Large pore ordered mesoporous silica materials with 3D cubic Ia3d structure: a comprehensive gas adsorption study

Large pore ordered mesoporous silica materials with 3D cubic Ia3d structure: a comprehensive gas adsorption study

From Zeolites to Porous MOF Materials – the 40th Anniversary of International Zeolite Conference R. Xu, Z. Gao, J. Chen and W. Yan (Editors) © 2007 El...

839KB Sizes 1 Downloads 9 Views

From Zeolites to Porous MOF Materials – the 40th Anniversary of International Zeolite Conference R. Xu, Z. Gao, J. Chen and W. Yan (Editors) © 2007 Elsevier B.V. All rights reserved.

1843

Large pore ordered mesoporous silica materials with 3D cubic Ia3d structure: a comprehensive gas adsorption study Freddy Kleitza*, François Bérubé a, Chia-Min Yangb and Matthias Thommesc* a

Université Laval, Department of Chemistry, Quebec G1K 7P4, Canada. Tel.: +1 418 656 7812; Fax: +1 418 656 7916; E-mail: [email protected]

b

Department of Chemistry, National Tsing-Hua University, Hsinchu

c

Quantachrome Instruments, 1900 Corporate Drive, Boynton Beach, FL 33426. Tel: +1 561 731 4999; Fax: +1 561 732 9888; E-Mail: [email protected]

ABSTRACT The paper deals with the systematic gas adsorption investigation and structural characterization of novel ordered mesoporous silica materials with three-dimensional (3-D) cubic Ia3d structure (KIT-6). The pore condensation and hysteresis behavior of nitrogen (at 77.4 K) and argon (at 77.4 K and 87.3 K) was studied in KIT-6 silicas of different porosities and various mean pore diameters (from 5 nm up to 11 nm). We compared the sorption and phase behavior of nitrogen and argon confined to these 3-D cubic porous systems with their behavior in pseudo 1-D pore systems (e.g. SBA-15). Our results provide information on how pore connectivity and network effects influence pore condensation and hysteresis, and the validity of the single pore model for the pore size analysis of materials consisting in pore networks is discussed. 1. INTRODUCTION Recently, a novel type of large-pore mesoporous silica with a cubic Ia3d structure was synthesized by using triblock copolymer Pluronic P123 as structure-directing agent [1]. This mesoporous silica material is composed of two interwoven mesoporous networks similar as in MCM-48, but can be synthesized with much larger mean pore diameters. In addition to promising applications in catalysis, adsorption, separation etc., these novel silica materials have the potential to serve as model systems for evaluating the phenomena of adsorption and phase behavior of fluids in highly ordered pore networks. Even though some progress was achieved in the understanding of the sorption and phase behavior of fluids in materials consisting of single pores (e.g. MCM-41) [ref. 2 and references therein], the process of pore condensation and hysteresis in pore networks is still under investigation [2-4]. To address this problem, we performed, in addition to a characterization by XRD, systematic gas adsorption experiments of pure fluids in a series of highly ordered mesoporous KIT-6 silica materials of different porosities and mean pore size (from 5 nm - 11 nm) complemented by a comparative study with hexagonal SBA-15 silicas.

1844 2. MATERIALS AND EXPERIMENTAL 2.1 Materials Large quantities of high quality mesoporous KIT-6 silica material were easily obtained following the method reported by Kleitz et al. [1] Briefly, 9 g of P123 were dissolved in 325 g of distilled water and 17.4 g HCl (37 %) under vigorous stirring. After complete dissolution, 9 g of butanol were added. The mixture was left under stirring at 35°C for 1 h, after which 19.35 g of TEOS were added at once to the homogeneous clear solution. The molar composition of the starting reaction mixture is TEOS/P123/HCl/H2O/BuOH = 1/0.017/1.83/195/1.31 in mole ratio. This mixture was further left under stirring at 35 °C for 24 h, followed by aging at 50, 80, 100 or 130°C for 24 h under static conditions. The KIT-6 materials aged at different temperatures are designated KIT-6-y, where y represents the aging temperature. The solid product was filtered and dried for 48 h at 95°C. For template removal, the as-synthesized silica powders were first shortly slurried in an ethanol-HCl mixture and were subsequently calcined at 550°C for 2 hours. For comparison, different 2-D hexagonal SBA-15 samples were synthesized following the method proposed by Choi et al. [5]. Briefly, 13.9 g of Pluronic P123 were dissolved in 252 g of distilled water and 7.7 g HCl (37%). After complete dissolution, 25.0 g of TEOS were added at once. The mixture was left under stirring at 35°C for 24 h, followed by a hydrothermal treatment at 100°C (may change as previous case of KIT-6) for 24 h under static conditions. The solid product was filtered, dried and finally calcined at 550°C as reported. 2.2 Characterization Powder X-ray diffractograms of the calcined samples were recorded on a Stoe STADI P T-T X-ray diffractometer in reflection geometry. High resolution nitrogen (77.4 K) and argon (77.4 K, and 87.3 K) adsorption/desorption isotherm measurements were performed with an Autosorb-I-MP adsorption instrument (Quantachrome Instruments) in the relative pressure range from 1 x 10-6 to 1. Pore Size analysis was performed by applying proper non-local density functional theory (NLDFT) methods. The NLDFT calculations were carried out according to ref. [10] by using the Quantachrome Autosorb 1.52 software provided by Quantachrome Instruments, USA.

3.RESULTS AND DISCUSSION The series of cubic Ia3d KIT-6 silica materials were first characterized by powder X-ray diffraction in order to ascertain the nature of the mesophases. The XRD patterns of all the assynthesized materials indicate excellent structural order with the symmetry being commensurate with the body-centered cubic Ia3d space group, and the highly ordered structure was retained after template removal by calcination. Fig. 1a illustrates the XRD patterns measured for calcined sample synthesized with the hydrothermal aging step performed at 50°C (KIT-6-50) and at 130°C (KIT-6-130). The unit cell sizes, calculated from the (211) reflexion of the cubic Ia3d phase, are measured to be 19.5 nm and 24.1 nm for 50°C and 130°C, respectively. These values are substantially larger than that of other cubic analogues (e.g. MCM-48). The exact assignment to the Ia3d symmetry for this new family of mesoporous materials was also confirmed using transmission electron microscopy and

1845

Intensity (a.u.)

reported elsewhere [1]. Most importantly, the highly resolved XRD patterns as well as the electron microscopy studies suggest high phase purity without structural distortion. The pore system of cubic KIT-6 materials consists in two continuous subframeworks separated by a silica wall that follows the infinite periodic minimal surface (IPMS) designated as the Gyroid G surface. The synthesis KIT-6-130 of KIT-6 employing n-butanol has several advantages over other reported synthesis strategies leading to Ia3d materials [6,7]: in particular, the addition of n-butanol at low adic concentration affords excellent reproducibility, and it is costKIT-6-50 effective and easily scaled-up. The high structural order of the large pore cubic 1.0 1.5 2.0 2.5 3.0 Ia3d material is also visible in the gas adsorption 2 theta (°) data. For example, Fig. 2a shows nitrogen Fig. 1: Powder XRD patterns for the physisorption isotherms at 77.4 K measured for mesostructured KIT-6 silica materials three selected KIT-6 silica samples, KIT-6-50, KITsynthesized with aging at 50°C (KIT6-80 and KIT-6-130, which were synthesized at 6-50) and 130 °C (KIT-6-130) for 24 various aging temperatures, i.e. 50qC, 80qC and hours. 130qC, respectively. As can be seen, all the isotherms are typical Type IV isotherms exhibiting a pronounced capillary condensation step at high relative pressures characteristic of narrow distribution of large mesopores. With increasing aging temperature, a shift of the capillary condensation step to higher relative pressures is evident, indicating substantial enlargement of the pore size. Furthermore, the application of higher hydrothermal aging temperatures produces materials with significantly increased total adsorption capacity consistent with previous data obtained for similar large pore ordered mesoporous materials [1,8]. Furthermore, the adsorption isotherms reveal a plateau-like region after the condensation step indicating extremely low external surface area. 1100

Volume [cm3 g-1] STP

900 800

0.22

KIT-6 (aged at 50 C) KIT-6 (aged at 80 C) KIT-6 (aged at 130 C)

0.16

700

0.14

600

0.12

500 400

0.1

0.08

300

0.06

200

0.04

100

0.02

0 0

KIT-6 (aged at 50 C) KIT-6 ( aged at 80 C) KIT- 6 (aged at 130 C)

0.2 0.18

Volume [cm3 g-1] STP

1000

0.2

0.4

P/P0

0.6

0.8

1

0

20

40

60

80

100 120 140 Pore Diameter [Å]

160

180

200

Fig. 2 (a) Nitrogen physisorption isotherms measured at 77.3 K on different KIT-6 samples (KIT-6-50, KIT-6-80 and KIT-6-130). (b) NLDFT pore size distributions calculated from the desorption branches of the isotherms shown in Fig. 2a.

1846 The specific surface areas of the three samples as obtained from the nitrogen adsorption data are 500 m2/g (KIT-6-50), 630 m2/g (KIT-6-80), and 580 m2/g (KIT-6-130). Pore size distributions generated from non-local density functional theory (NLDFT) method are shown in Fig. 5 for KIT-6-50, KIT-6-80 and KIT-6-130, which confirm the narrow distribution of mesopores present in these materials. It appears evident that the simple variation of the hydrothermal aging temperature permits precise tailoring of the mesopore dimension of the KIT-6 silicas from 5 nm in diameter up to values above 10 nm. We applied the equilibrium N2/cylindrical pore NLDFT kernel on the desorption branches of the isotherms [9], by assuming a cylindrical pore model. In earlier work, [10] the assumption of a cylindrical pore geometry for cubic Ia3d materials had been justified for MCM-48 silica, hence we are considering it in a first approach as also being acceptable also for KIT-6. 10

0.07

0.6

N2(77K) KIT- 6 (561-03, aged at 50 C)

KIT-6 (561-03, aged at 50 C)

Pore Volume [cm3 g-1]

0.05

0.4

6

Dv(cm3/Å/g]

Volume [cm3 g-1]

561 Ar-(ads) 561-Ar-(des) 561-N2 (ads) 561-N2 (des)

0.06

0.5

8

0.04

0.3

4

0.03

0.2

2

0 5 10-6

0.02

0.1

5 10-5

5 10-4

5 10-3

P/P0

5 10-2

5 10-1

5 100

0 5

0.01

13

21

29

37

45

53

Pore Diameter [Å]

61

69

77

0 0

10

20

30

40

50

60

Pore Diameter [Å]

70

80

90

100

Fig. 3 (a) Semi-logarithmic plot of the nitrogen physisorption isotherm measured at 77.3 K for KIT-650 over a wide range of relative pressures (i.e. 10-6 to 1). (b) NLDFT cumulative pore volume plot calculated from the adsorption branch of the isotherm measured for KIT-6-50. (c) Pore size distribution curves obtained for same KIT-6-50 sample. The NLDFT PSD curves are calculated from the nitrogen physisorption isotherm at 77.4 K and an argon physisorption isotherm at 87.3 K.

The application of the NLDFT method allows to calculate the pore size distribution over the complete range of micro- and mesopores. It appears that materials synthesized at low aging temperatures (e.g. 50°C) contain an appreciable amount of microporosity (see Fig. 3). This is demonstrated clearly in the NLDFT cumulative pore volume plot (Fig. 3b), which shows the pore volume distribution over the complete range of micro- and mesopores. Fig. 3c shows the good agreement between the NLDFT mesopore size distribution curves obtained from argon (87.3 K) (see ref. [4]) and nitrogen (77.4 K ) data. In contrast to the KIT-6 materials synthesized at low aging temperatures, the materialsynthesized at high aging temperatures do not reveal any microporosity. This is demonstrated for KIT-6-130 which reveals a high degree of order as indicated from the almost vertical adsorption/desorption branches of the type H1 hysteresis loop (see nitrogen adsorption results at 77.4 K in Fig. 2). The shape of the high resolution argon adsorption (87.3 K) isotherms obtained over a wide range of rel. pressures, i.e. 10-6 to 1, sensing the micro, meso- and macropore range of this sample (see Fig. 4a), does (in contrast to the N2 isotherm on KIT-6-50 shown in Fig. 3a) not indicate any appreciable microporosity. This is also clearly shown in the corresponding NLDFT pore size analysis which is depicted in Fig. 4b. In addition, it is important to underline that the argon and nitrogen sorption data are again fully consistent, i.e. the pore size distribution curves obtained from the argon and nitrogen isotherms are in perfect agreement. Moreover, the NLDFT mode pore diameter of 10.13 nm (calculated from the desorption branch of the nitrogen isotherm) is in very good agreement with the pore diameter (10.2 nm) obtained using a geometrical model based on unit cell (a0 =

1847 24.1 nm), and wall thickness (2 nm) as derived from XRD modeling. The XRD modeling was performed following the method reported by Solovyov et al. [11]. The excellent agreement between the pore size data obtained with the NLDFT method and the XRD modeling approach suggests that the independent pore model (which has been confirmed for pseudo 1D pore systems such as MCM-41 and SBA-15 [2]) appears also to be applicable to KIT-6 materials (at least if the desorption branch is chosen for the pore size calculation). The high degree of order of the KIT-6 samples allows to compare the sorption and phase behavior of pure fluids in a well-defined 3-D pore network with the behavior in pseudo 1-D pore geometries such as they can be found in MCM-41 and SBA-15 silica. It is most important to perform such studies because currently available approaches for pore size analysis are purely based on the assumption of an independent pore model. Some work concerning this problem had been done in the past by comparing the sorption and phase behavior of argon in a series of MCM-48 and MCM-41 silica samples with pore diameters up to ca. 5 nm [12]. These experimentals studies revealed indeed some differences in the pore condensation and hysteresis behavior, i.e. for a given pore size it appeared that the width of the argon 77 K hysteresis loop was more narrow for MCM-48 as compared to MCM-41. We are currently extending these investigations to much larger pore sizes (i.e. up to 12 nm) by using these highly ordered KIT-6 materials and ordered SBA-15 samples. 0.25

1200

Argon adsorption at 87.3 K Argon desorption at 87.3 K

0.15

720

480

0.1

0.05

240

0

NLDFT Nitrogen (77 K) NLDFT Argon (87 K)

0.2 Dv(d)[cm3/Å/g]

Volume [cm3 g-1] STP

960

5 10-5

5 10-4

5 10-3

5 10-2

Relative Pressure P/P0

5 10-1

5 100

0 5

25

45

65

85 105 125 Pore Diameter [Å]

145

165

185

205

Fig. 4 (a) Argon physisorption isotherm measured at 87.3 K (log scale). (b) Pore size distribution curves for same KIT-6-130 sample obtained from the desorption branch. The NLDFT PSD curves are calculated from the nitrogen physisorption isotherm at 77.3 K and the argon physisorption isotherm at 87.3 K The desorption branch was taken for chosen for the calculation of the pore size distribution in the pressure range where hysteresis occurs.

Fig. 5a compares the pore condensation and hysteresis behavior of nitrogen (77.4 K) in a SBA-15 silica with pore size of 7.3 nm with the behavior in a proper KIT-6 silica sample (KIT-6-80). The nitrogen isotherm of the SBA-15 sample reveals a well-pronounced hysteresis loop of type H1. The pore condensation and hysteresis behavior in this SBA-15 sample can be completely described within the independent pore model. This is demonstrated clearly in Fig. 5b. The pore size distribution curves obtained from the adsorption branch (by applying the NLDFT-kernel of metastable adsorption isotherms) and desorption branch (by applying the NLDFT equilibrium method) are in perfect agreement for SBA-15. This confirms that sorption hysteresis in this SBA-15 silica sample is more or less entirely caused by delayed condensation (i.e., by metastable states of the pore fluid occurring during adsorption/ condensation) [2,9,10].

1848 700

0.14

SBA-15 (7.3 nm) KIT-6

0.12

Volume [cm3 g-1] STP

600 500

Dv(d) [cm3/Å/g]

0.1

400 300

0.08 0.06

200

0.04

100

0.02

0

0

0.2

0.4

P/P0

0.6

0.8

1

KIT-6(Ads) KIT-6(Des) SBA-15(Ads: 7.3 nm) SBA-15(Des: 7.3 nm)

0 20

40

60

80

Pore Diameter [Å]

100

120

Fig. 5 Comparative mesopore analysis of KIT-6-80 and SBA-15 silicas exhibiting the same pore dimension (7.3 nm). a) Nitrogen physisorption isotherms measured at 77.3 K. b) NLDFT PSD calculated from nitrogen physisorption isotherms at 77.3 K

The perfect agreement between the pore size distribution curves obtained from the adsorption and desorption branches also confirms that the microporosity (or microporous corona [13]) of the SBA-15 sample does not affect the pore condensation and hysteresis behavior. The nitrogen sorption isotherm measured for KIT-6 shows some major differences in the hysteresis loop region. On the one hand, both desorption branches agree very well for the two types of structures, but adsorption, on the other hand, occurs at a lower relative pressure, which suggests an advanced pore condensation process in the case of the 3-D cubic structure. This could indicate that the pore connectivity in this 3-D network may reduce the range over which metastable pore fluid exists, which leads to advanced capillary condensation that should be facilitated in the highly interconnected open 3-D pore network. The pore size distribution (PSD) of KIT-6 calculated from the desorption branch agrees well with PSDs obtained for SBA-15, whereas the maximum of the PSD curve of KIT-6 calculated using the adsorption branch is shifted to smaller pore size. Similar observations (i.e. difference between the PSD’s calculated from adsorption and desorption branches) were also made for large pore materials obtained at high temperature (e.g. 130qC ), but now both isotherms of SBA-15 and KIT-6 seem to be consistent with each other (with regard to the width of the hysteresis loop), supporting the hypothesis that SBA-15 materials aged at high temperatures become highly interconnected in agreement with previous reports [8]. The detailed study on a series of KIT-6 and SBA-15 silica’s (pore diameter range from ca. 5 to 11 nm) suggests that there are differences in the hysteresis behavior of SBA-15 and KIT-6 materials in the pore diameter range < 9 nm, i.e. the hysteresis loop for KIT-6 is slightly (but clearly detectable) narrower as compared to appropriate SBA-15 samples. These differences in the width of hysteresis between SBA-15 and KIT-6 vanish for pore diameters > ca. 9 nm. Details of these results and their interpretation, as well as consequences for pore size characterization will be further discussed elsewhere [4]. 4. CONCLUSION We report the characterization of the pore structure of a novel a series of highly ordered large pore silicas with 3-D interconnected network structure (large pore MCM-48-like). The cubic Ia3d KIT-6 materials are easily synthesized with high phase purity in the presence of a co-

1849 surfactant (n-butanol) at reduced HCl concentration compared to conventional SBA-15-type silicas. XRD diffraction and gas adsorption studies confirm the excellent structural quality of this new family of sorbents. Importantly, the NLDFT method based on cylindrical pore model allows one to calculate accurately the pore size distribution, the total pore volume, micropore volume and surface area. Excellent agreement is found between the pore size data obtained with the NLDFT method (equilibrium transistion model) and data extracted independently from XRD modeling. We are currently performing a comparison of the pore condensation and hysteresis behavior of argon and nitrogen in a series of SBA-15 and KIT-6 materials. Preliminary results described herein suggest that, in the pore diameter range below 9-10 nm, the adsorption/desorption hysteresis loop for KIT-6 is slightly narrower as compared to appropriate SBA-15 samples. This may be attributed to the fact that the highly connected 3-D pore network could induce a smaller delay of pore condensation caused by metastability. Differences between SBA-15 and KIT-6 with respect to pore condensation delay seem to vanish for materials prepared with hydrothermal treatment performed at sufficiently high temperature (130°C). ACKNOWLEDGMENTS Freddy Kleitz wishes to thank the Canadian Government for the Canada Research Chair on Functional nanostructured materials, and financial support from NSERC, the Canadian Fundation for Innovation and the Province of Québec (FQRNT). C.M. Yang thanks the National Science Council of the Republic of China for financial support under the contract no. NSC 93-2119-M-007-004. REFERENCES [1] [2]

[3] [4] [5] [6 ] [7] [8] [9] [10]

[11] [12] [13]

a) F. Kleitz, S. H. Choi and R. Ryoo, Chem. Commun., (2003) 2136. b) T.-W. Kim, F. Kleitz, B. Paul and R. Ryoo, J. Am. Chem. Soc., 127 (2005) 7601. a) M. Thommes in Nanoporous Materials; Science and Engineering; G. Q. Lu, X. S. Zhao, Eds.; Imperial College Press: London, U.K., 2004, pp 317-364. b) M. Thommes, B. Smarsly, M. Groenewolt, P I. Ravikovitch and A. V. Neimark, Langmuir, 22 (2006) 756. K. Morishige and N. Tarui, J. Phys. Chem. C, 111 (2006) 280. F. Kleitz, C. M. Yang and M. Thommes, manuscript in preparation, (2006). M. Choi, W. Heo, F. Kleitz and R. Ryoo, Chem. Commun., (2003) 1340. X. Liu, B. Tian, C. Yu, F. Gao, S. Xie, B. Tu, R. Che, L.-M. Peng and D. Zhao, Angew. Chem. Int. Ed., 41 (2002) 3876. K. Flodström, V. Alfredsson and N. Källrot, J. Am. Chem. Soc., 125 (2003) 4402. A. Galarneau, H. Cambon, F. DiRenzo, R. Ryoo, M. Choi and F. Fajula, New J. Chem., 27 (2003) 73. P. L. Ravikovitch and A. V. Neimark, J. Phys. Chem. B, 105 (2001) 6817. A. V. Neimark and P. I. Ravikovitch, Microporous Mesoporous Mater., 44 (2001) 697. (b) P. I. Ravikovitch and A. V. Neimark, Colloids and Surfaces A: Physicochemical and Engineering Aspects, 187 (2001) 11. L. A. Solovyov, O. V. Belousov, R. E. Dinnebier, A. N. Shmakov and S. D. Kirik, J. Phys. Chem. B, 109 (2005) 3233. M. Thommes, R. Koehn and M. Fröba, Studies in Surface Science and Catalysis, 142 (2002) 1695. M. Impéror-Clerc, P. Davidson and A. Davidson, J. Am. Chem. Soc., 122 (2000) 11925.