Z. Zhou, X.S. Zhao, Langmuir 20 (2004) 1524. [15] V. Kitaev, G.A. Ozin, Adv. Mater. 15 (2003) 75. [16] Y.-H. Ye, T. Mayer, I.-C. Khoo, I.B. Divliansky, N. Abrams, T.E. Mallouk, J. Mater. Chem. 12 (2002) 3637. [17] Y. Yin, Z.-Y. Li, Y. Xia, Langmuir 19 (2003) 622. [18] K. Wostyn, Y. Zhao, G. de Schaetzen, L. Hellemans, N. Matsuda, K. Clays, A. Persoons, Langmuir 19 (2003) 4465. [19] H. Cong, W. Cao, Langmuir 19 (2003) 8177. [20] M.A. McLachlan, N.P. Johnson, R.M. De La Rue, D.W. McComb, J. Mater. Chem. 14 (2004) 144. [21] M. Campbell, D.N. Sharp, M.T. Harrison, R.G. Denning, A.J. Turberfield, Nature 404 (2000) 53.
478
S. Reculusa et al. / Journal of Colloid and Interface Science 279 (2004) 471–478
[22] H.-B. Sun, S. Matsuo, H. Misawa, Appl. Phys. Lett. 74 (1999) 786. [23] I. Divliansky, T.S. Mayer, K.S. Holliday, V.H. Crespi, Appl. Phys. Lett. 82 (2003) 1667. [24] S. Wong, V. Kitaev, G.A. Ozin, J. Am. Chem. Soc. 125 (2003) 15589. [25] S.H. Im, M.H. Kim, O.O. Park, Chem. Mater. 15 (2003) 1797. [26] S. Reculusa, S. Ravaine, Chem. Mater. 15 (2003) 598. [27] C.E. Reese, S.A. Asher, J. Colloid Interface Sci. 248 (2002) 41. [28] W. Wang, B. Gu, L. Liang, W.A. Hamilton, J. Phys. Chem. B 107 (2003) 12113. [29] I. Soten, H. Miguez, S.M. Yang, S. Petrov, N. Coombs, N. Tetreault, N. Matsuura, H.E. Ruda, G.A. Ozin, Adv. Funct. Mater. 12 (2002) 71. [30] M. Szekeres, O. Kamalin, R.A. Schoonheydt, K. Wostyn, K. Clays, A. Persoons, I. Dekany, J. Mater. Chem. 12 (2002) 3268.
[31] M. Szekeres, O. Kamalin, P.G. Grobet, R.A. Schoonheydt, K. Wostyn, K. Clays, A. Persoons, I. Dekany, Colloids Surf. A 227 (2003) 77. [32] S. Huang, K. Minami, H. Sakaue, S. Shingubara, T. Takahagi, Langmuir 20 (2004) 2274. [33] W.-K. Shung, Y.C. Tsai, Phys. Rev. B 48 (1993) 11265. [34] D.M. Mittleman, J.F. Bertone, P. Jiang, K.S. Hwang, V.L. Colvin, J. Chem. Phys. 111 (1999) 345. [35] P. Jiang, G.N. Ostojic, R. Narat, D.M. Mittleman, V.L. Colvin, Adv. Mater. 13 (2001) 389. [36] K.P. Velikov, C.G. Christova, R.P.A. Dullens, A. van Blaaderen, Science 296 (2002) 106. [37] R.R. Bhave, in: Inorganic Membranes: Synthesis, Characteristics and Applications, Van Nostrand Reinhold, New York, 1991.