Organosulfonic acid functionalised silica mesostructured cellular foams - efficient acidic catalysts for reactions of esterification

Polish Academy of Sciences, Institute of Chemical Engineering From the SelectedWorks of Janusz J. Malinowski 2008 Organosulfonic acid functionalised silica mesostructured cellular foams - efficient acidic catalysts for reactions of esterification Janusz J. Malinowski J. Mrowiec-Bialon K. Maresz A. B. Jarzebski Available at: http://works.bepress.com/janusz_malinowski/5/

CHEMICAL AND PROCESS ENGINEERING 29, 701 711 (2008) JULITA MROWIEC-BIAŁOŃ 1*, KATARZYNA MARESZ 1, JANUSZ J. MALINOWSKI 1, ANDRZEJ B. JARZĘBSKI 1,2 ORGANOSULFONIC ACID FUNCTIONALISED SILICA MESOSTRUCTURED CELLULAR FOAMS EFFICIENT ACIDIC CATALYSTS FOR RECATIONS OF ESTERIFICATION 1 Institute of Chemical Engineering, Polish Academy of Sciences, ul. Bałtycka 5, 44-100 Gliwice, Poland 2 Departament of Chemical and Process Engineering, Silesian University of Technology, ul. M. Strzody 7, 44-100 Gliwice, Poland Structural and catalytic properties of sulfonic acid functionalised silica mesostructured cellular foams (MCFs) prepared using different precursors of sulfonic acid groups have been investigated. Catalytic activity tested in the esterification of butanol with acetic acid was found to decrease in the order MCF-et-ph-SO 3 H> MCF-ph-SO 3 H> MCF-propyl-SO 3 H. Owing to excellent structural stability, the catalysts developed can be used repeatedly without loss of activity during several cycles. Zbadano właściwości strukturalne i katalityczne mezostrukturalnych krzemionkowych pianek komórkowych (MCF) funkcjonalizowanych różnymi grupami organosulfonowymi. Stwierdzono, że aktywność katalityczna w reakcji estryfikacji butanolu kwasem octowym zmniejsza się w kolejności MCF-etph-SO 3 H> MCF-ph-SO 3 H> MCF-propyl-SO 3 H. Otrzymane materiały charakteryzują się dobrą stabilnością strukturalną. Mogą być użyte wielokrotnie jako katalizatory bez zmniejszenia aktywności. 1. INTRODUCTION Porous materials functionalised with sulfonic acid groups are promising solid acidic catalyst attracting much attention because they overcome traditional drawbacks of strong mineral acids such as difficulties in separation, corrosiveness, formation of toxic wastes and hazards in handling [1 3]. In this way, they meet the objectives set for green chemical technologies. Sulfonic acid groups can be attached to silica surface support via various organic spacers such as propyl, phenyl and ethylphenyl. There are two strategies of anchoring these groups. One is the post synthesis grafting of silica materials with organofunctionalised trialkoxysilanes ((R 1 O) 3 Si R) and the other consists in a direct co- * Corresponding author, e-mail: j.bialon@iich.gliwice.pl

702 J. MROWIEC-BIAŁOŃ et al. condensation of these precursors with tetraalkoxysilane during synthesis of silica support. Usually, tetraethoxysilane (TEOS) is used for this purpose. Mercaptopropyltrimethoxysilane (MPTMS) was the most studied precursor of organosulfonic acid sites which are obtained after oxidation of marcapto entities with acidified water solutions of H 2 O 2 [1, 4, 5]. Also phenyl groups anchored to silica surface can be sulfonated by oleum or chlorosulfonic acid to obtain acid sites [7]. Application of 2-(4-chlorosulfonylphenyl)ethyltrimethoxysilane (CSPTMS) allows one to avoid an unselective oxidation or sulfonation [5, 8] and also shortens the time of synthesis. Many types of porous silica materials were used as supports for organosulfonic acid groups but mesostructured materials of MCM and SBA-15 families synthesized by the templating method were of special interest because of their high specific surface area and large uniform pore sizes [4, 9 11]. Most recently, we have demonstrated huge potentials of siliceous mesostructured cellular foams (MCF) to afford excellent heterogeneous catalysts and biocatalysts for selective oxidation of organics and various biochemical reactions [12 15]. In contrast to hexagonally ordered MCM and SBA-15, the silica of MCF family has spherical pores with larger diameters in the range of 25 40 nm, connected by windows (diameters of 10 15 nm) and large specific surface area (500 1000 m 2 g 1 ) [16]. The accessibility to silica surface is significantly higher in these materials compared to other families; it stems from their uniquely open structure and twice as large porosity. Therefore, it seemed natural to test the potentials of MCF, with the surface modified with sulfonic groups as acidic catalysts. In the paper, we report on the catalytic and structural properties of organosulfonic acid functionalised silica mesostructured cellular foams (MCF R SO 3 H) obtained by post-synthesis grafting with three precursors: mercaptopropyltrimethoxysilane, phenyltriethoxysilane and 2-(4-chlorosulfonylfenyl)ethyltrichlorosilane. The use of grafting stems from our previous studies which clearly showed that cellular structure of silica is very well preserved after modification with functional groups, in contrast to one step synthesis in which this structure was barely seen. Phenyl functionalised MCFs were sulfonated with SO 3 or chlorosulfonic acid. Additionally, the silica surface was modified with methyl groups to impart week hydrophobic properties. Catalytic activity in esterification of butanol with acetic acid was compared with those obtained for sulfuric acid and a popular solid acid catalyst Amberlyst-15. Stability of the catalysts was tested in a number of catalytic cycles. 2. EXPERIMENTAL Preparation of the catalysts. Organosulfonic acid functionalised silica mesostructured cellular foams (MCF R SO 3 H) were prepared by two step procedure. First, pure siliceous foams were synthesized using the method proposed by Stucky at al. [16]. In briefly, surfactant Pluronic P 123 (0.4 mmol) was dissolved in 1.6 M HCl (75 cm 3 ) at

Silica mesostructured cellular foams 703 room temperature. 1,3,5-trimethylbenzene (17 mmol) and NH 4 F (0.6 mmol) were added under vigorous stirring and the mixture was heated to 40 C. Following 1 h of stirring, tetraethoxysilane (TEOS) was added (4.4 g). The mixture was stirred for 1 h and subsequently stored at 40 C for 20 h and at 100 C for 24 h. After cooling to room temperature, the precipitate was isolated by filtration, washed with water, dried at room temperature for 4 days and calcined at 500 C for 8 h. Before grafting, the MCFs were contacted with water vapour for 2 h and subsequently calcined at 200 C for 2 h to obtain surface concentration of hydroxyl groups of ca. 3 OH nm 2 [17]. Next, the sample was grafted under reflux for 24 h with suitable precursor of functional groups dissolved in organic solvent. The solvent was removed by evaporation. The following precursors were used for grafting: mercaptopropyltrimethoxysilane (MPTMS), phenyltriethoxysilane (PTES) (dissolved in hexane) and 2-(4-chlorosulfonylfenyl)ethyltrichlorosilane (CSPETCS) (dissolved in acetonitrile). The nominal content of organic groups was 1.5 mmol g 1. Mercaptopropyl groups were oxidised with H 2 O 2 [6] whereas phenyl groups were sulfonated with chlorosulfonic acid [7] or in the gas phase using 20% oleum (SO 3 /H 2 SO 4 ) vapours [7]. To improve the hydrophobicity of the catalysts, methyl groups were introduced by grafting as described above using chlorotrimethylsilane (CTMS). Nominal content of methyl groups was 1 mmol g 1. Details on surface properties of the samples are given in Table 1. Table 1. Surface properties of the MCF R SO 3 H samples Sample S1 S2 a S3 b S4 S5 c MCF Organic spacer (R) propyl phenyl phenyl phenylethyl phenylethyl S BET [m 2 g 1 ] 518 524 494 357 260 643 a Sulfonation by chlorosulfonic acid. b Sulfonation by SO 3. c Sample with CH 3 groups. V p [m 3 g 1 ] 1.80 2.10 2.00 1.60 1.26 2.30 d s [nm] 18 23 23 25 25 32 d w [nm] 9 11 11 12 12 14 [H] [mmol g 1 ] 0.27 0.58 0.48 1.1 1.1 Characterization. The textural characteristics of the samples were determined from nitrogen adsorption isotherms measured at 196 C (Micromeritics ASAP 2000). The specific surface area was determined by the BET method in the relative pressure range of 0.05 0.20. The pore size distribution was determined using the BJH model. The pore diameters of spheres were taken off from the adsorption curve while the diameter of windows from the desorption curve. Thermogravimetric measurements were made with a Mettler Toledo thermobalance (Star 851, LF/1100) using 150 μl standard platinum crucibles and sample weight ca. 15 mg. Samples were heated at the rate of 10 deg min 1 from 25 C to 800 C in

704 J. MROWIEC-BIAŁOŃ et al. nitrogen flow of 50 cm 3 min 1. The apparatus was linked to the FTIR spectrometer (Nicolet 6700) that allowed on-line analysis of evolving gases. The amount of protons [H + ] was determined based on the amount of chemisorbed ammonia obtained using Micromeritics ASAP 2010 Chemi System apparatus. After outgassing the samples at 200 C for 3 h and cooling to 100 C, first adsorption isotherm of NH 3 was measured (physically and chemically bound NH 3 ). Then, the sample was outgassed to remove physically adsorbed NH 3 and after that the second adsorption isotherm of ammonia was measured (physically bound NH 3 ). Catalytic tests. Catalytic activity of materials was tested in esterification of butanol with acetic acid in a stirred batch system at 75 C. The molar ratio of acetic acid to butanol was 1:1. Directly before using the catalysts were dried at 200 C. Their amount in the reaction was 1 wt. %. The products of reaction were analyzed by a GC HP 5890 Series II with a TCD detector. GC and 1 H NMR analysis confirmed that no by-products were formed. 3. RESULTS AND DISCUSSION Nitrogen adsorption isotherm of pure silica MCF was of type IV according to IUPAC classification (Fig. 1). The hysteresis loop at relative pressure in the range 0.8-0.9 indicates the presence of large uniform mesopores. Fig. 1. Nitrogen adsorption isotherms on pure silica MCF, functionalised with ethylphenyl-so 3 H (S4) and hydrophobised with methyl entity (S5) The pore size distributions (PDS) of spherical pores and connecting windows are shown in Fig. 2. The locus of maximum in the PDS is about 32 nm for spheres (d s )

Silica mesostructured cellular foams 705 and ca. 14 nm for windows (d w ). The mesopores volume achieved 2.3 cm 3 g 1 and the specific surface area of the sample was ca. 640 m 2 g 1. After grafting, the shape of the isotherm did not change, only the volume of adsorbed nitrogen (V p ) and the specific surface area (S BET ) decreased slightly and the PDS shifted to somewhat lower values (Table 1, Fig. 1). This confirms that cellular structure of MCFs was preserved after grafting. The decrease of S BET was caused mainly by the disappearance of small pores after their fixing with the silicates. However, the surface of larger pores was covered with functional groups which caused the decrease of their diameters and volume. The scale of changes depended on the type of the functional group. The largest decrease in S BET and V p was observed for S4 sample grafted with ethylphenylsulfonic acid groups. After its hydrophobisation with methyl groups (sample S5) additional decrease in S BET and V p was observed. The same tendency was recorded for the samples with propyl and phenyl sulfonic acid groups (data not shown here). Fig. 2. Window (a) and pore size (b) distributions in MCF, S4 and S5 samples functionalised with ethylphenyl SO 3 H (S4) and hydrophobised with methyl entity (S5) The thermal stability of a catalyst is of primary importance in terms of its catalytic properties. In this work, thermal stability of sulfonic acid functionalised MCFs was investigated by thermogravimetry. Figure 3 shows TG and DTG curves from S4. On the TG curve a mass loss in the range of 25 150 C is visible due to desorption of water and remaining solvent. Above 200 C a continuous mass loss with various rates

706 J. MROWIEC-BIAŁOŃ et al. is observed which can be better interpreted from the DTG curve. We can distinguish two well resolved regions of mass loss: i) between 200 C and 350 C attributed to removal of organic groups not tightly bound to the surface, ii) above 350 C due to thermal decomposition of ethylphenylsulfonic acid groups, which was the highest at 550 C. Fig. 3. Weight loss curve (TG) and derivative plot (DTG) for sample S4 Fig. 4. Weight loss curve (TG) and derivative plot (DTG) for sample S5 For sample S5 additional mass loss was recorded in the range of 340 460 C due to decomposition of methyl groups (Fig. 4). FTIR on line investigation of the gases released during thermal analysis confirmed the mechanisms of catalyst decomposition (data not shown here). Samples S2 and S3 containing sulfonic acid sites anchored to silica surface

Silica mesostructured cellular foams 707 by phenyl groups showed slightly higher thermal stability than sample S4. The decomposition of strongly bonded phenylsulfonic groups started at 400 C and it was the highest at 630 C. Sample S1 containing propyl groups began to decompose at temperature about 50 lower than S4 and the highest rate of the decomposition was recorded at 480 C. The data obtained by thermogravimetric analysis for sulfonic acid functionalised MCFs coincides with those obtained for other silica materials containing the same organosulfonic acid groups [8, 6]. They also indicate that sulfonic acid functionalised MCFs can be used at significantly higher temperatures than the conventional solid acid catalyst Amberlyst- 15, for which the maximum operating temperature is limited to 120 C [18]. Fig. 5. Ammonia adsorption isotherms for sample S4 The surface contents of sulfonic acid sites were measured by two step adsorption of ammonia. Figure 5 shows typical ammonia adsorption isotherms for sample S4. The isotherm obtained in first cycle comprise of physically and chemically adsorbed ammonia. During second cycle only physically bound ammonia was detected. The linear part of two adsorption isotherms extrapolated to zero pressure and the difference between the corresponding terminal adsorption values give the values of surface acid sites [H + ] listed in Table 1. The highest content of acid sites was recorded for sample S4 (1.1 mmol g 1 ). For this sample, sulfonic acid groups were directly introduced using CSPETCS, while in the case of samples S1, S2 and S3, these groups were obtained during further chemical modification of previously attached mercaptopropyl and phenyl groups. Oxidation of mercaptopropyl groups with H 2 O 2 gave the lowest value of acid sites (0.27 mmol g 1, S1). Sulfonation of fenyl groups with chlorosulfonic acid was more efficient (0.58 mmol g 1, S2) than using SO 3 (0.48 mmol g 1, S3). The catalytic activities of the samples in the esterification of acetic acid with butanol very well correlate with the measured content of sulfonic acid sites (Table 1,

708 J. MROWIEC-BIAŁOŃ et al. Fig. 6). The catalytic activity expressed as a conversion of acetic acid decreased in order S4 > S2 > S3 > S1. For sample S4 the maximum value of conversion after 6 h was ca. 65%. The conversion for sample S4 was only slightly lower than that obtained for sulfuric acid but it was about 6% higher than for Amberlyst-15, most likely due to the larger specific surface area and better accessibility of the active sites (S BET of Amberlyst-15 equals about 45 m 2 g 1 [18] and hence it is almost an order of magnitude lower than for MCFs). Fig. 6. Conversion of acetic acid in the esterification with butanol for various catalysts Fig. 7. Conversion of acetic acid for S4 and additionally hydrophobised with methyl groups (S5)

Silica mesostructured cellular foams 709 For sample S5 hydrophobised with methyl groups, the efficiency of esterification increased but only by ca. 3% (Fig. 7). It is noteworthy that water released during reaction was not removed from the reaction mixture and hence the conversion values obtained after 6 h draw on a thermodynamically limited conversion level. Fig. 8. Change in conversion of acetic acid in a batch process (6 h) with repeated use of S4 The catalytic stability was tested for S4 during repeated esterification reactions. After each cycle, the catalyst was washed with ethanol, subsequently with acidified water, then pure water and finally dried at 200 C. The catalytic activity decreased slightly after following cycles and it was about 10% lower after 6 cycles (Fig. 8). Texture properties of the catalysts were well preserved. ACKNOWLEDGEMENTS This work was partially supported by the Polish Ministry of Science and Higher Education under grant N208 020334. 4. CONCLUSIONS Structure and catalytic properties of sulfonic acid functionalised mesostructured cellular foams depend on the type of organosulfonic acid group attached to the surface. The highest catalytic activity was achieved for materials with ethylphenyl spacer which was comparable to sulfuric acid and it was better than for the commercial solid acid catalyst Amberlyst-15. Materials obtained showed appropriate structure and catalytic stability and hence can be used in repeatedly.

710 J. MROWIEC-BIAŁOŃ et al. SYMBOLS d w d s diameter of windows, nm diameter of sphere, nm 1 S BET specific surface area, m 2 g 1 V p mesopore volume, cm 3 g REFERENCES [1] MELERO J.A., VAN GRIEKEN R., MORALES G., Chem. Rev., 2006,106, 3790. [2] PETERS T.A., BENES N.E., HOLMEN A., KEURENTJES J.T.F., Appl. Catal. A: Gen., 2006, 297, 182. [3] SIRIL P.F., DAVISON A.D., RANDHAWA.J.K., BROWM D.R., J. Mol. Catal. A: Chem., 2007, 267, 72. [4] YANG L.M., WANG Y.J., LUO S.G., DAI Y.Y., Micropor. Mesopor. Mater., 2005, 84, 275. [5] MORALES G., ATHENS G., CHMELKA B.F., VAN GRIEKEN R., MELERO J.A., J. Catal., 2008, 254, 205. [6] SHYLESH S., SHARMA S., MIRAJKAR S.P., SINGH A.P., J. Mol. Catal. A: Chem., 2004, 212, 219. [7] MOHINO F., PEREZ-PARIENTE J., SASTRE E., Stud. Surf. Sci Catal., 2002, 142, 1275. [8] HAMOUDI S., ROYER S., KALIAQUINE S., Micropor. Mesopor. Mater., 2004, 71, 17. [9] DIAZ I., MARQUEZ-ALVAREZ C., MOHINO F., PEREZ-PARIENTE J., SASTRE E., J. Catal., 2000, 193, 283. [10] MBARAKA I.K., RADU D.R., LIN V.S.-Y, SHANKS B.H., J. Catal., 2003, 219, 329. [11] DAS D., LEE J.-F., CHENG S., J. Catal., 2004, 223, 152. [12] MELGUNOV M.S., CHESALOV YU.A., MROWIEC-BIAŁOŃ J., JARZĘBSKI A.B., O.A.KHOLDEEVA O.A., J. Catal., 2007, 246, 241. [13] MAKSIMCHUK N.V., MELGUNOV A M.S., MROWIEC-BIAŁOŃ J., JARZĘBSKI A.B., KHOLDEEVA O.A., J. Catal., 2005, 235, 175. [14] SZYMAŃSKA K., BRYJAK J., MROWIEC-BIAŁOŃ J., JARZĘBSKI A.B., Micropor. Mesopor. Mater., 2007, 99, 167. [15] SZYMAŃSKA K., BRYJAK J., MROWIEC-BIAŁOŃ J., JARZĘBSKI A.B., Catal. Today, 2007, 124, 2. [16] SCHMIDT-WINKEL P., LUKENS JR.W.W., YANG P., MARGOLESE D.I., LETTOW J.S., YING J.Y., STUCKY G.D., Chem. Mater. 12, 686-696 (2000). [17] MROWIEC-BIAŁOŃ J., Thermochim. Acta, 2006, 443, 49. [18] www.rohmhaas-polska.com/produkty. JULITA MROWIEC-BIAŁOŃ, KATARZYNA MARESZ, JANUSZ J. MALINOWSKI, ANDRZEJ B. JARZĘBSKI FUNKCJONALIZOWANE KRZEMIONKOWE MEZOSTRUKTURALNE PIANKI KOMÓRKOWE EFEKTYWNE KATALIZATORY DO ESTRYFIKACJI Przedstawiono wyniki badań właściwości strukturalnych i katalitycznych mezostrukturalnych pianek komórkowych funkcjonalizowanych grupami sulfonowymi, przyłączonymi do powierzchni krzemionki poprzez organiczne grupy propylowe, fenylowe oraz etylofenylowe (MCF R SO 3 H). Materiały otrzymano metodą impregnowania krzemionek typu MCF następującymi organotrialkoksysilanami: merkaptoprolylotrimetoksysilanem, fenylotrietoksysilanem oraz 2-(4-chlorosulfonylofenylo)trichlorosilanem. Grupy merkaptopropylowe utleniano za pomocą nadtlenku wodoru, grupy fenylowe sulfonowano kwasem chlorosulfonowym lub oleum. W celu zwiększenia hydrofobowości materiałów do ich powierzchni przyłączono dodatkowo grupy metylowe. Stwierdzono, że po wprowadzeniu grup funkcyjnych następowało zmniejszenie powierzchni właściwej krzemionkowych MCF oraz zmniejszenie objętości porów i ich rozmiarów. Zmniejszenie powierzchni właściwej było głównie spowodowane zanikiem małych porów na skutek przyłączenia do ich powierzchni grup funkcyjnych. Największe zmiany wielkości powierzchni

Silica mesostructured cellular foams 711 właściwej i objętości porów stwierdzono dla próbek funkcjonalizowanych grupą etylofenylosulfonową, największą pod względem rozmiarów wśród wprowadzanych grup. Otrzymane materiały charakteryzują się lepszą stabilnością termiczną niż kwasowe katalizatory polimerowe stosowane w przemyśle, np. Amberlyst-15. Największą stabilnością termiczną charakteryzowały się materiały z grupą fenylową, ich rozkład zachodził powyżej 400 C. Rozkład grup etylofenylosulfonowych rozpoczynał się natomiast w temperaturze ok. 350 C, grup propylosulfonowych zaś ok. 300 C. Materiały, w których grupy sulfonowe otrzymano w wyniku dodatkowej obróbki chemicznej uprzednio wprowadzonych grup funkcyjnych, charakteryzowały się mniejszym stężeniem kwasowych centrów, niż próbki, do których bezpośrednio wprowadzono grupy etylofenylosulfonowe. Dla próbek tych otrzymano największe wartości stopnia konwersji kwasu octowego w reakcji estryfikacji z alkoholem butylowym, zbliżone do wartości uzyskanych dla kwasu siarkowego, ale większe niż otrzymano, stosując Amberlyst-15. Zwiększenie hydrofobowości materiałów powodowało nieznaczne zwiększenie aktywności badanych materiałów. Otrzymane materiały charakteryzowały się bardzo dobrą stabilnością właściwości strukturalnych i katalitycznych w kolejnych cyklach; aktywność katalityczna w reakcji testowej zmniejszyła się o ok. 10% po sześciokrotnym użyciu. Received 14 July 2008