- Email: [email protected]
Formation of mesoporous Co3O4 replicas of different mesostructures with different pore sizes
Microporous and Mesoporous Materials 123 (2009) 314–323
Contents lists available at ScienceDirect
Microporous and Mesoporous Materials journal homepage: www.elsevier.com/locate/micromeso
Formation of mesoporous Co3O4 replicas of different mesostructures with different pore sizes Peng Shu a,1, Juanfang Ruan b, Chuanbo Gao a,b, Huachun Li a, Shunai Che a,* a
School of Chemistry and Chemical Engineering, State Key Laboratory of Metal Matrix Composite Materials, Shanghai Jiao Tong University, 800 Dongchuan Road, Shanghai 200240, China b Structural Chemistry, Arrhenius Laboratory, Stockholm University, S-10691 Stockholm, Sweden
a r t i c l e
i n f o
Article history: Received 17 February 2009 Received in revised form 9 April 2009 Accepted 14 April 2009 Available online 21 April 2009 Keywords: Mesoporous cobalt oxide Hard templating Replica Mesostructure Pore size
a b s t r a c t Mesoporous metal oxides Co3O4 are prepared via hard templating synthesis method by using various mesoporous silicas with different pore size as templates. The pore size of the mesoporous silicas with and Pn3m have been the symmetry of two-dimensional (2d)-hexagonal p6mm, bicontinuous cubic Ia3d controlled in the range of 6.6–10.7, 4.2–7.5 and 5.1–6.7 nm, respectively, by choosing different surfactants and co-surfactants and by adjusting either the aging temperature or the ionization degree of the surfactant. The pore size of the silica template has been considered to be an important factor that determines the mesostructure of the resulting metal oxides. It has been found that for p6mm, it is easier to at large-pore size two sets of replicate the mesoporous symmetry at large size of mesopores. For Ia3d, leads to replibicontinuous meso-channels are replicated into mesoporous Co3O4, while small-pore Ia3d cation of both one set and two sets of meso-channels. Co3O4 can replicate both one set and two sets of meso-channels at all pore sizes that can be obtained (5.1–6.7 nm), indicating the exisbicontinuous Pn3m tence of ordered complementary micropores within the silica walls. Ó 2009 Elsevier Inc. All rights reserved.
1. Introduction Mesoporous metal oxides are of great interest for applications in catalysis since they can be regarded as self-supported catalysts with large specific surface areas and shape selective properties, which find utilities in areas of catalysis, chemical sensing, superconductivity, colossal magneto-resistance, piezoelectricity, etc. [1–4]. Their nano-sized forms differ greatly in many properties from the bulk materials [5]. These materials can be synthesized by using either organic soft templates [6], similar to the synthesis of mesoporous silica, or mesoporous silica as a hard template [7– 10]. While soft templating normally results in amorphous materials, compared to silica, the surfactant/oxide composite precursors are often more susceptible to suffer from lack of condensation, redox reaction, and phase transitions accompanied by thermal breakdown of the structural integrity [6,11]. The hard templating method was firstly introduced by Ryoo and co-workers for synthesizing mesoporous carbon [7,12–14], and several mesoporous metals [15,16], metal sulfides [17], nitride [18] and oxides [19–21] have been fabricated, although the wall of most of these materials was still disordered and the formation mechanisms of these new * Corresponding author. Tel.: +86 21 5474 2852; fax: +86 21 5474 1297. E-mail address: [email protected] (S. Che). 1 Present address: Department of Material Science and Engineering, Stanford University, USA. 1387-1811/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.micromeso.2009.04.017
materials were not discussed extensively. Recently, researchers reported the synthesis of mesoporous metal oxides by using hard templating method, some of which are in single crystalline state [22–29]. Mesoporous cobalt oxide Co3O4 is one of the metal oxides that have been studied extensively. Zhao et al. and Schüth et al. prepared mesoporous Co3O4 with functional mesoporous silicas as hard templates by using impregnating method [20,23]. Zhou et al. prepared mesoporous Co3O4 by using a solid–liquid route, and products were maintained to be single crystal [25]. The formation mechanism of mesoporous Co3O4 has been studied, and the factors determining the porous structure of replica are considered to be crystallization conditions, pore size of template, interconnectivity between mesopores and precursor loading levels [22,27]. Recently, Schüth et al. visualized the pore topology of mesoporous Co3O4 by high-resolution scanning electron microscopy [28]. Various mesoporous silicas can be used as hard template for synthesizing mesoporous metals or metal oxides. Two typical examples are mesoporous silica SBA-15 with hexagonally closepacked straight mesopores connected by some smaller channels symmetry containing and micropores [30], and KIT-6 with Ia3d three-dimensional (3d) bicontinuous meso-channel networks [31]. Typically, such methods employ mesoporous silicas as the hard templates, into which a solution-based precursor of the desired metal or metal oxide is introduced, followed by heating or hydrogenation to form the desired metal oxide or metal and silica
P. Shu et al. / Microporous and Mesoporous Materials 123 (2009) 314–323
composites [10]. The silica template is dissolved away to leave a replica mesoporous structure of the target compound. The hard templating route has opened a way to great varieties of mesoporous transition metal oxides. The relationship between the pore size of the silica template and that of the obtained metal oxide replicas is an interesting topic. Here, the mesostructures including 2d-hexagonal p6mm, cubic and Pn3m that possess rod-type pore system have been seIa3d lected for studying the relationship, because, in the case of cagetype mesoporous silica, it is well known that the small apertures make it difficult for the precursors to completely fill the cages and to form rigid carbon or metal oxide bridges between their nanoparticles prepared in silica cages. Here, we present a report on the study of the effect of pore size and pore symmetry on the formation of mesoporous Co3O4 by preparing the mesoporous co crystal structure) using mesoporous silicas balt oxide Co3O4 (Fd3m and SBA-15 (2d-hexagonal p6mm), KIT-6 (bicontinuous cubic Ia3d) AMS-10 (bicontinuous Pn3m) with different pore size as hard templates. The dependence of the structure of replica on pore size and symmetry of the silica template will be discussed in details. Different from earlier literatures, in this study, we will show that solely the pore size of silica template cannot determine the replication extent, and the discussion is valid only when considering the mesopore symmetry. Structural properties of the mesoporous silicas, especially newly developed AMS-10, are also derived from inverse replication.
2. Experimental section 2.1. Synthesis 2.1.1. Synthesis of mesoporous silica The mesoporous silica templates SBA-15 [23], KIT-6 [24], and AMS-10 [32] were prepared according to the procedure described previously. 2.1.1.1. Synthesis of 2d-hexagonal p6mm structured mesoporous silica (SBA-15) with different pore size. SBA-15 mesoporous silica was synthesized by using the triblock copolymer Pluronic P123 (EO20PO70EO20, BASF) as surfactant and tetraethoxylsilicon (TEOS) as silica source. In a typical synthesis experiment, TEOS (20.8 g) was added to a mixture of P123 (10 g), HCl (62.57 g, 35 wt.%), and deionized water (319 g) under stirring at 35 °C. After the mixture was stirred for 24 h, the mesostructured products thus formed were aged at 100 °C for additional 24 h. The products were filtered, dried without washing, and calcined at 550 °C. The pore size was controlled by adjusting aging temperature in the range of 40– 100 °C. The samples with three pore sizes have been synthesized at aging temperature of 40, 80 and 100 °C. These samples were denoted as SBA-15-40, SBA-15-80, and SBA-15-100, respectively. structured mesoporous silica 2.1.1.2. Synthesis of bicontinuous Ia3d mesopor(KIT-6) with different pore size. The large-pore cubic Ia3d ous silica KIT-6 was synthesized by using triblock copolymer Pluronic P123 as surfactant and TEOS as silica source. In a typical synthesis, 6 g of P123 was dissolved in 217 g of distilled water and 11.8 g of HCl (35 wt.%). To this, 6 g of butanol was added under stirring at 35 °C. After stirring for 1 h, 12.9 g of TEOS was added at 35 °C (TEOS:P123:HCl:H2O:BuOH = 1:0.017:1.83:195:1.31 mol ratio). The mixture was left under stirring for 24 h at 35 °C, and subsequently aged for 24 h at 100 °C under static conditions in a closed polypropylene bottle. The products were filtered, dried without washing, and calcined at 550 °C. In the similar manner of SBA-15, the pore size was controlled by adjusting aging temperature in the range of 40–100 °C. The samples with three pore sizes
315
have been synthesized at different aging temperature of 40, 80, and 100 °C. These samples were denoted as KIT-6-40, KIT-6-80, and KIT-6-100, respectively. structured mesoporous silica 2.1.1.3. Synthesis of bicontinuous Pn3m (AMS-10) with different pore size. The bicontinuous cubic Pn3m mesoporous silica (AMS-10) was synthesized by using N-myristoyl-L-glutamic acid sodium salt (C14GluAS, one of the carboxylic acid groups of the glutamic acid was neutralized by sodium hydroxide) as surfactant, N-trimethoxylsilylpropyl-N,N,N-trimehylammonium chloride (TMAPS) as co-structure-directing agent (CSDA) and TEOS as silica source. In a typical synthesis experiment, HCl (0.417 g of 1 mol L 1 solution) was added to a solution of C14GluAS (0.379 g, 1 mmol) in deionized water (36 g) under stirring at 80 °C. After the solution became homogeneous, a mixture of TMAPS (0.773 g, 50 wt.% in methanol, 1.5 mmol) and TEOS (3.12 g, 15 mmol) was added, and the mixture was stirred at 80 °C for 10 min. The molar composition of the final gel was C14GluAS/TMAPS/TEOS/H2O/HCl = 1:1.5:15:2000:0.42. The precipitate was then aged at 80 °C for 2 days, filtered, and calcined at 550 °C for 6 h to remove the surfactants. The pore size of AMS-10 was controlled by adjusting the HCl/C14GluAS molar ratio in the range of 0.42–0.48. The samples with four pore sizes were synthesized with different HCl/C14GluAS molar ratios of 0.42, 0.44, 0.45 and 0.48. They were denoted as AMS-10-0.42, AMS-10-0.44, AMS-10-0.45 and AMS-10-0.48, respectively. 2.1.2. Synthesis of mesoporous Co3O4 replica In a typical synthesis of mesoporous Co3O4, 0.75 g of Co(NO3)26H2O (98% Aldrich) was dissolved in 16 mL of ethanol followed by addition of 0.5 g of mesoporous silica templates. The mixture was stirred at room temperature until nearly dry powder had been obtained; the sample was then heated slowly to 300 °C and calcined at the same temperature for 3 h to pyrolyze the nitrate. The impregnation procedure was repeated twice with 0.4 g of Co(NO3)26H2O and 0.25 g Co(NO3)26H2O dissolved in 16 mL ethanol, followed by calcination at 550 °C for 5 h with a ramp 1 °C min 1 for metal oxide to crystallize. The resulting samples were treated with 10% HF to remove the silica template, centrifuged, washed with water and ethanol, and then dried at 60 °C in air. They were denoted as Co3O4-SBA-15-X, Co3O4-KIT-6-X and Co3O4-AMS-10-Y (X: aging temperature of silica template in °C; Y: molar ratio of HCl/C14GluAS in the synthesis of AMS-10). 2.2. Characterizations X-ray diffraction (XRD) patterns were recorded on a Rigaku RINT 2000/PC Multiplex instrument using Cu Ka radiation (k = 0.15406 nm), operated at 40 kV and 20 mA (0.8 kW) at the rate of 1.0° min 1 over the range of 0.8–6.0° (2h). N2 adsorption–desorption isotherms were measured at 196 °C on a Quantachrome Nova 4200e volumetric adsorption analyzer. Before the adsorption measurements, all samples were outgassed at 200 °C in the port of the adsorption analyzer for 4 h. Specific surface areas were calculated via the Brunauer–Emmett–Teller (BET) model in regions applicable to the derivation of the model between P/P0 values of 0.05–0.3. The total pore volume was determined from the uptake of nitrogen at a relative pressure of P/P0 0.99. Pore size distribution curves were obtained from the adsorption branch by using the Barrett–Joyner–Halenda (BJH) method. Scanning electron microscopy (SEM) features of all the samples were observed by SEM (JEOL JSM-7401F). The samples were observed without any metal coating. High-resolution transmission electron microscopy (HRTEM) images were taken from thin edges of particles supported on a porous carbon grid, using JEOL JEM3010 electron microscope (point resolution 1.7 Å, Cs = 0.6 mm)
316
P. Shu et al. / Microporous and Mesoporous Materials 123 (2009) 314–323
operated at 300 kV. Samples were crushed in an agate mortar, suspended in ethanol using ultrasonics and then dropped onto the carbon grid. The wall thickness (w) of mesoporous silicas was calculated combining the unit cell parameter (a) derived from XRD pattern and pore diameter (d) derived from N2 adsorption–desorption. For 2d-hexagonal p6mm symmetry, w was calculated as a–d. For bicontinuous cubic Ia-3d symmetry, w was calculated as a/ 3.0919-d/2 [33,34]. For bicontinuous cubic Pn-3m symmetry, w was determined as d110-d, i.e., w = a/1.4142-d (structure model, Ref. [32]).
3. Results and discussion 3.1. Mesoporous Co3O4 synthesized by SBA-15 with p6mm structure as hard template Triblock copolymer Pluronic P123 was used as a template to synthesize SBA-15. It was reported that the pore size can be modulated by aging temperature [23]. PPO moieties are hydrophobic while PEO moieties are hydrophilic. PEO moieties are expected to interact with the inorganic species more strongly and closely associated with the inorganic wall than the hydrophobic PPO. As temperature rises, the PEO moieties become more hydrophobic, resulting in increased hydrophobic domain volume and shorter length of the PEO segments associated with the inorganic wall, thus the pore size increases. All the SBA-15-X (X: aging temperature in °C) samples show three well-resolved XRD peaks in the region of 2h = 0.5–2.5°, which correspond to (10), (11) and (20) reflections based on the 2d-hexagonal system (Fig. S1). In addition, the higher hydrothermal aging temperature gives rise to 2d-hexagonal p6mm silica with larger pores and improved structural integrity. The resultant products show the increasing d-spacing indicated from the fact that the (10) reflection peak shifts to a smaller angle with increasing hydrothermal aging temperature from 40 to 100 °C. N2 adsorption and desorption isotherms for SBA-15-X (Fig. S2) are of type IV according to the IUPAC classification [35]. Significant H1 type hysteresis loop is observed for all isotherms, and the capillary condensation occurs at increasing relative pressure with increasing aging temperature, indicating increasing size of cylindrical mesopores. In addition, the capillary condensation step becomes narrowed with increasing synthesis aging temperature, which indicates the degree of pore uniformity is increased. Interestingly, the pore size distribution curves, BJH plots of derivative of the pore volume per unit weight with respect to the pore diameter (dV/dD) versus the pore diameter (D), show a clear shift of pore diameter from 6.6 to 10.7 nm as the aging temperature is increased from 40 to 100 °C (Fig. S2). Table 1 summarizes the structural properties of the SBA-15-X mesoporous silicas. All the SBA-15-X samples have BET surface area larger than 700 m2 g 1, and the pore volume increases from 0.70 to 1.18 cm3 g 1 with higher aging temperatures. The wall thickness has also been determined on the basis of unit cell parameter and pore diameter, and it can be seen that the wall thickness
decreases from 2.4 to 0.7 nm when the aging temperature increases from 40 to 100 °C in the synthesis. Fig. 1 shows the XRD patterns of mesoporous Co3O4 after removal of silica, using SBA-15 with different pore sizes as hard templates. After cobalt oxide is embedded within the mesopores of SBA-15, the intensities of the (10), (11) and (20) reflections are dramatically reduced. The obtained mesoporous cobalt oxides synthesized by large-pore samples SBA-15-80 and 100, show wellresolved (10) reflection while lose the (20) reflection peak. It indicates that the mesostructures are well retained but the structural order decreases during replication process. Sample Co3O4-SBA15-40 templated by mesoporous silica with the smallest pore size gives no significant X-ray reflection peak in the range of 2h = 0.5– 4°. It suggests that after the removal of silica template, the hexagonal arrangements of nano-wires collapse, which is confirmed by later TEM characterization. From the large-angle XRD pattern, it can be concluded that the Co3O4 products are well crystallized (not shown). and the crystal structure is cubic Fd3m As is seen in the SEM images (Fig. S3), the morphology of the Co3O4 replica is close to that of the mesoporous silica template (SBA-15-100). Cobalt nitrate was deposited in the straight channels of templates. Nuclei were formed simultaneously in the channels with the temperature increase, followed by the growth of the cobalt oxide. The confinement of the straight channels of the SBA15 allowed the cobalt oxide to grow in one direction and completely fill the pores, keeping the morphology. Fig. 2 shows the N2 adsorption–desorption isotherms of the silica-free Co3O4 mesoporous replicas synthesized by using SBA-1580 and 100. The structural properties are summarized in Table 2. The isotherms are of type IV classification, which have typical hysteresis loops of mesoporous materials. With SBA-15-100 mesoporous silica as template, the silica-free mesoporous Co3O4 sample has a surface area of 86 m2 g 1 and a pore volume of 0.17 cm3 g 1. It is noticeable that its isotherm is different from that of SBA-15-100, giving a pore size distribution centered at 2.3 nm, indicating the expected pore dimension, which is larger but is still in scale with the wall thickness of SBA-15 as shown in Table 1. Similarly, the pore size of mesoporous Co3O4 templated by SBA-15-80 is also close to the wall thickness of silica template. It suggests that the SBA-15-100 and 80 are successfully replicated. Fig. 3a and b are TEM images of the Co3O4 mesoporous replicas synthesized with small-pore SBA-15-40 (6.6 nm) and large-pore SBA-15-100 (10.7 nm), respectively, as hard templates. The length of Co3O4 nano-wires ranged from several tens to several hundreds of nanometers, and the average diameters of Co3O4 nano-wires are about 6.3 and 9.8 nm, respectively, which match the pore sizes of the corresponding SBA-15 templates very well. It can be clearly seen from Fig. 3b that mesoporous Co3O4 replica well retains the ordered 2d-hexagonal p6mm structure of the large-pore silica template SBA-15-100. However, the mesoporous Co3O4 replicas templated by small-pore SBA-15-40 (Fig. 3a) are isolated nano-wires without interconnections to retain the symmetry of silica template. Notably, the corresponding HRTEM of Fig. 3a and b marked by arrows are given on the right. From the HRTEM images it can be seen that the mesoporous Co3O4 replicas are crystalline, and each
Table 1 Structural properties of the mesoporous silica SBA-15-X (X: aging temperature in °C) with 2d-hexagonal p6mm symmetry. Sample
Unit cell parameter (nm)
S (m2 g
SBA-15-40 SBA-15-80 SBA-15-100
9.0 10.0 11.3
794 711 714
1
)
V (cm3 g 0.70 0.96 1.18
1
)
Pore diameter (nm)
Wall thickness (nm)
6.6 8.9 10.7
2.4 1.1 0.7
317
P. Shu et al. / Microporous and Mesoporous Materials 123 (2009) 314–323 Table 2 Structural properties of the mesoporous Co3O4 templated by SBA-15-X.
10
11
Intensity (a.u.)
Co3O4-SBA-15-100
Co3O4-SBA-15-80
Co3O4-SBA-15-40
1
2
3
4
2 theta (degree) Fig. 1. XRD patterns of the mesoporous Co3O4 templated by SBA-15-X.
of them can be regarded as a single crystal although the crystal orientations of the two adjacent Co3O4 nano-wires are different. In summary, SBA-15 with large-pores (larger than 8.9 nm) can be fully replicated by Co3O4, retaining the mesostructure of silica template, while replication of small-pore SBA-15 (6.6 nm) results in separated nano-wires. It is probably due to the microporous defects in the walls of SBA-15. If the pore dimension gets larger, the pore wall becomes thinner, and the pore-forming P123 in adjacent micelles would share hydrated PEO and therefore forms micropores that connect mesopores, which plays a vital role in retaining the template symmetry during the inverse replication [36]. 3.2. Mesoporous Co3O4 synthesized by KIT-6 with bicontinuous Ia3d structure as hard template Similar to the case of SBA-15, the pore size of bicontinuous cu mesoporous silica KIT-6 can also be modulated by changbic Ia3d
Sample
S (m2 g
Co3O4-SBA-15-80 Co3O4-SBA-15-100
64 86
1
)
V (cm3 g 0.16 0.17
1
)
Pore diameter (nm) 2.2 2.3
ing the aging temperature. All the KIT-6-X (X: aging temperature in °C) materials show three well-resolved XRD peaks in the region of 2h = 0.8–2.8° (Fig. S4), which correspond to (2 1 1), (2 2 0) and (3 2 1) reflections, based on a cubic system. The bicontinuous cubic symmetry is confirmed by combining the XRD pattern and Ia3d HRTEM images (not shown). For KIT-6-X, higher hydrothermal aging temperature gives mesoporous silicas with larger pores and improved structural integrity. The resultant products show increasing d-spacing with the (2 1 1) reflection peak shifting to a smaller angle, when the hydrothermal treatment temperature gets higher. The N2 adsorption–desorption isotherms of the materials (Fig. S5) are type IV with a sharp capillary condensation step at high relative pressures and a H1 type hysteresis loop, indicative of large channel-like pores in a narrow range of size. The pore size distributions (Fig. S5) illustrate the pore size changes from 4.2 to 7.5 nm when the aging temperature rises from 40 to 100 °C. Table 3 summarizes the structural properties of KIT-6-X. From this table it can be seen that all the materials have good mesoporosity. As the aging temperature gets higher, the pore diameter increases, while the wall thickness decreases from 4.2 to 3.4 nm. Fig. 4 shows the small-angle XRD patterns of the mesoporous Co3O4 synthesized by using mesoporous silica KIT-6-X with Ia3d structure and different pore sizes as hard templates. The obtained mesoporous cobalt oxides clearly show well-resolved (2 1 1) reflections, showing that the pore symmetry is successfully replicated. The intensities of the reflection peaks of mesoporous Co3O4 increase with the pore size of silica template. And as the pore size of silica template rises, fine structural details become more evident and (2 2 0) reflection can be well observed in the XRD pattern of Co3O4-KIT-6-100. It indicates that large-pore KIT6 facilitate the replication of mesostructure by metal oxide. Notably, from the large-angle XRD pattern (Fig. 5), it can be seen that the Co3O4 products are well-crystalline and they possess Fd3m structure.
140
100 80
Co3O4-SBA-15-100
dV/dD
Adsorbed amount(cm3g-1 STP)
120
60 40 20
Co3O4-SBA-15-80
0
0.0
0.2
0.4
0.6
0.8
Relative pressure P/P0
1.0
2
4
6
8
10
Pore size (nm)
Fig. 2. N2 adsorption–desorption isotherms and pore size distributions of mesoporous Co3O4 replicas templated by SBA-15-X. The isotherm of Co3O4-SBA-15-100 was moved vertically by 20 cm3 g 1.
318
P. Shu et al. / Microporous and Mesoporous Materials 123 (2009) 314–323
Fig. 3. TEM images of mesoporous Co3O4 replicas prepared by (a) SBA-15-40 (pre size: 6.6 nm) and (b) SBA-15-100 (pore size: 10.7 nm) taken along the channel direction. The corresponding HRTEM images are given on the right.
Table 3 Structural properties of the mesoporous silica KIT-6-X (X: aging temperature in °C) with bicontinuous cubic Ia-3d symmetry. Sample
Unit cell parameter (nm)
S (m2 g
KIT6-40 KIT6-80 KIT6-100
19.5 21.4 22.1
630 569 666
1
)
Intensity (a.u.)
211
Co3O4-KIT-6-100
Co3O4-KIT-6-80
Co3O4-KIT-6-40
1
2
3
4
2 theta (degree) Fig. 4. XRD patterns of mesoporous Co3O4 templated by KIT-6-X.
Fig. 6 shows the N2 adsorption–desorption isotherms of silicafree mesoporous Co3O4 replicas (Co3O4-KIT-6-X). The structural
V (cm3 g 0.56 0.71 0.96
1
)
Pore diameter (nm)
Wall thickness (nm)
4.2 6.6 7.5
4.2 3.6 3.4
properties are summarized in Table 4. The isotherms are of type IV classification, which have typical hysteresis loops of mesoporous materials. With KIT-6-100 as template, the silica-free mesoporous Co3O4 sample has a surface area of 64 m2 g 1 and a pore volume of 0.14 cm3 g 1. It is noticeable that its isotherm is different from that of template KIT-6-100, giving pore size distribution centered at 3.5 nm, reflecting exactly the wall thickness of KIT-6100 (3.4 nm). This value is in good agreement with the pore diameters anticipated for a replica structure of KIT-6, indicating that the two sets of mesopores are successfully fabricated. As for the mesoporous Co3O4 templated by KIT-6-40 and 80, the pore sizes are not equal to the wall thickness of silica template. The pore size distribution curves centered at 10.7 and 7.5 nm for Co3O4-KIT-6-40 and Co3O4-KIT-6-80, respectively, much larger than the wall thickness of the corresponding silica template, 4.2 nm for KIT-6-40 and 3.6 nm for KIT-6-80. This might come from partial replication of only one set of the bicontinuous meso-channels. And this conclusion is further proved by TEM observations. Fig. 7a–c shows TEM images of mesoporous Co3O4 replicas synthesized using KIT-6 with different pore sizes (4.2, 6.6 and 7.5 nm, respectively) as the hard template, taken along the [1 0 0], [1 1 0] and [1 1 1] directions. It is clear from TEM images that all the Co3O4 replicas maintain a 3d structure and the channels are arranged in a well-ordered bicontinuous cubic structure. This
319
P. Shu et al. / Microporous and Mesoporous Materials 123 (2009) 314–323
20
30
40
50
Sample
S (m2 g
Co3O4-KIT-6-40 Co3O4-KIT-6-80 Co3O4-KIT-6-100
78 170 64
1
V (cm3 g
)
0.23 0.40 0.14
1
)
Pore diameter (nm) 10.7 7.6 3.5
440
511 422
222
400
220 111
Intensity(a.u.)
311
Table 4 Structural properties of the mesoporous Co3O4 templated by KIT-6-X.
60
70
2theta (degree) Fig. 5. Large-angle XRD pattern of mesoporous replica Co3O4-KIT-6-100.
indicates that when the cobalt nitrate was introduced into KIT-6 and reduced, the Co3O4 replica was formed inside the channels of KIT-6 without destroying the channel geometry. It is well known space group symmetry and possesses two nonthat KIT-6 has Ia3d intersecting channel networks separated by the silica wall which follows Gyroid minimal surface. From Fig. 7a (Co3O4-KIT-6-40) and 7b (Co3O4-KIT-6-80), it can be clearly seen that only one set of the nonintersecting channel network is replicated at the edge of the Co3O4 replica crystal, however, in the thick areas both nonintersecting channel networks are replicated and the symmetry is slightly destroyed due to the shift of the two networks. It also explains the difference of pore size distributions between replicas and the corresponding silica templates. Fig. 7c (Co3O4-KIT-6-100) clearly shows that two nonintersecting channel networks are completely replicated and the Co3O4 replica keeps the same space group symmetry as KIT-6. From Fig. 7a to c, it can be concluded that the pore size of KIT-6 template determines the extent of the replication, and two sets of channel networks can be completely replicated when the pore size is above or equal to 7.5 nm. Furthermore, when the pore size of KIT-6 template is small (66.6 nm),
Co3O4 can only replicate one set of pores at the edge of the particles and the area of one set replication is larger when the template pore size is smaller. Results of TEM observations are consistent with and strongly support the N2 adsorption data. structure facilitates the In summary, large-pore KIT-6 with Ia3d replication of both sets of the bicontinuous networks, and partial replication of one set network occurs when using small-pore KIT6 as the hard template. The disordered micropores connecting adjacent mesopores is among the reasons, similar to the case of replication of SBA-15. In addition, another possibility also answers structure for this fact. It is known that when the mesopore of Ia3d gets larger, ordered complementary micropores are present on the thinnest part of the pore walls, which connect the two independent mesopore networks [36]. Through the disordered micropores and ordered complementary micropores, the cobalt precursor diffuses between the two disconnected bicontinuous networks and there networks is achieved. fore replication of both sets of Ia3d 3.3. Mesoporous Co3O4 synthesized by AMS-10 with bicontinuous structure as hard template Pn3m In our previous work, we have found that both the mesostructure and pore size strongly depend on the ionization degree of the two-headed anionic surfactant [38]. The mesoporous materials synthesized with the surfactant neutralized with low HCl amount having larger effective head group areas often show smaller pore sizes, regardless of the type of CSDA. It has been proposed that the hydrophobic ‘‘tail” of the surfactant with larger effective head group area would ‘‘coil” to fill the space and thus maximize their van der Waals interactions, leading to lowering of the overall
300
200
dV/dD
Adsorbed amount (cm3g-1 STP)
400
Co3O4-KIT-6-80
100 Co3O4-KIT-6-40
Co3O4-KIT-6-100 0 0.0
0.2
0.4
0.6
0.8
Relative pressure P/P0
1.0
2
4
6
8
10
12
14
16
18
Pore size (nm)
Fig. 6. N2 adsorption–desorption isotherms and pore size distributions of the mesoporous Co3O4 replicas templated by KIT-6-X. The isotherms of Co3O4-KIT-6-40 and 80 were moved vertically by 50, 100 cm3 g 1, respectively.
320
P. Shu et al. / Microporous and Mesoporous Materials 123 (2009) 314–323
Fig. 7. TEM images of mesoporous Co3O4 replicas prepared by (a) KIT-6-40 (pore size: 4.2 nm), (b) KIT-6-80 (pre size: 6.6 nm) and (c) large-pore KIT-6-100 (pore size: 7.5 nm) taken along [1 0 0], [1 1 0] and [1 1 1] directions, respectively.
energy. Lower surfactant packing density leads to more tightly coiled surfactant hydrocarbon tails and smaller pore size. It has been found that the unit cell parameter and pore size are significantly different even between the same mesostructures. Mesoporous silicas AMS-10 with different pore sizes are synthesized by adding different amount of HCl into the C14GluAS solution. The synthesis molar composition is: C14GluAS/TMAPS/TEOS/ H2O/HCl = 1:1.5:15:2000:Y. Two X-ray reflection peaks of each samples can be distinguished with a reciprocal spacing ratio of about 1:1.225 and can be indexed to (1 1 0) and (1 1 1) reflections of bicontinuous cubic Pn3m mesophase (Fig. S6). The HRTEM images (not shown) confirm that all of the space groups of these The resultant mesoporous materials are bicontinuous cubic Pn3m. products show increasing d-spacing with the reflection peak [1 1 0] shifting to the smaller angle from 1.39° to 1.14° with increasing HCl/C14GluAS molar ratio from 0.42 to 0.48. The N2 adsorption–desorption isotherms of the samples AMS10-Y (Y = HCl/C14GluAS) (Fig. S7) show type IV features, typical for mesoporous materials, and three well-distinguished regions of the adsorption isotherm are observed including monolayer– multilayer adsorption, capillary condensation, and multilayer adsorption on the outer surface. All the samples clearly show the hysteresis loop of H1 type and the capillary condensation occurs at increasing relative pressure with increasing HCl/C14GluAS molar ratio, indicating increasing size of cylindrical mesopores. Interestingly, the pore size distribution curves, BJH plots of derivative of the pore volume per unit weight with respect to the pore diameter (dV/dD) versus the pore diameter (D), show a clear shift of pore
diameter from 5.1 to 6.7 nm, as the molar ratio of HCl/C14GluAS changes from 0.42 to 0.48 (Fig. S7). The porous properties of AMS-10-Y are summarized in Table 5. All the samples show large surface areas (larger than 600 m2 g 1) and pore volumes (larger than 0.8 cm3 g 1), indicating high mesoporosity of the samples. As the HCl/C14GluAS molar ratio increases, the pore size increases significantly. However, the wall thickness is almost the same around 1.2 nm for all the samples. Fig. 8 shows the XRD patterns of Co3O4 mesoporous replicas synthesized with AMS-10-Y as the template. After Co3O4 is embedded in the mesopores and silica is removed, the XRD patterns show no significant peaks within the range of 0.8–6.0° (2h). Only one reflection peak with low intensity is observed from each pattern, which can be reasonably assigned to the (1 1 0) reflection, showing the d-spacing close to that of the silica template. These peaks can be explained as the X-ray reflections from the mesoporous Co3O4 that is obtained by replication of two sets of the bicontinuous mesopore networks, which shows the same unit cell dimensions as the silica template. The low intensity of the X-ray reflection indicates that replication of two sets of bicontinuous mesopore network by Co3O4 occurs partially. Replication of only one set of the mesochannels may also occur, which doubles the unit cell dimensions Ref. [32]) and therefore the X-ray reflec(structural model of Pn3m, tions shift to a much smaller angle (2h). Fig. 9 shows the N2 adsorption–desorption isotherms and pore size distributions of silica-free Co3O4 mesoporous replicas templated by AMS-10-Y. The porous properties are shown in Table 6. The isotherms are of type IV classification, which have typical
321
P. Shu et al. / Microporous and Mesoporous Materials 123 (2009) 314–323 Table 5 Structural properties of mesoporous silica AMS-10-Y (Y: molar ratio of HCl/C14GluNa) with bicontinuous cubic Pn-3m symmetry. Unit cell parameter (nm)
S (m2 g
AMS-10-0.42 AMS-10-0.44 AMS-10-0.45 AMS-10-0.48
9.0 9.5 9.8 11.0
649 624 640 640
Intensity (a.u.)
Sample
1
V (cm3 g
)
Co3O4-AMS-10-0.44 Co3O4-AMS-10-0.42
3
4
5
Pore diameter (nm)
Wall thickness (nm)
5.1 5.5 5.8 6.7
1.3 1.2 1.1 1.1
the wall thickness of silica template, which might be due to the replication of only one set of the bicontinuous meso-channels. The large mesopore of Co3O4 is close to the expected value, twice of the wall thickness plus pore diameter of silica template, on condition that only one set of meso-channels is replicated. Therefore, it can be concluded that, within the full range of pore size of AMS10, Co3O4 replicates both one set and two sets of the bicontinuous meso-channels. It could be further proved by TEM observations. Fig. 10a and b shows TEM images of mesoporous Co3O4 replicas synthesized from AMS-10 with different pore sizes (5.1 and 6.7 nm, respectively) taken along [1 1 0] and [1 1 1] directions. It can be clearly seen that all the Co3O4 replicas have a 3d structure and the channels show a well-ordered bicontinuous cubic structure. This implies that the Co3O4 replica is formed inside the channels of AMS-10 like in the channel of KIT-6 described above. Similar to KIT-6, AMS-10 has two nonintersecting channel networks as well. However, those two channel networks are separated by the silica wall which follows Diamond minimal surface and space group symmetry. From Fig. 10a AMS-10 possesses Pn3m (Co3O4-AMS-10-0.42) and 10b (Co3O4-AMS-10-0.48), it is clear that only one set of the nonintersecting channel network was replicated in the edge of the Co3O4 replica crystal, but in the thick areas both nonintersecting channel networks were replicated and further the contrast between two channel networks is different from each other which results in the break of the symmetry. The presence of the replication of one set of channel network illustrates the difference of pore size distributions between both replicas and the corresponding templates. Therefore, it can be summarized that mesoporous Co3O4 prepared by using AMS-10 with pore size of 5.1–6.7 nm is obtained by replication of both one set and two sets of bicontinuous meso-channel networks. AMS-10 is synthesized by using anionic surfactant as the template, and therefore no micropore is present
Co3O4-AMS-10-0.45
2
)
0.87 0.90 0.94 1.04
Co3O4-AMS-10-0.48
1
1
6
2theta (degree) Fig. 8. XRD patterns of mesoporous Co3O4 templated by AMS-10-Y.
hysteresis loops of mesoporous materials. The pore size distributions show similar feature that the materials have bimodal porous structure, with two maxima of dV/dD within the pore size regions of 2–5 nm and 5–15 nm. Considering that the pore size and wall thickness of silica template are 5.1–6.7 nm and 1.2 nm, respectively, it is reasonable that the pore size distribution of mesoporous Co3O4 within 2–5 nm is attributed to the replication of two sets of the bicontinuous mesopores of AMS-10. One the other hand, the pore size of mesoporous Co3O4 within 5–12 nm is much larger than
400 Co3O4-AMS-10-0.48
300 Co3O4-AMS-10-0.45
200
dV/dD
Adsorbed amount (cm3g-1, STP)
500
Co3O4-AMS-10-0.44
100 Co3O4-AMS-10-0.42
0 0.0
0.2
0.4
0.6
2 theta (degree)
0.8
1.0
2
4
6
8 10 12 14 16 18 20
Pore size (nm)
Fig. 9. N2 adsorption–desorption isotherms and pore size distributions of the mesoporous Co3O4 replicas templated by AMS-10-Y. The isotherms of Co3O4-AMS-10-0.44, 0.45 and 0.48 were moved vertically by 100, 200 and 300 cm3 g 1, respectively.
322
P. Shu et al. / Microporous and Mesoporous Materials 123 (2009) 314–323
Table 6 Structural properties of the mesoporous Co3O4 templated by AMS-10-Y. Sample
S (m2 g
Co3O4-AMS-10-0.42 Co3O4-AMS-10-0.44 Co3O4-AMS-10-0.45 Co3O4-AMS-10-0.48
134 119 151 141
1
)
V (cm3 g 0.23 0.27 0.34 0.26
1
)
Pore diameter (nm) 1.9, 3.1, 3.6, 2.0,
7.6 8.8 10.6 6.6
in the silica wall growing following the Diamond minimal surface. A reasonable explanation is the existence of ordered complementary micropores at the thinnest part of the silica wall, similar to the case of KIT-6 [37]. These ordered micropores in the silica wall facilitate the diffusion of cobalt precursor between two sets of bicontinuous meso-channels. As a result, in the thick part of the particles, a complete replication of two sets of meso-channels is achieved. Metal oxide inverse replication provides additional structural information of the silica template AMS-10. The bicontinuous networks are visualized by this means. And as discussed above, the materials have thin silica wall (1.2 nm) and at the thinnest parts there are ordered complementary micropores which connect the two independent bicontinuous meso-channels. 3.4. Comparison of pore size effect in the inverse replication As has been discussed in details above, it is clear that the pore size of template silica will greatly influence the structure of inverse Co3O4 replica. However, this discussion of the effect of pore size is valid when taking the pore symmetry into consideration. SBA-15 (p6mm) normally have large-pore size and small wall thickness. When the pore size is 6.6 nm or smaller, only separated
nano-wires of Co3O4 replicas are obtained, due to a lack of micropore connections between adjacent mesopores. In the case of KIT-6 (Ia3d), when the pore size is 6.6 nm or smaller, both replication of one set and two sets of bicontinuous mesopores are replicated, indicating mesopore connections by micropores in the wall. It is symmetry worth noting that the wall thickness of KIT-6 with Ia3d (>3.6 nm) is much larger than that of SBA-15 with p6mm symmetry (2.4 nm), which may have less micropore connections without strucconsidering the pore symmetry. However, because the Ia3d ture has an inhomogeneous wall thickness, some part of the silica wall is fairly thin though the mean wall thickness is high, which may have micropore connections. Besides, ordered complementary structure. Theremicropore connections might also exist in Ia3d fore, different pore symmetry differs much in inverse replication. AMS-10 (Pn3m) materials have similar pore size to KIT-6 (Ia3d), but they have a large porosity. The change of pore size does not change the wall thickness significantly, and all of materials possess ordered micropore connections between mesopores. Therefore, the effect of pore size is ignoring, and all the replication process includes replication of one set and two sets of the bicontinuous mesopore networks. 4. Conclusion Pn3m Mesoporous silica with different pore size of p6mm, Ia3d, symmetries are employed as hard templates to synthesize mesoporous Co3O4 replica. It is found that for p6mm, large-pores facili and Pn3m, tate the replication and that for Ia3d the pore size determines the extent of the replication. For p6mm, at small-pore sizes, the structure of Co3O4 replica collapses and isolated nanowires have been obtained; At large-pore sizes, complete replication can be achieved, retaining the original symmetry of silica template.
Fig. 10. TEM images of mesoporous Co3O4 replicas prepared by (a) small-pore AMS-10-0.42 (pore size: 5.1 nm) and (b) large-pore AMS-10-0.48 (pore size: 6.7 nm) taken along [1 1 0] and [1 1 1] directions.
P. Shu et al. / Microporous and Mesoporous Materials 123 (2009) 314–323
structure, at large-pore sizes two sets of the mesopores are For Ia3d replicated, and at small-pore sizes, both two sets and one set of the bicontinuous mesopore networks are replicated by Co3O4. For structure, unlike Ia3d, within the pore size range of 5.1– Pn3m 6.7 nm, both one set and two sets of the bicontinuous meso-channel networks are replicated. The discussion of pore size effect on replication is valid only when pore symmetry of silica template is specified. In the inverse replication, disordered micropores in the walls of SBA-15 and KIT-6 play an important role, and ordered complementary micropores in the walls of KIT-6 and AMS-10 are the vital factors in determining the extent of replication. Inverse replication also provides structural information of silica template. It is proposed that for AMS-10, there are ordered complementary micropores at the thinnest parts of the wall, which connect the two independent bicontinuous meso-channels.
[7] [8] [9] [10] [11] [12]
Acknowledgments
[24]
This work was supported by the National Natural Science Foundation of China (Grant Nos. 20890121 and 20521140450) and 973 project (2009CB930403) of China. Appendix A. Supplementary data
[13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23]
[25] [26] [27] [28] [29] [30]
Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.micromeso.2009.04.017. References [1] P. Behrens, Angew. Chem. Int. Ed. 35 (1996) 515–518. [2] X. He, D. Antonelli, Angew. Chem. Int. Ed. 41 (2002) 214–229. [3] E. Rosencher, A. Fiore, B. Vinter, V. Berger, P. Bios, J. Nagle, Science 271 (1996) 168–173. [4] M. Tiemann, Chem. Eur. J. 13 (2007) 8376–8388. [5] (a) F. Schüth, Chem. Mater. 13 (2001) 3184–3195; (b) G.J.de A.A. Soler-II-lia, C. Sanchez, B. Lebeau, J. Patarin, Chem. Rev. 102 (2002) 4093–4138. [6] P. Kluson, P. Kacer, T. Cajthaml, M. Kalaji, J. Mater. Chem. 11 (2001) 644–651.
[31] [32] [33] [34] [35] [36] [37] [38]
323
R. Ryoo, S.H. Joo, S.J. Jun, Phys. Chem. B 103 (1999) 7743–7746. F. Schüth, Angew. Chem. Int. Ed. 42 (2003) 3604–3622. H. Yang, D. Zhao, J. Mater. Chem. 15 (2005) 1217–1231. M. Tiemann, Chem. Mater. 20 (2008) 961–971. B. Ohtani, Y. Ogawa, S. Nishimoto, J. Phys. Chem. 101 (1997) 3746–3752. S. Jun, S.H. Joo, R. Ryoo, M. Kruk, M. Jaroniec, Z. Liu, T. Ohsuna, O. Terasaki, J. Am. Chem. Soc. 122 (2000) 10712–10713. T.-W. Kim, I.-S. Park, R. Ryoo, Angew. Chem. 115 (2003) 4511–4515. S.H. Joo, S.J. Choi, I. Oh, J. Kwak, Z. Liu, O. Terasaki, R. Ryoo, Nature 412 (2001) 169–172. H.J. Shin, R. Ryoo, J. Mater. Chem. 11 (2001) 260–261. H.J. Shin, R. Ryoo, Z. Liu, O. Terasaki, J. Am. Chem. Soc. 123 (2001) 1246–1247. M.J. Maclachlan, G.A. Ozin, J. Am. Chem. Soc. 121 (1999) 12005–12017. Y. Shi, Y. Wan, R. Zhang, D. Zhao, Adv. Funct. Mater. 18 (2008) 2436–2443. S.C. Laha, R. Ryoo, Chem. Commun. 17 (2003) 2138–2139. Y.Q. Wang, C.M. Yang, W. Schmidt, B. Spliethoff, E. Bill, F. Schüth, Adv. Mater. 17 (2005) 53–56. J. Roggenbuck, M. Tiemann, J. Am. Chem. Soc. 127 (2005) 1096–1097. C. Dickinson, W. Zhou, R. Hodgkins, Y. Shi, D. Zhao, H. He, Chem. Mater. 18 (2006) 3088–3095. B. Tian, X. Liu, H. Yang, S. Xie, C. Yu, B. Tu, D. Zhao, Adv. Mater. 15 (2003) 1370– 1374. H. Yang, Q. Shi, B. Tian, Q. Lu, F. Gao, S. Xie, J. Fan, C. Yu, B. Tu, D. Zhao, J. Am. Chem. Soc. 125 (2003) 4724–4725. W. Yue, W. Zhou, Chem. Mater. 19 (2007) 2359–2363. K. Zhu, B. Yue, W. Zhou, H. He, Chem. Commun. (2003) 98–99. A. Rumplecker, F. Kleitz, E.-L. Salabas, Ferdi Schüth, Chem. Mater. 19 (2007) 485–496. H. Tüysüz, C.W. Lehmann, H. Bongard, B. Tesche, R. Schmidt, F. Schüth, J. Am. Chem. Soc. 130 (2008) 11510–11511. B. Yue, H. Tang, Z. Kong, K. Zhu, C. Dickinson, W. Zhou, H. He, Chem. Phys. Lett. 407 (2005) 83–86. D. Zhao, J. Feng, Q. Huo, N. Melosh, G.H. Fredrickson, B.F. Chmelka, G.D. Stucky, Science 279 (1998) 548–552. F. Kleitz, S.H. Choi, R. Ryoo, Chem. Commun. 17 (2003) 2136–2137. C. Gao, Y. Sakamoto, K. Sakamoto, O. Terasaki, S. Che, Angew. Chem. Int. Ed. 118 (2005) 4401–4404. P.I. Ravikovitch, A.V. Neimark, Langmuir 16 (2000) 2419–2423. K. Schumacher, P.I. Ravikovitch, A. Du Chesne, A.V. Neimark, K.K. Unger, Langmuir 16 (2000) 4648–4654. K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquerol, T. Siemieniewska, Pure Appl. Chem. 57 (1985) 603–619. A. Galarneau, H. Cambon, F. Di Renzo, R. Ryoo, M. Choi, F. Fajula, New J. Chem. 27 (2003) 73–79. Y. Sakamoto, T-W. Kim, R. Ryoo, O. Terasaki, Angew. Chem. Int. Ed. 43 (2004) 5231–5234. C. Gao, H. Qiu, W. Zeng, Y. Sakamoto, O. Terasaki, K. Sakamoto, Q. Chen, S. Che, Chem. Mater. 18 (2006) 3904–3914.
Contents lists available at ScienceDirect
Microporous and Mesoporous Materials journal homepage: www.elsevier.com/locate/micromeso
Formation of mesoporous Co3O4 replicas of different mesostructures with different pore sizes Peng Shu a,1, Juanfang Ruan b, Chuanbo Gao a,b, Huachun Li a, Shunai Che a,* a
School of Chemistry and Chemical Engineering, State Key Laboratory of Metal Matrix Composite Materials, Shanghai Jiao Tong University, 800 Dongchuan Road, Shanghai 200240, China b Structural Chemistry, Arrhenius Laboratory, Stockholm University, S-10691 Stockholm, Sweden
a r t i c l e
i n f o
Article history: Received 17 February 2009 Received in revised form 9 April 2009 Accepted 14 April 2009 Available online 21 April 2009 Keywords: Mesoporous cobalt oxide Hard templating Replica Mesostructure Pore size
a b s t r a c t Mesoporous metal oxides Co3O4 are prepared via hard templating synthesis method by using various mesoporous silicas with different pore size as templates. The pore size of the mesoporous silicas with and Pn3m have been the symmetry of two-dimensional (2d)-hexagonal p6mm, bicontinuous cubic Ia3d controlled in the range of 6.6–10.7, 4.2–7.5 and 5.1–6.7 nm, respectively, by choosing different surfactants and co-surfactants and by adjusting either the aging temperature or the ionization degree of the surfactant. The pore size of the silica template has been considered to be an important factor that determines the mesostructure of the resulting metal oxides. It has been found that for p6mm, it is easier to at large-pore size two sets of replicate the mesoporous symmetry at large size of mesopores. For Ia3d, leads to replibicontinuous meso-channels are replicated into mesoporous Co3O4, while small-pore Ia3d cation of both one set and two sets of meso-channels. Co3O4 can replicate both one set and two sets of meso-channels at all pore sizes that can be obtained (5.1–6.7 nm), indicating the exisbicontinuous Pn3m tence of ordered complementary micropores within the silica walls. Ó 2009 Elsevier Inc. All rights reserved.
1. Introduction Mesoporous metal oxides are of great interest for applications in catalysis since they can be regarded as self-supported catalysts with large specific surface areas and shape selective properties, which find utilities in areas of catalysis, chemical sensing, superconductivity, colossal magneto-resistance, piezoelectricity, etc. [1–4]. Their nano-sized forms differ greatly in many properties from the bulk materials [5]. These materials can be synthesized by using either organic soft templates [6], similar to the synthesis of mesoporous silica, or mesoporous silica as a hard template [7– 10]. While soft templating normally results in amorphous materials, compared to silica, the surfactant/oxide composite precursors are often more susceptible to suffer from lack of condensation, redox reaction, and phase transitions accompanied by thermal breakdown of the structural integrity [6,11]. The hard templating method was firstly introduced by Ryoo and co-workers for synthesizing mesoporous carbon [7,12–14], and several mesoporous metals [15,16], metal sulfides [17], nitride [18] and oxides [19–21] have been fabricated, although the wall of most of these materials was still disordered and the formation mechanisms of these new * Corresponding author. Tel.: +86 21 5474 2852; fax: +86 21 5474 1297. E-mail address: [email protected] (S. Che). 1 Present address: Department of Material Science and Engineering, Stanford University, USA. 1387-1811/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.micromeso.2009.04.017
materials were not discussed extensively. Recently, researchers reported the synthesis of mesoporous metal oxides by using hard templating method, some of which are in single crystalline state [22–29]. Mesoporous cobalt oxide Co3O4 is one of the metal oxides that have been studied extensively. Zhao et al. and Schüth et al. prepared mesoporous Co3O4 with functional mesoporous silicas as hard templates by using impregnating method [20,23]. Zhou et al. prepared mesoporous Co3O4 by using a solid–liquid route, and products were maintained to be single crystal [25]. The formation mechanism of mesoporous Co3O4 has been studied, and the factors determining the porous structure of replica are considered to be crystallization conditions, pore size of template, interconnectivity between mesopores and precursor loading levels [22,27]. Recently, Schüth et al. visualized the pore topology of mesoporous Co3O4 by high-resolution scanning electron microscopy [28]. Various mesoporous silicas can be used as hard template for synthesizing mesoporous metals or metal oxides. Two typical examples are mesoporous silica SBA-15 with hexagonally closepacked straight mesopores connected by some smaller channels symmetry containing and micropores [30], and KIT-6 with Ia3d three-dimensional (3d) bicontinuous meso-channel networks [31]. Typically, such methods employ mesoporous silicas as the hard templates, into which a solution-based precursor of the desired metal or metal oxide is introduced, followed by heating or hydrogenation to form the desired metal oxide or metal and silica
P. Shu et al. / Microporous and Mesoporous Materials 123 (2009) 314–323
composites [10]. The silica template is dissolved away to leave a replica mesoporous structure of the target compound. The hard templating route has opened a way to great varieties of mesoporous transition metal oxides. The relationship between the pore size of the silica template and that of the obtained metal oxide replicas is an interesting topic. Here, the mesostructures including 2d-hexagonal p6mm, cubic and Pn3m that possess rod-type pore system have been seIa3d lected for studying the relationship, because, in the case of cagetype mesoporous silica, it is well known that the small apertures make it difficult for the precursors to completely fill the cages and to form rigid carbon or metal oxide bridges between their nanoparticles prepared in silica cages. Here, we present a report on the study of the effect of pore size and pore symmetry on the formation of mesoporous Co3O4 by preparing the mesoporous co crystal structure) using mesoporous silicas balt oxide Co3O4 (Fd3m and SBA-15 (2d-hexagonal p6mm), KIT-6 (bicontinuous cubic Ia3d) AMS-10 (bicontinuous Pn3m) with different pore size as hard templates. The dependence of the structure of replica on pore size and symmetry of the silica template will be discussed in details. Different from earlier literatures, in this study, we will show that solely the pore size of silica template cannot determine the replication extent, and the discussion is valid only when considering the mesopore symmetry. Structural properties of the mesoporous silicas, especially newly developed AMS-10, are also derived from inverse replication.
2. Experimental section 2.1. Synthesis 2.1.1. Synthesis of mesoporous silica The mesoporous silica templates SBA-15 [23], KIT-6 [24], and AMS-10 [32] were prepared according to the procedure described previously. 2.1.1.1. Synthesis of 2d-hexagonal p6mm structured mesoporous silica (SBA-15) with different pore size. SBA-15 mesoporous silica was synthesized by using the triblock copolymer Pluronic P123 (EO20PO70EO20, BASF) as surfactant and tetraethoxylsilicon (TEOS) as silica source. In a typical synthesis experiment, TEOS (20.8 g) was added to a mixture of P123 (10 g), HCl (62.57 g, 35 wt.%), and deionized water (319 g) under stirring at 35 °C. After the mixture was stirred for 24 h, the mesostructured products thus formed were aged at 100 °C for additional 24 h. The products were filtered, dried without washing, and calcined at 550 °C. The pore size was controlled by adjusting aging temperature in the range of 40– 100 °C. The samples with three pore sizes have been synthesized at aging temperature of 40, 80 and 100 °C. These samples were denoted as SBA-15-40, SBA-15-80, and SBA-15-100, respectively. structured mesoporous silica 2.1.1.2. Synthesis of bicontinuous Ia3d mesopor(KIT-6) with different pore size. The large-pore cubic Ia3d ous silica KIT-6 was synthesized by using triblock copolymer Pluronic P123 as surfactant and TEOS as silica source. In a typical synthesis, 6 g of P123 was dissolved in 217 g of distilled water and 11.8 g of HCl (35 wt.%). To this, 6 g of butanol was added under stirring at 35 °C. After stirring for 1 h, 12.9 g of TEOS was added at 35 °C (TEOS:P123:HCl:H2O:BuOH = 1:0.017:1.83:195:1.31 mol ratio). The mixture was left under stirring for 24 h at 35 °C, and subsequently aged for 24 h at 100 °C under static conditions in a closed polypropylene bottle. The products were filtered, dried without washing, and calcined at 550 °C. In the similar manner of SBA-15, the pore size was controlled by adjusting aging temperature in the range of 40–100 °C. The samples with three pore sizes
315
have been synthesized at different aging temperature of 40, 80, and 100 °C. These samples were denoted as KIT-6-40, KIT-6-80, and KIT-6-100, respectively. structured mesoporous silica 2.1.1.3. Synthesis of bicontinuous Pn3m (AMS-10) with different pore size. The bicontinuous cubic Pn3m mesoporous silica (AMS-10) was synthesized by using N-myristoyl-L-glutamic acid sodium salt (C14GluAS, one of the carboxylic acid groups of the glutamic acid was neutralized by sodium hydroxide) as surfactant, N-trimethoxylsilylpropyl-N,N,N-trimehylammonium chloride (TMAPS) as co-structure-directing agent (CSDA) and TEOS as silica source. In a typical synthesis experiment, HCl (0.417 g of 1 mol L 1 solution) was added to a solution of C14GluAS (0.379 g, 1 mmol) in deionized water (36 g) under stirring at 80 °C. After the solution became homogeneous, a mixture of TMAPS (0.773 g, 50 wt.% in methanol, 1.5 mmol) and TEOS (3.12 g, 15 mmol) was added, and the mixture was stirred at 80 °C for 10 min. The molar composition of the final gel was C14GluAS/TMAPS/TEOS/H2O/HCl = 1:1.5:15:2000:0.42. The precipitate was then aged at 80 °C for 2 days, filtered, and calcined at 550 °C for 6 h to remove the surfactants. The pore size of AMS-10 was controlled by adjusting the HCl/C14GluAS molar ratio in the range of 0.42–0.48. The samples with four pore sizes were synthesized with different HCl/C14GluAS molar ratios of 0.42, 0.44, 0.45 and 0.48. They were denoted as AMS-10-0.42, AMS-10-0.44, AMS-10-0.45 and AMS-10-0.48, respectively. 2.1.2. Synthesis of mesoporous Co3O4 replica In a typical synthesis of mesoporous Co3O4, 0.75 g of Co(NO3)26H2O (98% Aldrich) was dissolved in 16 mL of ethanol followed by addition of 0.5 g of mesoporous silica templates. The mixture was stirred at room temperature until nearly dry powder had been obtained; the sample was then heated slowly to 300 °C and calcined at the same temperature for 3 h to pyrolyze the nitrate. The impregnation procedure was repeated twice with 0.4 g of Co(NO3)26H2O and 0.25 g Co(NO3)26H2O dissolved in 16 mL ethanol, followed by calcination at 550 °C for 5 h with a ramp 1 °C min 1 for metal oxide to crystallize. The resulting samples were treated with 10% HF to remove the silica template, centrifuged, washed with water and ethanol, and then dried at 60 °C in air. They were denoted as Co3O4-SBA-15-X, Co3O4-KIT-6-X and Co3O4-AMS-10-Y (X: aging temperature of silica template in °C; Y: molar ratio of HCl/C14GluAS in the synthesis of AMS-10). 2.2. Characterizations X-ray diffraction (XRD) patterns were recorded on a Rigaku RINT 2000/PC Multiplex instrument using Cu Ka radiation (k = 0.15406 nm), operated at 40 kV and 20 mA (0.8 kW) at the rate of 1.0° min 1 over the range of 0.8–6.0° (2h). N2 adsorption–desorption isotherms were measured at 196 °C on a Quantachrome Nova 4200e volumetric adsorption analyzer. Before the adsorption measurements, all samples were outgassed at 200 °C in the port of the adsorption analyzer for 4 h. Specific surface areas were calculated via the Brunauer–Emmett–Teller (BET) model in regions applicable to the derivation of the model between P/P0 values of 0.05–0.3. The total pore volume was determined from the uptake of nitrogen at a relative pressure of P/P0 0.99. Pore size distribution curves were obtained from the adsorption branch by using the Barrett–Joyner–Halenda (BJH) method. Scanning electron microscopy (SEM) features of all the samples were observed by SEM (JEOL JSM-7401F). The samples were observed without any metal coating. High-resolution transmission electron microscopy (HRTEM) images were taken from thin edges of particles supported on a porous carbon grid, using JEOL JEM3010 electron microscope (point resolution 1.7 Å, Cs = 0.6 mm)
316
P. Shu et al. / Microporous and Mesoporous Materials 123 (2009) 314–323
operated at 300 kV. Samples were crushed in an agate mortar, suspended in ethanol using ultrasonics and then dropped onto the carbon grid. The wall thickness (w) of mesoporous silicas was calculated combining the unit cell parameter (a) derived from XRD pattern and pore diameter (d) derived from N2 adsorption–desorption. For 2d-hexagonal p6mm symmetry, w was calculated as a–d. For bicontinuous cubic Ia-3d symmetry, w was calculated as a/ 3.0919-d/2 [33,34]. For bicontinuous cubic Pn-3m symmetry, w was determined as d110-d, i.e., w = a/1.4142-d (structure model, Ref. [32]).
3. Results and discussion 3.1. Mesoporous Co3O4 synthesized by SBA-15 with p6mm structure as hard template Triblock copolymer Pluronic P123 was used as a template to synthesize SBA-15. It was reported that the pore size can be modulated by aging temperature [23]. PPO moieties are hydrophobic while PEO moieties are hydrophilic. PEO moieties are expected to interact with the inorganic species more strongly and closely associated with the inorganic wall than the hydrophobic PPO. As temperature rises, the PEO moieties become more hydrophobic, resulting in increased hydrophobic domain volume and shorter length of the PEO segments associated with the inorganic wall, thus the pore size increases. All the SBA-15-X (X: aging temperature in °C) samples show three well-resolved XRD peaks in the region of 2h = 0.5–2.5°, which correspond to (10), (11) and (20) reflections based on the 2d-hexagonal system (Fig. S1). In addition, the higher hydrothermal aging temperature gives rise to 2d-hexagonal p6mm silica with larger pores and improved structural integrity. The resultant products show the increasing d-spacing indicated from the fact that the (10) reflection peak shifts to a smaller angle with increasing hydrothermal aging temperature from 40 to 100 °C. N2 adsorption and desorption isotherms for SBA-15-X (Fig. S2) are of type IV according to the IUPAC classification [35]. Significant H1 type hysteresis loop is observed for all isotherms, and the capillary condensation occurs at increasing relative pressure with increasing aging temperature, indicating increasing size of cylindrical mesopores. In addition, the capillary condensation step becomes narrowed with increasing synthesis aging temperature, which indicates the degree of pore uniformity is increased. Interestingly, the pore size distribution curves, BJH plots of derivative of the pore volume per unit weight with respect to the pore diameter (dV/dD) versus the pore diameter (D), show a clear shift of pore diameter from 6.6 to 10.7 nm as the aging temperature is increased from 40 to 100 °C (Fig. S2). Table 1 summarizes the structural properties of the SBA-15-X mesoporous silicas. All the SBA-15-X samples have BET surface area larger than 700 m2 g 1, and the pore volume increases from 0.70 to 1.18 cm3 g 1 with higher aging temperatures. The wall thickness has also been determined on the basis of unit cell parameter and pore diameter, and it can be seen that the wall thickness
decreases from 2.4 to 0.7 nm when the aging temperature increases from 40 to 100 °C in the synthesis. Fig. 1 shows the XRD patterns of mesoporous Co3O4 after removal of silica, using SBA-15 with different pore sizes as hard templates. After cobalt oxide is embedded within the mesopores of SBA-15, the intensities of the (10), (11) and (20) reflections are dramatically reduced. The obtained mesoporous cobalt oxides synthesized by large-pore samples SBA-15-80 and 100, show wellresolved (10) reflection while lose the (20) reflection peak. It indicates that the mesostructures are well retained but the structural order decreases during replication process. Sample Co3O4-SBA15-40 templated by mesoporous silica with the smallest pore size gives no significant X-ray reflection peak in the range of 2h = 0.5– 4°. It suggests that after the removal of silica template, the hexagonal arrangements of nano-wires collapse, which is confirmed by later TEM characterization. From the large-angle XRD pattern, it can be concluded that the Co3O4 products are well crystallized (not shown). and the crystal structure is cubic Fd3m As is seen in the SEM images (Fig. S3), the morphology of the Co3O4 replica is close to that of the mesoporous silica template (SBA-15-100). Cobalt nitrate was deposited in the straight channels of templates. Nuclei were formed simultaneously in the channels with the temperature increase, followed by the growth of the cobalt oxide. The confinement of the straight channels of the SBA15 allowed the cobalt oxide to grow in one direction and completely fill the pores, keeping the morphology. Fig. 2 shows the N2 adsorption–desorption isotherms of the silica-free Co3O4 mesoporous replicas synthesized by using SBA-1580 and 100. The structural properties are summarized in Table 2. The isotherms are of type IV classification, which have typical hysteresis loops of mesoporous materials. With SBA-15-100 mesoporous silica as template, the silica-free mesoporous Co3O4 sample has a surface area of 86 m2 g 1 and a pore volume of 0.17 cm3 g 1. It is noticeable that its isotherm is different from that of SBA-15-100, giving a pore size distribution centered at 2.3 nm, indicating the expected pore dimension, which is larger but is still in scale with the wall thickness of SBA-15 as shown in Table 1. Similarly, the pore size of mesoporous Co3O4 templated by SBA-15-80 is also close to the wall thickness of silica template. It suggests that the SBA-15-100 and 80 are successfully replicated. Fig. 3a and b are TEM images of the Co3O4 mesoporous replicas synthesized with small-pore SBA-15-40 (6.6 nm) and large-pore SBA-15-100 (10.7 nm), respectively, as hard templates. The length of Co3O4 nano-wires ranged from several tens to several hundreds of nanometers, and the average diameters of Co3O4 nano-wires are about 6.3 and 9.8 nm, respectively, which match the pore sizes of the corresponding SBA-15 templates very well. It can be clearly seen from Fig. 3b that mesoporous Co3O4 replica well retains the ordered 2d-hexagonal p6mm structure of the large-pore silica template SBA-15-100. However, the mesoporous Co3O4 replicas templated by small-pore SBA-15-40 (Fig. 3a) are isolated nano-wires without interconnections to retain the symmetry of silica template. Notably, the corresponding HRTEM of Fig. 3a and b marked by arrows are given on the right. From the HRTEM images it can be seen that the mesoporous Co3O4 replicas are crystalline, and each
Table 1 Structural properties of the mesoporous silica SBA-15-X (X: aging temperature in °C) with 2d-hexagonal p6mm symmetry. Sample
Unit cell parameter (nm)
S (m2 g
SBA-15-40 SBA-15-80 SBA-15-100
9.0 10.0 11.3
794 711 714
1
)
V (cm3 g 0.70 0.96 1.18
1
)
Pore diameter (nm)
Wall thickness (nm)
6.6 8.9 10.7
2.4 1.1 0.7
317
P. Shu et al. / Microporous and Mesoporous Materials 123 (2009) 314–323 Table 2 Structural properties of the mesoporous Co3O4 templated by SBA-15-X.
10
11
Intensity (a.u.)
Co3O4-SBA-15-100
Co3O4-SBA-15-80
Co3O4-SBA-15-40
1
2
3
4
2 theta (degree) Fig. 1. XRD patterns of the mesoporous Co3O4 templated by SBA-15-X.
of them can be regarded as a single crystal although the crystal orientations of the two adjacent Co3O4 nano-wires are different. In summary, SBA-15 with large-pores (larger than 8.9 nm) can be fully replicated by Co3O4, retaining the mesostructure of silica template, while replication of small-pore SBA-15 (6.6 nm) results in separated nano-wires. It is probably due to the microporous defects in the walls of SBA-15. If the pore dimension gets larger, the pore wall becomes thinner, and the pore-forming P123 in adjacent micelles would share hydrated PEO and therefore forms micropores that connect mesopores, which plays a vital role in retaining the template symmetry during the inverse replication [36]. 3.2. Mesoporous Co3O4 synthesized by KIT-6 with bicontinuous Ia3d structure as hard template Similar to the case of SBA-15, the pore size of bicontinuous cu mesoporous silica KIT-6 can also be modulated by changbic Ia3d
Sample
S (m2 g
Co3O4-SBA-15-80 Co3O4-SBA-15-100
64 86
1
)
V (cm3 g 0.16 0.17
1
)
Pore diameter (nm) 2.2 2.3
ing the aging temperature. All the KIT-6-X (X: aging temperature in °C) materials show three well-resolved XRD peaks in the region of 2h = 0.8–2.8° (Fig. S4), which correspond to (2 1 1), (2 2 0) and (3 2 1) reflections, based on a cubic system. The bicontinuous cubic symmetry is confirmed by combining the XRD pattern and Ia3d HRTEM images (not shown). For KIT-6-X, higher hydrothermal aging temperature gives mesoporous silicas with larger pores and improved structural integrity. The resultant products show increasing d-spacing with the (2 1 1) reflection peak shifting to a smaller angle, when the hydrothermal treatment temperature gets higher. The N2 adsorption–desorption isotherms of the materials (Fig. S5) are type IV with a sharp capillary condensation step at high relative pressures and a H1 type hysteresis loop, indicative of large channel-like pores in a narrow range of size. The pore size distributions (Fig. S5) illustrate the pore size changes from 4.2 to 7.5 nm when the aging temperature rises from 40 to 100 °C. Table 3 summarizes the structural properties of KIT-6-X. From this table it can be seen that all the materials have good mesoporosity. As the aging temperature gets higher, the pore diameter increases, while the wall thickness decreases from 4.2 to 3.4 nm. Fig. 4 shows the small-angle XRD patterns of the mesoporous Co3O4 synthesized by using mesoporous silica KIT-6-X with Ia3d structure and different pore sizes as hard templates. The obtained mesoporous cobalt oxides clearly show well-resolved (2 1 1) reflections, showing that the pore symmetry is successfully replicated. The intensities of the reflection peaks of mesoporous Co3O4 increase with the pore size of silica template. And as the pore size of silica template rises, fine structural details become more evident and (2 2 0) reflection can be well observed in the XRD pattern of Co3O4-KIT-6-100. It indicates that large-pore KIT6 facilitate the replication of mesostructure by metal oxide. Notably, from the large-angle XRD pattern (Fig. 5), it can be seen that the Co3O4 products are well-crystalline and they possess Fd3m structure.
140
100 80
Co3O4-SBA-15-100
dV/dD
Adsorbed amount(cm3g-1 STP)
120
60 40 20
Co3O4-SBA-15-80
0
0.0
0.2
0.4
0.6
0.8
Relative pressure P/P0
1.0
2
4
6
8
10
Pore size (nm)
Fig. 2. N2 adsorption–desorption isotherms and pore size distributions of mesoporous Co3O4 replicas templated by SBA-15-X. The isotherm of Co3O4-SBA-15-100 was moved vertically by 20 cm3 g 1.
318
P. Shu et al. / Microporous and Mesoporous Materials 123 (2009) 314–323
Fig. 3. TEM images of mesoporous Co3O4 replicas prepared by (a) SBA-15-40 (pre size: 6.6 nm) and (b) SBA-15-100 (pore size: 10.7 nm) taken along the channel direction. The corresponding HRTEM images are given on the right.
Table 3 Structural properties of the mesoporous silica KIT-6-X (X: aging temperature in °C) with bicontinuous cubic Ia-3d symmetry. Sample
Unit cell parameter (nm)
S (m2 g
KIT6-40 KIT6-80 KIT6-100
19.5 21.4 22.1
630 569 666
1
)
Intensity (a.u.)
211
Co3O4-KIT-6-100
Co3O4-KIT-6-80
Co3O4-KIT-6-40
1
2
3
4
2 theta (degree) Fig. 4. XRD patterns of mesoporous Co3O4 templated by KIT-6-X.
Fig. 6 shows the N2 adsorption–desorption isotherms of silicafree mesoporous Co3O4 replicas (Co3O4-KIT-6-X). The structural
V (cm3 g 0.56 0.71 0.96
1
)
Pore diameter (nm)
Wall thickness (nm)
4.2 6.6 7.5
4.2 3.6 3.4
properties are summarized in Table 4. The isotherms are of type IV classification, which have typical hysteresis loops of mesoporous materials. With KIT-6-100 as template, the silica-free mesoporous Co3O4 sample has a surface area of 64 m2 g 1 and a pore volume of 0.14 cm3 g 1. It is noticeable that its isotherm is different from that of template KIT-6-100, giving pore size distribution centered at 3.5 nm, reflecting exactly the wall thickness of KIT-6100 (3.4 nm). This value is in good agreement with the pore diameters anticipated for a replica structure of KIT-6, indicating that the two sets of mesopores are successfully fabricated. As for the mesoporous Co3O4 templated by KIT-6-40 and 80, the pore sizes are not equal to the wall thickness of silica template. The pore size distribution curves centered at 10.7 and 7.5 nm for Co3O4-KIT-6-40 and Co3O4-KIT-6-80, respectively, much larger than the wall thickness of the corresponding silica template, 4.2 nm for KIT-6-40 and 3.6 nm for KIT-6-80. This might come from partial replication of only one set of the bicontinuous meso-channels. And this conclusion is further proved by TEM observations. Fig. 7a–c shows TEM images of mesoporous Co3O4 replicas synthesized using KIT-6 with different pore sizes (4.2, 6.6 and 7.5 nm, respectively) as the hard template, taken along the [1 0 0], [1 1 0] and [1 1 1] directions. It is clear from TEM images that all the Co3O4 replicas maintain a 3d structure and the channels are arranged in a well-ordered bicontinuous cubic structure. This
319
P. Shu et al. / Microporous and Mesoporous Materials 123 (2009) 314–323
20
30
40
50
Sample
S (m2 g
Co3O4-KIT-6-40 Co3O4-KIT-6-80 Co3O4-KIT-6-100
78 170 64
1
V (cm3 g
)
0.23 0.40 0.14
1
)
Pore diameter (nm) 10.7 7.6 3.5
440
511 422
222
400
220 111
Intensity(a.u.)
311
Table 4 Structural properties of the mesoporous Co3O4 templated by KIT-6-X.
60
70
2theta (degree) Fig. 5. Large-angle XRD pattern of mesoporous replica Co3O4-KIT-6-100.
indicates that when the cobalt nitrate was introduced into KIT-6 and reduced, the Co3O4 replica was formed inside the channels of KIT-6 without destroying the channel geometry. It is well known space group symmetry and possesses two nonthat KIT-6 has Ia3d intersecting channel networks separated by the silica wall which follows Gyroid minimal surface. From Fig. 7a (Co3O4-KIT-6-40) and 7b (Co3O4-KIT-6-80), it can be clearly seen that only one set of the nonintersecting channel network is replicated at the edge of the Co3O4 replica crystal, however, in the thick areas both nonintersecting channel networks are replicated and the symmetry is slightly destroyed due to the shift of the two networks. It also explains the difference of pore size distributions between replicas and the corresponding silica templates. Fig. 7c (Co3O4-KIT-6-100) clearly shows that two nonintersecting channel networks are completely replicated and the Co3O4 replica keeps the same space group symmetry as KIT-6. From Fig. 7a to c, it can be concluded that the pore size of KIT-6 template determines the extent of the replication, and two sets of channel networks can be completely replicated when the pore size is above or equal to 7.5 nm. Furthermore, when the pore size of KIT-6 template is small (66.6 nm),
Co3O4 can only replicate one set of pores at the edge of the particles and the area of one set replication is larger when the template pore size is smaller. Results of TEM observations are consistent with and strongly support the N2 adsorption data. structure facilitates the In summary, large-pore KIT-6 with Ia3d replication of both sets of the bicontinuous networks, and partial replication of one set network occurs when using small-pore KIT6 as the hard template. The disordered micropores connecting adjacent mesopores is among the reasons, similar to the case of replication of SBA-15. In addition, another possibility also answers structure for this fact. It is known that when the mesopore of Ia3d gets larger, ordered complementary micropores are present on the thinnest part of the pore walls, which connect the two independent mesopore networks [36]. Through the disordered micropores and ordered complementary micropores, the cobalt precursor diffuses between the two disconnected bicontinuous networks and there networks is achieved. fore replication of both sets of Ia3d 3.3. Mesoporous Co3O4 synthesized by AMS-10 with bicontinuous structure as hard template Pn3m In our previous work, we have found that both the mesostructure and pore size strongly depend on the ionization degree of the two-headed anionic surfactant [38]. The mesoporous materials synthesized with the surfactant neutralized with low HCl amount having larger effective head group areas often show smaller pore sizes, regardless of the type of CSDA. It has been proposed that the hydrophobic ‘‘tail” of the surfactant with larger effective head group area would ‘‘coil” to fill the space and thus maximize their van der Waals interactions, leading to lowering of the overall
300
200
dV/dD
Adsorbed amount (cm3g-1 STP)
400
Co3O4-KIT-6-80
100 Co3O4-KIT-6-40
Co3O4-KIT-6-100 0 0.0
0.2
0.4
0.6
0.8
Relative pressure P/P0
1.0
2
4
6
8
10
12
14
16
18
Pore size (nm)
Fig. 6. N2 adsorption–desorption isotherms and pore size distributions of the mesoporous Co3O4 replicas templated by KIT-6-X. The isotherms of Co3O4-KIT-6-40 and 80 were moved vertically by 50, 100 cm3 g 1, respectively.
320
P. Shu et al. / Microporous and Mesoporous Materials 123 (2009) 314–323
Fig. 7. TEM images of mesoporous Co3O4 replicas prepared by (a) KIT-6-40 (pore size: 4.2 nm), (b) KIT-6-80 (pre size: 6.6 nm) and (c) large-pore KIT-6-100 (pore size: 7.5 nm) taken along [1 0 0], [1 1 0] and [1 1 1] directions, respectively.
energy. Lower surfactant packing density leads to more tightly coiled surfactant hydrocarbon tails and smaller pore size. It has been found that the unit cell parameter and pore size are significantly different even between the same mesostructures. Mesoporous silicas AMS-10 with different pore sizes are synthesized by adding different amount of HCl into the C14GluAS solution. The synthesis molar composition is: C14GluAS/TMAPS/TEOS/ H2O/HCl = 1:1.5:15:2000:Y. Two X-ray reflection peaks of each samples can be distinguished with a reciprocal spacing ratio of about 1:1.225 and can be indexed to (1 1 0) and (1 1 1) reflections of bicontinuous cubic Pn3m mesophase (Fig. S6). The HRTEM images (not shown) confirm that all of the space groups of these The resultant mesoporous materials are bicontinuous cubic Pn3m. products show increasing d-spacing with the reflection peak [1 1 0] shifting to the smaller angle from 1.39° to 1.14° with increasing HCl/C14GluAS molar ratio from 0.42 to 0.48. The N2 adsorption–desorption isotherms of the samples AMS10-Y (Y = HCl/C14GluAS) (Fig. S7) show type IV features, typical for mesoporous materials, and three well-distinguished regions of the adsorption isotherm are observed including monolayer– multilayer adsorption, capillary condensation, and multilayer adsorption on the outer surface. All the samples clearly show the hysteresis loop of H1 type and the capillary condensation occurs at increasing relative pressure with increasing HCl/C14GluAS molar ratio, indicating increasing size of cylindrical mesopores. Interestingly, the pore size distribution curves, BJH plots of derivative of the pore volume per unit weight with respect to the pore diameter (dV/dD) versus the pore diameter (D), show a clear shift of pore
diameter from 5.1 to 6.7 nm, as the molar ratio of HCl/C14GluAS changes from 0.42 to 0.48 (Fig. S7). The porous properties of AMS-10-Y are summarized in Table 5. All the samples show large surface areas (larger than 600 m2 g 1) and pore volumes (larger than 0.8 cm3 g 1), indicating high mesoporosity of the samples. As the HCl/C14GluAS molar ratio increases, the pore size increases significantly. However, the wall thickness is almost the same around 1.2 nm for all the samples. Fig. 8 shows the XRD patterns of Co3O4 mesoporous replicas synthesized with AMS-10-Y as the template. After Co3O4 is embedded in the mesopores and silica is removed, the XRD patterns show no significant peaks within the range of 0.8–6.0° (2h). Only one reflection peak with low intensity is observed from each pattern, which can be reasonably assigned to the (1 1 0) reflection, showing the d-spacing close to that of the silica template. These peaks can be explained as the X-ray reflections from the mesoporous Co3O4 that is obtained by replication of two sets of the bicontinuous mesopore networks, which shows the same unit cell dimensions as the silica template. The low intensity of the X-ray reflection indicates that replication of two sets of bicontinuous mesopore network by Co3O4 occurs partially. Replication of only one set of the mesochannels may also occur, which doubles the unit cell dimensions Ref. [32]) and therefore the X-ray reflec(structural model of Pn3m, tions shift to a much smaller angle (2h). Fig. 9 shows the N2 adsorption–desorption isotherms and pore size distributions of silica-free Co3O4 mesoporous replicas templated by AMS-10-Y. The porous properties are shown in Table 6. The isotherms are of type IV classification, which have typical
321
P. Shu et al. / Microporous and Mesoporous Materials 123 (2009) 314–323 Table 5 Structural properties of mesoporous silica AMS-10-Y (Y: molar ratio of HCl/C14GluNa) with bicontinuous cubic Pn-3m symmetry. Unit cell parameter (nm)
S (m2 g
AMS-10-0.42 AMS-10-0.44 AMS-10-0.45 AMS-10-0.48
9.0 9.5 9.8 11.0
649 624 640 640
Intensity (a.u.)
Sample
1
V (cm3 g
)
Co3O4-AMS-10-0.44 Co3O4-AMS-10-0.42
3
4
5
Pore diameter (nm)
Wall thickness (nm)
5.1 5.5 5.8 6.7
1.3 1.2 1.1 1.1
the wall thickness of silica template, which might be due to the replication of only one set of the bicontinuous meso-channels. The large mesopore of Co3O4 is close to the expected value, twice of the wall thickness plus pore diameter of silica template, on condition that only one set of meso-channels is replicated. Therefore, it can be concluded that, within the full range of pore size of AMS10, Co3O4 replicates both one set and two sets of the bicontinuous meso-channels. It could be further proved by TEM observations. Fig. 10a and b shows TEM images of mesoporous Co3O4 replicas synthesized from AMS-10 with different pore sizes (5.1 and 6.7 nm, respectively) taken along [1 1 0] and [1 1 1] directions. It can be clearly seen that all the Co3O4 replicas have a 3d structure and the channels show a well-ordered bicontinuous cubic structure. This implies that the Co3O4 replica is formed inside the channels of AMS-10 like in the channel of KIT-6 described above. Similar to KIT-6, AMS-10 has two nonintersecting channel networks as well. However, those two channel networks are separated by the silica wall which follows Diamond minimal surface and space group symmetry. From Fig. 10a AMS-10 possesses Pn3m (Co3O4-AMS-10-0.42) and 10b (Co3O4-AMS-10-0.48), it is clear that only one set of the nonintersecting channel network was replicated in the edge of the Co3O4 replica crystal, but in the thick areas both nonintersecting channel networks were replicated and further the contrast between two channel networks is different from each other which results in the break of the symmetry. The presence of the replication of one set of channel network illustrates the difference of pore size distributions between both replicas and the corresponding templates. Therefore, it can be summarized that mesoporous Co3O4 prepared by using AMS-10 with pore size of 5.1–6.7 nm is obtained by replication of both one set and two sets of bicontinuous meso-channel networks. AMS-10 is synthesized by using anionic surfactant as the template, and therefore no micropore is present
Co3O4-AMS-10-0.45
2
)
0.87 0.90 0.94 1.04
Co3O4-AMS-10-0.48
1
1
6
2theta (degree) Fig. 8. XRD patterns of mesoporous Co3O4 templated by AMS-10-Y.
hysteresis loops of mesoporous materials. The pore size distributions show similar feature that the materials have bimodal porous structure, with two maxima of dV/dD within the pore size regions of 2–5 nm and 5–15 nm. Considering that the pore size and wall thickness of silica template are 5.1–6.7 nm and 1.2 nm, respectively, it is reasonable that the pore size distribution of mesoporous Co3O4 within 2–5 nm is attributed to the replication of two sets of the bicontinuous mesopores of AMS-10. One the other hand, the pore size of mesoporous Co3O4 within 5–12 nm is much larger than
400 Co3O4-AMS-10-0.48
300 Co3O4-AMS-10-0.45
200
dV/dD
Adsorbed amount (cm3g-1, STP)
500
Co3O4-AMS-10-0.44
100 Co3O4-AMS-10-0.42
0 0.0
0.2
0.4
0.6
2 theta (degree)
0.8
1.0
2
4
6
8 10 12 14 16 18 20
Pore size (nm)
Fig. 9. N2 adsorption–desorption isotherms and pore size distributions of the mesoporous Co3O4 replicas templated by AMS-10-Y. The isotherms of Co3O4-AMS-10-0.44, 0.45 and 0.48 were moved vertically by 100, 200 and 300 cm3 g 1, respectively.
322
P. Shu et al. / Microporous and Mesoporous Materials 123 (2009) 314–323
Table 6 Structural properties of the mesoporous Co3O4 templated by AMS-10-Y. Sample
S (m2 g
Co3O4-AMS-10-0.42 Co3O4-AMS-10-0.44 Co3O4-AMS-10-0.45 Co3O4-AMS-10-0.48
134 119 151 141
1
)
V (cm3 g 0.23 0.27 0.34 0.26
1
)
Pore diameter (nm) 1.9, 3.1, 3.6, 2.0,
7.6 8.8 10.6 6.6
in the silica wall growing following the Diamond minimal surface. A reasonable explanation is the existence of ordered complementary micropores at the thinnest part of the silica wall, similar to the case of KIT-6 [37]. These ordered micropores in the silica wall facilitate the diffusion of cobalt precursor between two sets of bicontinuous meso-channels. As a result, in the thick part of the particles, a complete replication of two sets of meso-channels is achieved. Metal oxide inverse replication provides additional structural information of the silica template AMS-10. The bicontinuous networks are visualized by this means. And as discussed above, the materials have thin silica wall (1.2 nm) and at the thinnest parts there are ordered complementary micropores which connect the two independent bicontinuous meso-channels. 3.4. Comparison of pore size effect in the inverse replication As has been discussed in details above, it is clear that the pore size of template silica will greatly influence the structure of inverse Co3O4 replica. However, this discussion of the effect of pore size is valid when taking the pore symmetry into consideration. SBA-15 (p6mm) normally have large-pore size and small wall thickness. When the pore size is 6.6 nm or smaller, only separated
nano-wires of Co3O4 replicas are obtained, due to a lack of micropore connections between adjacent mesopores. In the case of KIT-6 (Ia3d), when the pore size is 6.6 nm or smaller, both replication of one set and two sets of bicontinuous mesopores are replicated, indicating mesopore connections by micropores in the wall. It is symmetry worth noting that the wall thickness of KIT-6 with Ia3d (>3.6 nm) is much larger than that of SBA-15 with p6mm symmetry (2.4 nm), which may have less micropore connections without strucconsidering the pore symmetry. However, because the Ia3d ture has an inhomogeneous wall thickness, some part of the silica wall is fairly thin though the mean wall thickness is high, which may have micropore connections. Besides, ordered complementary structure. Theremicropore connections might also exist in Ia3d fore, different pore symmetry differs much in inverse replication. AMS-10 (Pn3m) materials have similar pore size to KIT-6 (Ia3d), but they have a large porosity. The change of pore size does not change the wall thickness significantly, and all of materials possess ordered micropore connections between mesopores. Therefore, the effect of pore size is ignoring, and all the replication process includes replication of one set and two sets of the bicontinuous mesopore networks. 4. Conclusion Pn3m Mesoporous silica with different pore size of p6mm, Ia3d, symmetries are employed as hard templates to synthesize mesoporous Co3O4 replica. It is found that for p6mm, large-pores facili and Pn3m, tate the replication and that for Ia3d the pore size determines the extent of the replication. For p6mm, at small-pore sizes, the structure of Co3O4 replica collapses and isolated nanowires have been obtained; At large-pore sizes, complete replication can be achieved, retaining the original symmetry of silica template.
Fig. 10. TEM images of mesoporous Co3O4 replicas prepared by (a) small-pore AMS-10-0.42 (pore size: 5.1 nm) and (b) large-pore AMS-10-0.48 (pore size: 6.7 nm) taken along [1 1 0] and [1 1 1] directions.
P. Shu et al. / Microporous and Mesoporous Materials 123 (2009) 314–323
structure, at large-pore sizes two sets of the mesopores are For Ia3d replicated, and at small-pore sizes, both two sets and one set of the bicontinuous mesopore networks are replicated by Co3O4. For structure, unlike Ia3d, within the pore size range of 5.1– Pn3m 6.7 nm, both one set and two sets of the bicontinuous meso-channel networks are replicated. The discussion of pore size effect on replication is valid only when pore symmetry of silica template is specified. In the inverse replication, disordered micropores in the walls of SBA-15 and KIT-6 play an important role, and ordered complementary micropores in the walls of KIT-6 and AMS-10 are the vital factors in determining the extent of replication. Inverse replication also provides structural information of silica template. It is proposed that for AMS-10, there are ordered complementary micropores at the thinnest parts of the wall, which connect the two independent bicontinuous meso-channels.
[7] [8] [9] [10] [11] [12]
Acknowledgments
[24]
This work was supported by the National Natural Science Foundation of China (Grant Nos. 20890121 and 20521140450) and 973 project (2009CB930403) of China. Appendix A. Supplementary data
[13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23]
[25] [26] [27] [28] [29] [30]
Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.micromeso.2009.04.017. References [1] P. Behrens, Angew. Chem. Int. Ed. 35 (1996) 515–518. [2] X. He, D. Antonelli, Angew. Chem. Int. Ed. 41 (2002) 214–229. [3] E. Rosencher, A. Fiore, B. Vinter, V. Berger, P. Bios, J. Nagle, Science 271 (1996) 168–173. [4] M. Tiemann, Chem. Eur. J. 13 (2007) 8376–8388. [5] (a) F. Schüth, Chem. Mater. 13 (2001) 3184–3195; (b) G.J.de A.A. Soler-II-lia, C. Sanchez, B. Lebeau, J. Patarin, Chem. Rev. 102 (2002) 4093–4138. [6] P. Kluson, P. Kacer, T. Cajthaml, M. Kalaji, J. Mater. Chem. 11 (2001) 644–651.
[31] [32] [33] [34] [35] [36] [37] [38]
323
R. Ryoo, S.H. Joo, S.J. Jun, Phys. Chem. B 103 (1999) 7743–7746. F. Schüth, Angew. Chem. Int. Ed. 42 (2003) 3604–3622. H. Yang, D. Zhao, J. Mater. Chem. 15 (2005) 1217–1231. M. Tiemann, Chem. Mater. 20 (2008) 961–971. B. Ohtani, Y. Ogawa, S. Nishimoto, J. Phys. Chem. 101 (1997) 3746–3752. S. Jun, S.H. Joo, R. Ryoo, M. Kruk, M. Jaroniec, Z. Liu, T. Ohsuna, O. Terasaki, J. Am. Chem. Soc. 122 (2000) 10712–10713. T.-W. Kim, I.-S. Park, R. Ryoo, Angew. Chem. 115 (2003) 4511–4515. S.H. Joo, S.J. Choi, I. Oh, J. Kwak, Z. Liu, O. Terasaki, R. Ryoo, Nature 412 (2001) 169–172. H.J. Shin, R. Ryoo, J. Mater. Chem. 11 (2001) 260–261. H.J. Shin, R. Ryoo, Z. Liu, O. Terasaki, J. Am. Chem. Soc. 123 (2001) 1246–1247. M.J. Maclachlan, G.A. Ozin, J. Am. Chem. Soc. 121 (1999) 12005–12017. Y. Shi, Y. Wan, R. Zhang, D. Zhao, Adv. Funct. Mater. 18 (2008) 2436–2443. S.C. Laha, R. Ryoo, Chem. Commun. 17 (2003) 2138–2139. Y.Q. Wang, C.M. Yang, W. Schmidt, B. Spliethoff, E. Bill, F. Schüth, Adv. Mater. 17 (2005) 53–56. J. Roggenbuck, M. Tiemann, J. Am. Chem. Soc. 127 (2005) 1096–1097. C. Dickinson, W. Zhou, R. Hodgkins, Y. Shi, D. Zhao, H. He, Chem. Mater. 18 (2006) 3088–3095. B. Tian, X. Liu, H. Yang, S. Xie, C. Yu, B. Tu, D. Zhao, Adv. Mater. 15 (2003) 1370– 1374. H. Yang, Q. Shi, B. Tian, Q. Lu, F. Gao, S. Xie, J. Fan, C. Yu, B. Tu, D. Zhao, J. Am. Chem. Soc. 125 (2003) 4724–4725. W. Yue, W. Zhou, Chem. Mater. 19 (2007) 2359–2363. K. Zhu, B. Yue, W. Zhou, H. He, Chem. Commun. (2003) 98–99. A. Rumplecker, F. Kleitz, E.-L. Salabas, Ferdi Schüth, Chem. Mater. 19 (2007) 485–496. H. Tüysüz, C.W. Lehmann, H. Bongard, B. Tesche, R. Schmidt, F. Schüth, J. Am. Chem. Soc. 130 (2008) 11510–11511. B. Yue, H. Tang, Z. Kong, K. Zhu, C. Dickinson, W. Zhou, H. He, Chem. Phys. Lett. 407 (2005) 83–86. D. Zhao, J. Feng, Q. Huo, N. Melosh, G.H. Fredrickson, B.F. Chmelka, G.D. Stucky, Science 279 (1998) 548–552. F. Kleitz, S.H. Choi, R. Ryoo, Chem. Commun. 17 (2003) 2136–2137. C. Gao, Y. Sakamoto, K. Sakamoto, O. Terasaki, S. Che, Angew. Chem. Int. Ed. 118 (2005) 4401–4404. P.I. Ravikovitch, A.V. Neimark, Langmuir 16 (2000) 2419–2423. K. Schumacher, P.I. Ravikovitch, A. Du Chesne, A.V. Neimark, K.K. Unger, Langmuir 16 (2000) 4648–4654. K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquerol, T. Siemieniewska, Pure Appl. Chem. 57 (1985) 603–619. A. Galarneau, H. Cambon, F. Di Renzo, R. Ryoo, M. Choi, F. Fajula, New J. Chem. 27 (2003) 73–79. Y. Sakamoto, T-W. Kim, R. Ryoo, O. Terasaki, Angew. Chem. Int. Ed. 43 (2004) 5231–5234. C. Gao, H. Qiu, W. Zeng, Y. Sakamoto, O. Terasaki, K. Sakamoto, Q. Chen, S. Che, Chem. Mater. 18 (2006) 3904–3914.