Template synthesis of large pore ordered mesoporous carbon

Template synthesis of large pore ordered mesoporous carbon

Microporous and Mesoporous Materials 80 (2005) 117–128 www.elsevier.com/locate/micromeso Template synthesis of large pore ordered mesoporous carbon A...

462KB Sizes 0 Downloads 19 Views

Microporous and Mesoporous Materials 80 (2005) 117–128 www.elsevier.com/locate/micromeso

Template synthesis of large pore ordered mesoporous carbon An-Hui Lu, Wen-Cui Li, Wolfgang Schmidt, Ferdi Schu¨th

*

Max-Planck-Institut fu¨r Kohlenforschung, Kaiser-Wilhelm-Platz 1, D-45470 Mu¨lheim, Germany Received 21 September 2004; received in revised form 6 December 2004; accepted 7 December 2004 Available online 20 January 2005

Abstract Nanocast carbon (NCC-1) with large pores and ordered structure was synthesized via a nanocasting process using aluminumcontaining SBA-15 as template and furfuryl alcohol (FA) as carbon precursor. This carbon has several interesting features, such as two steps with distinguished hystereses in the nitrogen sorption isotherm, high surface area of 2000 m2/g and large pore volumes of 3.0 cm3/g. It was found that the key factors in the synthesis of such carbons are the aging temperature of the SBA-15 template, the concentration of furfuryl alcohol (dissolved in trimethylbenzene), and the carbonization temperature. The optimal conditions for materials with high surface area and pore volumes are SBA-15 starting materials aged at 140 °C, 25 vol% of FA solution and 850– 1100 °C carbonization temperatures. Moreover, it has been demonstrated that such nanocast carbon can be synthesized in a more facile way than previously reported. Purely siliceous SBA-15 without the need of Al3+-incorporation can be directly used as template. In this case, the polymerization catalyst—oxalic acid and FA were simultaneously introduced into the pore space of SBA-15. Ó 2004 Elsevier Inc. All rights reserved. Keywords: Mesoporous carbon; Template; Bimodal; Adsorption

1. Introduction Porous carbon materials are very important in various fields of science and technology, including catalysis, purification, separation and energy storage [1–4]. Many synthetic strategies and a wide range of precursors were employed to prepare porous carbon materials with remarkable properties, such as a wide range of surface areas up to around 3000 m2/g and pore sizes ranging ˚ ngstro¨ms to hundreds of nanometers. In from several A the past decades, numerous studies have been carried out with respect to the control of the surface area and pore structure of porous carbons. These carbons are essentially still disordered on the mesoscopic scale, although on the atomic scale graphitic structures may be present [1]. Porous carbon prepared by using zeolite *

Corresponding author. Tel.: +49 208 306 2367; fax: +49 208 306 2995. E-mail address: [email protected] (F. Schu¨th). 1387-1811/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.micromeso.2004.12.007

Y as template shows a similar morphology as the template, although on the nanometer scale the structure was much less ordered than the zeolite [5]. In 1999, Ryoo et al. have demonstrated that mesoporous carbon with ordered structure can be synthesized through a nanocasting pathway using MCM-48 as hard template [6]. From then on, the door was opened for this concept in the synthesis of other structures and compositions of ordered porous carbon. Many scientists have contributed to this field, and thus great progress has been made in the synthesis of porous carbons with different structures [7–14]. The preparation of porous carbons with structural regularity on the meso- and/or macroscale is still a great challenge. Although it has been attempted to synthesize ordered porous carbon with structural regularity on the mesoscale using surfactants as structural-directing agents, which has been shown to work well for the synthesis of ordered porous silica, it seems, at present, very difficult to exert the same extent of control in the case of carbon. This is due to the complicated structural

118

A.-H. Lu et al. / Microporous and Mesoporous Materials 80 (2005) 117–128

evolution from polymer to carbon; the structure formed around the micelle easily collapses when it is exposed to the thermal treatment to induce carbonization [15]. Thus, the best method for the synthesis of mesoscopically ordered carbon materials is the replication from the hard template silica with well-developed interconnectivity [7]. For this replication, the pore system of the template is infiltrated with a carbon precursor which is then carbonized at high temperature. Ordered carbon can be recovered after template removal with either acid or base leaching. The hard template thus acts as a mold, in which the carbon precursor is confined during the carbonization to form the rigid carbonaceous structure. Generally, the pore size of a carbon replica is determined by the wall thickness of the parent silica [7]. Therefore, in order to vary the pore diameter of the carbon, one straightforward way is the synthesis of silicas with different wall thickness. An attempt has been made to use a mixture of binary surfactants to vary the wall thickness of silica. Using these silicas as templates, the pore sizes of the carbon replica can be tuned in the range of 2.2–3.3 nm [16]. As an alternative way, one can use large pore SBA-15 in combination with the surface-templating strategy to create mesoporous carbon (CMK-5 type). The inner diameter of the hexagonally arranged carbon tubes can reach around 5 nm [13,17–19]. The synthesis of ordered porous carbon with wider pores is a great challenge, although the availability of such materials would be very interesting both for fundamental research and practical applications. Recently, we have communicated a synthesis of ordered mesoporous carbon with well-developed bimodal pore system by using SBA-15 as template and furfuryl alcohol (FA) as carbon source via a nanocasting pathway [9]. The diameter of the large pore exceeds 10 nm. Essentially, this carbon can be considered to be of the CMK-5 type, but since it has some special features, it was denoted as NCC-1 (nanocast carbon) [9]. The previous investigation confirmed that the most critical parameter to form NCC-1 type carbon is the concentration of the furfuryl alcohol solution. The objective of this work is a more comprehensive understanding of the role of the different synthesis parameters. We systematically investigated the effect of the template structure, FA concentration, carbonization temperature, heating ramp and polymerization catalyst. This will give a deeper insight in the formation of such nanocast carbon materials and allow more precise tuning of the properties of these materials.

2. Experimental 2.1. Synthesis of the SBA-15 template The synthesis of SBA-15 is described in detail elsewhere [20]. Typically, the starting composition was 8.5

g tetraethoxysilane (TEOS, Aldrich): 4 g Pluronic P123 (BASF): 105 ml H2O: 20 ml HCl (37% aqueous solution, Fluka). After the reaction at 40 °C for 4 h, the white milky solution was transferred into an autoclave and aged for 3 days at 140 °C, unless different aging temperatures are given in the text. The solid was filtered off without washing, dried at 80 °C, and calcined in air at 550 °C for 6 h to obtain SBA-15. Afterwards, the SBA-15 was impregnated with AlCl3 in ethanol, followed by a calcination step at 550 °C for 5 h in air to produce alumosilicate SBA-15 (AlSBA-15) with a Si/Al ratio of 100. The introduction of Al3+ can provide acidic sites to catalyze the polymerization of the carbon source [9]. 2.2. Synthesis of NCC-1 The pore space of AlSBA-15 was filled with furfuryl alcohol (Fluka) dissolved in trimethylbenzene (TMB, Aldrich) solution by incipient wetness impregnation. The concentration of FA in TMB was varied from 25– 100 vol%. AlSBA-15 containing furfuryl alcohol was heated to 80 °C under vacuum for 12 h and then to 150 °C for 6 h to induce the polymerization of the furfuryl alcohol. Subsequently, the composite was heated under argon to 300 °C with a heating ramp of 1 °C/min, then the temperature was increased to the desired final temperatures with a heating ramp of 5 °C/min and maintained at this temperature for 4 h. After removing the silica template by leaching with aqueous HF solution, the carbon product was recovered after filtration, washing with water and ethanol, and drying at 90 °C. The resulting mesoporous carbons are denoted as NCC-1 series. 2.3. Characterization Low angle X-ray diffraction patterns were recorded with a Stoe STADI P diffractometer in Bragg–Brentano (reflection) geometry. The step width was 0.02° at an acquisition time of 8 s per step. Nitrogen adsorption isotherms were recorded with an ASAP2010 adsorption analyzer (Micromeritics) at liquid nitrogen temperature. Prior to the measurements, the samples were degassed at a temperature of 250 °C for 6 h. The BET surface area was calculated from the adsorption data in the relative pressure interval from 0.04–0.2. Pore sizes and pore size distribution curves were calculated by the BJH method from the desorption branch (from the adsorption branch, if the isotherm showed indications of network percolation effects). The total pore volume was estimated from the amount adsorbed at a relative pressure of 0.99. TEM images were obtained with a Hitachi HF2000 microscope equipped with a cold field emission gun. The acceleration voltage was 200 kV. Samples were prepared dry on a lacey carbon grid. EDX (Energy Dis-

A.-H. Lu et al. / Microporous and Mesoporous Materials 80 (2005) 117–128

119

persive X-ray Analysis) showed that the silicon content in the mesostructured carbons is below 1 wt%.

3. Results and discussion 3.1. Effect of SBA-15 template structure The pore structures of carbon replicas are usually determined by the structural parameters of the silica templates, for instance, pore wall thickness, pore size, and structural ordering [7]. We therefore primarily investigated the template effect on the final structure of the nanocast carbon. It is well known that hydrothermal treatment leads to SBA-15 with enlarged pore size [21– 24]. To vary the pore size of SBA-15, the milky mixture was hydrothermally aged at 110 °C, 130 °C and 140 °C, respectively. Since the decomposition temperature of the surfactant Pluronic P123 is ca. 145 °C [20], a higher aging temperature was not considered. The obtained SBA-15 was denoted as SBA-15-x, where x corresponds to the aging temperature. In the selected temperature range, the obtained SBA-15 still maintains the structural ordering, as shown in Fig. 1a. All the XRD patterns show well-resolved reflections, which can readily be assigned to (10), (11) and (20) reflections, indicating a 2-D hexagonal symmetry. The calculated unit cell parameter ÔaÕ almost remains constant; it only slightly changes from 11.5 nm (110 °C) to 11.7 nm (140 °C) with temperature. The aging effect on the structure of SBA-15 can easily be detected by nitrogen adsorption. As shown in Fig. 1b, the SBA-15 samples show type IV isotherms, with the hysteresis loop shifting to higher partial pressure, indicating an increase in pore size as the aging temperature is increased from 110 °C to 140 °C. In addition, the capillary condensation step becomes less steep, which indicates some broadening of the pore size distribution. As shown in Table 1, the calculated BET surface area and micropore volume decrease with increasing aging temperature. The total pore volume remains almost identical. The decrease of the micropore volume and surface area implies that the micropores in the silica wall are lost at higher aging temperature. Under such conditions, they are expanded to mesoporous tunnels (mesotunnels) resulting in an improved interconnectivity

Fig. 1. XRD patterns of (a) SBA-15 aged at various temperatures and (b) the corresponding nitrogen sorption isotherms. Sample AlSBA-15140 was obtained from SBA-15-140 by introduction of Al3+. The isotherms of SBA-15-140, SBA-15-130 and SBA-15-110 were offset vertically by 300, 600 and 900 cm3 g 1 STP, respectively.

between the adjacent mesoporous channels [21]. After introduction of Al3+ into the SBA-15 channel system, the obtained AlSBA-15 series still exhibits well-distinguished hexagonal ordering. As an example, the XRD pattern of AlSBA-15-140 is shown in Fig. 1a, and it is quite similar to that of SBA-15. The isotherms of the AlSBA-15 series also maintain the profiles of the corresponding SBA-15 series, as shown in Fig. 1b. However, due to structural shrinkage during the second calcination step, the pore size of the AlSBA-15 series is slightly smaller than that of the corresponding SBA-15 (Table 1). TEM observations corroborate these conclusions.

Table 1 Texture parameters of SBA-15 silicas used as molds Sample

Aging temp (°C)

SBET (m2g 1)

Vmic (cm3 g 1)

Vtot (cm3 g 1)

DMax (nm)

a (nm)

SBA-15-140 AlSBA-15-140 SBA-15-130 SBA-15-110

140 140 130 110

370 350 417 610

0.008 0.010 0.010 0.016

1.13 1.04 1.11 1.13

11.2 10.8 8.6 7.3

11.7 11.3 11.5 11.5

SBET: apparent surface area calculated by BET method; Vmic: micropore volume calculated by t-plot method; Vtot: total pore volume at p/p0 = 0.99; DMax: pore sizes at maxima of PSDs (desorption branch); a: unit cell parameter.

120

A.-H. Lu et al. / Microporous and Mesoporous Materials 80 (2005) 117–128

ing, however, that there is a substantial uncertainty in determining the exact value. As communicated previously [9], using 25 vol% of FA solution as carbon precursor, nanocast carbon with extremely high surface area and large pore size was obtained. In order to assess the synthesis space of such materials, we used SBA-15 aged at different temperatures and pretreated with Al3+ as templates, under the same conditions to investigate the structural development of the obtained carbon. After infiltration of FA and following polymerization (80 °C), carbonization (850 °C) and removal of the silica, nanocast carbons (NCC-1) were obtained. Samples were denoted as NCC-1/1 (templated from AlSBA-15-110), NCC-1/2 (from AlSBA-15-130) and NCC1/5 (from AlSBA-15-140), respectively. Their structures were characterized by low-angle XRD and nitrogen sorption technique. As shown in Fig. 3a, the XRD patterns of NCC-1/1 and NCC-1/2 have no reflections at low angle. They are thus not well ordered. In contrast, the XRD pattern of NCC1-/5 shows well-resolved XRD reflections, indicating structural ordering with similar symmetry as the starting SBA-15. The nitro-

Fig. 2. TEM images of AlSBA-15-140.

As shown in Fig. 2, AlSBA-15-140 shows the typical structural features and morphology of SBA-15. In the view parallel to the channel direction, one can see that the pores are arranged in a hexagonal fashion, which is consistent with the XRD results. Many visible mesotunnels are formed in the silica walls due to the high aging temperature, which has been discussed in considerable detail in previous reports [13,21,24]. Zhao et al. [20] and Kruk [22] have shown that the silica wall thickness decreases with the increase in the aging temperature. If one uses the pore size calculated from the desorption branch of 10.8 nm as the reference value, the wall thickness of SBA-15-140 would be estimated to 0.5 nm, which is extremely small. In contrast, the estimated value from the TEM image is ca. 4 nm. Evidently the silica wall thickness calculated from the sorption isotherm does not reflect the true value. This is due to the formation of the mesotunnels in the silica walls, which increase the apparent pore size of SBA-15 synthesized at high temperature, if the pore size is calculated from the sorption isotherm. Thus, to avoid misinterpretations in understanding the formation of the pores of nanocast carbon, the silica wall thickness estimated from the TEM measurements will be used in the following, know-

Fig. 3. (a) XRD patterns of carbons (NCC-1/1, NCC-1/2 and NCC-1/ 5) nanocast from AlSBA-15 with 25 vol% of FA solution. (b) Nitrogen sorption isotherms of nanocast carbon (NCC-1/1, NCC-1/2 and NCC1/5). The isotherms of NCC-1/1 and NCC-1/5 were offset vertically by 50 and 300 cm3 g 1 STP, respectively.

A.-H. Lu et al. / Microporous and Mesoporous Materials 80 (2005) 117–128

121

gen sorption isotherms (Fig. 3b) of all three carbons prove the mesoporous character of the samples. However, NCC-1/1 and NCC-1/2 show no pronounced pore filling steps with hystereses as NCC-1/5 does with the two very clear hysteresis loops, which are characteristic of NCC-1 type carbon [9]. These results confirm that the use of aluminum-containing SBA-15 aged at 110 °C and 130 °C as templates does not result in the formation of NCC-1 type carbon. Under the conditions used here, aging temperatures of 140 °C are required to induce the formation of the NCC-1 structure. It is noteworthy to point out that these carbons show very high surface areas of 1800 m2/g and large pore volumes of around 3.0 cm3/g. 3.2. Carbonization temperature In a carbonization process, polymer molecules undergo a specific decomposition and rearrangement to form a certain carbon framework in the corresponding temperature range. The chemical composition and the structural evolution of carbon materials are highly dependent on the carbonization temperature. Therefore, we have investigated the influence of the carbonization temperature on the final structure of the nanocast carbons. AlSBA-15-140 was selected as silica template, and the carbonization temperatures were varied from 600 °C to 1100 °C. The other factors such as FA concentration (25 vol%), heating ramp, and holding time were fixed. Nanocast carbons from NCC-1/3 to NCC1/7 were obtained in this series of experiments. The mesostructures

Fig. 4. XRD patterns of NCC-1/3 to NCC-1/7, obtained from AlSBA15-140 as a template with 25 vol% of FA solution after carbonization at different temperatures.

of these samples were analyzed using low-angle XRD reflection. As seen in Fig. 4, sample NCC-1/3 carbonized at 600 °C shows no visible low angle reflection. When the carbonization temperature exceeds 750 °C the XRD patterns of NCC-1/4 to NCC-1/7 show wellresolved reflections corresponding to the hexagonal symmetry, which indicates the enhanced structural ordering of the nanocast carbons obtained at higher carbonization temperatures. The calculated unit cell parameters ÔaÕ (in Table 2) are in the range of

Table 2 Texture parameters of nanocast carbons and the corresponding carbon/SBA-15 composites Sample

CFA (vol%)

Temp (°C)

SBET (m2 g 1)

Vmic (cm3 g 1)

Vtot (cm3 g 1)

DMax (nm)

a (nm)

NCC-1/1 NCC-1/2 NCC-1/3 NCC-1/4 NCC-1/5 NCC-1/6 NCC-1/7 NCC-1/8 NCC-1/9 NCC-1/10 NCC-1/11 NCC-1/12 SiO2/NCC-1/12 NCC-1/13 SiO2/NCC-1/13 NCC-1/14 SiO2/NCC-1/14 NCC-1/15 CMK-3-1 NCC-1/16 CMK-3-2

25 25 25 25 25 25 25 100 50 40 25 100 100 25 25 25 25 30 100 30 100

850 850 600 750 850 950 1100 850 850 850 850 850 850 850 850 1100 1100 850 850 850 850

1811 1753 1234 1630 1797 1710 1642 904 1361 1449 1947 728 385 1795 360 1659 278 2013 839 1906 1639

0.09 0.16 0.05 – 0.02 – 0.01 0.15 0.06 0.03 – 0.10

1.60 1.25 1.08 2.27 3.10 2.95 3.00 0.76 1.57 2.04 3.77 0.76 0.24 2.82 0.61 3.25 0.53 2.49 0.81 2.38 1.38

– – – 3.7; 5.1; 5.0; 4.8; 6.0 5.8 5.7 4.8; 6.9

– – – 11.7 11.2 11.3 11.1 10.9 11.1 11.2 11.4 11.3

4.8 10.6 11.0 10.3

9.8

4.8; 8.1

10.8

4.9; 9.2

10.8

3.4; 6.2

10.8

3.3; 6.1

10.8

CFA: concentration of furfuryl alcohol in trimethylbenzene; SBET: apparent surface area calculated by BET method; Vmic: micropore volume calculated from as-plot method; Vtot: total pore volume at p/p0 = 0.99; DMax: pore sizes at maxima of PSDs; a: unit cell parameter.

122

A.-H. Lu et al. / Microporous and Mesoporous Materials 80 (2005) 117–128

11.1–11.7 nm, which is quite close to that of the AlSBA15-140 template. Surprisingly, the unit cell values (Table 2) of NCC-1/5 and NCC-1/7 show negligible changes compared to that of their parent AlBSA-15-140 template even though the carbonization was carried out at high temperatures between 850 and 1100 °C. This unusual phenomenon demonstrates the unique structural stability of such carbons against high temperature sintering and the stabilization of the silica mold by the supporting carbon filling the voids. These carbons were further investigated using nitrogen sorption. All the isotherms in Fig. 5a are of type IV

Fig. 5. (a) Nitrogen sorption isotherms of NCC-1/3 to NCC-1/7 carbons synthesized using AlSBA-15-140 as a template. The isotherms of NCC-1/5, NCC-1/6 and NCC-1/7 were offset vertically by 400, 1200 and 1800 cm3 g 1 STP, respectively. (b) Pore size distributions of these NCC-1 carbons.

and exhibit clear hysteresis loops. The micropore volumes of these samples listed in Table 2 were estimated by the as-plot method using Cabot BP 280 as Ref. [25]. From sample NCC-1/3 to NCC-1/7, the micropore volumes are less than 5% of the total pore volume and in some cases even not detectable, indicating that these carbons are essentially mesoporous. The development of these isotherms is highly related to the carbonization temperature. As seen in Fig. 5a, the isotherm of sample NCC-1/3 carbonized at 600 °C differs significantly from those of the other samples carbonized at higher temperatures, and shows only one hysteresis loop in the relative pressure range of 0.4–0.8. In contrast to this, all other samples clearly show two hysteresis loops with closing points around p/p0 = 0.4 and 0.65. When the carbonization temperature is higher than 850 °C, the isotherm shape does not change any more up to a temperature of 1100 °C, indicating a wide range of applicable synthesis conditions for such NCC-1 type carbons. Although sample NCC-1/4 (750 °C) is hexagonally ordered according to the XRD pattern (Fig. 4), the isotherm of NCC-1/4 shows some deviation compared to samples NCC-1/5 (850 °C) to NCC-1/7 (1100 °C). Summarizing this section, high quality nanocast carbon was obtained at carbonization temperatures exceeding 850 °C. The quality of the material with respect to order and textural properties is significantly lower at a carbonization temperature of 750 °C. Below this temperature, amorphous porous carbon was obtained after removal of the silica template. This is due to the incomplete carbonization of the FA, which therefore cannot form a sufficiently rigid skeleton to resist the capillary forces during a drying stage. Interestingly, the NCC-1 type carbons show very high nitrogen uptake at a relative pressure close to unity, and thus have very high pore volumes of up to 3.0 cm3/g and high surface areas of 1600–1800 m2/g (see Table 2 for the detailed summary of the textural parameters). Moreover, sample NCC-1/7 carbonized at 1100 °C shows a pore volume and surface area comparable to those of the samples treated below this temperature. It is well known that high carbonization temperature results in the decrease of the pore volume and surface area of porous carbons due to structural shrinkage. However, this is not the case for NCC-1 type carbons, which again verifies the structural stability of such carbon materials, consistent with the XRD results. We tentatively explain this high stability by the presence of the well connected and well-developed 3-dimensional framework, which provides connections of similar stability throughout the structure. Honeycombs are also known to be highly stable structures on the macroscopic scale, and it has been shown that the mechanical stability of hexagonally ordered mesoporous materials can be explained by formalisms developed for macroscopic honeycomb structures [26].

A.-H. Lu et al. / Microporous and Mesoporous Materials 80 (2005) 117–128

Fig. 5b shows the pore size distributions of NCC-1/3 to NCC-1/7 (desorption branch). Except for sample NCC-1/3, all samples show a bimodal distribution with the first maximum at a pore size of 4.8–5.1 nm, and the second maximum at 10.3–11.0 nm. For NCC-1/4, the bimodal pore system is not as clear as for the other samples. As we reported before [9], the pores with the size of 4.8–5.1 nm result from the dissolution of the silica wall of SBA-15. This pore size is larger than the value calculated from the SBA-15 unit cell parameter and pore diameter, but quite close to the silica wall thickness measured from TEM. Combining the observations of various authors [21–23], it seems that the wall thickness of SBA-15 does not decrease, but rather reaches a constant value when the aging temperature is higher than 130 °C. However, mesotunnels in the silica wall are formed. Additionally, it is noticeable that the pore size of NCC-1/7 is reduced compared to NCC-1/5 and NCC1/6 due to the calcination temperature exceeding 1000 °C [27]. The structures and morphologies of the representative samples NCC-1/3 (600 °C) and NCC-1/6 (950 °C) have been investigated using high-resolution TEM. As shown in Fig. 6, one can see that NCC-1/3 inherits the morphology of its parent silica template although it loses the structural ordering. By carefully examining the pore structures, one can conclude that NCC-1/3 exhibits more or less homogeneous worm-like structures, and no distinguishable carbon wall as formed in CMK-5 can be observed. This is ascribed to the fairly low carbonization degree of poly-FA at 600 °C, which results in a low condensation degree of the carbon

123

framework, and correspondingly a low stability. On the other hand, the TEM images of NCC-1/6 show that this carbon copies the morphology of its parent AlSBA15 with its discernable individual particles. The high-resolution images show that this carbon has much more clearly discernable carbon walls compared to NCC-1/ 3. The carbon skeleton is amorphous, consisting of several layers of graphene sheets. Between adjacent channels, the interconnectivities are visible. These features are responsible for the enlarged pore size and the high surface area of such NCC-1 carbons. The TEM investigation shows that the formation of NCC-1 carbon mainly relies on the carbonization temperature. Higher carbonization temperature facilitates the formation of carbons with enhanced framework and structural ordering. 3.3. Heating ramp and concentration of FA The heating ramp is often considered as an important factor to determine the fine structure of many carbon materials. This is due to the fact that chemical and structural changes are already taking place during the heating to the final curing temperature. Changes in the heating ramp will thus result in changes of the relative kinetics of the processes occurring at elevated temperatures. In general, lower heating rates allow the molecules within the structure to orient, condense and decompose gently which results in the formation of a homogeneous structure. The former series of experiments was carried out with a heating rate of 5 °C/min from 300 °C on. In order to check the influence of the heating rate, it was adjusted

Fig. 6. TEM images of NCC-1/3 (600 °C) and NCC-1/6 (950 °C) nanocast from AlSBA-15-140.

124

A.-H. Lu et al. / Microporous and Mesoporous Materials 80 (2005) 117–128

to 1 °C/min. The final carbonization temperature was fixed at 850 °C in all cases. The concentration of FA solution was also varied in order to compare the results at lower heating rate with our previous results [9]. The obtained samples from NCC-1/8 to NCC-1/11 are discussed in the following. Low-angle XRD patterns of all the samples (Fig. 7), clearly show distinguishable reflections, which correspond to the hexagonal symmetry, indicating the structural ordering of the nanocast carbons. In the cases of the nanocast carbons were prepared with the concentration of the FA solution is lower than 50 vol%, the XRD patterns of NCC-1/9, NCC-1/10 and NCC-1/11 are almost identical. In the case of using pure FA, the XRD pattern of NCC-1/8 shows a distinguishable (1 0 0) peak. The intensities of the (1 1 0) and (2 0 0) peaks, however, are relatively small. The calculated unit cell parameters are in the range of 10.9 nm to 11.4 nm (Table 2). Obviously, nanocast carbon obtained at 1 °C/min heating rate does not show any clear difference compared to the sample obtained at 5 °C/min. It is worth pointing out that the use of pure FA (NCC-1/8) results in the formation of the carbon having the smallest unit cell parameter. This is probably due to the volume shrinkage of the poly-FA during carbonization. More polymer with increased interconnectivity and more contact points to the framework seems to result in higher compressive forces which are exerted on the silica framework, leading to structural shrinkage. In contrast, lower amounts of poly-FA cause less compressive force on the silica framework, leading to a lower degree of structural shrinkage, as seen in the case of samples

Fig. 7. XRD patterns of NCC-1/8 to NCC-1/11 carbons synthesized using AlSBA-15-140 as a template at a heating rate of 1 °C/min to the final temperature of 850 °C prepared from different FA concentrations.

NCC-1/7 to NCC-1/9. This effect of higher structural shrinkage with higher degree of pore filling of the silica with poly-FA has consistently been observed for many samples. In addition, also in the synthesis of monolithic samples strong shrinkage of the whole monolith is observed at high filling degrees. The nitrogen sorption isotherms (Fig. 8) show a very clear trend: with the increase of the FA concentration, the hysteresis loop at p/p0 > 0.7 gradually disappears. This corresponds to the gradual filling of the pore space of the silica template, i.e. to the transition from surfacetemplating to volume-templating [17]. Since the hysteresis is always occurring at p/p0 = 0.4–0.7, no matter how much carbon precursor was loaded, this filling step can be assigned to the capillary condensation in the pores

Fig. 8. (a) Nitrogen sorption isotherms of NCC-1/8 to NCC-1/11. The isotherms of NCC-1/10 and NCC-1/11 were offset vertically by 400 and 600 cm3 g 1 STP, respectively. (b) Pore size distributions of these NCC-1 carbons.

A.-H. Lu et al. / Microporous and Mesoporous Materials 80 (2005) 117–128

generated after the removal of the silica walls, in agreement with our previous report [9]. Fig. 8b shows the PSDs of the samples NCC-1/8 to NCC-1/11. For comparison, the PSDs of these samples both determined from adsorption and desorption branches are shown in this figure. In the cases of NCC-1/8 to NCC-1/10, the pore size calculated from the desorption branches may, to some extent, be related to a network percolation effect, resulting in a possible misrepresentation of the true PSD of these samples, because the closure of the hysteresis around p/p0 = 0.42 may result from the instability of the meniscus [28]. Hence, in these cases it is not clear whether the adsorption branch or the desorption branch better reflects the PSDs. However, for sample NCC-1/11, the desorption branch is the appropriate one, coming closer to the real values of the pore sizes, because the desorption branch is considered to correspond better to the equilibrium transition [29]. One can see the development from monomodal to bimodal PSDs with the concentration of FA solution decreasing from 100 vol% to 25 vol%. NCC-1/ 10, obtained from 40 vol% FA solution, is at the transition where the mechanism switches from surface to volume templating. From Table 2 one can see that the pore size related to the silica wall thickness gradually decreases with the FA concentration, which implies some structural change after removal of the silica support. Compared to the other samples prepared at 5 °C/min, the pore structures of the nanocast carbons prepared at 1 °C/min do not show detectable differences by XRD and nitrogen sorption. FA concentration, on the other hand, as expected has a major effect on the carbon structure. 3.4. Oxalic acid as polymerization catalyst Recently, we have demonstrated that SBA-15 can be directly used as template for the synthesis of CMK-5 type carbon without aluminum modification, and the procedure of introducing the aluminum species into the SBA-15 system can be avoided. Instead of aluminum, oxalic acid is used as polymerization catalyst, and oxalic acid and FA solution were added into the SBA-15 pore system at the same time [30]. Analogous to that synthesis, we also attempted using purely siliceous SBA-15 as template and oxalic acid as catalyst to synthesize carbons of the NCC-1 series. In this case, the FA concentration and carbonization temperature were varied, respectively. Since the heating ramp does not result in substantial differences in the nanostructure of NCC-1, we fixed the heating ramp at 5 °C/min. The obtained samples, from NCC-1/12 to NCC-1/14, were characterized as before. As seen in Fig. 9, the XRD patterns of samples NCC1/12 to NCC-1/14 prove structural ordering in the case of either using pure FA or 25 vol% FA solution as car-

125

Fig. 9. XRD patterns of NCC-1/12 to NCC-1/14 carbons synthesized using SBA-15-140 as template and oxalic acid as catalyst.

bon precursors, which is basically identical to the carbons prepared with AlSBA-15. This demonstrates that nanocast carbon can be synthesized using SBA-15 as template without pre-treatment. The nanocast carbons show slight differences in the peak intensity. Lower concentration of FA and higher carbonization temperature result in the formation of nanocast carbon with wellresolved XRD reflections. NCC-1/12 to NCC-1/14 as well as their composites were further analyzed by nitrogen sorption. The isotherms of these samples are shown in Fig. 10a. They have a type IV shape although they are differing from each other. In the case of using pure FA, NCC-1/12 shows one step with a hysteresis loop in the partial pressure range of 0.6–0.7. From the analysis of the isotherms of the carbon/silica composite in Fig. 10a, it can be deduced that NCC-1/12 is corresponding to the volumetemplated carbon since most pores of SBA-15 are blocked using pure FA solution. The removal of silica leads to the formation of the pores in the range of 6.9 nm, as shown in Fig. 10b and Table 2. The pore sizes exceed the wall thickness of the silica due to shrinkage of the carbon rods during carbonization. Using 25 vol% of FA solution as precursor, NCC-1/13 and NCC-1/14 were obtained at carbonization temperatures of 850 °C and 1100 °C, respectively. As seen in Fig. 10a, the isotherms of the composite SiO2/NCC-1/13 and SiO2/ NCC-1/14 are almost identical. However, higher carbonization temperature facilitates the formation of NCC-1 with two clearly distinguishable capillary condensation steps (NCC-1/14). The bimodal PSD is obvious is Fig. 10b. Conclusively, nanocast carbons prepared from AlSBA-15 and SBA-15/oxalic acid show identical structures, judging both from XRD patterns and nitrogen sorption analyses. However, at the same carbonization temperature, nanocast carbons

126

A.-H. Lu et al. / Microporous and Mesoporous Materials 80 (2005) 117–128

Fig. 10. (a) Nitrogen sorption isotherms of carbon/SBA-15 composites and NCC-1/12 to NCC-1/14 carbons synthesized using SBA-15-140 as template and oxalic acid as catalyst. The isotherms of NCC-1/14, SiO2/NCC-1/14 and SiO2/NCC-1/13 were offset vertically by 350, 200 and 50 cm3 g 1 STP, respectively. (b) Pore size distributions of these NCC-1 carbons.

NCC-1/14 and NCC-1/5, derived from SBA-15 and AlSBA-15, respectively, show slight differences in the position of the hysteresis loop. We assume that this difference results from the slightly smaller unit cell size of AlSBA-15 as compared to SBA-15 caused by the second calcination step needed in the synthesis of AlSBA-15, as revealed by the XRD patterns and the data in Table 2. In case of the oxalic acid catalyst, the silica template effect was also investigated. SBA-15-110 and SBA-15130 were used as templates to synthesize nanocast carbon. We found that carbons with ordered structure nanocast from SBA-15-110 and SBA-15-130 were obtained only when the concentration of FA solution is above 30 vol%. However, when pure FA is used to nanocast carbon from the SBA-15 template with oxalic acid as polymerization catalyst, CMK-3 type carbon is formed. This agrees quite well with our recent results [30]. Four selected representatives, templated from SBA-15-110 and SBA-15-130 at FA concentrations of 30 vol% and 100 vol% were characterized by XRD and nitrogen sorption. As seen in Fig. 11, their XRD patterns show well-resolved reflections, revealing hexagonal symmetry. The isotherms (Fig. 12a) of NCC-1/15 (from SBA-15-130) and NCC-1/16 (from SBA-15-110) clearly show the two capillary condensation steps in the relative pressure ranges of below 0.5 and 0.65–0.7. By comparing the isotherms of CMK-3-1 (from SBA-15-130) and CMK-3-2 (from SBA-15-110), one can conclude that

the uptake at p/p0 < 0.5 can be assigned to the capillary condensation in the pores generated from the space where the silica wall has been. The second uptake is ascribed to the capillary condensation in the inner space generated by the surface-templating of the SBA-15 pores. The PSDs of NCC-1/15 and NCC-1/16 nanocast from 30 vol% of FA solution exhibit the expected bimo-

Fig. 11. XRD patterns of NCC-1/15 to NCC-1/16 carbons and CMK3 synthesized using SBA-15-130 and SBA-15-110 as templates and oxalic acid as catalyst.

A.-H. Lu et al. / Microporous and Mesoporous Materials 80 (2005) 117–128

127

Fig. 12. (a) Nitrogen sorption isotherms of NCC-1/15 and NCC-1/16 and CMK-3 carbons synthesized using SBA-15-130 and SBA-15-110 as templates and oxalic acid as catalyst. (b) Pore size distributions of NCC-1 and CMK-3 carbons.

dal pore system (Fig. 12b). By comparing the PSD of their corresponding CMK-3 carbons, one realizes that in the case of SBA-15 aged at 130 °C the pore sizes of NCC-1/15 and CMK-3-1 corresponding to the silica wall thickness do not match each other. The pore size of CMK-3-1 is shifted to a larger value than expected from the wall thickness. This phenomenon was also observed in the case of the nanocast carbon from SBA-15140. However, in the case of SBA-15-110, the pore sizes corresponding to the silica wall thickness of NCC-1/16 and CMK-3-2 are quite close to each other. Therefore, the structural differences in the silica templates are also reflected in the nanocast carbons obtained from them. The structure of the silica template has significant influence on the pore structure of the nanocast carbon. By varying the loading amount of the carbon precursor, it is possible to obtain templated carbons with various pore structures and textural parameters.

maintains the high pore volume and the unit cell value as carbons obtained at lower carbonization temperatures. This demonstrates the special characteristics of such nanocast carbons, which we attribute to the regular 3-dimensional connectivity, allowing a very homogeneous distribution of the load of FA over the SBA-15 structure.

4. Conclusions

References

Larger pore size (10 nm) nanocast carbons can be prepared using SBA-15 as template and FA solution as carbon precursor. The concentration of the FA solution is very critical for the formation of the nanocast carbon, and also the structure of the SBA-15 mold affects the final structure of the carbon. Such nanocast carbons have very high structural stability. Even at a carbonization temperature of 1100 °C, no obvious structural shrinkage was observed, and the material almost

Acknowledgment We would like to thank B. Spliethoff for the TEM analysis. A.-H. Lu is grateful to the Alexander von Humboldt Foundation for a fellowship. The authors would like to thank the Leibniz-program and the FCI for support in addition to the basic funding provided by the Institute.

[1] T. Kyotani, Carbon 38 (2000) 269. [2] J. Lee, S. Han, T. Hyeon, J. Mater. Chem. 14 (2004) 478. [3] A.-H. Lu, W. Schmidt, N. Matoussevitch, H. Bo¨nnermann, B. Spliethoff, B. Tesche, E. Bill, W. Kiefer, F. Schu¨th, Angew. Chem., Int. Ed. 43 (2004) 4303. [4] H. Marsh, E.A. Heintz, F. Rodriguez-Reinoso, Introduction to Carbon Technology, Universidad de Alicante, Secretariado de Publications, Alicante, Spain, 1997. [5] T. Kyotani, T. Nagai, S. Inoue, A. Tomita, Chem. Mater. 9 (1997) 609. [6] R. Ryoo, S.H. Joo, S.J. Jun, Phys. Chem. B. 103 (1999) 7743.

128

A.-H. Lu et al. / Microporous and Mesoporous Materials 80 (2005) 117–128

[7] R. Ryoo, S.H. Joo, M. Kruk, M. Jaroniec, Adv. Mater. 13 (2001) 677. [8] T.-W. Kim, I.-S. Park, R. Ryoo, Angew. Chem., Int. Ed. 42 (2003) 4375. [9] A.-H. Lu, W. Schmidt, B. Spliethoff, F. Schu¨th, Adv. Mater. 15 (2003) 1602. [10] F. Kleitz, S.H. Choi, R. Ryoo, Chem. Commun. (2003) 2136. [11] C.H. Kim, D.-K. Lee, T.J. Pinnavaia, Langmuir 20 (2004) 5157. [12] J.Y. Kim, S.B. Yoon, J.-S. Yu, Chem. Mater. 15 (2003) 1932. [13] A.B. Fuertes, Micropor. Mesopor. Mater. 67 (2004) 273. [14] Z. Li, M. Jaroniec, J. Phys. Chem. B. 108 (2004) 824. [15] Z. Li, W. Yan, S. Dai, Carbon 42 (2004) 767. [16] J. Lee, S.H. Joo, R. Ryoo, J. Am. Chem. Soc. 124 (2002) 1156. [17] M. Kruk, M. Jaroniec, T.-W. Kim, R. Ryoo, Chem. Mater. 15 (2003) 2815. [18] W.H. Zhang, C. Liang, H. Sun, Z. Shen, Y. Guan, P. Ying, C. Li, Adv. Mater. 14 (2002) 1776. [19] S.N. Che, K. Lund, T. Tatsumi, S. Iijima, S.H. Joo, R. Ryoo, O. Terasaki, Angew. Chem., Int. Ed. 42 (2003) 2182.

[20] D. Zhao, J. Feng, Q. Huo, N. Melosh, G.H. Fredrickson, B.F. Chmelka, G.D. Stucky, Science 279 (1998) 548. [21] J. Fan, C. Yu, L. Wang, B. Tu, D. Zhao, Y. Sakamoto, O. Terasaki, J. Am. Chem. Soc. 123 (2001) 12113. [22] M. Kruk, M. Jaroniec, Chem. Mater. 12 (2000) 1961. [23] A. Galarneau, H. Cambon, F. Di Renzo, F. Fajula, Langmuir 17 (2001) 8328. [24] A. Galarneau, H. Cambon, F. Di Renzo, R. Ryoo, M. Choi, F. Fajula, New J. Chem. 27 (2003) 73. [25] M. Kruk, M. Jaroniec, K.P. Gadkaree, J. Colloid Interf. Sci. 192 (1997) 250. [26] D. Desplantier-Giscard, A. Galarneau, F. Di Renzo, F. Fajula, Mater. Sci. Eng. C 23 (2003) 727. [27] R. Ryoo, C.H. Ko, M. Kruk, V. Antochshuk, M. Jaroniec, J. Phys. Chem. B 104 (2000) 11465. [28] M. Kruk, M. Jaroniec, Chem. Mater. 13 (2001) 3169. [29] P.I. Ravikovitch, A.V. Neimark, Langmuir 18 (2002) 9830. [30] A.-H. Lu, W. Li, W. Schmidt, W. Kiefer, F. Schu¨th, Carbon 42 (2004) 2939.