- Email: [email protected]
Porous Texture and Surface Character of Dehydroxylated and Rehydroxylated MCM-41 Mesoporous Silicas—Analysis of Adsorption Isotherms of Nitrogen Gas and Water Vapor
Journal of Colloid and Interface Science 225, 411–420 (2000) doi:10.1006/jcis.2000.6777, available online at http://www.idealibrary.com on
Porous Texture and Surface Character of Dehydroxylated and Rehydroxylated MCM-41 Mesoporous Silicas—Analysis of Adsorption Isotherms of Nitrogen Gas and Water Vapor Hiromitu Naono,1 Masako Hakuman, Takashi Tanaka, Nobuki Tamura, and Kazuyuki Nakai2 Department of Chemistry, School of Science, Kwansei Gakuin University, Nishinomiya 662-0891, Japan; and Technical Department, Bel Japan, Inc., 2-11-27 Shinkitano, Yodogawa-ku, Osaka 532-0025, Japan Received October 11, 1999; accepted February 10, 2000
Four samples of MCM-41 mesoporous silicas whose average pore diameters are 2.4, 2.8, 3.2, and 3.6 nm were prepared using sodium orthosilicate and cationic surfactants of [CH3 (CH2 )n N(CH3 )3 ]X (n = 11, 13, 15, 17). These four samples were calcined at 1123 K in vacuo to obtain the dehydroxylated samples, which were further rehydroxylated at 298 K to obtain the rehydroxylated samples. The adsorption isotherms of nitrogen gas (77 K) for the 12 MCM-41 mesoporous silicas are of Type IVc, giving no adsorption hysteresis. On the other hand, the first adsorption isotherms of water vapor (298 K) for the dehydroxylated MCM-41 samples are quite different from those of nitrogen gas, giving the remarkable adsorption hysteresis. The second water isotherms for the rehydroxylated MCM-41 samples are of Type IV, showing slight hysteresis. Using the nitrogen isotherms, the relation between the pore size and carbon chain length of the surfactant has been determined, and the effect of dehydroxylation and rehydroxylation on the porous texture has been examined. Using the first and second water isotherms, the adsorption model of physisorbed waters adsorbed on the surface silanol groups has been proposed. From the pore size distribution curves of nitrogen and water, the presence of constrictions in the cylindrical pores has been predicted. °C 2000 Academic Press Key Words: MCM-41; cationic surfactant; nitrogen isotherm; water isotherm; dehydroxylation; rehydroxylation; adsorption hysteresis; pore size distribution curve.
INTRODUCTION
Porous textures of MCM-41 and FSM-16 have been extensively investigated by means of electron microscopy, X-ray powder diffraction, and adsorption isotherms of various adsorptives (1–4). It has been clarified from these studies that both MCM-41 and FSM-16 have the same porous texture; namely, the cylindrical pores having nearly uniform pore size are hexagonally arranged perpendicular to the c-axis of MCM-41 or FSM-16. In our previous paper, we have reported that these unique meso1 To whom correspondence should be addressed. E-mail: naono@kwansei. ac.jp. 2 E-mail: [email protected].
porous silicas can be utilized as the model adsorbents for analysis of the pore size distribution (5). The important feature of MCM41 or FSM-16 is that their pore sizes can be widely changed by using the cationic surfactants having different hydrocarbon chain lengths (1–3). So far, the adsorption isotherms of many adsorptives such as nitrogen, argon, benzene, and carbon tetrachloride have been measured (4). However, no systematic measurements of the nitrogen and water isotherms for a series of the dehydroxylated and rehydroxylated MCM-41 or FSM-16 samples have been carried out until now (6, 7). Brunauer and his co-workers (8–11) have emphasized that the water isotherms make possible an independent check on the pore size distribution determined by the nitrogen isotherms. In addition to the analysis of porous texture, the water isotherms play an important role in characterization of the pore surface of the MCM-41 or FSM-16 mesoporous silica (12–14). The purpose of the present paper is to measure systematically the adsorption–desorption isotherms of nitrogen gas (77 K) and water vapor (298 K) for a series of the MCM-41 mesoporous silicas with different pore sizes. The MCM-41 mesoporous silicas used in the present work are the following 12 samples. At first, four kinds of the MCM-41 mesoporous silicas were synthesized using sodium orthosilicate and cationic surfactants of ([CH3 (CH2 )n N(CH3 )3 ]X (n = 11, 13, 15, 17)). These four samples were calcined at 1123 K for 2 h in vacuo to prepare the dehydroxylated MCM-41 samples, which were further rehydroxylated at 298 K under exposure of liquid water to obtain the rehydroxylated MCM-41 samples. The porous texture and the surface character of the MCM-41 mesoporous silicas thus prepared were studied by measuring the nitrogen isotherms and the water isotherms. The t or αs method (15–18) was applied for analysis of the nitrogen isotherms to determine the pore area and the average pore size of the 12 MCM-41 samples. In addition, the pore size distribution curves of the rehydroxylated MCM-41 samples were calculated using the nitrogen isotherms, where the calculations were performed by our new method described previously (5). The first water isotherms were used for the study of the hydrophobic character of the dehydroxylated MCM-41 silica
411
0021-9797/00 $35.00
C 2000 by Academic Press Copyright ° All rights of reproduction in any form reserved.
412
NAONO ET AL.
surface, whereas the second water isotherms were used for the study of the hydrophilic character of the rehydroxylated MCM41 surface. With the use of the first and second water isotherms, the relationship between the physisorbed waters and the surface silanol groups was investigated. Furthermore, the second water isotherms were applied for the calculation of the pore size distribution curves of the rehydroxylated MCM-41 samples. The pore size distribution curves obtained from the nitrogen isotherms and the second water isotherms were utilized for the study of the pore structure of the rehydroxylated MCM-41 mesoporous silicas. EXPERIMENTAL
Materials Na4 SiO4 · nH2 O (water content = ca. 40% in weight) (Kishida Chem. Co. Ltd.) was used as the silica source of MCM41. Cationic surfactants of [C12 H25 N(CH3 )3 ]Br (purity = 99%), [C14 H29 N(CH3 )3 ]Br (purity = 98%), [C16 H33 N(CH3 )3 ]Cl (purity = 95%), and [C18 H37 N(CH3 )3 ]Br (purity = 98%) (Tokyo Kasei Co. Ltd.) were utilized as templates. These reagents were used without further purification. An analytical grade reagent of hydrochloric acid was used in the pH titration. Preparation of Solution A sodium orthosilicate solution of ca. 0.5 MW was prepared by dissolving Na4 SiO4 · nH2 O into distilled water, where MW is the weight molarity, which is defined as the number of moles of a reagent in 1 kg of solution (19). In the present experiment, all the solutions were prepared using the stationary-pan digital electronic balance, instead of the volumetric vessel. Thus, the concentration was expressed as the weight molarity (MW ). After preparation of the orthosilicate solution, the solution was kept several days and then was filtered to remove the undissolved substances. The concentration of orthosilicate was deter-
mined from the equivalent point of the pH titration curve after the 0.5 MW solution had been diluted by one-tenth. pH titration was carried out by the automatic gravimetric titrator constructed in our laboratory (20). The cationic surfactant solution of 0.1 MW was prepared by dissolving each surfactant into distilled water. In calculation of the concentration of the cationic surfactant, the purity mentioned above was taken into account. Preparation of MCM-41 In the preparation of PS-C12, PS-C14, and PS-C16 (cf. Table 1), 50 g of the 0.5 MW orthosilicate solution was mixed with 100 g of the 0.1 MW cationic surfactant solution at room temperature. In the preparation of PS-C18, 50 g of the 0.5 MW orthosilicate solution was mixed with 100 g of the 0.05 MW cationic surfactant solution at room temperature. After the temperature of the mixed solution rose to 343 K, the mixed solution was stirred for several hours. The pH of the mixed solution was found to be 12 at 343 K. Then, 2.0 MW HCl was gradually added to the mixed solution. Near pH 10, the white precipitate (silicate/surfactant complex, MCM-41 precursor) was rapidly formed (20). The HCl solution was added further until pH 9. After the precipitate aged for 16 h at 343 K, it was washed several times with distilled water and then dried at the desiccator for several days. The dried precipitate was calcined at 823 K under a nitrogen flow for 6 h and then under an air flow for 2 h. In this process, the surfactant was removed from the precipitate, and the MCM-41 mesoporous silica was produced. Four kinds of the MCM-41 samples with different pore sizes (PS-C12, PSC14, PS-C16, and PS-C18) were synthesized using the cationic surfactants of [CH3 (CH2 )n N(CH3 )3 ]X (n = 11, 13, 15, 17). Adsorption Isotherms of Nitrogen Gas and Water Vapor The adsorption–desorption isotherms of nitrogen gas (77 K) and water vapor (298 K) were measured on the MCM-41
TABLE 1 Characterization of MCM-41 Mesoporous Silicas by Nitrogen Isotherms and X-Ray Powder Diffraction Pretreatment
Sample name
Temp. (K)
Time (h)
SBET (m2 g−1 )
St (m2 g−1 )
Sex (m2 g−1 )
Sp (m2 g−1 )
Vp (cm3 g−1 )
rp(av) (nm)
d(100) (nm)
PS-C12 PS-C12-D PS-C12-R PS-C14 PS-C14-D PS-C14-R PS-C16 PS-C16-D PS-C16-R PS-C18 PS-C18-D PS-C18-R
298 1123 298 298 1123 298 298 1123 298 298 1123 298
4 2 4 4 2 4 4 2 4 4 2 4
1021 876 850 947 789 790 954 846 785 997 872 817
1064 895 903 981 835 856 980 881 815 1012 895 841
80 81 60 113 125 112 94 94 107 130 125 121
984 814 843 868 710 744 886 787 708 882 770 720
0.58 0.50 0.43 0.60 0.50 0.49 0.70 0.61 0.51 0.79 0.69 0.61
1.19 1.23 1.02 1.38 1.41 1.32 1.59 1.56 1.45 1.79 1.79 1.69
3.32 — — 3.55 — — 3.89 — — 4.24 — —
Note. The final letters of D and R given in the sample name mean dehydroxylated and rehydoxylated (see text).
POROUS TEXTURE AND SURFACE CHARACTER OF MCM-41
413
mesoporous silicas by means of the automatic volumetric adsorption apparatuses constructed in our laboratory (14, 21–24). The synthesized MCM-41 samples were pretreated in vacuo at 298 K for 4 h, whose sample names are designated as PSC12, PS-C14, PS-C16, and PS-C18. The dehydroxylated MCM41 samples (PS-C12-D, PS-C14-D, PS-C16-D, and PS-C18-D) were obtained by calcining the synthesized MCM-41 samples (PS-C12, PS-C14, PS-C16, and PS-C18) at 1123 K for 2 h in vacuo. The rehydroxylated MCM-41 samples (PS-C12-R, PSC14-R, PS-C16-R, and PS-C18-R) were obtained, after the first water isotherm for the dehydroxylated MCM-41 sample had been finished. Pretreatment conditions of the MCM-41 samples are listed in Table 1. X-Ray Powder Diffraction of MCM-41 The X-ray powder diffraction measurements were carried out for the four MCM-41 samples of PS-C12, PS-C14, PS-C16, and PS-C18 under the conditions of 30 kV, 30 mA, and the CuK α radiation. RESULTS AND DISCUSSION
Adsorption Isotherms of Nitrogen Gas of MCM-41 Mesoporous Silicas
FIG. 2. Adsorption–desorption isotherms of nitrogen gas at 77 K for MCM41 samples (PS-C16, PS-C18), dehydroxylated MCM-41 samples (PS-C16-D, PS-C18-D), and rehydroxylated MCM-41 samples (PS-C16- R, PS-C18-R).
The nitrogen adsorption–desorption isotherms for the 12 MCM-41 mesoporous silicas are shown in Figs. 1 and 2. According to the new classification proposed recently by Rouquerol, Rouquerol, and Sing (25), the adsorption isotherms in Figs. 1 and 2 may be designated as a Type IVc. It is evident from Figs. 1 and 2 that the adsorption branches of all nitrogen isotherms are completely superimposed on the desorption ones, indicating the reversible pore filling and emptying. One of the important features of the MCM-41 mesoporous silicas is that their adsorption isotherms rise steeply within a very narrow pressure range (cf. P/P 0 = 0.38 ± 0.02 in the case of PS-C18, Fig. 2), suggesting the uniformity of pore size (5). Furthermore, the steep rise in the isotherm is successively shifted to the higher relative pressure with an increase of the carbon number in the long hydrocarbon chain of cationic surfactants (cf. Figs. 1 and 2). This experimental fact clearly shows that the pore size of the MCM-41 mesoporous silica increases with the chain length of the surfactant. The relationship between the pore size and the carbon number will be given below (cf. Fig. 5a). Analysis of Nitrogen Adsorption Isotherms by the t or αs Method
FIG. 1. Adsorption–desorption isotherms of nitrogen gas at 77 K for MCM41 samples (PS-C12, PS-C14), dehydroxylated MCM-41 samples (PS-C12-D, PS-C14-D), and rehydroxylated MCM-41 samples (PS-C12- R, PS-C14-R).
t plots or αs plots of the MCM-41 samples are given in Figs. 3 and 4, where the t curve or αs curve for nonporous silica measured in our laboratory was used as the standard (23, 26). As is seen from Figs. 3 and 4, t plots (αs plots) of all samples show the upper deviation from the linear line (A–B). Such deviation is caused by the capillary condensation of nitrogen molecules into mesopores of the MCM-41 samples. Analysis of the t plot or
414
NAONO ET AL.
αs plot was carried out according to the method reported by de Boer and co-workers and Sing and co-workers (15–18). St (total surface area) was calculated from the slope of the A–B line, while Sex (external surface area) was computed from the slope of the C–D line. Sp (pore surface area) was estimated from the difference between St and Sex . Vp (pore volume) was evaluated from the extrapolating point E. The values of St , Sex , Sp , and Vp obtained from the t plots of Figs. 3 and 4 are listed in Table 1 together with the values of SBET . rp(av) (average pore radius) was calculated by Eq. [1], where the pore shape and the pore surface are assumed to be cylindrical and smooth, respectively: rp(av) = 2Vp /Sp .
[1]
The St values are in good agreement with the SBET values, suggesting both sets of data to be reliable. It is found from Table 1 that Sex is in the range of ca. 7–14% of St . Surface Roughness of the Pore Wall of MCM-41 Mesoporous Silicas
FIG. 3. t plots and αs plots of nitrogen isotherms for MCM-41 samples (PS-C12, PS-C14), dehydroxylated MCM-41 samples (PS-C12-D, PS- C14-D), and rehydroxylated MCM-41 samples (PS-C12-R, PS-C14-R).
When the MCM-41 samples were calcined at 1123 K in vacuo, their pore volumes decrease appreciably, as was seen from Vp of Table 1. Vp of PS-C12-D, PS-C14-D, PS-C16-D, and PSC18-D decreases by 16, 17, 13, and 13% in comparison with Vp of PS-C12, PS-C14, PS-C16, and PS-C18. When the pore volume decreases, it is expected that the pore size decreases by shrinkage of the pore, but the calculated rp(av) values given in Table 1 do not decrease. Such an inconsistency may be caused by neglecting the surface roughness of the pore wall. If the pore wall is assumed to be smooth, a decrease in the pore surface area (1Sp(calc) ) is calculated to be 8.4, 8.9, 6.7, and 6.7% for PS-C12-D, PS-C14-D, PS-C16-D, and PS-C18-D, respectively. However, the experimental data of Sp given in Table 1 show that a decrease in Sp for the above four samples (1Sp(exp) ) is 17, 18, 11, and 13%, respectively. The experimental value (1Sp(exp) ) is significantly larger than the calculated value (1Sp(calc) ). The difference between 1Sp(exp) and 1Sp(calc) may be attributed to surface roughness of the MCM-41 samples. The pore wall of the MCM- 41 mesoporous silica changes from an uneven surface to a smooth surface by pretreatment in vacuo at 1123 K. This speculation is supported by the following experimental fact: namely, when PS-C16 is pretreated in vacuo at 1273 K for 9 h, the pore volume decreases by 65% together with the drastic sintering (5). From the above discussion, it is reasonable to conclude that the two effects (i.e., change in surface roughness and shrinkage in pore size) lead to a large decrease in the pore surface area. Pore Models of PS-C12, PS-C14, PS-C16, and PS-C18
FIG. 4. t plots and αs plots of nitrogen isotherms for MCM-41 samples (PS-C16, PS-C18), dehydroxylated MCM-41 samples (PS-C16-D, PS- C18-D), and rehydroxylated MCM-41 samples (PS-C16-R, PS-C18-R).
In Fig. 5a, the average pore diameters (2rp(av) ) of PS-C12, PS-C14, PS-C16, and PS-C18 are plotted as a function of the carbon number of the long hydrocarbon chain of the cationic surfactant. The relation between the carbon number and the d(100) value is also shown in Fig. 5a. The d(100) value is calculated from
415
POROUS TEXTURE AND SURFACE CHARACTER OF MCM-41
smooth (cf. Eq. [1]). As was discussed in the previous section, the MCM-41 samples have the uneven surface. If we assume that the difference between 1Sp(exp) and 1Sp(calc) is attributed to the roughness of the pore wall, it is possible to correct the calculated rp(av) value given in Table 1. The corrected pore radius is to be 1.29, 1.50, 1.65, and 1.90 nm for PS-C12, PS-C14, PS-C16, and PS-C18, respectively. In the present work, the corrected pore radius is tentatively used in the pore model. With use of the data of the corrected pore radii and the d(100) values (Table 2), the pore models of the four MCM-41 mesoporous silicas are constructed and shown in Fig. 5b. The unit cell dimension (a0 ) of the MCM-41 sample is calculated using a0 = (2/31/2 ) × d(100) . The difference between the unit cell dimension (a0 ) and 2 times the corrected pore radius is estimated to be 1.25, 1.10, 1.19, and 1.09 nm for PS-C12, PSC14, PS-C16, and PS-C18, respectively. This difference may be used as a parameter of the average thickness of the silica layer of the MCM-41 mesoporous silicas. It is evident from Fig. 5a that the silica layer or the pore wall of the MCM-41 samples is quite thin (1.1–1.2 nm), which leads to the very large pore area of 882–985 m2 /g (cf. Table 1). Surface Character of Dehydroxylated and Rehydroxylated MCM-41 Mesoporous Silicas The first adsorption isotherms of water vapor on the dehydroxylated MCM-41 mesoporous silicas (PS-C12-D, PS-C14-D, PS-C16-D, and PS-C18-D), and the second adsorption isotherms of water vapor on the rehydroxylated MCM-41 mesoporous silicas (PS-C12-R, PS-C14-R, PS-C16-R, PS-C18-R) are shown in Figs. 6–9. In these figures, the adsorbed amounts of water are expressed as cm3 (STP)/g(left ordinate) and H2 O/nm2 (right ordinate), the latter data being calculated using the surface areas of the rehydroxylated MCM-41 samples (cf. Table 1). The third adsorption isotherm of water vapor was also measured on PSC12-R (cf. Fig. 6). The first water isotherms cannot be assigned
FIG. 5. (a) Relation between 2rp , d(100) of MCM-41 samples, and carbon number of long hydrocarbon chain of cationic surfactant. A, PS-C12; B, PS-C14; C, PS-C16; and D, PS-C18. (b) Pore models of MCM-41 samples (PS-C12, PSC14, PS-C16, and PS-C18).
the (100) peak of the X-ray powder diffraction (Table 2). The diffraction data given in Table 2 coincide with the hexagonal symmetry reported by the previous investigators (1–3). As is seen from Fig. 5a, the good linear relationship is obtained between the average pore diameter and the carbon number, and between the d(100) value and the carbon number. From the slope of Fig. 5a, an increase of the average pore diameter (2rp ) per –CH2 group is estimated to be 0.20 nm. The next step is the construction of the pore model of PS-C12, PS-C14, PS-C16, and PS-C18. Here, it should be mentioned that rp(av) is determined by assuming the pore wall to be perfectly
TABLE 2 X-Ray Powder Diffraction Data of MCM-41 Mesoporous Silicas 2θ (◦ )
d(hkl) (nm)
I /I0
(hkl)
PS-C18
2.08 3.62 4.21 5.54
4.24 2.44 2.10 1.59
100 21 12 6
(100) (110) (200) (210)
PS-C16
2.22 4.02 4.56 6.15
3.89 2.20 1.94 1.44
100 20 14 7
(100) (110) (200) (210)
PS-C14
2.49 4.37 4.78 6.77
3.55 2.02 1.85 1.30
100 19 16 6
(100) (110) (200) (210)
PS-C12
2.66 4.88
3.32 1.81
100 17
(100) (110)
416
NAONO ET AL.
FIG. 6. Adsorption–desorption isotherms of water vapor at 298 K for dehydroxylated MCM-41 sample (PS-C12-D) and rehydroxylated MCM-41 sample (PS-C12-R).
FIG. 8. Adsorption–desorption isotherms of water vapor at 298 K for dehydroxylated MCM-41 sample (PS-C16-D) and rehydroxylated MCM-41 sample (PS-C16-R).
to any type of the IUPAC classification of adsorption isotherms. On the other hand, the second and third adsorption isotherms of water vapor are classified as Type IV (25). The characteristic features of the first water isotherms on the dehydroxylated MCM-41 mesoporous silicas are summarized as follows (cf. Figs. 6–9). First, the adsorbed amount of water is extremely small at low relative pressure (P/P 0 = 0–0.3). Second, remarkably large hysteresis appears in the adsorption and desorption branches. And third, the desorption isotherms do not return to the origin. Inagaki et al. and Bambrough et al. reported a similar phenomenon for FSM-16 and MCM-41 mesoporous silicas (6, 7). Such characteristic features of the first water isotherms can be explained by the hydrophobic character of the dehydroxylated silica surface and slow rehydroxylation of the dehydroxylated silica (12, 13). It is well-known that most of the surface silanol groups on the pore wall of MCM-41 mesoporous silicas are removed by thermal treatment at 1123 K
in vacuo. The residual silanol groups have been estimated to be 0.4–0.8 SiOH/nm2 , when silica gel was calcined at 1123 K in vacuo (27, 28). Nearly the same amount of silanol groups may be present in the present MCM-41 mesoporous silicas. The BET method was tried to determine the monolayer capacity of physisorbed water on the dehydroxylated MCM-41 samples (PS-C12-D, PS-C14-D, PS-C16-D, and PS-C18-D). However, because of the extremely small amount of adsorbed water, it was difficult to obtain the reliable monolayer capacity of physisorbed water. The monolayer capacity estimated by the BET method is only an approximate value, which is in the range of 0.05–0.1 H2 O/nm2 . These water molecules are very small compared with the residual silanol groups mentioned above (0.4– 0.8 SiOH/nm2 ). From this result, we have judged that the silanol groups that remained on the 1123 K-treated MCM-41 mesoporous silica surface do not act as the primary adsorption sites of water molecules.
FIG. 7. Adsorption–desorption isotherms of water vapor at 298 K for dehydroxylated MCM-41 sample (PS-C14-D) and rehydroxylated MCM-41 sample (PS-C14-R).
FIG. 9. adsorption–desorption isotherms of water vapor at 298 K for dehydroxylated MCM-41 sample (PS-C18-D) and rehydroxylated MCM- 41 sample (PS-C18-R).
POROUS TEXTURE AND SURFACE CHARACTER OF MCM-41
417
TABLE 3 Chemisorption of Water and Regenerated Silanol Groups on the Pore Wall of Dehydroxylated MCM-41 Mesoporous Silicas
P/P 0
PS-C12-D PS-C14-D PS-C16-D PS-C18-D 10 10 10 10 (H2 O nm−2 ) (H2 O nm−2 ) (H2 O nm−2 ) (H2 O nm−2 )
0.05 0.10 0.15 0.20 0.25 0.30 Average chemisorbed waters (H2 O nm−2 ) Average regenerated silanols (OH nm−2 )
2.02 1.90 1.90 2.14 2.14 — 2.01
1.90 1.90 1.90 1.90 1.79 — 1.88
2.14 2.38 2.38 2.38 2.14 — 2.28
1.90 1.90 2.02 2.14 2.14 2.14 2.04
4.02
3.76
4.56
4.08
FIG. 10. Adsorption water isotherms at 283, 298, and 308 K for rehydroxylated MCM-41 sample (PS-C18-R).
Note. 10: Difference between 1st and 2nd desorption isotherms.
The vertical rise of the first adsorption isotherm at P/P 0 = 0.6–0.7 corresponds to the beginning of the capillary condensation of water vapor into the hydrophobic (dehydroxylated) mesopores. Slow rehydroxylation of the dehydroxylated silica surface occurs in the mesopores filled with water. A part of the adsorbed water molecules is consumed in the formation of silanol groups. As a result, the first desorption isotherms do not return to the origin. The amounts of chemisorbed water or the regenerated silanol groups were determined from the difference between the first and second desorption isotherm (12), whose result is given in Table 3. The regenerated silanol groups per nm2 are found to be 4.02, 3.76, 4.56, and 4.08 for PS-C12-D, PSC14-D, PS-C16-D, and PS-C18-D, respectively. These regenerated silanol groups act as the primary sites for the physisorption of water molecules. The formation of the silanol groups makes the MCM-41 mesoporous silica surface hydrophilic. To examine the surface character of the rehydroxylated MCM41 sample, the differential enthalpy of physisorbed water (1 H¯ a ) is calculated by the following equation: µ
∂ ln P ∂T
¶ 0
=−
1 H¯ a . RT 2
ters. As is seen from Fig. 11, the 1 H¯ a decreases gradually in the vicinity of the BET monolayer values. This means that the adsorption energy of water on the silica surface covered by silanol groups cannot be clearly differentiated from the adsorption energy of water on the surface covered by physisorbed water. Here, we propose the adsorption model shown in Fig. 12. However, the proposed model is considered to be an approximate model of the above discussion. Pore Size Distribution of Rehydroxylated MCM-41 by Adsorption Isotherms of Nitrogen Gas and Water Vapor In this section, the pore size distribution curves of the rehydroxylated MCM-41 samples (PS-C12-R, PS-C14-R,
[2]
Calculation of 1 H¯ a is carried out using the water isotherms shown in Fig. 10. In Fig. 11, the differential enthalpy of physisorbed water (1 H¯ a ) for PS-C18-R is plotted as a function of the adsorbed amount (0), where 1 H¯ a vs 0 for the nonporous hydrophilic silica gel (Aerosil-200) is also given. It is seen from Fig. 11 that the surface character of PS-C18-R is very similar to that of hydrophilic Aerosil-200. Next, the relation between rehydroxylated silanol groups and physisorbed waters is examined. The BET monolayer values of physisorbed water for PS-C18-R are approximately estimated to be 2.0 ± 0.3 H2 O/nm2 from three water isotherms of Fig. 10. Because of the small BET constant (10–15), it seems difficult to determine precisely the monolayer coverage of physisorbed wa-
FIG. 11. Differential adsorption enthalpies of physisorbed waters on rehydroxylated MCM-41 mesoporous silica (PS-C18-R) and hydroxylated nonporous silica gel (Aerosil-200).
418
NAONO ET AL.
FIG. 12. Chemisorption of water on the pore wall of dehydroxylated MCM41 mesoporous silica and physisorption of water on the pore wall of rehydroxylated MCM-41 mesoporous silica.
PS-C16-R, and PS-C18-R) are calculated using the second water isotherms given in Figs. 6–9. It should be mentioned that the adsorption branches of the first water isotherms cannot be used in the calculation of pore size distribution for the following reasons. First, the contact angle, θ, between the adsorbed film and the capillary condensate cannot be determined because the dehydroxylated MCM-41 samples are remarkably hydrophobic. Second, the slow chemisorption of water occurs in the first adsorption process. On the other hand, it is possible to use the desorption branches of the first water isotherm in pore size calculation. The first desorption branch overlaps with the second desorption one, if the first desorption branch is shifted to the definite distance equal to chemisorbed water along the ordinate. The second water isotherms (298 K) shown in Figs. 6–9 have slight hysteresis. Their hysteresis loops are classified as Type H1 according to the IUPAC classification (1985). It is known that the Type H1 hysteresis loop arises from adsorbents with a narrow distribution of uniform pores. As has been shown in Figs. 1 and 2, the nitrogen isotherms (77 K) have no hysteresis. In our previous paper, we pointed out that the lower closure point of the hysteresis loop is positioned at P/P 0 = 0.42 for nitrogen isotherms and at P/P 0 = 0.27 for water isotherms (29). As shown in Figs. 1 and 2, the capillary condensation of nitrogen molecules took place in the range of P/P 0 = 0.16 (PS-C12-R) and P/P 0 = 0.32 (PS-C18-R). Evidently, the capillary condensation of nitrogen molecules into the rehydroxylated MCM-41 samples occurs below the low closure point of hysteresis. On the other hand, as shown in Figs. 6–9, the capillary condensation of water molecules took place in the range of P/P 0 = 0.29–0.35 (PS-C12-R) and P/P 0 = 0.40–0.46 (PSC18-R). In the case of the water isotherms, the capillary conden-
sation occurs above the low closure point of hysteresis. Accordingly, an appreciable hysteresis loop was detected in the water isotherms. In the latter part of this paper, we shall use both of the nitrogen and water isotherms to calculate the pore size distribution curves of the rehydroxylated MCM-41 samples. The hysteresis loop in the water isotherms shifts to the higher relative pressure according to the increase of the carbon number of the cationic surfactant used in the synthesis of the MCM-41 precursor. It is seen from Figs. 6–9 that, in all the second isotherms, the low-pressure hysteresis was slightly detected at P/P 0 = 0–0.3. To judge whether the low-pressure hysteresis is attributed to the unsatisfactory rehydroxylation, the third water isotherm was measured at 298 K for the rehydroxylated sample (PS-C12-R). As was seen from Fig. 6, the second water isotherm is completely superimposed on the third one. This result indicates that complete rehydroxylation has occurred after the first desorption isotherm had been determined. The lowpressure hysteresis observed in the MCM-41 mesoporous silicas may be attributed to the presence of ultramicropores on the MCM-41 pore wall. In calculation of the pore size distribution using the water isotherm, it is necessary to determine the adsorbed thickness (t) and the Kelvin radius (rk ) as a function of the relative pressure. t curves of water vapor were reported by Hagymassy and co-workers (8) and Naono and Hakuman (29). In the present paper, we used the t curve determined from the water isotherm at 298 K on hydrophilic nonporous silica (Aerosil-200) (14, 29). The Kelvin radius, rk (i.e., core radius), of the cylindrical pore with the semispherical meniscus is calculated using the following equation, ¶µ ¶ µ µ ¶ 1 2γ VL P =− , [3] ln P0 RT rk where γ is the surface tension of water (72.6 mJ/m2 ), VL is the molar volume of water (18.07 × 10−6 m3 /mol), rk is the Kelvin radius, T is the absolute temperature (298 K), and R is the gas constant. When the Kelvin equation is utilized, the contact angle, θ , between the adsorbed film and the capillary condensate is assumed to be zero because the pore surface of the rehydroxylated MCM-41 samples becomes hydrophilic. The pore radius (rp ) of the cylindrical pore is given as the sum of rk and t. The pore size distribution curves were calculated by the BJH method (30). We used the cylindrical pore model whose one end was closed. Calculation of the pore size distribution curves using the nitrogen isotherms was carried out according to a new method reported in our previous paper (5). The t curve and the rp vs P/P 0 curve shown in Figs. 12 and 13 of Ref. (5) were used in the calculation. Kruk and Jaroniec (31) proposed the corrected form of the Kelvin equation, where they use the t curve and the rp vs P/P 0 curve given in our previous paper. The pore size distribution curves of PS-C12-R, PS-C14-R, PS-C16-R, and PS-C18-R are shown in Figs. 13 and 14. In the upper part of these figures, the distribution curves calculated from the adsorption and
419
POROUS TEXTURE AND SURFACE CHARACTER OF MCM-41
TABLE 4 Peak Position and Half-Width of Pore Size Distribution Curves of Rehydroxylated MCM-41 Mesoporous Silicas rp(peak) (nm (half-width/nm)) From H2 O isotherm
Sample name
From N2 isotherm
Ads. branch
Des. branch
PS-C12-R PS-C14-R PS-C16-R PS-C18-R
1.10(0.20) 1.36(0.18) 1.50(0.21) 1.67(0.15)
1.11(0.16) 1.40(0.18) 1.53(0.19) 1.74(0.18)
1.01(0.09) 1.26(0.15) 1.33(0.15) 1.49(0.19)
desorption branches of the second water isotherms are given, while the distribution curves obtained from the nitrogen isotherms are shown in the lower part of Figs. 13 and 14. In the following section, the distribution curves obtained from two adsorptives are discussed in light of the pore structure of the rehydroxylated MCM-41 samples. Pore Structure Evaluated from Pore Size Distribution Curves
FIG. 13. Pore size distribution curves of rehydroxylated MCM-41 samples (PS-C12-R and PS-C14-R) calculated from water isotherms and nitrogen isotherms.
FIG. 14. Pore size distribution curves of rehydroxylated MCM-41 samples (PS-C16-R and PS-C18-R) calculated from water isotherms and nitrogen isotherms.
The peak position (rp(peak) ) and the half-width are determined from the pore size distribution curves shown in Figs. 13 and 14. The result is given in Table 4. It is seen from the half-width of Table 4 that the majority of the pores (more than 90%) are distributed within the range of ±0.1 nm from rp(peak) , indicating the cylindrical pores of the rehydroxylated MCM-41 samples (PS-C12-R, PS-C14-R, PS-C16-R, and PS-C18-R) to be fairly uniform. Finally, the pore structure of the MCM-41 samples will be discussed on the basis of the results of Figs. 13 and 14 and Table 4. The present results clearly indicate that (1) the rp(peak) value calculated from the desorption branch of the water isotherm is slightly smaller than the rp(peak) value calculated from the adsorption branch, and (2) the rp(peak) value obtained from the nitrogen isotherm is positioned near the rp(peak) values obtained from the adsorption branch of the water isotherm. The difference between rp(peak) (ads. branch) and rp(peak) (des. branch) is found to be 9, 10, 13, and 14% for PS-C12-R, PS-C14-R, PS-C16R, and PS-C18-R, respectively. This difference strongly suggests that the cylindrical pores are not uniform, but there present constrictions in the cylindrical pores. The size of constrictions is estimated to be 9–14% less than the size of the main cylindrical pores. The presence of constrictions cannot be detected by the nitrogen isotherm because the capillary condensation of nitrogen molecules occurs below the lower closure point of hysteresis. In conclusion, we have determined the pore size distribution curves of the rehydroxylated MCM-41 mesoporous silicas using the two adsorptives whose physical property is remarkably different. The calculated distribution curves give valuable information about the porous texture. It has been clarified in the present work that the water isotherms not only make an independent check on the pore size distribution estimated by the nitrogen isotherm but also can detect the presence of constrictions.
420
NAONO ET AL.
ACKNOWLEDGMENTS We express our sincere thanks to Dr. S. Inagaki for X-ray diffraction measurements and to T. Shiono, C. Oohashi, and Y. Tsukahara for performing part of the experiments.
REFERENCES 1. Kresge, C. T., Leonowicz, M. E., Roth, W. J., Vartuli, J. C., and Beck, J. S., Nature 359, 710 (1992). 2. Beck, J. S., Vartuli, J. C., Roth, W. J., Leonowicz, M. E., Kresge, C. T., Schmitt, K. D., Chu, C. T.-W., Olson, D. H., Sheppard, E. W., McCullen, S. B., Higgins, J. B., and Schlenker, J. L., J. Am. Chem. Soc. 114, 10834 (1992). 3. Inagaki, S., Fukushima, Y., and Kuroda, K., J. Chem. Soc., Chem. Commun. 680 (1993). Inagaki, S., “Synthesis and Structure of Mesoporous Crystal, FSM-16,” Ph.D. thesis. Nagoya University, 1998. 4. Rouquerol, F., Rouquerol, J., and Sing, K., in “Adsorption by Powders & Porous Solids,” Chap. 12. Academic Press, San Diego, 1999. 5. Naono, H., Hakuman, M., and Shiono, T., J. Colloid Interface Sci. 186, 360 (1997). 6. Inagaki, S., Fukushima, Y., Kuroda, K., and Kuroda, K., J. Colloid Interface Sci. 180, 623 (1996). 7. Bambrough, C. M., Slade, R. C. T., Williams, R. T., Burkett, S. L., Sims, S. D., and Mann, S., J. Colloid Interface Sci. 201, 220 (1998). 8. Hagymassy, J., Jr., Brunauer, S., and Mikhail, R. Sh., J. Colloid Interface Sci. 29, 485 (1969). 9. Hagymassy, J., Jr., and Brunauer, S., J. Colloid Interface Sci. 33, 317 (1970). 10. Hagymassy, J., Jr., Odler, I., Yudenfreund, M., Skalny, J., and Brunauer, S., J. Colloid Interface Sci. 38, 20 (1972). 11. Odler, I., Hagymassy, J., Jr., Yudenfreund, M., Hanna, K. M., and Brunauer, S., J. Colloid Interface Sci. 38, 265 (1972).
12. Naono, H., Fujiwara, R., and Yagi, M., J. Colloid Interface Sci. 76, 74 (1980). 13. Gregg, S. J., and Sing, K. S. W., in “Adsorption, Surface Area and Porosity,” 2nd ed., Chap. 5. Academic Press, London, 1982. 14. Naono, H., and Hakuman, M., J. Colloid Interface Sci. 145, 405 (1991). 15. Lippens, B. C., Linsen, B. G., and de Boer, J. H., J. Catal. 3, 32 (1964). 16. Lippens, B. C., and de Boer, J. H., J. Catal. 4, 319 (1965). 17. de Boer, J. H., Lippens, B. C., Linsen, B. G., Broekhoff, J. C. P., van den Heuvel, A., and Osinga, Th. J., J. Colloid Interface Sci. 21, 405 (1966). 18. Rouquerol, F., Rouquerol, J., and Sing, K., “Adsorption by Powders & Porous Solids,” Chap. 6. Academic Press, San Diego, 1999. 19. Skoog, D. A., West, D. M., and Holler, F. J., in “Fundamentals of Analytical Chemistry,” 7th ed., Chap. 6, Sect. 6D. Saunders College Publishing, Fort Worth, 1997. 20. Naono, H., Hakuman, M., Tsunehisa, T., Tamura, N., and Nakai, K., J. Colloid Interface Sci. 224, 358 (2000). 21. Naono, H., and Nakai, K., J. Colloid Interface Sci. 128, 146 (1989). 22. Naono, H., J. Soc. Powder Tech. Jpn. 28, 34 (1991). 23. Naono, H., and Hakuman, M., Hyomen (Surface) 29, 362 (1991). 24. Naono, H., in “Development of Porous Ceramics” (M. Hattori, and S. Yamanaka, Eds.), p. 9. CMC, Tokyo, 2000. 25. Rouquerol, F., Rouquerol, J., and Sing, K., in “Adsorption by Powders & Porous Solids,” Chap. 13. Academic Press, San Diego, 1999. 26. Chemical Society of Japan, “Handbook of Chemistry (Kagaku Binran),” revised 4th ed., Vol. II-93. Maruzen, Tokyo, 1993. 27. Morimoto, T., and Naono, H., Bull. Chem. Soc. Jpn. 46, 2000 (1973). 28. Iler, R. K., “The Chemistry of Silica,” Chap. 6. Wiley, New York, 1979. Brinker, C. J., and Scherer, G. W., “Sol-Gel Science,” Chap. 10. Academic Press, Boston, 1990. 29. Naono, H., and Hakuman, M., J. Colloid Interface Sci. 158, 19 (1993). 30. Barrett, E. P., Joyner, L. G., and Halenda, P. P., J. Am. Chem. Soc. 73, 373 (1951). 31. Kruk, M., and Jaroniec, M., Langmuir 13, 6267 (1997).
Porous Texture and Surface Character of Dehydroxylated and Rehydroxylated MCM-41 Mesoporous Silicas—Analysis of Adsorption Isotherms of Nitrogen Gas and Water Vapor Hiromitu Naono,1 Masako Hakuman, Takashi Tanaka, Nobuki Tamura, and Kazuyuki Nakai2 Department of Chemistry, School of Science, Kwansei Gakuin University, Nishinomiya 662-0891, Japan; and Technical Department, Bel Japan, Inc., 2-11-27 Shinkitano, Yodogawa-ku, Osaka 532-0025, Japan Received October 11, 1999; accepted February 10, 2000
Four samples of MCM-41 mesoporous silicas whose average pore diameters are 2.4, 2.8, 3.2, and 3.6 nm were prepared using sodium orthosilicate and cationic surfactants of [CH3 (CH2 )n N(CH3 )3 ]X (n = 11, 13, 15, 17). These four samples were calcined at 1123 K in vacuo to obtain the dehydroxylated samples, which were further rehydroxylated at 298 K to obtain the rehydroxylated samples. The adsorption isotherms of nitrogen gas (77 K) for the 12 MCM-41 mesoporous silicas are of Type IVc, giving no adsorption hysteresis. On the other hand, the first adsorption isotherms of water vapor (298 K) for the dehydroxylated MCM-41 samples are quite different from those of nitrogen gas, giving the remarkable adsorption hysteresis. The second water isotherms for the rehydroxylated MCM-41 samples are of Type IV, showing slight hysteresis. Using the nitrogen isotherms, the relation between the pore size and carbon chain length of the surfactant has been determined, and the effect of dehydroxylation and rehydroxylation on the porous texture has been examined. Using the first and second water isotherms, the adsorption model of physisorbed waters adsorbed on the surface silanol groups has been proposed. From the pore size distribution curves of nitrogen and water, the presence of constrictions in the cylindrical pores has been predicted. °C 2000 Academic Press Key Words: MCM-41; cationic surfactant; nitrogen isotherm; water isotherm; dehydroxylation; rehydroxylation; adsorption hysteresis; pore size distribution curve.
INTRODUCTION
Porous textures of MCM-41 and FSM-16 have been extensively investigated by means of electron microscopy, X-ray powder diffraction, and adsorption isotherms of various adsorptives (1–4). It has been clarified from these studies that both MCM-41 and FSM-16 have the same porous texture; namely, the cylindrical pores having nearly uniform pore size are hexagonally arranged perpendicular to the c-axis of MCM-41 or FSM-16. In our previous paper, we have reported that these unique meso1 To whom correspondence should be addressed. E-mail: naono@kwansei. ac.jp. 2 E-mail: [email protected].
porous silicas can be utilized as the model adsorbents for analysis of the pore size distribution (5). The important feature of MCM41 or FSM-16 is that their pore sizes can be widely changed by using the cationic surfactants having different hydrocarbon chain lengths (1–3). So far, the adsorption isotherms of many adsorptives such as nitrogen, argon, benzene, and carbon tetrachloride have been measured (4). However, no systematic measurements of the nitrogen and water isotherms for a series of the dehydroxylated and rehydroxylated MCM-41 or FSM-16 samples have been carried out until now (6, 7). Brunauer and his co-workers (8–11) have emphasized that the water isotherms make possible an independent check on the pore size distribution determined by the nitrogen isotherms. In addition to the analysis of porous texture, the water isotherms play an important role in characterization of the pore surface of the MCM-41 or FSM-16 mesoporous silica (12–14). The purpose of the present paper is to measure systematically the adsorption–desorption isotherms of nitrogen gas (77 K) and water vapor (298 K) for a series of the MCM-41 mesoporous silicas with different pore sizes. The MCM-41 mesoporous silicas used in the present work are the following 12 samples. At first, four kinds of the MCM-41 mesoporous silicas were synthesized using sodium orthosilicate and cationic surfactants of ([CH3 (CH2 )n N(CH3 )3 ]X (n = 11, 13, 15, 17)). These four samples were calcined at 1123 K for 2 h in vacuo to prepare the dehydroxylated MCM-41 samples, which were further rehydroxylated at 298 K under exposure of liquid water to obtain the rehydroxylated MCM-41 samples. The porous texture and the surface character of the MCM-41 mesoporous silicas thus prepared were studied by measuring the nitrogen isotherms and the water isotherms. The t or αs method (15–18) was applied for analysis of the nitrogen isotherms to determine the pore area and the average pore size of the 12 MCM-41 samples. In addition, the pore size distribution curves of the rehydroxylated MCM-41 samples were calculated using the nitrogen isotherms, where the calculations were performed by our new method described previously (5). The first water isotherms were used for the study of the hydrophobic character of the dehydroxylated MCM-41 silica
411
0021-9797/00 $35.00
C 2000 by Academic Press Copyright ° All rights of reproduction in any form reserved.
412
NAONO ET AL.
surface, whereas the second water isotherms were used for the study of the hydrophilic character of the rehydroxylated MCM41 surface. With the use of the first and second water isotherms, the relationship between the physisorbed waters and the surface silanol groups was investigated. Furthermore, the second water isotherms were applied for the calculation of the pore size distribution curves of the rehydroxylated MCM-41 samples. The pore size distribution curves obtained from the nitrogen isotherms and the second water isotherms were utilized for the study of the pore structure of the rehydroxylated MCM-41 mesoporous silicas. EXPERIMENTAL
Materials Na4 SiO4 · nH2 O (water content = ca. 40% in weight) (Kishida Chem. Co. Ltd.) was used as the silica source of MCM41. Cationic surfactants of [C12 H25 N(CH3 )3 ]Br (purity = 99%), [C14 H29 N(CH3 )3 ]Br (purity = 98%), [C16 H33 N(CH3 )3 ]Cl (purity = 95%), and [C18 H37 N(CH3 )3 ]Br (purity = 98%) (Tokyo Kasei Co. Ltd.) were utilized as templates. These reagents were used without further purification. An analytical grade reagent of hydrochloric acid was used in the pH titration. Preparation of Solution A sodium orthosilicate solution of ca. 0.5 MW was prepared by dissolving Na4 SiO4 · nH2 O into distilled water, where MW is the weight molarity, which is defined as the number of moles of a reagent in 1 kg of solution (19). In the present experiment, all the solutions were prepared using the stationary-pan digital electronic balance, instead of the volumetric vessel. Thus, the concentration was expressed as the weight molarity (MW ). After preparation of the orthosilicate solution, the solution was kept several days and then was filtered to remove the undissolved substances. The concentration of orthosilicate was deter-
mined from the equivalent point of the pH titration curve after the 0.5 MW solution had been diluted by one-tenth. pH titration was carried out by the automatic gravimetric titrator constructed in our laboratory (20). The cationic surfactant solution of 0.1 MW was prepared by dissolving each surfactant into distilled water. In calculation of the concentration of the cationic surfactant, the purity mentioned above was taken into account. Preparation of MCM-41 In the preparation of PS-C12, PS-C14, and PS-C16 (cf. Table 1), 50 g of the 0.5 MW orthosilicate solution was mixed with 100 g of the 0.1 MW cationic surfactant solution at room temperature. In the preparation of PS-C18, 50 g of the 0.5 MW orthosilicate solution was mixed with 100 g of the 0.05 MW cationic surfactant solution at room temperature. After the temperature of the mixed solution rose to 343 K, the mixed solution was stirred for several hours. The pH of the mixed solution was found to be 12 at 343 K. Then, 2.0 MW HCl was gradually added to the mixed solution. Near pH 10, the white precipitate (silicate/surfactant complex, MCM-41 precursor) was rapidly formed (20). The HCl solution was added further until pH 9. After the precipitate aged for 16 h at 343 K, it was washed several times with distilled water and then dried at the desiccator for several days. The dried precipitate was calcined at 823 K under a nitrogen flow for 6 h and then under an air flow for 2 h. In this process, the surfactant was removed from the precipitate, and the MCM-41 mesoporous silica was produced. Four kinds of the MCM-41 samples with different pore sizes (PS-C12, PSC14, PS-C16, and PS-C18) were synthesized using the cationic surfactants of [CH3 (CH2 )n N(CH3 )3 ]X (n = 11, 13, 15, 17). Adsorption Isotherms of Nitrogen Gas and Water Vapor The adsorption–desorption isotherms of nitrogen gas (77 K) and water vapor (298 K) were measured on the MCM-41
TABLE 1 Characterization of MCM-41 Mesoporous Silicas by Nitrogen Isotherms and X-Ray Powder Diffraction Pretreatment
Sample name
Temp. (K)
Time (h)
SBET (m2 g−1 )
St (m2 g−1 )
Sex (m2 g−1 )
Sp (m2 g−1 )
Vp (cm3 g−1 )
rp(av) (nm)
d(100) (nm)
PS-C12 PS-C12-D PS-C12-R PS-C14 PS-C14-D PS-C14-R PS-C16 PS-C16-D PS-C16-R PS-C18 PS-C18-D PS-C18-R
298 1123 298 298 1123 298 298 1123 298 298 1123 298
4 2 4 4 2 4 4 2 4 4 2 4
1021 876 850 947 789 790 954 846 785 997 872 817
1064 895 903 981 835 856 980 881 815 1012 895 841
80 81 60 113 125 112 94 94 107 130 125 121
984 814 843 868 710 744 886 787 708 882 770 720
0.58 0.50 0.43 0.60 0.50 0.49 0.70 0.61 0.51 0.79 0.69 0.61
1.19 1.23 1.02 1.38 1.41 1.32 1.59 1.56 1.45 1.79 1.79 1.69
3.32 — — 3.55 — — 3.89 — — 4.24 — —
Note. The final letters of D and R given in the sample name mean dehydroxylated and rehydoxylated (see text).
POROUS TEXTURE AND SURFACE CHARACTER OF MCM-41
413
mesoporous silicas by means of the automatic volumetric adsorption apparatuses constructed in our laboratory (14, 21–24). The synthesized MCM-41 samples were pretreated in vacuo at 298 K for 4 h, whose sample names are designated as PSC12, PS-C14, PS-C16, and PS-C18. The dehydroxylated MCM41 samples (PS-C12-D, PS-C14-D, PS-C16-D, and PS-C18-D) were obtained by calcining the synthesized MCM-41 samples (PS-C12, PS-C14, PS-C16, and PS-C18) at 1123 K for 2 h in vacuo. The rehydroxylated MCM-41 samples (PS-C12-R, PSC14-R, PS-C16-R, and PS-C18-R) were obtained, after the first water isotherm for the dehydroxylated MCM-41 sample had been finished. Pretreatment conditions of the MCM-41 samples are listed in Table 1. X-Ray Powder Diffraction of MCM-41 The X-ray powder diffraction measurements were carried out for the four MCM-41 samples of PS-C12, PS-C14, PS-C16, and PS-C18 under the conditions of 30 kV, 30 mA, and the CuK α radiation. RESULTS AND DISCUSSION
Adsorption Isotherms of Nitrogen Gas of MCM-41 Mesoporous Silicas
FIG. 2. Adsorption–desorption isotherms of nitrogen gas at 77 K for MCM41 samples (PS-C16, PS-C18), dehydroxylated MCM-41 samples (PS-C16-D, PS-C18-D), and rehydroxylated MCM-41 samples (PS-C16- R, PS-C18-R).
The nitrogen adsorption–desorption isotherms for the 12 MCM-41 mesoporous silicas are shown in Figs. 1 and 2. According to the new classification proposed recently by Rouquerol, Rouquerol, and Sing (25), the adsorption isotherms in Figs. 1 and 2 may be designated as a Type IVc. It is evident from Figs. 1 and 2 that the adsorption branches of all nitrogen isotherms are completely superimposed on the desorption ones, indicating the reversible pore filling and emptying. One of the important features of the MCM-41 mesoporous silicas is that their adsorption isotherms rise steeply within a very narrow pressure range (cf. P/P 0 = 0.38 ± 0.02 in the case of PS-C18, Fig. 2), suggesting the uniformity of pore size (5). Furthermore, the steep rise in the isotherm is successively shifted to the higher relative pressure with an increase of the carbon number in the long hydrocarbon chain of cationic surfactants (cf. Figs. 1 and 2). This experimental fact clearly shows that the pore size of the MCM-41 mesoporous silica increases with the chain length of the surfactant. The relationship between the pore size and the carbon number will be given below (cf. Fig. 5a). Analysis of Nitrogen Adsorption Isotherms by the t or αs Method
FIG. 1. Adsorption–desorption isotherms of nitrogen gas at 77 K for MCM41 samples (PS-C12, PS-C14), dehydroxylated MCM-41 samples (PS-C12-D, PS-C14-D), and rehydroxylated MCM-41 samples (PS-C12- R, PS-C14-R).
t plots or αs plots of the MCM-41 samples are given in Figs. 3 and 4, where the t curve or αs curve for nonporous silica measured in our laboratory was used as the standard (23, 26). As is seen from Figs. 3 and 4, t plots (αs plots) of all samples show the upper deviation from the linear line (A–B). Such deviation is caused by the capillary condensation of nitrogen molecules into mesopores of the MCM-41 samples. Analysis of the t plot or
414
NAONO ET AL.
αs plot was carried out according to the method reported by de Boer and co-workers and Sing and co-workers (15–18). St (total surface area) was calculated from the slope of the A–B line, while Sex (external surface area) was computed from the slope of the C–D line. Sp (pore surface area) was estimated from the difference between St and Sex . Vp (pore volume) was evaluated from the extrapolating point E. The values of St , Sex , Sp , and Vp obtained from the t plots of Figs. 3 and 4 are listed in Table 1 together with the values of SBET . rp(av) (average pore radius) was calculated by Eq. [1], where the pore shape and the pore surface are assumed to be cylindrical and smooth, respectively: rp(av) = 2Vp /Sp .
[1]
The St values are in good agreement with the SBET values, suggesting both sets of data to be reliable. It is found from Table 1 that Sex is in the range of ca. 7–14% of St . Surface Roughness of the Pore Wall of MCM-41 Mesoporous Silicas
FIG. 3. t plots and αs plots of nitrogen isotherms for MCM-41 samples (PS-C12, PS-C14), dehydroxylated MCM-41 samples (PS-C12-D, PS- C14-D), and rehydroxylated MCM-41 samples (PS-C12-R, PS-C14-R).
When the MCM-41 samples were calcined at 1123 K in vacuo, their pore volumes decrease appreciably, as was seen from Vp of Table 1. Vp of PS-C12-D, PS-C14-D, PS-C16-D, and PSC18-D decreases by 16, 17, 13, and 13% in comparison with Vp of PS-C12, PS-C14, PS-C16, and PS-C18. When the pore volume decreases, it is expected that the pore size decreases by shrinkage of the pore, but the calculated rp(av) values given in Table 1 do not decrease. Such an inconsistency may be caused by neglecting the surface roughness of the pore wall. If the pore wall is assumed to be smooth, a decrease in the pore surface area (1Sp(calc) ) is calculated to be 8.4, 8.9, 6.7, and 6.7% for PS-C12-D, PS-C14-D, PS-C16-D, and PS-C18-D, respectively. However, the experimental data of Sp given in Table 1 show that a decrease in Sp for the above four samples (1Sp(exp) ) is 17, 18, 11, and 13%, respectively. The experimental value (1Sp(exp) ) is significantly larger than the calculated value (1Sp(calc) ). The difference between 1Sp(exp) and 1Sp(calc) may be attributed to surface roughness of the MCM-41 samples. The pore wall of the MCM- 41 mesoporous silica changes from an uneven surface to a smooth surface by pretreatment in vacuo at 1123 K. This speculation is supported by the following experimental fact: namely, when PS-C16 is pretreated in vacuo at 1273 K for 9 h, the pore volume decreases by 65% together with the drastic sintering (5). From the above discussion, it is reasonable to conclude that the two effects (i.e., change in surface roughness and shrinkage in pore size) lead to a large decrease in the pore surface area. Pore Models of PS-C12, PS-C14, PS-C16, and PS-C18
FIG. 4. t plots and αs plots of nitrogen isotherms for MCM-41 samples (PS-C16, PS-C18), dehydroxylated MCM-41 samples (PS-C16-D, PS- C18-D), and rehydroxylated MCM-41 samples (PS-C16-R, PS-C18-R).
In Fig. 5a, the average pore diameters (2rp(av) ) of PS-C12, PS-C14, PS-C16, and PS-C18 are plotted as a function of the carbon number of the long hydrocarbon chain of the cationic surfactant. The relation between the carbon number and the d(100) value is also shown in Fig. 5a. The d(100) value is calculated from
415
POROUS TEXTURE AND SURFACE CHARACTER OF MCM-41
smooth (cf. Eq. [1]). As was discussed in the previous section, the MCM-41 samples have the uneven surface. If we assume that the difference between 1Sp(exp) and 1Sp(calc) is attributed to the roughness of the pore wall, it is possible to correct the calculated rp(av) value given in Table 1. The corrected pore radius is to be 1.29, 1.50, 1.65, and 1.90 nm for PS-C12, PS-C14, PS-C16, and PS-C18, respectively. In the present work, the corrected pore radius is tentatively used in the pore model. With use of the data of the corrected pore radii and the d(100) values (Table 2), the pore models of the four MCM-41 mesoporous silicas are constructed and shown in Fig. 5b. The unit cell dimension (a0 ) of the MCM-41 sample is calculated using a0 = (2/31/2 ) × d(100) . The difference between the unit cell dimension (a0 ) and 2 times the corrected pore radius is estimated to be 1.25, 1.10, 1.19, and 1.09 nm for PS-C12, PSC14, PS-C16, and PS-C18, respectively. This difference may be used as a parameter of the average thickness of the silica layer of the MCM-41 mesoporous silicas. It is evident from Fig. 5a that the silica layer or the pore wall of the MCM-41 samples is quite thin (1.1–1.2 nm), which leads to the very large pore area of 882–985 m2 /g (cf. Table 1). Surface Character of Dehydroxylated and Rehydroxylated MCM-41 Mesoporous Silicas The first adsorption isotherms of water vapor on the dehydroxylated MCM-41 mesoporous silicas (PS-C12-D, PS-C14-D, PS-C16-D, and PS-C18-D), and the second adsorption isotherms of water vapor on the rehydroxylated MCM-41 mesoporous silicas (PS-C12-R, PS-C14-R, PS-C16-R, PS-C18-R) are shown in Figs. 6–9. In these figures, the adsorbed amounts of water are expressed as cm3 (STP)/g(left ordinate) and H2 O/nm2 (right ordinate), the latter data being calculated using the surface areas of the rehydroxylated MCM-41 samples (cf. Table 1). The third adsorption isotherm of water vapor was also measured on PSC12-R (cf. Fig. 6). The first water isotherms cannot be assigned
FIG. 5. (a) Relation between 2rp , d(100) of MCM-41 samples, and carbon number of long hydrocarbon chain of cationic surfactant. A, PS-C12; B, PS-C14; C, PS-C16; and D, PS-C18. (b) Pore models of MCM-41 samples (PS-C12, PSC14, PS-C16, and PS-C18).
the (100) peak of the X-ray powder diffraction (Table 2). The diffraction data given in Table 2 coincide with the hexagonal symmetry reported by the previous investigators (1–3). As is seen from Fig. 5a, the good linear relationship is obtained between the average pore diameter and the carbon number, and between the d(100) value and the carbon number. From the slope of Fig. 5a, an increase of the average pore diameter (2rp ) per –CH2 group is estimated to be 0.20 nm. The next step is the construction of the pore model of PS-C12, PS-C14, PS-C16, and PS-C18. Here, it should be mentioned that rp(av) is determined by assuming the pore wall to be perfectly
TABLE 2 X-Ray Powder Diffraction Data of MCM-41 Mesoporous Silicas 2θ (◦ )
d(hkl) (nm)
I /I0
(hkl)
PS-C18
2.08 3.62 4.21 5.54
4.24 2.44 2.10 1.59
100 21 12 6
(100) (110) (200) (210)
PS-C16
2.22 4.02 4.56 6.15
3.89 2.20 1.94 1.44
100 20 14 7
(100) (110) (200) (210)
PS-C14
2.49 4.37 4.78 6.77
3.55 2.02 1.85 1.30
100 19 16 6
(100) (110) (200) (210)
PS-C12
2.66 4.88
3.32 1.81
100 17
(100) (110)
416
NAONO ET AL.
FIG. 6. Adsorption–desorption isotherms of water vapor at 298 K for dehydroxylated MCM-41 sample (PS-C12-D) and rehydroxylated MCM-41 sample (PS-C12-R).
FIG. 8. Adsorption–desorption isotherms of water vapor at 298 K for dehydroxylated MCM-41 sample (PS-C16-D) and rehydroxylated MCM-41 sample (PS-C16-R).
to any type of the IUPAC classification of adsorption isotherms. On the other hand, the second and third adsorption isotherms of water vapor are classified as Type IV (25). The characteristic features of the first water isotherms on the dehydroxylated MCM-41 mesoporous silicas are summarized as follows (cf. Figs. 6–9). First, the adsorbed amount of water is extremely small at low relative pressure (P/P 0 = 0–0.3). Second, remarkably large hysteresis appears in the adsorption and desorption branches. And third, the desorption isotherms do not return to the origin. Inagaki et al. and Bambrough et al. reported a similar phenomenon for FSM-16 and MCM-41 mesoporous silicas (6, 7). Such characteristic features of the first water isotherms can be explained by the hydrophobic character of the dehydroxylated silica surface and slow rehydroxylation of the dehydroxylated silica (12, 13). It is well-known that most of the surface silanol groups on the pore wall of MCM-41 mesoporous silicas are removed by thermal treatment at 1123 K
in vacuo. The residual silanol groups have been estimated to be 0.4–0.8 SiOH/nm2 , when silica gel was calcined at 1123 K in vacuo (27, 28). Nearly the same amount of silanol groups may be present in the present MCM-41 mesoporous silicas. The BET method was tried to determine the monolayer capacity of physisorbed water on the dehydroxylated MCM-41 samples (PS-C12-D, PS-C14-D, PS-C16-D, and PS-C18-D). However, because of the extremely small amount of adsorbed water, it was difficult to obtain the reliable monolayer capacity of physisorbed water. The monolayer capacity estimated by the BET method is only an approximate value, which is in the range of 0.05–0.1 H2 O/nm2 . These water molecules are very small compared with the residual silanol groups mentioned above (0.4– 0.8 SiOH/nm2 ). From this result, we have judged that the silanol groups that remained on the 1123 K-treated MCM-41 mesoporous silica surface do not act as the primary adsorption sites of water molecules.
FIG. 7. Adsorption–desorption isotherms of water vapor at 298 K for dehydroxylated MCM-41 sample (PS-C14-D) and rehydroxylated MCM-41 sample (PS-C14-R).
FIG. 9. adsorption–desorption isotherms of water vapor at 298 K for dehydroxylated MCM-41 sample (PS-C18-D) and rehydroxylated MCM- 41 sample (PS-C18-R).
POROUS TEXTURE AND SURFACE CHARACTER OF MCM-41
417
TABLE 3 Chemisorption of Water and Regenerated Silanol Groups on the Pore Wall of Dehydroxylated MCM-41 Mesoporous Silicas
P/P 0
PS-C12-D PS-C14-D PS-C16-D PS-C18-D 10 10 10 10 (H2 O nm−2 ) (H2 O nm−2 ) (H2 O nm−2 ) (H2 O nm−2 )
0.05 0.10 0.15 0.20 0.25 0.30 Average chemisorbed waters (H2 O nm−2 ) Average regenerated silanols (OH nm−2 )
2.02 1.90 1.90 2.14 2.14 — 2.01
1.90 1.90 1.90 1.90 1.79 — 1.88
2.14 2.38 2.38 2.38 2.14 — 2.28
1.90 1.90 2.02 2.14 2.14 2.14 2.04
4.02
3.76
4.56
4.08
FIG. 10. Adsorption water isotherms at 283, 298, and 308 K for rehydroxylated MCM-41 sample (PS-C18-R).
Note. 10: Difference between 1st and 2nd desorption isotherms.
The vertical rise of the first adsorption isotherm at P/P 0 = 0.6–0.7 corresponds to the beginning of the capillary condensation of water vapor into the hydrophobic (dehydroxylated) mesopores. Slow rehydroxylation of the dehydroxylated silica surface occurs in the mesopores filled with water. A part of the adsorbed water molecules is consumed in the formation of silanol groups. As a result, the first desorption isotherms do not return to the origin. The amounts of chemisorbed water or the regenerated silanol groups were determined from the difference between the first and second desorption isotherm (12), whose result is given in Table 3. The regenerated silanol groups per nm2 are found to be 4.02, 3.76, 4.56, and 4.08 for PS-C12-D, PSC14-D, PS-C16-D, and PS-C18-D, respectively. These regenerated silanol groups act as the primary sites for the physisorption of water molecules. The formation of the silanol groups makes the MCM-41 mesoporous silica surface hydrophilic. To examine the surface character of the rehydroxylated MCM41 sample, the differential enthalpy of physisorbed water (1 H¯ a ) is calculated by the following equation: µ
∂ ln P ∂T
¶ 0
=−
1 H¯ a . RT 2
ters. As is seen from Fig. 11, the 1 H¯ a decreases gradually in the vicinity of the BET monolayer values. This means that the adsorption energy of water on the silica surface covered by silanol groups cannot be clearly differentiated from the adsorption energy of water on the surface covered by physisorbed water. Here, we propose the adsorption model shown in Fig. 12. However, the proposed model is considered to be an approximate model of the above discussion. Pore Size Distribution of Rehydroxylated MCM-41 by Adsorption Isotherms of Nitrogen Gas and Water Vapor In this section, the pore size distribution curves of the rehydroxylated MCM-41 samples (PS-C12-R, PS-C14-R,
[2]
Calculation of 1 H¯ a is carried out using the water isotherms shown in Fig. 10. In Fig. 11, the differential enthalpy of physisorbed water (1 H¯ a ) for PS-C18-R is plotted as a function of the adsorbed amount (0), where 1 H¯ a vs 0 for the nonporous hydrophilic silica gel (Aerosil-200) is also given. It is seen from Fig. 11 that the surface character of PS-C18-R is very similar to that of hydrophilic Aerosil-200. Next, the relation between rehydroxylated silanol groups and physisorbed waters is examined. The BET monolayer values of physisorbed water for PS-C18-R are approximately estimated to be 2.0 ± 0.3 H2 O/nm2 from three water isotherms of Fig. 10. Because of the small BET constant (10–15), it seems difficult to determine precisely the monolayer coverage of physisorbed wa-
FIG. 11. Differential adsorption enthalpies of physisorbed waters on rehydroxylated MCM-41 mesoporous silica (PS-C18-R) and hydroxylated nonporous silica gel (Aerosil-200).
418
NAONO ET AL.
FIG. 12. Chemisorption of water on the pore wall of dehydroxylated MCM41 mesoporous silica and physisorption of water on the pore wall of rehydroxylated MCM-41 mesoporous silica.
PS-C16-R, and PS-C18-R) are calculated using the second water isotherms given in Figs. 6–9. It should be mentioned that the adsorption branches of the first water isotherms cannot be used in the calculation of pore size distribution for the following reasons. First, the contact angle, θ, between the adsorbed film and the capillary condensate cannot be determined because the dehydroxylated MCM-41 samples are remarkably hydrophobic. Second, the slow chemisorption of water occurs in the first adsorption process. On the other hand, it is possible to use the desorption branches of the first water isotherm in pore size calculation. The first desorption branch overlaps with the second desorption one, if the first desorption branch is shifted to the definite distance equal to chemisorbed water along the ordinate. The second water isotherms (298 K) shown in Figs. 6–9 have slight hysteresis. Their hysteresis loops are classified as Type H1 according to the IUPAC classification (1985). It is known that the Type H1 hysteresis loop arises from adsorbents with a narrow distribution of uniform pores. As has been shown in Figs. 1 and 2, the nitrogen isotherms (77 K) have no hysteresis. In our previous paper, we pointed out that the lower closure point of the hysteresis loop is positioned at P/P 0 = 0.42 for nitrogen isotherms and at P/P 0 = 0.27 for water isotherms (29). As shown in Figs. 1 and 2, the capillary condensation of nitrogen molecules took place in the range of P/P 0 = 0.16 (PS-C12-R) and P/P 0 = 0.32 (PS-C18-R). Evidently, the capillary condensation of nitrogen molecules into the rehydroxylated MCM-41 samples occurs below the low closure point of hysteresis. On the other hand, as shown in Figs. 6–9, the capillary condensation of water molecules took place in the range of P/P 0 = 0.29–0.35 (PS-C12-R) and P/P 0 = 0.40–0.46 (PSC18-R). In the case of the water isotherms, the capillary conden-
sation occurs above the low closure point of hysteresis. Accordingly, an appreciable hysteresis loop was detected in the water isotherms. In the latter part of this paper, we shall use both of the nitrogen and water isotherms to calculate the pore size distribution curves of the rehydroxylated MCM-41 samples. The hysteresis loop in the water isotherms shifts to the higher relative pressure according to the increase of the carbon number of the cationic surfactant used in the synthesis of the MCM-41 precursor. It is seen from Figs. 6–9 that, in all the second isotherms, the low-pressure hysteresis was slightly detected at P/P 0 = 0–0.3. To judge whether the low-pressure hysteresis is attributed to the unsatisfactory rehydroxylation, the third water isotherm was measured at 298 K for the rehydroxylated sample (PS-C12-R). As was seen from Fig. 6, the second water isotherm is completely superimposed on the third one. This result indicates that complete rehydroxylation has occurred after the first desorption isotherm had been determined. The lowpressure hysteresis observed in the MCM-41 mesoporous silicas may be attributed to the presence of ultramicropores on the MCM-41 pore wall. In calculation of the pore size distribution using the water isotherm, it is necessary to determine the adsorbed thickness (t) and the Kelvin radius (rk ) as a function of the relative pressure. t curves of water vapor were reported by Hagymassy and co-workers (8) and Naono and Hakuman (29). In the present paper, we used the t curve determined from the water isotherm at 298 K on hydrophilic nonporous silica (Aerosil-200) (14, 29). The Kelvin radius, rk (i.e., core radius), of the cylindrical pore with the semispherical meniscus is calculated using the following equation, ¶µ ¶ µ µ ¶ 1 2γ VL P =− , [3] ln P0 RT rk where γ is the surface tension of water (72.6 mJ/m2 ), VL is the molar volume of water (18.07 × 10−6 m3 /mol), rk is the Kelvin radius, T is the absolute temperature (298 K), and R is the gas constant. When the Kelvin equation is utilized, the contact angle, θ , between the adsorbed film and the capillary condensate is assumed to be zero because the pore surface of the rehydroxylated MCM-41 samples becomes hydrophilic. The pore radius (rp ) of the cylindrical pore is given as the sum of rk and t. The pore size distribution curves were calculated by the BJH method (30). We used the cylindrical pore model whose one end was closed. Calculation of the pore size distribution curves using the nitrogen isotherms was carried out according to a new method reported in our previous paper (5). The t curve and the rp vs P/P 0 curve shown in Figs. 12 and 13 of Ref. (5) were used in the calculation. Kruk and Jaroniec (31) proposed the corrected form of the Kelvin equation, where they use the t curve and the rp vs P/P 0 curve given in our previous paper. The pore size distribution curves of PS-C12-R, PS-C14-R, PS-C16-R, and PS-C18-R are shown in Figs. 13 and 14. In the upper part of these figures, the distribution curves calculated from the adsorption and
419
POROUS TEXTURE AND SURFACE CHARACTER OF MCM-41
TABLE 4 Peak Position and Half-Width of Pore Size Distribution Curves of Rehydroxylated MCM-41 Mesoporous Silicas rp(peak) (nm (half-width/nm)) From H2 O isotherm
Sample name
From N2 isotherm
Ads. branch
Des. branch
PS-C12-R PS-C14-R PS-C16-R PS-C18-R
1.10(0.20) 1.36(0.18) 1.50(0.21) 1.67(0.15)
1.11(0.16) 1.40(0.18) 1.53(0.19) 1.74(0.18)
1.01(0.09) 1.26(0.15) 1.33(0.15) 1.49(0.19)
desorption branches of the second water isotherms are given, while the distribution curves obtained from the nitrogen isotherms are shown in the lower part of Figs. 13 and 14. In the following section, the distribution curves obtained from two adsorptives are discussed in light of the pore structure of the rehydroxylated MCM-41 samples. Pore Structure Evaluated from Pore Size Distribution Curves
FIG. 13. Pore size distribution curves of rehydroxylated MCM-41 samples (PS-C12-R and PS-C14-R) calculated from water isotherms and nitrogen isotherms.
FIG. 14. Pore size distribution curves of rehydroxylated MCM-41 samples (PS-C16-R and PS-C18-R) calculated from water isotherms and nitrogen isotherms.
The peak position (rp(peak) ) and the half-width are determined from the pore size distribution curves shown in Figs. 13 and 14. The result is given in Table 4. It is seen from the half-width of Table 4 that the majority of the pores (more than 90%) are distributed within the range of ±0.1 nm from rp(peak) , indicating the cylindrical pores of the rehydroxylated MCM-41 samples (PS-C12-R, PS-C14-R, PS-C16-R, and PS-C18-R) to be fairly uniform. Finally, the pore structure of the MCM-41 samples will be discussed on the basis of the results of Figs. 13 and 14 and Table 4. The present results clearly indicate that (1) the rp(peak) value calculated from the desorption branch of the water isotherm is slightly smaller than the rp(peak) value calculated from the adsorption branch, and (2) the rp(peak) value obtained from the nitrogen isotherm is positioned near the rp(peak) values obtained from the adsorption branch of the water isotherm. The difference between rp(peak) (ads. branch) and rp(peak) (des. branch) is found to be 9, 10, 13, and 14% for PS-C12-R, PS-C14-R, PS-C16R, and PS-C18-R, respectively. This difference strongly suggests that the cylindrical pores are not uniform, but there present constrictions in the cylindrical pores. The size of constrictions is estimated to be 9–14% less than the size of the main cylindrical pores. The presence of constrictions cannot be detected by the nitrogen isotherm because the capillary condensation of nitrogen molecules occurs below the lower closure point of hysteresis. In conclusion, we have determined the pore size distribution curves of the rehydroxylated MCM-41 mesoporous silicas using the two adsorptives whose physical property is remarkably different. The calculated distribution curves give valuable information about the porous texture. It has been clarified in the present work that the water isotherms not only make an independent check on the pore size distribution estimated by the nitrogen isotherm but also can detect the presence of constrictions.
420
NAONO ET AL.
ACKNOWLEDGMENTS We express our sincere thanks to Dr. S. Inagaki for X-ray diffraction measurements and to T. Shiono, C. Oohashi, and Y. Tsukahara for performing part of the experiments.
REFERENCES 1. Kresge, C. T., Leonowicz, M. E., Roth, W. J., Vartuli, J. C., and Beck, J. S., Nature 359, 710 (1992). 2. Beck, J. S., Vartuli, J. C., Roth, W. J., Leonowicz, M. E., Kresge, C. T., Schmitt, K. D., Chu, C. T.-W., Olson, D. H., Sheppard, E. W., McCullen, S. B., Higgins, J. B., and Schlenker, J. L., J. Am. Chem. Soc. 114, 10834 (1992). 3. Inagaki, S., Fukushima, Y., and Kuroda, K., J. Chem. Soc., Chem. Commun. 680 (1993). Inagaki, S., “Synthesis and Structure of Mesoporous Crystal, FSM-16,” Ph.D. thesis. Nagoya University, 1998. 4. Rouquerol, F., Rouquerol, J., and Sing, K., in “Adsorption by Powders & Porous Solids,” Chap. 12. Academic Press, San Diego, 1999. 5. Naono, H., Hakuman, M., and Shiono, T., J. Colloid Interface Sci. 186, 360 (1997). 6. Inagaki, S., Fukushima, Y., Kuroda, K., and Kuroda, K., J. Colloid Interface Sci. 180, 623 (1996). 7. Bambrough, C. M., Slade, R. C. T., Williams, R. T., Burkett, S. L., Sims, S. D., and Mann, S., J. Colloid Interface Sci. 201, 220 (1998). 8. Hagymassy, J., Jr., Brunauer, S., and Mikhail, R. Sh., J. Colloid Interface Sci. 29, 485 (1969). 9. Hagymassy, J., Jr., and Brunauer, S., J. Colloid Interface Sci. 33, 317 (1970). 10. Hagymassy, J., Jr., Odler, I., Yudenfreund, M., Skalny, J., and Brunauer, S., J. Colloid Interface Sci. 38, 20 (1972). 11. Odler, I., Hagymassy, J., Jr., Yudenfreund, M., Hanna, K. M., and Brunauer, S., J. Colloid Interface Sci. 38, 265 (1972).
12. Naono, H., Fujiwara, R., and Yagi, M., J. Colloid Interface Sci. 76, 74 (1980). 13. Gregg, S. J., and Sing, K. S. W., in “Adsorption, Surface Area and Porosity,” 2nd ed., Chap. 5. Academic Press, London, 1982. 14. Naono, H., and Hakuman, M., J. Colloid Interface Sci. 145, 405 (1991). 15. Lippens, B. C., Linsen, B. G., and de Boer, J. H., J. Catal. 3, 32 (1964). 16. Lippens, B. C., and de Boer, J. H., J. Catal. 4, 319 (1965). 17. de Boer, J. H., Lippens, B. C., Linsen, B. G., Broekhoff, J. C. P., van den Heuvel, A., and Osinga, Th. J., J. Colloid Interface Sci. 21, 405 (1966). 18. Rouquerol, F., Rouquerol, J., and Sing, K., “Adsorption by Powders & Porous Solids,” Chap. 6. Academic Press, San Diego, 1999. 19. Skoog, D. A., West, D. M., and Holler, F. J., in “Fundamentals of Analytical Chemistry,” 7th ed., Chap. 6, Sect. 6D. Saunders College Publishing, Fort Worth, 1997. 20. Naono, H., Hakuman, M., Tsunehisa, T., Tamura, N., and Nakai, K., J. Colloid Interface Sci. 224, 358 (2000). 21. Naono, H., and Nakai, K., J. Colloid Interface Sci. 128, 146 (1989). 22. Naono, H., J. Soc. Powder Tech. Jpn. 28, 34 (1991). 23. Naono, H., and Hakuman, M., Hyomen (Surface) 29, 362 (1991). 24. Naono, H., in “Development of Porous Ceramics” (M. Hattori, and S. Yamanaka, Eds.), p. 9. CMC, Tokyo, 2000. 25. Rouquerol, F., Rouquerol, J., and Sing, K., in “Adsorption by Powders & Porous Solids,” Chap. 13. Academic Press, San Diego, 1999. 26. Chemical Society of Japan, “Handbook of Chemistry (Kagaku Binran),” revised 4th ed., Vol. II-93. Maruzen, Tokyo, 1993. 27. Morimoto, T., and Naono, H., Bull. Chem. Soc. Jpn. 46, 2000 (1973). 28. Iler, R. K., “The Chemistry of Silica,” Chap. 6. Wiley, New York, 1979. Brinker, C. J., and Scherer, G. W., “Sol-Gel Science,” Chap. 10. Academic Press, Boston, 1990. 29. Naono, H., and Hakuman, M., J. Colloid Interface Sci. 158, 19 (1993). 30. Barrett, E. P., Joyner, L. G., and Halenda, P. P., J. Am. Chem. Soc. 73, 373 (1951). 31. Kruk, M., and Jaroniec, M., Langmuir 13, 6267 (1997).