Incorporation of zinc into silica mesoporous molecular sieves

Incorporation of zinc into silica mesoporous molecular sieves

Microporous and Mesoporous Materials 44±45 (2001) 283±293 www.elsevier.nl/locate/micromeso Incorporation of zinc into silica mesoporous molecular si...

519KB Sizes 0 Downloads 22 Views

Microporous and Mesoporous Materials 44±45 (2001) 283±293

www.elsevier.nl/locate/micromeso

Incorporation of zinc into silica mesoporous molecular sieves ski, Anna Mackowiak Stanisµaw Kowalak *, Krystian Stawin Faculty of Chemistry, A. Mickiewicz University, Grunwaldzka 6, 60-780 Pozna n, Poland Received 16 February 2000; accepted 16 May 2000

Abstract The direct syntheses of Zn/Si mesoporous materials [S. Kowalak et al., First ISMMS, Baltimore (1998), Poster Session] allowed to introduce only a limited number of Zn, and the mesostructures became less ordered at higher Zn loading. Therefore, we have attempted to graft zinc into the well ordered and stable silica structures of MCM-41, SBA3, HMS. The properties of the resulting samples depended on the silica parent structure and on the number of zinc introduced. The structure of MCM-41 was a€ected drastically even by small amounts of zinc, whereas HMS and particularly SBA-3 retained their structure despite a relatively high zinc loading (Zn/Si ˆ 0.15). Such Zn loading was not reached for the structures obtained by means of the direct synthesis. The samples with grafted zinc showed ionexchange properties and considerable catalytic activity for 2-propanol decomposition. This was in contrast to the directly prepared samples for the series SBA-3. Ó 2001 Elsevier Science B.V. All rights reserved. Keywords: MCM-41; SBA-3; HMS; Silica mesoporous materials; Modi®cation with Zn; Grafting of Zn; Zincosilicate

1. Introduction The mesoporous materials (M41S) show high speci®c surface areas and well ordered pore systems [2,3]. The mesostructures of chemical composition involving two or more oxides usually display less ordered structures. A substantial increase in loading of the second or third component subsequently leads to total amorphization of the structure. We noticed in our earlier study on Zn±Si mesoporous materials, that only a limited amount of Zn could be introduced upon the direct synthesis into the structures of MCM-41, HMS and SBA-3 [1]. Increasing the zinc loading in the system resulted in a decreased speci®c surface area *

Corresponding author.

and in lowering of the product stability. The samples of Zn/Si ratio higher than 0.1 were always totally amorphous. In the case of the SBA-3 the mesophase was achieved only below the ratio Zn/ Si ˆ 0.04. The main reason for modifying the chemical composition is to search for catalytically active solids. Very low loadings of certain components might not a€ect the structure or its stability, but the in¯uence of the introduced component on catalytic activity could also be negligible. The main objective of this study was the post-synthesis incorporation of zinc into silica mesoporous materials (MCM-41, HMS, SBA-3), by means of grafting. We would like to introduce a considerable amount of zinc without any noticeable deterioration of the well ordered original mesostructure. It was conceivable that zinc could be linked mostly to the inner surface, without any

1387-1811/01/$ - see front matter Ó 2001 Elsevier Science B.V. All rights reserved. PII: S 1 3 8 7 - 1 8 1 1 ( 0 1 ) 0 0 1 9 4 - 9

284

S. Kowalak et al. / Microporous and Mesoporous Materials 44±45 (2001) 283±293

serious e€ect on the integrity of the pore walls. The properties of the products have been compared with those of directly synthesized zincosilicate mesostructures. The catalytic activity of the samples for 2-propanol decomposition was examined. 2. Experimental The silica mesoporous molecular sieves were prepared with either cetyltrimethylammonium bromide (CTAB) or octadecylamine (ODA) as surfactant templates. The preparation procedures were analogous to those applied for obtaining the Zn±Si materials of chosen structures [1]. The material Si-MCM-41 was prepared in the presence of CTAB (Lancaster) at pH > 10, using water glass (Prayon Rupel; SiO2 -27.3 wt.%; Na2 O13.6 wt.%) as the main substrate. The aqueous solution (25 wt.%) of surfactant was obtained by dissolving CTAB in water at 40°C. Water glass and the aqueous solution (25%) of tetramethylammonium hydroxide (TMAOH) (Lancaster) were mixed in a beaker and then admitted gradually to the surfactant solution. The resulting gel was magnetically stirred for 2 h at room temperature. The molar composition of the gel was the following: 1SiO2 :0.48Na2 O:0.25CTAB:0.08TMAOH:125H2 O. The gel was transferred to the autoclave and heated at 100°C for 72 h. Mesostructured Si-SBA-3 was also synthesized in the presence of CTAB, but TEOS (Lancaster) was used as the silicon source. CTAB was dissolved in hydrochloric acid (POCh, Poland). TEOS was mixed with a 25% aqueous solution of TMAOH. The resulting solution was added stepwise to the surfactant solution. The pH of the formed gel was below 1, i.e. below the isoelectric point of silica (2). The gel composition was as follows: 1TEOS:9HCl:0.25CTAB:0.08TMAOH:125H2 O. The initial mixture was stirred at room temperature for 2 h and then it was moved to an autoclave and heated at 100°C for 72 h. Si-HMS was prepared at pH ˆ 7 using TEOS, ODA (Riedel de Haen) and ethanol±water as the solvent. ODA was dissolved in ethanol at 50°C and then water was added. The molar proportions of the mixture were as follows: 1TEOS:0.3ODA:

32EtOH:35H2 O. The synthesis was conducted at room temperature for 72 h. The resulting samples were washed with water, extracted with ethanol and eventually calcined at 530°C for 3 h with a temperature increase of 1°C/min. The parent silica mesoporous molecular sieves were impregnated with an ethanol solution (30%) of zinc acetylacetonate (Zn(acac)2 ) (Lancaster) similarly as in the work of Inubishi et al. [4]. The amount of introduced Zn corresponded to Zn/Si ratios: 0.04; 0.1; 0.15, respectively. After evaporation of the solvent the modi®ed samples were calcined in air at 530°C for 3 h. The temperature rise was very slow (1°C/min). The obtained samples were characterized by means of standard techniques such as low angle XRD (TUR  IR (FTIR Bruker, Vector M-62, CuKa ˆ 1:54 A), 22), TEM (JOEL, JEM 1200EX II), SEM (Phillips SEM 515), elemental analysis (AAS Pye-Unicam SP90A). The adsorption measurements (Micromeritics ASAP-2010) were conducted after evacuation at 350°C. Nitrogen adsorption isotherm was measured at 196°C (77 K). Selected samples were examined for their ion-exchange properties and catalytic activity for 2-propanol decomposition. The ion-exchange was carried out at room temperature with aqueous solution (0.1 M) of CuCl2 . The 20 cm3 aliquots of solution per 1 g of the sample were employed, the procedure was repeated three times. Then the samples were washed with water until chlorides were detected in the ®ltrate. The content of copper in the sample was determined by means of atomic absorption analysis. The samples examined in catalytic tests were ®rst treated with a NH4 Cl aqueous solution and then calcined at 350°C. The tests were conducted in a pulse microreactor attached to the gas chromatograph. 3. Results and discussion The XRD data illustrated in Fig. 1 show that the initial structures HMS and SBA-3 are well preserved after incorporation of considerable amounts of zinc (Zn/Si ˆ 0.15). In the case of MCM-41, degradation of the initial structure is noticeable already for the sample of low Zn

S. Kowalak et al. / Microporous and Mesoporous Materials 44±45 (2001) 283±293

285

Fig. 1. XRD patterns of (a) the initial silica mesostructures and (b) the materials modi®ed with grafted zinc.

loading (i.e. for Zn/Si ˆ 0.04). The reason for the lower stability of MCM-41 is not quite clear for us, although the pore walls in this structure are always thinner than those in the other structures under study which make them perhaps, more vulnerable for transformation upon thermal reaction with zinc compounds. The TEM pictures con®rm the well preserved hexagonal pore system of HMS and SBA-3 modi®ed with zinc (Fig. 2a and b), which is not the case for the MCM-41 structure. Table 1 presents the in¯uence of the zinc introduced on adsorption

properties of the modi®ed samples. The changes in speci®c surface area are most distinctive for the MCM-41 series, particularly when calculated by means of the BJH method, which considers exclusively the cylindrical mesopores (20% of the initial value for the sample of the highest Zn loading). The reliability of the BJH data is also limited because part of the pores is below the threshold of mesopores (1.7±1.8 nm) [5]. We are trying to apply the DFT method for a more reliable estimation [10]. The speci®c surface area decrease calculated by means of the BET equation is

286

S. Kowalak et al. / Microporous and Mesoporous Materials 44±45 (2001) 283±293

ucts resulting from Zn-grafting on MCM-41 still show some porosity, although the pore system is not well ordered. The decrease in surface area with the Zn loading is also considerable for the series HMS. The speci®c surface area attains about 40% of the initial value for the Zn-richest sample. The pore diameters are also shifted towards the micropore range after the zinc grafting. The in¯uence of the zinc introduced is relatively low for the SBA-3 samples. The maximum loading of Zn causes a reduction of the speci®c surface area to the level of about 70% of the initial value. The pore size distribution is much less a€ected by modi®cation with zinc than the other series. The correlation between the Zn content and the speci®c surface area (Fig. 4a±c) indicates a drastic declining of area values for the samples grafted with small contents of Zn (Zn/Si ˆ 0.04), regardless of the parent structure. Further increase in zinc loading always results in a milder diminishing of the speci®c surface areas. In the case of the directly synthesized samples the decrease in the surface area is less pronounced at low zinc content but a higher zinc loading always results in forming of totally amorphous products (surface areas below 100 m2 /g). In the case of the

Fig. 2. TEM pictures of (a) the parent silica SBA-3 and (b) the sample modi®ed with grafted zinc (Zn/Si ˆ 0.15).

less drastic (40% of the original value) which results from a possible contribution of micropores generated upon a degradation of the original structure MCM-41. The pore size distribution is very much shifted towards the lower values (Fig. 3). Despite the amorphization (XRD), the prod-

Table 1 Adsorption properties of the selected samples Sample (Zn/Si)

V (m3 /g) pore volume

S (m2 /g) speci®c surface area

 D (A) pore average

 Dpred (A) predominant pores

BET

BJH…ads†

BET

BJH…ads†

BET

BJH…ads†

Zn/Si-SBA-3(0) Zn/Si-SBA-3(0.04) Zn/Si-SBA-3(0.10) Zn/Si-SBA-3(0.15)

1 0.7 0.6 0.6

1250 890 870 820

1050 850 820 780

31 30 29 27

33 31 30 28

25.5 19 21 21.5

Zn/Si-MCM-41(0) Zn/Si-MCM-41(0.04) Zn/Si-MCM-41(0.10) Zn/Si-MCM-41(0.15)

0.5 0.3 0.2 0.15

850 250a 180a 170a

1000 480 440 430

29 30a 34a 35a

29 29 22 21

27 18.1a 18.1a 18.1a

Zn/Si-HMS(0) Zn/Si-HMS(0.04) Zn/Si-HMS(0.10) Zn/Si-HMS(0.15)

0.35 0.3 0.3 0.3

760 420a 340a 190a

830 640 340 300

27 27a 26a 24a

25 24 23 22

19 18.9a 18.9a 18.9a

D ± average pore diameter; Dpred ± predominant pore diameter appointed from pore diameter distribution (adsorption branch, BJH method); S ± speci®c surface area. a Non applicable range of pore diameter for BJH method (under border of meso/micropores), geometrical parameter is not preserved (cylindrical model of pores).

S. Kowalak et al. / Microporous and Mesoporous Materials 44±45 (2001) 283±293

287

Fig. 3. Pore distribution (BJHads ) for the indicated parent silica mesophases (±±) and for the samples modi®ed with grafted Zn (Zn/ Si ˆ 0.15).

288

S. Kowalak et al. / Microporous and Mesoporous Materials 44±45 (2001) 283±293

Fig. 4. Correlation between the zinc content and speci®c surface area for the samples synthesized directly [1] and for the samples obtained by the post-synthesis grafting (±±); (a) SBA-3, (b) HMS, (c) MCM-41.

S. Kowalak et al. / Microporous and Mesoporous Materials 44±45 (2001) 283±293

directly synthesized SBA-3 series, the mesoporous structure was not achieved already for the Zn/Si ratios above 0.02 [1]. The pore volumes is very much a€ected by the incorporation of Zn into MCM-41, whereas in the case of SBA-3 and particularly in the one of HMS the introduced Zn does not in¯uence the Zngrafted samples very distinctively. The adsorption/desorption isotherms of nitrogen (Fig. 5) for the Zn modi®ed samples MCM-41 di€er very much from the parent materials and from the directly synthesized zincosilicate MCM41 [1]. The in¯exion of the isotherm noticed for the all-silica MCM-41 and for the directly synthesized zincosilicates at the p=p0 of about 0.3 is not seen for the post-synthesis modi®ed materials. The capillary condensation starts already at very low p=p0 values (0.15). For the parent silica HMS as well as for the zinc modi®ed samples isotherms of type IV show speci®c broad hysteresis loops of type H4, which can result from broad range of pore distribution and from the cage-like arrangement of pores [6]. Similarly as for previous series the introduction of zinc causes a shift of the in¯exion resulting from capillary condensation towards lower values of partial pressure. The adsorption isotherms for the structure SBA-3 are of type IV, regardless of the presence of zinc. The hysteresis loops of type H2 suggest an evidence of necking pores [7,9]. The shift in the in¯ection of the curve is almost negligible. In the case of the zincosilicate materials obtained by the direct synthesis, the isotherms are reminiscent of those of zinc-free samples. Only in the case of MCM-41 modi®ed with grafted Zn the isotherm indicates changes in the hysteresis loop from the type H4 (suggesting a slit-like pore system for the zinc-free sample) to combined system of H4 and H1, for the zinc-bearing materials. The latter can result from contribution of some macropores [8,9]. Similarly, an even more distinctive contribution of both H4 and H1 hysteresis loops is noticed for the MCM41 zincosilicates prepared directly. The IR spectra do not provide any remarkable information on the nature of zinc species in modi®ed samples. The Si±O stretching band at 1100 cm 1 is always the most distinctive one, regardless of the preparation procedure and the Zn loading.

289

The mesoporous zincosilicate materials obtained by means of our direct syntheses [1] displayed considerable ion-exchange properties for the samples MCM-41 and HMS and a lack of these properties in the case of samples SBA-3. The ®rst two series showed ion-exchange capacities proportional to the zinc content. Such properties supported the hypothesis of a tetragonal coordination for the introduced zinc and its contribution to the zincosilicate framework. The samples SBA-3 obtained in acidic medium may rather be considered as zinc silicates. It is very interesting to ®nd that the silica SBA-3 modi®ed with grafted zinc indicates very distinctive ion-exchange ability (Fig. 6). This means, that contrary to the direct synthesis in severe acidic medium, the grafting of zinc on the silica SBA-3 surface results in formation of a zincosilicate system, in which the tetragonally bonded zinc atoms play the role of anionic sites. It is likely that the acidic medium of the initial gel makes impossible the generation of anionic zinc sites, whereas the zinc-free silica SBA-3 is impregnated with zinc acetylacetonate after washing and calcination, which had made the surface neutral. Further calcination allows zinc to be incorporated into the framework. The role of cation in the system is most likely played by a proton or even by some Zn cations. The ion-exchange ability of the HMS samples obtained by the post-synthesis grafting is somewhat lower than for the directly synthesized zincosilicate. The di€erence is more distinct for the MCM-41, where grafting results in deterioration of the structure. This indicates that some part of the zinc does not form the tetra-coordinated anions combined with silicate tetrahedra. The catalytic tests for 2-propanol decomposition (Fig. 7) show an increase in activity as a function of zinc loading. The conversions over the directly prepared samples MCM-41 are distinctively higher than those of the grafted series. The lower activity of the latter is not surprising, given the deterioration of the ordered structure. It is interesting that the grafted samples still show a considerable activity, that is comparable to the other grafted series (SBA-3, HMS). The HMS samples of similar zinc loading show similar activity, regardless of the preparation procedure (either direct or post-synthesis grafting).

290

S. Kowalak et al. / Microporous and Mesoporous Materials 44±45 (2001) 283±293

Fig. 5. Nitrogen adsorption/desorption isotherms for the parent silica mesophases and for the Zn-grafted samples (Zn/Si ˆ 0.15); (a) MCM-41, (b) HMS, (c) SBA-3.

S. Kowalak et al. / Microporous and Mesoporous Materials 44±45 (2001) 283±293

291

Fig. 6. Ion-exchange capacity of the indicated samples, expressed as the Cu/Zn ratio.

The di€erences are very spectacular for the SBA-3 series, which was completely inactive when prepared directly. The appearance of activity after Zn grafting con®rms the incorporation of zinc into the framework and the subsequent generation of some acid sites. The preliminary ammonia TPD and the IR measurements with adsorbed pyridine indicate the presence of some, mostly of the Lewis type acid sites of intermediate strength [10]. All the examined samples drive the reaction towards the dehydration and propene is the main product. In some cases small amount of diisopropyl ether was detectable (particularly for the HMS series). It is worth mentioning that no acetone was found in the products which may result from a lack of a noticeable number of lone zinc oxide. 4. Conclusions Modi®cation of silica mesoporous materials MCM-41, HMS and SBA-3 allows to introduce zinc into their structures by means of impregnation with zinc acetylacetonate and subsequent calcination. The properties of the zinc grafted samples depend on the structure of the parent silica mesophase and on zinc loading. The materials HMS and SBA-3 retain their structures after

modi®cation with much higher number of zinc (Zn/Si ˆ 0.015) than the number attainable upon direct synthesis (Zn/Si below 0.1). The structure MCM-41 undergoes, on the other hand, a noticeable degradation even after introduction of a low amount of zinc. Although the XRD and TEM data suggest well preserved mesostructures for the Zn-grafted HMS and SBA-3, the adsorption properties of the resulting modi®ed samples are a€ected. The speci®c surface areas usually decline with zinc loading. The decrease for the grafted low Zn-loading samples is always more distinctive than for the respective directly synthesized zincosilicates. A higher Zn loading brings less pronounced changes. The pore size distribution after zinc grafting is usually shifted towards lower values and it approaches the level of micropores. The nitrogen adsorption isotherms of the modi®ed HMS and SBA-3 are very much like the parent materials. The di€erences are very distinctive for the MCM-41 series because of structure deterioration. The samples modi®ed with zinc indicate noticeable ion-exchange properties, which is in contrast to the parent silica and this con®rms a tetragonal coordination of zinc in the framework. It is particularly signi®cant that the Zn-grafted SBA-3 show no such properties when prepared directly. The catalytic activity noticed for 2-propanol dehydration re¯ects the presence of some

292

S. Kowalak et al. / Microporous and Mesoporous Materials 44±45 (2001) 283±293

Fig. 7. Correlation between Zn content and activity for 2-propanol decomposition. The reaction was conducted in a pulse microreactor at 230°C; (a) MCM-41, (b) SBA-3, (c) HMS.

S. Kowalak et al. / Microporous and Mesoporous Materials 44±45 (2001) 283±293

acid sites in the modi®ed samples (including deteriorated MCM-41). The further investigation of the presented materials is underway and will be published elsewhere [10].

References [1] S. Kowalak, K. Stawi nski, A. Matµoka, First ISMMS, Baltimore 1998, Poster Session. [2] C.T. Kresge, M.E. Leonowicz, W.J. Roth, J.C. Vartuli, J.S. Beck, Nature 359 (1992) 710.

293

[3] S. Inagaki, Y. Fukushima, K. Kuroda, J. Chem. Soc., Chem. Commun. (1993) 680. [4] Y. Inubishi, R. Takami, M. Iwasaki, H. Tada, S. Ito, J. Coll. Surf. Sci. 200 (1998) 220. [5] E.P. Barrett, L.G. Joyner, P.H. Halenda, J. Am. Chem. Soc. 73 (1951) 373. [6] A. Sayari, P. Liu, M. Kruk, M. Jaroniec, Chem. Mater. 9 (1997) 2499. [7] Q. Huo, R. Leon, P.M. Petro€, G.D. Stucky, Science 268 (1995) 1324. [8] C.N. Wu, T.S. Tsai, Ch.N. Liao, K.J. Chao, Micropor. Mater. 7 (1996) 173. [9] M. Kruk, M. Jaroniec, A. Sayari, J. Phys. Chem. B 101 (1997) 583. [10] S. Kowalak, K. Stawi nski, to be published.