IL“!!- "I... to}. ...r . . 115535.531, A. ‘0‘. “a... .‘ A .4 2 52.31%. NW4... 1.8..” fimfinfii . (:4;— . m-.. “dun: g, Eda...» _ I1. $18.19!?! ”4d Siam? . . luff”?! . gig .1. to" 9%.? . ihnwxriflfi. . .3. ungwittswlif! . ., .3 ., , . . ._...fla...c.:s 2...?qu . e .. . u 23.5....anaéfiffs: thg .flfiflfiiAumafi...§&i 3 3 . .. ; Miss A A l 4 IUIMI‘H“WW”WINNIEMWWWI 3 1293 017893 LIBRARY Michigan State University This is to certify that the dissertation entitled A Neutral Templating Route to MeSOporous Molecular Sieves and Their Catalytic Application for Peroxide Oxidation of Large Aromatic Molecules. presented by Peter Tanev Tanev has been accepted towards fulfillment of the requirements for A _ ' [3L7 awn 5 l‘r P D degree In /7 t/ 7%..W Wt professor 7744/; Mr / a ’ MSU i: an Affirmative Action/Equal Opportunity Institution 0-12771 PLACE IN REI‘URN Box to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECAU.ED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 1M chIRC/DatoDmpfiS-p.“ A NEUTRAL TEMPLATING ROUTE TO MESOPOROUS MOLECULAR SIEVES AND THEIR CATALYTIC APPLICATION FOR PEROXIDE OXIDATION OF LARGE AROMATIC MOLECULES By Peter Tanev Tanev A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemisz 1995 ABSTRACT A NEUTRAL TEMPLATING ROUTE TO MESOPOROUS MOLECULAR SIEVES AND THEIR CATALYTIC APPLICATION FOR PEROXIDE OXIDATION OF LARGE AROMATIC MOLECULES . By Peter Tanev Tanev Because of their uniform pore size and ability to "sieve" molecules on the basis of their size and shape microporous molecular sieves and zeolites (with pores of < 2.0 nm) are widely used in a number of adsorption, ion - exchange and catalytic processes. The recently discovered hexagonal mesoporous molecular sieves MCM-41 (with uniform pore size in the range of 2.0 - 10.0 nm) by the scientists at Mobil 1 provide an unique opportunity for a logical extension of these applications toward processing and transforming of much larger molecules. The preparation of MCM-41 was originally accomplished by a templating mechanism involving electrostatic interactions between assemblies of positively charged quaternary ammonium cation surfactants (5+) and anionic inorganic species (I') as framework precursors. Recently, Stucky, and his (:0 - workers 2 extended this electrostatic approach by providing three complementary templating pathways, namely, the charge - reversed S' I+ and the counterion (X‘ or M+) - mediated 5+ X’ I+ or, S' M+ I' routes. However, these electrostatic templating pathways afford mesoporous molecular sieves with low degree of framework cross - linking, limited framework wall thickness, large crystallite size and little or no textural mesoporosity which does not contribute to improving thermal Stability and to accessing the framework - confined meSOpores. In addition, due to the strong electrostatic interactions and charge matching the cationic template is Strongly bound to the framework and difficult to recover by non-dEStI'UCtive methods. 0,1: 81' .30“ if I. ‘.I h . HyaWd": 'I " H‘- .5...‘ v 1 , .A..I ' ”R “It "a.- A t ‘ . ( I ‘-"".'.1.‘ S uh- . ).“' I :2 lack c n'~¢‘.fl 0‘ D L,‘u.A.-4. . l . T1523. 3 U4 utA-i (1 ‘ r "‘ '7550 m; ;. ..~_ - ( 1.2m“- ."N that .QI A Our efforts to circumvent these limitations have materialized into a neutral templating approach 3 (8° 1°) which allows for the preparation of hexagonal mesoporous molecular sieves (denoted HMS) with: (i) more completely cross - linked framework; (ii) thicker pore walls; (iii) superior thermal stability; (iv) small crystallite size and complementary textural mesoporosity for better access of the framework - confined mesopores. Due to the lack of electrostatic interactions the method allows for the effective and environmentally benign recovery and recycling of the neutral template by simple solvent extraction. A new I° S°-S° I° biomimetic templating approach to the synthesis of lamellar silicas with high degree of framework cross-linking, exceptional thermal stability, sorption properties typical of a pillared lamellar material, and specific surface area and pore volume similar to that of the MCM-41 and HMS is demonstrated. The simultaneous biomimetic templating of silica layers and intragallery pillars occurs in the intralayer regions of multilamellar vesicles of neutral diamine surfactant. This biomimetic method also provides for the efficient and environmentally benign recycling of the neutral template by simple solvent extraction. The new 8° 1° templating approach has been used to prepare transition metal - substituted HMS derivatives. 4'5 These new mesoporous metallosilicates exhibit exceptional catalytic activity for peroxide oxidation of bulky aromatics with kinetic diameters that are too large (larger than 0.6 nm) to access the pore structure of the conventional microporous transition metal - substituted molecular sieves such as titano- and vanadosilicates. 4'5 Due to its complementary textural mesoporosity Ti-HMS exhibits superior catalytic activity for peroxide oxidation of the bulky 2,6 -di-tert-butyl-phenol relative to the electrostatic Ti-MCM-41 counterparts. ; Sect I. Stilt] Bugs. Li's; Q; Frail if? P. References 1. Beck, 1. 5.; Vartuli, 1. C.; Roth, W. 1.; Leonowitz, M. E.; Kresge, C. T.; Schmitt, K. D.; Chu, C. T-W.; Olson, D. H.; Sheppard, E. W.; McCullen, S. B.; Higgins, 1. B.; Schlenker, 1. L. I. Amer. Chem. Soc. 1992, 114, 10834. 2. Huo, Q.; Margolese, D. I.; Ciesla, U.; Feng, P; Gier, T.; Sieger, P.; Leon, R.; Petroff, P. M.; Schiith, F.; Stucky, G. D. Nature 1994, 368, 317. 3. Tanev, P. T.; Pinnavaia, T. 1. Science, 1995, 267, 865. 4. Tanev, P. T.; Chibwe, M.; Pinnavaia, T. 1. Nature, 1994, 368, 321. 5. Pinnavaia, T. 1.; Tanev, P. T.; Wang, 1.; Zhang, W. In Advances in Porous Materials; Komarneni, Sh.; Beck, 1. S; Smith, D. M., Eds, Mat. Res. Soc. Symp. Proc., vol. 371, (Pittsburgh, MRS, 1995) pp. 53 - 62. To my Wife, my Son, and my Parents "‘ '1‘“) 02.7.7-1." ACKNOWLEDGMENTS I would like to acknowledge Dr. T. 1. Pinnavaia for his guidance throughout my graduate studies here at Michigan State University. I am grateful for the many, many exiting and inspiring discussions we have had and I hope that we will still have. Most of all I would like to thank him for giving me the freedom to do the research that I wanted to do. I hope he does not regret that now. I also wish to say that his help and understanding are very much appreciated. I would also like to express my gratitude to Dr. M. G. Kanatzidis for his quidance and suggestions in editing this dissertation as well as Dr. Ledford and Dr. Miller for their critical suggestions. Drs. Stan Flegler and Karen Klomparens are gratefully acknowledged for teaching me the essence of SEM and TEM and for giving me the opportunity to teach Scanning Electron Microscopy graduate classes at the Center for Electron Optics. Without this knowledge and experience I would have not be able to make some of the most important points in this work that "obvious". I would like to thank Professors 1. Wang, W. Zhang and Dr. Malama Chibwe for their friendship and for the fruitful collaboration. Without their efforts and help the catalytic part of this work would not have been the same. Financial support from the National Science Foundation, through a research assistantship (1993 - 1995) is greatly appreciated. I would like to also express my deepest gratitude to my wife Nelly for her constant support, encouragement and endless patience during my long nights at the Department. Finally, I would like to thank my Parents and my son Stan for their support and encouragement. vi 1510i 1. ROPE “SOFA 1am tom; mm 1m 5th TABLE OF CONTENTS page LIST OF TABLES .................................................................................................... xi LIST OF FIGURES .................................................................................................. xiii LIST OF ABBREVIATIONS ................................................................................. xx CHAPTER ONE MESOPOROUS MOLECULAR SIEVES, THEIR PREPARATION, PROPERTIES AND CATALYTIC APPLICATIONS FOR TRANSFORMATION OF LARGE ORGANIC MOLECULES, A REVIEW .............................................................................................................. 1 A. Introduction ...... . ................................................................................................ 2 1. Classification of porous materials .......................................................... 2 2. Zeolites and molecular sieves ................................................................. 2 3. Templating approaches to microporous molecular sieves ............... 5 4. Applications of zeolites and molecular sieves .................................... 7 B. Recent advances in the field of microporous molecular sieves ............ 10 C. Mesoporous molecular sieves ....................................................................... 12 1. Assemblies of surfactant molecules as templates - a milestone to the synthesis of mesoporous molecular sieves ................................................. 12 2. Mobil’s M418 family of mesoporous molecular sieves - preparation ........................................................................................................ 14 3. Templating pathways to mesoporous molecular sieves .................... 18 4. Properties of the mesoporous molecular sieves prepared by different templating approaches ................................................................... 23 vii 1 mm: myoqa 1. been solemn: ‘ Cara}; getes ...... litmus... INTER 7 Milli 1315 ...... start--. i lnl'IOduc 3' Aperim '1 Results .' 5- Condus hittinceg D. Catalytic applications of mesoporous molecular sieves involving large organic molecules .................................................................................. 1. Important catalytic applications of microporous zeolites and molecular sieves ............................................................................................... 2. Catalytic applications of metal - substituted mesoporous molecular sreves--- .................................................................................................. References ................................................................................................................ CHAPTER TWO A NEUTRAL TEMPLATING ROUTE TO MESOPOROUS MOLECULAR SIEVES - - - .......................................................................................... Abstract -- - - ......................................................................................... A. Introduction. ..................................................................................................... B. Experimental ..................................................................................................... C. Results and Discussion .................................................................................... D. Conclusion - ...................................................................... References and Notes ............................................................................................ CHAPTER THREE A COMPARISON OF MESOPOROUS MOLECULAR SIEVES PREPARED BY IONIC AND NEUTRAL SURFACT ANT TEMPLATING APPROACHES - - - ...................................................... Abstract - .......................................................... A. Introduction ...................................................................................................... B. Experimental ..................................................................................................... 1. Materials ....................................................................................................... 2. Synthesis .............................................................................. 3. Template removal ..................................................................................... 4. Analytical techniques ................................................................................ viii 31 31 38 43 45 46 52 53 55 56 57 58 58 58 59 59 Islam-42.3 ”Km «in...» ‘ I v . C. Results and Discussion .................................................................................... 60 1. Structure and morphology ....................................................................... 60 2. Sorption properties and framework wall thickness ............................ 68 3. 295i MAS NMR ............................................................................................ 76 4. Template removal and thermal stability ............................................... 78 5. Chemical analysrs ............ 85 6. Mechanistic considerations ....................................................................... 86 D. Conclusions ....................................................................................................... 91 References ................................................................................................................ 92 CHAPTER FOUR BIOMIMETIC TEMPLATING OF A POROUS LAMELLAR SILICA BY MULTILAMELLAR VESICLES OF A NEUTRAL DIAMINE SURFACTANT ................................................................. 94 Abstract ........... - ..................................................................................... 95 A. Introduction ...................................................................................................... 96 B. Experimental ..................................................................................................... 97 C. Results and Discussion .................................................................................... - 99 D. Conclusions -- - ............................................................................... 116 References ................................................................................................................ 117 CHAPTER FIVE NEW Ti - SUBSTITUTED MESOPOROUS MOLECULAR SIEVES FOR CATALYTIC OXIDATION OF AROMATIC COMPOUNDS ....................... 119 Abstract ..................................................................................................................... 120 A. Introduction ...................................................................................................... 121 B. Experimentd ..... -- - . ..................................................................... 121 C. Results and Discussion .................................................................................... 122 References - .......................................................... 129 ix Er! CHAPTERSIX Ti - SUBSTITUTED MESOPOROUS MOLECULAR SIEVES FOR CATALYTIC OXIDATION OF LARGE AROMATIC COMPOUNDS PREPARED BY NEUTRAL TEMPLATING ROUTE ...................................... 130 Abstract ..................................................................................................................... 131 A. Introduction ...................................................................................................... 132 B. Experimental ..................................................................................................... 133 1. Materials ........................................................................................................ 133 2. Characterization ........................................................................................... 134 3. Catalytic evaluation ..................................................................................... 135 C. Results and Discussion .................................................................................... 135 References and Notes ............................................................................................. 143 LIST OF TABLES page Table 1. Major commercial zeolite processes .................................................. 9 Table 2. Properties of calcined mesoporous silicas assembled by different templating routes ................................................................................... 50 Table 3. Properties of calcined mesoporous silicas assembled by different templating routes ............................................................. 73 Table 4. Sorption properties of calcined mesoporous silicas assembled by different templating routes .......................................................... 75 Table 5. Chemical analysis data for as - synthesized and ethanol extracted mesoporous molecular sieves prepared by different templating pathways from Cup“ and C12° surfactants ................. 85 Table 6. Properties of the TPLM samples prepared by neutral templating 114 Table 7. Specific surface areas and catalytic oxidation properties of Ti- I-IMS, Ti-MCM-41 and related materials ......................................... 127 Table 8. Effect of the nominal Ti-loading on the catalytic activity for peroxide oxidation of 2,6-DTBP over mesoporous molecular sieves ....... - xi Table 9. Effect of the substrate : H202 ratio on the catalytic activity of 10 "/0 Ti-HMS .............................................................................................. 140 Table 10. Effect of the method of template removal on the catalytic activity of 1% Ti-substituted mesoporous molecular sieves ..................................................................................................... 141 xii LIST OF FIGURES page Figure 1. A schematic representation of the formation of the aluminosilicate members of sodalite family of zeolites. The different frameworks are obtained by putting together sodalite cages and other polyhedron units in a different spatial arrangement ......................................................................... 3 Figure 2. A mechanism of formation of high silica microporous zeolite ZSM-S ................................................................................................... 6 Figure 3. Uniform pore size and framework structure of some of the well known microporous zeolites and recently developed molecular sieves ................................................................................. 11 Figure 4. Schematic illustration of the rearrangement of a layered silicic acid host (kanemite) into a ”ribbon - candy” - like structure in the presence of assemblies of long chain quaternary ammonium cations as templates .................................................... 13 Figure 5. (A) Mobil S“ I' mechanistic routes for the formation of the hexagonal MCM-41: route (1) liquid crystal phase initiated and route (2) silicate anion initiated. (B) Transmission electron micrographs of MCM-41 samples with uniform mesopore sizes of (a) 2.0, (b) 4.0, (c) 6.5, and (d) 10.0 nm ........... 15 xiii Figure 6. Powder X - ray diffraction patterns of calcined (A) hexagonal MCM-41, (B) cubic Ia3d phase and (C) as - synthesized unstable lamellar phase prepared by S+ I'templating mechanism ........................................................................................... Figure 7. A mechanism for the 8*“ 1‘ formation of MCM—41 proposed by Davis et al ............................................................................................. Figure 8. Schematic representation of the four complementary electrostatic templating pathways to ordered mesostructures. Figure 9. A neutral 5° 1° templating mechanism to mesoporous molecular sieves ................................................................................ Figure 10. N2 adsorption - desorption isotherms for (A) amorphous silica, zeolite NaY, MCM-41 prepared by the S+ 1' pathway, 8 and (B) HMS prepared by the neutral 5° 1° templating route. Figure 11. (A) Schematic representation of the TS-l solid framework (left) and corresponding pore network (right). (B) A formation of an active Ti - peroxocomplex upon addition of aqueous H202 to site - isolated tetrahedral Ti atoms in the framework of TS -1 .......................................................................... Figure 12. Powder XRD patterns of (A) HMS, (B) HMS-C, and (C) HMS- ER vvv'vv'v ‘33‘ ‘ vv v vvvvvv v ““‘ xiv 17 19 29 32 47 Figure 13. Nitrogen adsorption (—) and desorption (- -) isotherms for (A) HMS-C and (B) HMS-EB. Insert: The corresponding Horvath - Kawazoe pore size distribution curves; dW/dR is the derivative of the normalized nitrogen volume adsorbed with respect to the pore diameter of the adsorbent ................... Figure 14. Powder XRD patterns of calcined mesoporous molecular sieves templated with neutral (Cn°) and cationic (Cn+) surfactants of different alkyl chain lengths (n = 8, 10, 12, 14, 16, and 18). (A) HMS silicas prepared by 8° 1° templating method, (B) MCM-41 silicas prepared by 8+ 1' route, and (C) MCM-41 silicas prepared by the acidic S+ X' 1+ pathway .............................................................................................. Figure 15. Powder XRD diffraction patterns of (A) as - synthesized and (B) calcined (650°C) 5*" I' MCM-41 samples prepared from _ C13+ surfactant .................................................................................... Figure 16. Representative TEM micrographs for calcined (A) 8° 1° HMS and (B) 8* X' 1*“ MCM-41 samples .................................................. Figure 17. Representative SEM micrographs of calcined (A) 8° 1° HMS, (B) St I' MCM-41, and (C) S+ X' It MCM-41 samples ................. Figure 18. N2 adsorption - desorption isotherms for calcined 5° 1° HMS samples prepared with neutral primary amine surfactants (Cn°) of different chain lengths (n: 8, 10, 12, 14, 16, 18) 000000000000000000000000000000000000000000 XV 48 62 63 65 67 69 1"”! LI...” I'LL-'4- .J Figure 19. Figure 20. Figure 21. Figure 22. Figure 23. Figure 24. N2 adsorption - desorption isotherms for calcined S+ 1' MCM- 41 samples prepared with quaternary ammonium cationic surfactants (Cn+) of different chain lengths (n= 8, 10, 12, 14, 16) ......................................................................................................... N2 adsorption - desorption isotherms for calcined St X' I+ MCM-41 samples prepared with quaternary ammonium cationic surfactants (Cn+) of different chain lengths (n= 14, 16, 18) .................................................................................................. 298i MAS NMR spectra of as - synthesized (A) 5° 1° HMS, (B) 5* I' MCM-41, (C) 8* X‘ 1+ MCM-41, and calcined (D) 5° 1° HMS, (E) 8* I' MCM-41, (F) 8+ X“ 1+ MCM-41 mesoporous molecular sieves ................................................................................ Powder XRD patterns of as - synthesized, ethanol - extracted, ethanol - extracted and calcined mesoporous molecular sieves (A) 8° 1° HMS, (B) 8+ 1' MCM-41, and (C) 5* X' It MCM-41 .............................. TGA curves of as - synthesized, ethanol - extracted, ethanol - extracted and calcined mesoporous molecular sieves (A) 5° 1° HMS, (B) 3+ 1‘ MCM-41, and (C) 8* X' I+ MCM-41 ................. Nitrogen adsorption (—) and desorption (- -) isotherms for (A) calcined HMS and (B) ethanol - extracted HMS. Insert: The corresponding Horvath - Kawazoe pore size distribution curves; dW/dR is the derivative of the normalized nitrogen volume adsorbed with respect to the pore diameter of the adsorbent- ..... ‘ OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO xvi 71 72 79 81 Figure 25. Powder XRD patterns of (A) 5° 1° HMS, (B) St I‘ MCM-41, and (C) 5+ X‘ 1* MCM-41 samples calcined in air at 900°C for 4 hr. 84 Figure 26. Schematic representation of the 5° 1° templating mechanism of formation of HMS mesoporous molecular sieves ................ 88 Figure 27. 14N NMR spectra of (A) aqueous solution of intentionally protonated DDA (molar composition 0.27 DDA : 0.054 HCl : . 9.09 EtOH : 29.6 H20), (B) the supernatant liquid separated from the HMS+ reaction product prepared from templating - solution (A), (C) the wet HMS+ solid product separated from " ’ the reaction product obtained from templating solution (A), (D) our neutral templating solution (molar composition 0.27 DDA : 9.09 EtOH : 29.6 H20), and (E) fresh HMS reaction product prepared from templating solution (D) ....................... 89 Figure 28. Powder X - ray diffraction patterns of (A) as - synthesized, (B) ethanol - extracted, and (C) ethanol - extracted and calcined TPLM - . ........................................ 100 Figure 29. Transmission electron micrographs of the as - synthesized TPLM sample clearly showing regions of (A and B) multilamellar vesicles and (B and C) lattice fringes of well - ordered lamellar TPLM phase ........................................................ 101 Figure 30. Scanning electron micrographs of (A and B) the as - synthesized, and (C) calcined TPLM sample ............................ 105 xvii Figure 31. Figure 32. Figure 33. Figure 34. Proposed biomimetic I° S°-S° I° templating mechanism of formation of TPLM ........................................................................... 107 29Si MAS NMR spectra of (A) as - synthesized, (B) ethanol - extracted, and (C) calcined TPLM ................................................... 109 TGA curves of (A) as - synthesized, (B) ethanol - extracted, and (C) ethanol - extracted and calcined TPLM .......................... 110 N2 adsorption - desorption isotherms for (A) calcined A1 - pillared montmorillonite, (B) calcined TPLM, (C) ethanol - extracted TPLM, (D) ethanol - extracted and calcined TPLM, and (E) calcined MCM-41 ................................................................. 112 Figure 35. Horvath - Kawazoe pore size distribution curves for (A) Figure 36. Figure 37. calcined Al - pillared montrnorrilonite, (B) calcined TPLM, (C) ethanol - extracted TPLM, (D) ethanol - extracted and calcined TPLM. Insert: the Horvath - Kawazoe pore size distribution of calcined mesoporous MCM-41 ........................... 113 X - ray powder diffraction patterns of mesoporous molecular sieves -- - ................. . ................................................................ 123 (A) Transmission electron micrograph of the hexagonal pore structure of Ti-MCM—41. (B) Electron diffraction pattern of Ti-HMS ................................................................................................ 124 xviii ‘. "2 JWF.-1'2‘T“ Figure 38. Nitrogen adsorption - desorption isotherms for mesoporous molecular sieves (Ti-HMS, HMS, and Ti-MCM-41), microporous titanium silicalite (TS-1), and nonporous anatase (TiOz). Insert: The corresponding Horvath - Kawazoe pore size distribution curves .......................... -. ............ 125 Figure 39. Powder XRD pattern of 1% Ti-HMS sample calcined at 650°C for 4 h ................................................................................................... 135 Figure 40. Nitrogen adsorption-desorption isotherms for (A) calcined 1% Ti-HMS and (B) ethanol-extracted 1% Ti-HMS sample... 136 Figure 41. Horvath-Kawazoe pore size distribution curves for (A) calcined 1% Ti-HMS and (B) ethanol - extracted 1% Ti-HMS sample ............................................................................................... 138 Scheme 1. Reactions involving large organic molecules catalyzed by metal - substituted MCM-41 and HMS as reported to date ..... 36 LIST OF THE ABBREVIATIONS IUPAC - International Union of Pure and Applied Chemistry. TEOS - tetraethyl orthosilicate: Si(OC2H5)4. TEOT - tetraethyl orthotitanate: Ti(OC2H5)4. TIPOT - tetraisopropyl orthotitanate: Ti[OCH(CH3)2]4. 'TMAOH - tetramethylammonium hydroxide: (CH3)4NOH. TEAOH - tetraethylammonium hydroxide: (C2H5)4NOH. TPAOH - tetrapropylammonium hydroxide: (C3H7)4NOH. TPA+ - tetrapropylammonium cation: (C3H7)4N+. TBA+ - tetrabutylammonium cation: (C4H9)4N+. CTMABr - cetyltrimethylammonium bromide: [C16H33(CH3)3N]Br. DDA - dodecylamine: C12H25NH2. DADD - 1,12-diaminododecane. Cu+ - quaternary ammonium cation surfactant with alkyl chain length n = 8, 10, 12, 14, 16, 18. Cn° - neutral primary amine surfactant with alkyl chain length n = 8, 10, 12, 14, 16, 18. i - PrOH - isopropyl alcohol: (CH3)2CHOH. EtOH - ethyl alcohol: C2H50H. H - bonding - hydrogen bonding. XRD - X - ray diffraction. TEM - transmission electron microscopy. SEM - scanning electron microscopy. TGA- thermogravimetric analysis. NMR- nuclear magnetic resonance. SBET - specific surface area in m2/ g obtained from the linear part of the Brunauer - Emmett - Teller equation. BET - Brunauer - Emmett - Teller. Vt - total volume of pores in cm3/ g. Pi/Po - relative pressure. Pi is the equilibrium pressure of the adsorbate and P0 is the saturation pressure of the adsorbate at the temperature of the adsorbent, volume adsorbed is at standard temperature and pressure. -IX'. nan-*1 Ira-M,- ‘ . .rb nlfl‘ HK - Horvath - Kawazoe pore size distribution. FID - flame ionization detector. l-D - one - dimensional. 2-D - two - dimensional. 3-D - three - dimensional. AlPO - aluminophosphate molecular sieve. VAPO - vanadium phosphate molecular sieve. M418 - broad family of mesoporous silica - based molecular sieves with lamellar, hexagonal or cubic structure. MCM-41 - Mobil Composition of Matter number 41 possessing long range hexagonal order (a member of the M415 family). MCM-48 - Mobil Composition of Matter number 48 possessing cubic symmetry (a member of the M413 family). HMS - mesoporous molecular sieve prepared by neutral templating possessing short-range hexagonal order. TPLM - templated porous lamellar material prepared by biomimetic assembly in multilamellar vesicles of neutral diamine surfactant. TS-l - microporous titanium - substituted silica molecular sieve with MFI topology (analogous to ZSM-S). TS—2 - microporous titanium- substituted silica molecular sieve with MEL topology (analogous to ZSM—ll). TSG - titanium - substituted silica gel. Ti-ZSM-48 - microporous titanium - substituted silica molecular sieve with ZSM-48 topology. ETS-10 - Engelhard Corporation titanosilicate molecular sieve (microporous) with titanium in tetrahedral and octahedral coordination. VS-l - microporous vanadium - substituted silica molecular sieve with MFI topology. VS-2 - microporous vanadium - substituted silica molecular sieve with MEL topology. Al-MCM-41- aluminum - substituted analog of MCM-41. Ti-MCM-41 - titanium - substituted analog of MCM-41. V-MCM-41 - vanadium - substituted analog of MCM—41. Ti-HMS - titanium - substituted hexagonal mesoporous molecular sieve prepared by neutral surfactant templating method. 2.6 - DTBP - 2,6 - di- tert - butylphenol. CHAPTER ONE MESOPOROUS MOLECULAR SIEVES, THEIR PREPARATION, PROPERTIES AND CATALYTIC APPLICATIONS FOR TRANSFORMATION OF LARGE ORGANIC MOLECULES, A REVIEW A. Introduction 1. Classification of Porous Materials Porous materials created by nature or by synthetic design have found great utility in all aspects of human activity. Their pore structure is usually formed in the stages of crystallization or subsequent treatment and consists of isolated or interconnected pores that may have similar or different shapes and sizes. 1 The pore shape can be roughly approximated to cylindrical, slit - shaped, ink - bottled or cone - shaped. Depending on the predominant pore size, the solid materials are classified by IUPAC as: (i) microporous, having pore sizes below 2.0 nm; (ii) macroporous, with pore sizes exceeding 50.0 nm; and (iii) mesoporous, with intermediate pore sizes between 2.0 and 50.0 nm. 2 The use of macroporous solids as adsorbents and catalysts is relatively limited due to their low surface area and large non - uniform pores. Microporous and mesoporous solids, however, are widely used in adsorption, separation technology and catalysis. 3'5 Owing to the need for higher accessible surface area and pore volume for efficient chemical processes, there is a growing demand for new highly stable mesoporous materials. Porous materials can be structurally amorphous, paracrystalline, or crystalline. Amorphous materials, such as silica gel or alumina gel, do not possess long range order, whereas paracrystalline solids, such as y- or n- A1203 are quasiordered as evidenced by the broad peaks on their X - ray diffraction patterns. 5'5 Both classes of materials exhibit broad distribution of pores predominantly in the mesoporous range. This broad pore size distribution limits the shape - Selectivity and the effectiveness of the adsorbents, ion - exchanges and catalysts Prepared from amorphous and paracrystalline solids. The only class of porous materials possessing narrow pore size distributions or uniform pore sizes includes crystalline zeolites and related molecular sieves. 2. Zeolites and Molecular sieves Zeolites are naturally occurring or synthetic highly crystalline aluminosilicates with a general chemical formula [Mx/n(AlOz)]x. [SiOZ]y . rd‘ 1-‘ _ Junie. 3"“ mH20. Their frameworks are negatively charged due to the replacement of Si 4+ by Al 3+. In natural zeolites the charge is balanced by M monovalent or polyvalent metal cations such as Na+ or Ca2+. Here n represents the formal positive charge on the cation. In synthetic zeolites the charge could also be balanced by an organic cation, for example, tetrapropylammonium ion. In both cases the cation not only compensates the charge of the framework but also plays an important role as a space filling agent, structure directing agent or template. According to their Si/Al framework ratio the zeolites are classified as high silica zeolites Si/Al > 5, intermediate 5 2 Si/A1> 2, and low silica zeolites 1 S Si/Al S 2. 7 As the Si/Al ratio increases the zeolite framework becomes more hydrophobic. Thus, high silica zeolites exhibit high adsorption affinity for organic moieties, whereas low silica zeolites (because of their Lewis and Bransted acidity) adsorb predominantly water. Figure 1 illustrates the formation of a typical zeolite lattice from elementary framework building blocks. Wpe A Socialite Faujaslte (Time X. Y) Figure 1. A schematic representation of the formation of the aluminosilicate n"'len'ibers of sodalite family of zeolites. The different frameworks are obtained by putting together sodalite cages and other polyhedron units in a different spatial arrangement. 8 The straight lines here represent oxygen bridges and the vertices are occupied by the Al or Si atoms. The blocks are usually polyhedron units that are formed during synthesis by cross - linking of the T04 tetrahedra (where T is Al and Si). The spatial arrangement of these framework units is orchestrated by the cationic template. For example, the formation of the sodalite cage occurs by cross - linking of the A102' and SiOz tetrahedra around the cationic template. This gives rise to the corresponding sodalite truncated octahedron building block (Figure 1). As shown in Figure 1 further cross - linking between these sodalite cages (by sharing the apical oxygen atoms) can generate different 3-D zeolite frameworks such as that of sodalite, zeolite A or faujasite (zeolite X, Y). However, the formation of the last two structures is accomplished trough the participation of smaller building blocks denoted by zeolite chemists as double 4 ring (D4R) or double 6 ring (D6R) building blocks. 3 All known zeolites and molecular sieves are formed in a similar fashion by' putting together polyhedron units with different shapes and sizes. 9:10:13 The pore network of a typical zeolite, which is confined by the solid framework, consists 0f cavities and connecting windows of uniform size occupied by cations and Water molecules. Because of their uniform pore size and ability to ”sieve” molecules by shape and size, zeolites are considered as a subclass of molecular sieves. Molecular sieves are also crystalline framework materials but with non - aluminosilicate nature. The tetrahedral positions in the framework of molecular sieves are occupied solely by Si or by combinations of other T atoms such as Al, Ga, Ti, V, and P. The pure SiOz molecular sieves are neutral due to the charge compensation between the +4 silicon and the four -2 O atoms (each of these 0 atoms is shared between two Si tetrahedra). In the aluminophosphate based mOIecular sieves, for example, the ratio of Al / P is kept at 1 and the framework is again neutral due to charge compensation between the A104; and P04+ te’tl‘ahedra. There are many natural zeolites, few natural molecular sieves and large and continuously growing number of synthetic zeolites and molecular sieves with or without natural counterparts. 13 an: 3. Templating Approaches to Microporous Molecular Sieves Synthetic zeolites and molecular sieves are prepared under hydrothermal conditions from aluminosilicate or metallophosphate gels. The synthesis of the first zeolite without synthetic counterpart was accomplished by Barrer in 1948. 11 This date marks the beginning of an extensive use of alkali - metal aluminosilicate gels to crystallize zeolite structures. It has been postulated that the hydrated alkali metal cations play a structure directing role. The first synthesis of a zeolite (zeolite A) using an organic template, in particular the triInethylammonium cation, was reported by Barrer and Danny in 1961. 12 Since that pioneering work a myriad of different organic directing agents (with one or more functional groups) were put to test in the search of new and more exotic zeolite structures. It didn’t take very long to realize that the proper selection of template is of extreme importance for the preparation of a Particular framework and pore network. Thus, today we have hundreds of molecular sieves with more than 85 different frameworks that are classified and well systematized by Meier and Olson in an Atlas of Zeolite Structure Types. 13 Excellent up to date reviews of the use of various organic templates and their corresponding structures, as well as the mechanism of structure directing are given elsewhere. 14'” Here, we will briefly illustrate some of the conventional preparation approaches to high silica zeolites and pure silica molecular sieves. The synthesis of high silica zeolites usually involves the addition of charged organic molecules (mostly quaternary ammonium cations) to the aluminosilicate or silica gel. The obtained reaction mixture is then placed in an autoclave and heated at high temperature (from 100 to 200°C) for a Pmlonged period of time (1 to 180 days) to afford the crystalline product. A detailed review on the preparation approaches to high silica microporous ZEOIites was made available by Peter 1acobs and Johan Martens. 18 According to Davis 15 the role of the organic molecules in the synthesis of high silica ZeOlites is that of (i) space - filling species, (ii) structure - directing agents, or (iii) templates. A typical example of space - filling function is illustrated by the fact that 22 and 13 different organic molecules can be used to synthesize zeOlites ZSM-S and ZSM-48, respectively. 19 When a specific structure is Prepared in the presence of a particular organic molecule we talk about a structure - directing role of the organic molecule. For example, the preparation of the high silica zeolite SSZ-26 was accomplished by the structure - directing role of an exotic organic ion N,N,N,N’,N',N’- hexamethyl-8,11- [4.3.3.0] dodecane diammonium cation. 20 The “templating” function also involves crystallization of the molecular sieve in the presence of single organic species. However, the absence of a free rotation of the organic guest in the pore network of the inorganic host is considered to be specific for a templating function. For example, the preparation of ZSM-18 was accomplished in the presence of a tri-quat ammonium cation (C13H36N+). The cage of this molecular sieve was found to have the same tri - fold rotational symmetry as the organic guest. 15 The lack of rotational freedom of the tri- quat molecule in the cavity of zeolite ZSM-18 was attributed to the true templating role of this organic guest. The preparation of high silica zeolites such as ZSM-5, ZSM-ll, ZSM-12 and ZSM-23 is usually performed in the presence of small quaternary ammonium ”templates” such as tetrapropyl, tetrabutyl or tetraethyl ammonium ions and di- trimethyl- diammonium heptane, respectively. Very recently, Davis et al. elaborated on the mechanism of formation of high silica microporous zeolites, in particular ZSM-S (see Figure 2). 21 gig“! 2. A mechanism of formation of high silica microporous zeolite ZSM- . 1 According to the proposed mechanism the organic cations tend to organize water molecules in their vicinity. The free energy of this templating solution is minimized by hydrophobic hydration of single organic cations. Upon the addition of a silicate precursor, the negatively charged silicate species (at high pH the silica species are negatively charged 5) begin to condense and polymerize on the surface of the single organic cations giving rise to a long - range ordered zeolite framework. Thus, the geometry of the single organic molecules is transferred into the particular ZSM-S inorganic framework. It is important to emphasize that the cationic template arranges the T04 tetrahedra in the framework walls of the microporous molecular sieves in very ordered fashion. This is evidenced by the multiple reflections observed on the corresponding X - ray diffraction patterns. 22 Primary amines and diamines such as propylamine, i— propylamine, 23 diaminopentane, diaminohexane and diaminododecane 24 also were found to direct the synthesis of the ZSM-S structure. However, Hearmon et al. have p0inted out 25 that the above framework assembly process (under the specific reaction conditions) is directed by the protonated forms of these amines. On the other hand, the use of neutral primary amines of short alkyl chain lengths afforded predominantly microporous molecular sieves. In summary, most of the existing microporous high silica zeolites and molecular sieves were assembled by electrostatic templating using single quaternary ammonium cations or protonated forms of amines or diamines, prolonged reaction times and hydrothermal reaction conditions. The few cases involving the use of neutral primary amines as templates also afforded microporous high silica molecular sieves. This again could be attributed to the crystallization of the neutral framework around single surfactant molecules. 4- Applications of Zeolites and Molecular sieves The industrial application of zeolites was launched in late 1950’s following the remarkable accomplishment made by Milton with the preparation of zeOlite A 25 and zeolite X. 27 The following years witnessed rapid growth in the field of zeolite and molecular sieve synthesis. Currently, thousands of “40 )1 Lg? patents and publications in a number of respected scientific magazines, such as Zeolites, Microporous Materials, Advanced Materials, Chemistry of Materials, Catalysis Letters, Journal of Catalysis, journal of the Chemical Society, Chemical Communications, and even Nature and Science are devoted (per year) to the synthesis and the properties of existing and newly discovered zeolites and molecular sieves. Because of their polar nature, uniform pore size and significant pore volume zeolites are widely used as adsorbents for removing water and polar molecules from gases. 27 Recently, they have been used to separate glucose - fructose mixtures 28 and even antibiotics. 29 The zeolites with small Si/Al ratios or high ion - exchange capacity, such as zeolite A, are primarily used as water softeners in detergents. 30 Almost two - thirds of the world demand for zeolites is based on the need for detergent builders. However, there is little doubt that the enormous growth of the field of zeolites and molecular sieves was due to the discovery of their catalytic petential in fluid catalytic cracking (FCC) of heavy petroleum fractions and other refining processes. 30,31 The acidic aluminosilicate zeolite Y (H’r form) was found to be effective FCC catalyst for conversion of the ”middle distillates” to gasoline. Currently, this particular zeolite is used on a very large scale in the oil - refining industry. However, due to the small pore size of this microporous framework (~ 0.74 nm) the large hydrocarbon molecules from the ”bottom of the barrel” can not penetrate the pore volume and, hence, can not be converted to gasoline. The high silica zeolites, such as ZSM-5 or its pure silica analog silicalite, exhibit high affinity toward organic molecules and the ability to selectively adsorb organic pollutants from waste waters or rivers. However, the small pore size of these molecular sieves (approximately 0.57 nm) precludes the possibility for adsorption and separation of the toxic polyaromatic Chlorohydrocarbons from the waste or drinking waters. Thus, high silica zeolites and molecular sieves with larger uniform pore size are extremely dCSirable. Zeolites and molecular sieves are also very useful as shape - selective catalysts and catalyst supports. The first example of shape selective catalysis Was illustrated by Weitz and Frillete. 32 In an elegant experiment they showed that both primary and secondary alcohols can diffuse and undergo dehydration in the larger pore (~0.74 nm) zeolite X, whereas only primary alcohols were accommodated and dehydrated in the small pore size zeolite A (~ 0.43 nm). The uniform micropore size not only limits the shape and the size of the reactants that could penetrate the framework but also influences the reaction selectivity by ruling out the shape and the size of the corresponding reaction products. These unusual properties of zeolites were quickly realized and a number of important catalytic processes were developed in the last three decades. Some of the commercial catalytic processes involving zeolites are summarized in Table 1. Table 1. Major commercial zeolite processes. Process Zeolite Product $/ ton 1| Catalytic cracking faujasite gasoline, fuel oil 1.5 - 3000 Hydrocracking faujasite kerosene, jet fuel, 12,000 Pt " benzene, toluene, xylene Hydroisomerisation mordenite i-hexane, heptane 12,000 Pt " (octane enhancer) iso/n-paraffin Ca-A pure n-paraffins 5,000 separation . Dewaxing ZSM-S low pour point 60,000 Pt " mordenite lubes 14,000 Pt " Olefin drying K-A polyolefin feed 4,000 Benzene alkylation ZSM-S styrene 60,000 60,000' X lene isomerisation ZSM-S parafllene Pt does not include the price of the recoverable Pt and Pd component, which may vary between 100 and 300 troy ounces pet ton. Costs are difficult to determine because of combinations with other licensing SerVices. Adopted from Ref. 8. It did not take long to realize that the potential of these microporous molecular sieves is strictly limited by the small pore size of their frameworks. 10 B. Recent Advances in the Field of Microporous Molecular Sieves Because of the above limitations a myriad of new organic directing agents, aluminosilicate compositions and reaction variables were put to test in the last three decades in attempts to expand the uniform micropore size and prepare new and stable frameworks with useful properties. In‘ spite of this considerable effort, until 1988 the larger pore size available in zeolites and molecular sieves (see Figure 3) was still that of the synthetic faujasite analogs, zeolites X and Y (prepared as early as 1959). 33 The replacement of the aluminosilicate gels with aluminophosphate gels led in 1982 to the advent of the aluminophosphate molecular sieves or AlPO’s by Wilson and co - workers. 34 However, the first AlPO’s, namely AlPO-5 and AlPO-l 1, exhibited even smaller pore size than zeolite Y. Nevertheless, the attention was shifted toward preparation of aluminophosphates and in 1988 Davis and co - workers 35 disclosed the first 18 - membered ring aluminophosphate molecular sieve (denoted VPI-S) with an hexagonal arrangement of 1-D channels and uniform pore size of approximately 1.2 nm (see Figure 3). Three years after that Estermann and colleagues discovered 36 a 20 - membered ring gallophosphate molecular sieve - cloverite with 3-D channel system and uniform pore size of 1.3 nm. In 1992, Thomas and his co - workers reported 37 yet another 20 - membered ring aluminophosphate molecular sieve, denoted 1DF-20, having uniform pore size of 1.45 nm. Very recently, a preparation of vanadium phosphate molecular sieve (VAPO) with 1.84 nm lattice cavity was announced by Haushalter and colleagues. 33 The actual pore size of the latter two materials is still to be determined since sorption data are lacking. The thermal stability of VPI-5 and cloverite seems to be much lower than that of the high Silica zeolites and molecular sieves, 39.40 whereas that of the novel 1DF-20 and the VAPO is still to be determined. All of these newly developed microporous structures were again prepared “Sing single, cationic templates (primary ammonium, diammonium, tertiary and quaternary ammonium ions), electrostatic interactions and hydrothermal reaction conditions. In addition, the eventual processes that could be Performed over these materials would still be restricted to molecules of e‘lllivalent ”micropore” size. pu- 11 ('3 .. O) O) F Pro-1980 2.5 ” 1 992 1 990 1 988 1 991 Pore size, nm Molecular sieve Figure 3. Uniform pore size and framework structure of some of the well known microporous zeolites and recently developed molecular sieves. 33 12 C. Mesoporous Molecular Sieves 1. Assemblies of Surfactant Molecules as Templates - a Milestone to the Synthesis of Mesoporous Molecular Sieves The considerable synthetic effort toward expanding the uniform micropore size available in zeolites and molecular sieves met with limited success until 1992. This was mainly due to the use of single organic molecules as structure directing agents or templates. Simultaneously, clay scientists studied the behavior of intercalated assemblies of surfactant molecules in the galleries of their lamellar hosts. 41 This fact had an important impact on the discovery of the mesoporous molecular sieves. Thus, in 1990 while studying the ion - exchange reaction between single - layered silicic acid host (kanemite, NaHSi205.3HzO) and long chain quaternary ammonium cations, Kuroda and co - workers reported 42 that the layers seems to distort and cross - link around the cations to form a new porous structure. The driving force for this layer folding process is most likely the ion - pairing between the positively charged assemblies of quaternary ammonium cations (3+) and the negatively charged layered host (1'). Upon calcination of this organic - inorganic complex the surfactant was burned off and the 1-D mesopore network was made available to molecules. However, the unusual X - ray diffraction pattern of the cross - linked material. caused some confusion. 42 The pattern was very similar to these exhibited by the pillared lamellar solids and was essentially comprised by a single reflection at very low 26 angles. Nevertheless, the adsorption studies revealed that the cross - linked material possess a nearly uniform pore size centered at ~ 3.0 nm and the 29Si MAS NMR confirmed the hypothesis for significant cross - linking of the structure because of the large increase of the Q‘i/Q3 ratio. 42 The resulted material can now be described as ”ribbon - candy” - like structure as depicted in Figure 4. 13 SINGLE LAYERED SILICIC ACID (KANEMITE) NaHSi205.3H20 Surfactant CnH2n+1(CH3)3N+ l 'RIBBON CANDY“ STRUCTURE Calcination MESOPOROUS SILICA Figure 4. Schematic illustration of the rearrangement of a layered silicic acid host (kanemite) into a ”ribbon - candy" - like structure in the presence of assemblies of long chain quaternary ammonium cations as templates. In 1992 scientists at Mobil Oil Research and Development disclosed in a lengthy series of patent applications the preparation and some of the most important properties of the first family of mesoporous molecular sieves (denoted M41S). 43"“ A detailed summary of the Mobil patents on mesoporous molecular sieves is beyond the purpose of this review. Such a summary was provided very recently by Casci. 45 This review will focus mainly on the different templating approaches to mesoporous molecular sieves and on their catalytic applications. 14 2. Mobil’s M415 Family of Mesoporous Molecular Sieves - Preparation According to Mobil’s technology long chain quaternary ammonium surfactants minimize their energy in solution by assembling into micelles (Figure 5 A). 45 Under certain conditions these micelles can adopt a rod - like shape and spontaneously organize into long - range hexagonal arrays with the charged head groups pointing toward the solution and the long hydrocarbon chains (hydrophobic) pointing toward the center of the micelles. The ability of the long chain quaternary ammonium cations to form rod - like micelles and long - range ordered hexagonal arrays in aqueous solutions (with rod diameters in the mesopore range of 2.0 - 4.0 nm) has been known for a very long time after the pioneering work of Luzzati (1968). 47 The formation of the micellar rods and their organization into hexagonal arrays is strongly dependent on the surfactant’s alkyl chain length, concentration, the nature of the halide counterion, and temperature of the solution. 47"” Upon the addition of a silicate precursor, for example, sodium silicate, the negatively charged silica species (I') condense and polymerize on the surface of the positively charged micelles (5*) giving rise to the corresponding hexagonal S+ 1' organic - inorganic biphase array (see Figure 5 A). The calcination. of the complex revealed the hexagonal solid framework of this particular mesoporous molecular sieve denoted as MCM-41. Another important contribution of Mobil researchers is the disclosure that the uniform mesopore size of MCM-41 can be varied in the range from 13 to 10.0 nm by simply varying the surfactant alkyl chain length (form 6 to 16 carbon atoms) or by adding an auxiliary organic (e.g., trimethylbenzene) into the internal hydrophobic region of the micelles. Thus, a hexagonal MCM-41 silicas with uniform mesopore size of 2.0, 4.0, 6.5 and even 10.0 nm have been reported (see Figure 5 B). A typical Mobil preparation of MCM-41 involves the use of aqueous solution of cationic surfactant, such as cetyltrimethyl ammonium chloride, partially exchanged for OH' (C16H33(CH)3NOH/ C1) over an ion - exchange resin. The surfactant concentration of this solution was selected to be approximately 25 wt %, which is much larger than the critical micelle concentration (cmc) necessary for the formation of rod - like micelles (~11 - 20.5 wt %). 50 The sources of silica were varied from sodium silicate, (W Tr 7. 15 . - b. 5. . .- . a; o . C‘b I.'U._'<"J '1‘.)".‘ . .‘ ‘q , gr i .‘ o' ‘; " ~ whines-Mere u .0. .O V q ‘22-". -V ’uln‘vq'4';fi,~u.'~.~;’t - ' u,_":g.,""°.-, 9,): ‘(fi‘h/‘k”. i, - - i .‘s, 't «'43. nigh/mg ‘- ' RV»- a L on .' . . . Figure 5. (A) Mobil 5+ 1' mechanistic routes for the formation of the hexagonal MCM-41: route (1) liquid crystal phase initiated and route (2) silicate anion initiated. (B) Transmission electron micrographs of MCM-4l samples with uniform mesopore sizes of (a) 2.0, (b) 4.0, (c) 6.5, and (d) 10.0 nm. 46 16 amorphous fumed silica, colloidal silica or tetraethyl orthosilicate (TEOS). In some cases aluminum substituted analogs were prepared using alumina sources selected from the group of sodium aluminate, aluminum sulfate and pseudoboehmite. The following is a typical reaction mixture composition 43 for the preparation of MCM-41 (expressed in moles): 1.0 SiOz: 0.03 A1203 : 0.007 NazO : 0.183 (CTMA)zO : 0.156 (TMA)zO : 23.5 H20. The addition of tetramethylammonium hydroxide (TMAOH) was most likely dictated by the need for a base capable of dissolving the amorphous silica source. The preparation comprises mixing of the above inorganic precursors with the solution of template and autoclaving of the corresponding reaction mixture at temperatures from 100 to 150°C for a period of time from 4 to 144 hr. The crystalline product was recovered by filtration, air - dried and subjected to heating in N2 atmosphere at 550°C for 1 hr followed by calcination in air at the same temperature for 6 hr. 46 The use of different surfactant/ silica ratios afforded, in addition to the hexagonal MCM-41 phase, cubic and lamellar phases. Figure 6 illustrates the XRD patterns of the above materials. This entire family of mesoporous molecular sieve phases formed by 5+ 1' templating was denoted as M415. 46 Surfactant/ silica ratios lower than 1 were found to give the hexagonal MCM- 41 (pattern A), whereas ratios larger than 1 afforded the cubic (MCM-48) counterpart (pattern B). In addition, ratios much larger than 1 afforded the lamellar mesophase (pattern C). However, the lamellar phase was found to be unstable and collapsed upon calcination at elevated temperatures. 46 ETTA too so: no 22.0 zoo an no «.9 2 .- fi a «'5 2 T o F Ti I V Y no «AI (B) 211 no a no no '5 «'5 5 no «.1 a as: 17.: o g m m a 3 5 am in .5 a i a 3 o i to m (C) s m '- ioo so: aoo 11.9 aoo u.- a ' 4 ' 3 - o to Degrees (29) Figure 6. Powder X - ray diffraction patterns of calcined (A) hexagonal MCM- 41, (B) cubic Ia3d phase and (C) as - synthesized unstable lamellar phase prepared by 5+1“ templating mechanism. 46 18 3. Templating Pathways to Mesoporous Molecular Sieves Mobil’s mechanism of the formation of M415 materials involves strong electrostatic interactions and ion pairing between quaternary ammonium liquid crystal cations (5*), as structure directing agents, and anionic silicate oligomer species (I‘). The recently reported 48 preparation of related hexagonal mesoporous structures by rearrangement of a layered silicate host (kanemite) can also be considered a derivative of the above electrostatic approach to mesoporous molecular sieves. Originally, two possible variations of this 5+ 1' electrostatic mechanism were proposed by Beck et al. (see Figure 5 A). 46 According to route 1 the surfactant molecules in solution organize spontaneously into micellar liquid crystals. The addition of the silicate source leads to condensation and polymerization of the anionic silicate species on the surface of the positively charged micelles by a mechanism of ion - pairing. Further condensation of the inorganic precursor in the continuous solvent region (water) leads to the formation of the inorganic framework walls. The second mechanistic direction proposed by Mobil (route 2 in Figure 5 A) suggests that the addition of anionic silicate to the templating solution triggers the formation of the hexagonal liquid crystalline phase. Stucky and his co - workers also elaborated on the 5* I' mechanism and identified three closely coupled phenomena for the formation of the surfactant - silica mesophases. 49 These are: (i) multidentate binding of silicate oligomers to the surfactant head groups; (ii) preferred polymerization of silicate oligomers at the surfactant - silicate interface, and (iii) charge density matching across the surfactant - silicate interface. According to these workers, at high pH, the reaction mixture contains small silica oligomers such as single or double 3 and 4 rings of different negative charges (see Figure 1). These oligomeric anions act as multidentate ligands for the cationic head groups of the surfactant, leading to a strongly interacting surfactant - silicate interface. Perhaps, the most important implication of this work is that the framework wall thickness of the MCM-41 materials is limited. The number of anionic silicate oligomers that can be incorporated into the walls of MCM-41 materials by 5* I' templating is restricted by the strong electrostatic repulsions between the I' silicate species and the 5+ 1' charge matching interactions. Thus, the framework wall growth of MCM—41 is terminated when charge compensation is achieved. This hypothesis is supported by the observation that the 19 framework wall thickness (0.8 to 0.9 nm) of MCM-41 did not vary over a wide range of reaction conditions and surfactant chain lengths. 49 Davis and co - workers also have elaborated on the 5+ 1‘ templating mechanism. 50 Based on 14N NMR analyses of their quaternary ammonium surfactant solutions and corresponding MCM-41 products these authors concluded that the formation mechanism does not involve the liquid crystalline phase as suggested originally by Mobil (route 1). Instead they proposed that randomly ordered rod - like micelles form initially and interact via electrostatic interactions with the anionic silicate species to give polymerized monolayers of silica on the surface of the micelles (Figure 7). Further Condensation Condensation ”Hips: i E E l .32.. i. ”iv” ' A I’ if T u I (a) ‘ ‘: "I "l—4)> l: r vii»? ~ . .. iii iii iii Figure 7. A mechanism for the 5+ 1' formation of MCM-41 proposed by Davis et al. 50 These ”silica coated” micelles pack spontaneously into long range hexagonal array. The driving force for this process was believed to be the further condensation of the silicate species. Thus, this work provided a refinement of the route 2 originally proposed by Beck et al. (see Figure 5 B). 45 An important finding supported by 295i MAS NMR and elemental analysis data is that complete condensation of the anionic silicate species in the walls of the electrostatically templated MCM-41 is not possible because framework (SiO)4- xSiOxx‘ species are necessary for charge compensation of the occluded 20 quaternary ammonium template. 50 This implies that the cationic template in typical 5* I' MCM-4l preparations is strongly bound to the negatively charged framework and difficult to recover by non - destructive methods. Recently, Stucky and colleagues further extended the electrostatic assembly approach to mesoporous molecular sieves by proposing four complementary synthesis pathways (Figure 8). 51 Inorganic _ Solution Specres STI' Antimony Oxide ‘—" Tungsten Oxide (pH (7) Surfactant Examples Observed l) D... /' * M.,...“ hm'm Iron Oxide \2)- + ~81; beadeide Cationic S'X'F Silica (pH < 2) —> Zinc Phosphate Aluminum Oxide ~. _' ‘- 3) . . /' Cationic + Mediated ‘9" ‘3) Pathways S‘M"! \a) + —> Zinc Oxide (pH >115) l' O- 0 :- \I ”\U “\1 I," ‘Kl KI/o ‘ ~ . w. 1 r 5" 5" x O cI-az or, a or or, I. l'nfl-CN;.¢3N5.C3H1 o. n.cu,.n‘.(cn1),-so,-.can. . 9mm. cnz© x . 0. © . 9‘3- . G;- .. (un,n-)-©.@. eaNsO Figure 8. Schematic representation of the four complementary electrostatic templating pathways to ordered mesostructures. 51 Pathway 1 involved the direct co - condensation of anionic inorganic species (I') with a cationic surfactant (5+) to give assembled ion pairs (S+ 1'), the original synthesis of M415 silicas being the prime example. 46 In the charge reversed situation (Pathway 2) an anionic template (5') was used to direct the self - assembly of cationic inorganic species (1+) via 5'1+ ion pairs. The pathway 2 has been found to give a hexagonal iron and lead oxide and different lamellar lead and aluminum oxide phases. Pathways 3 and 4 involved counterion (X' or M+) mediated assemblies of surfactants and inorganic species of similar charge. These counterion - mediated pathways 21 afforded assembled solution species of type 8" X' 1+ (where X‘ = Cl‘ or Br ') or, S' M+ I’ (where M+= Na+ or 10‘), respectively. The viability of Pathway 3 was demonstrated by the synthesis of a hexagonal MCM-41 using a quaternary ammonium cation template and strongly acidic conditions (5 - 10 M HCl or HBr) in order to generate and assemble positively - charged framework precursors. ' In another example, provided by the same researchers, a condensation of anionic aluminate species was accomplished by alkali cation mediated (Na+, K“) ion pairing with an anionic template (C12H250P03'). The preparation of the corresponding lamellar Al(OH)3 phase in this case has been attributed to the fourth pathway (5’ M+ I'). 51 The contributions of Stucky and co - workers for the preparation of mesoporous molecular sieves with non — silicate composition, especially transition metal oxides, can not be overstated. Transition metal oxide mesoporous molecular sieves could be very important in a number of catalytic processes such as metathesis of alkenes, methane oxidation, and photocatalytic decomposition of large organic pollutants. Unfortunately, all of Stucky’s templated mesoporous metal oxides, including the lamellar alumina phase, were unstable to template removal by calcination or other methods. 51 Thus, a synthetic approach that will generate stable transition metal mesoporous phases is highly desired. Pathway 3 (8* X' 1*) afforded not only the preparation of hexagonal. MCM- 41 but also a Pm3n cubic and a lamellar phase. 51 A typical preparation involved the addition of TEOS to strongly acidic solution (5 - 10 M HCl or HBr) of quaternary ammonium surfactant and aging of the reaction mixture at ambient temperature for more than 30 min. 51 It has been postulated that the interactions between the cationic silica species ('='Si(OH2)+) and halide - cationic surfactant headgroups are mediated by the large excess of halide ions (X'). However, the main driving force for mesostructure formation is also electrostatic. The need for large excess of corrosive acidic reagent will require special reactor equipment and waste disposal considerations in potential industrial scale preparation. Very recently, we have reported 52 a neutral (8° 1°) templating route to mesoporous molecular sieves, which we denoted as Pathway 5. We postulated that the formation of our HMS mesostructures occurs through the organization of neutral primary amine surfactant molecules (S°) into neutral rod - like micelles (see Figure 9). 5‘5; "-./ Pathway 5 3 CnH2n+1NH2 + SI(OEt)4.x(OH)x > “N“ Spontaneous rod-like ‘3!!!" '“ HN 'H micelle formation /.| l (H . a: Ztn\\ film": H ‘H )0 it h S! \‘ ' 'afl’ma na’wH‘N'DSifo 0’ Gk HHNHHN" "N's“ ‘ ; \D _——O/ '01 m. ' N H 8 18‘" (Sm-.3 J/JJ 3‘ I /s."° 0" \ :X My". 05: a; mlIvSia‘l—Dlaw W". .9" m g 9} - 0" Figure 9 A neutral 5° 1° templating mechanism to mesoporous molecular sieves. 23 The addition of neutral inorganic precursor, for example TEOS, to the solution of template affords the hydrolyzed intermediate Si(OC2H5)4.x(OH)x species. We believe that these species participate in H - bonding interactions with the lone pairs on the surfactant head groups affording surfactant - inorganic complexes in which the surfactant part could be viewed as a hydrophobic tail and the inorganic precursor - as a bulky head group. This significantly changes the packing parameter of the obtained surfactant - inorganic complexes and most likely triggers the formation of rod - like micelles in our solutions of neutral primary amine surfactants. Further hydrolysis and condensation of the silanol groups on the micelle - solution interface afford short - range hexagonal packing of the micelles and framework wall formation. We can not also preclude the possibility of having slightly different pathway involving preorganized spherical micelles of surfactant molecules or even surfactant bilayer arrays in our initial ethanol - water solutions of template. However, we think that a rearrangement into rod - like micelles takes place upon addition of inorganic precursor that is again triggered by H - bonding interactions between the lone pairs on the surfactant head groups and the intermediate silica precursor species. A typical preparation of our HMS materials involves the addition of tetraethyl orthosilicate (TEOS) (1.0 mol) to a solution of neutral primary amine (0.27 mol) in ethanol (9.09 mol) and deionized water (29.6 mol). The reaction mixture is aged at ambient temperature for 18 hours, and the resulting HMS silica is recovered by filtration or air - drying. In contrast to the electrostatic templating pathways our neutral templating allows for a cost reduction in large scale preparations by employing mild reaction conditions and cheap primary amine surfactants that can be easily recovered and recycled by simple solvent extraction. 4. Properties of the Mesoporous Molecular Sieves Prepared by Different Templating Approaches The X - ray diffraction patterns of M415 materials prepared by the 5+ 1' templating pathway are presented in Figure 6. 46 The hexagonal MCM-41 phase exhibited just a few hkO reflections (100, 110, 200 and 210) and lacked reflections with non - zero 1 component. This has been attributed to the absence of local order in the walls of the hexagonal MCM-41. The observed Jfldiflilll 24 hkO reflections were assigned to scattering arising from the uniform hexagonally arranged mesopores. Very recently, Froba et al., performed SiK XANES measurements on a series of M415 samples, related high silica zeolites, lamellar silicate hosts, amorphous glasses and high temperature silica phases. 53 The comparison of the results revealed that M413 samples exhibit intermediate local wall order between the high temperature Silica phases and amorphous glasses, similar to that of the lamellar silicates. This led the authors to propose that the walls of M415 silicas closely resemble curved silicate sheets. However, the comparison of the 29Si NMR data reveals that the MCM-41 materials exhibit much broader and overlapping Q3 and Q4 peaks, whereas those of a crystalline lamellar silicate such as magadite are much sharper and well resolved. 55 This suggests larger variety of Si-O-Si bond angles in MCM-41 silicas than in the lamellar silicates. The XRD patterns of the MCM-41 prepared by the acidic 5* X' 1* templating (pathway 3) also possessed the hkO reflections characteristic of the hexagonal phase. 51 In contrast, the XRD patterns of HMS materials prepared by 8° 1° templating exhibited single dloo reflections accompanied with diffuse scattering centered and approximately 5° 26. 52 Higher order Bragg reflections of the hexagonal structure were not resolved in the patterns of our HMS materials. However, we and others have demonstrated 435254 that similar "single reflection" MCM-41 - type products, still possess short .- range hexagonal symmetry. Therefore, the diffuse scattering at ~ 5° is attributable to broadening of the remaining hkO reflections of the hexagonal phase. However, future studies aimed at optimizing the reaction conditions for our 5° 1° templating approach should ultimately afford long range ordered hexagonal, cubic and lamellar mesoporous materials. The framework wall thickness of MCM-41 materials prepared by the 5* I" pathway have been found to be limited by the charge compensation between the positively charged micelles and the negatively charged inorganic species. 49 According to the original report by Beck and co - workers, 46 the framework wall thickness varies from 0.8 to 1.3 nm with variation of the surfactant chain length. Stucky and co - workers observed even smaller variations of the framework wall thickness (0.8 to 0.9 nm) for MCM-41 silicas prepared by 5* I' templating under a wide range of reaction conditions and surfactant chain lengths. 49 This difference in wall thickness reported by the two groups could be attributed to the use of C1“ or Br' salts of the quaternary ammonium ions as 25 surfactants in the work of Stucky, whereas Beck et al. used partially exchanged Cl'/OH' solutions of surfactants. Davis et al. also found that the framework wall thickness of their 5* I' MCM-41 materials was in the range of 1.0 nm. 55 The framework wall thickness of MCM-41 samples prepared by the acidic 8+ X‘ I+ pathway also seems to be similar, 1.0 - 1.2 nm. 51:56 In general, the small framework wall thicknesses of the electrostatically templated MCM-41 materials could seriously limit their thermal and hydrothermal stability in adsorption and catalytic applications. In an attempt to improve the thermal stability of the 5“ I' MCM-41 aluminosilicates, Fajula and co - workers varied the amount of 8+ surfactant, sodium aluminate and water in their reaction mixtures and observed the effect of this variation on the framework wall thickness. 57 They stated that the alkalinity of the solution influenced most the pore wall thickness. In particular, six out of ten samples showed framework wall thicknesses of less than 1.0 nm, one showed ~ 1.3 nm, and only three exhibited thicknesses of from 1.5 to 1.6 nm. However, given the range of variation of the framework wall thickness observed by the above groups 46:49:55 for the same class of materials (from 0.8 to 1.3 nm) it is unlikely that the variation of wall thickness observed by Fajula et al. is due exclusively to differences in alkalinity. An important finding of this work was that MCM- 41 samples with thicker pore walls exhibit higher thermal stability. This was nicely illustrated by the comparison of the degree of pore volume contraction exhibited by thin walled and thicker walled MCM-41 samples after heat treatment in air at 550°C. 57 Very recently we have reported that the framework wall thickness of our HMS molecular sieves, prepared by 5° 1° templating with primary amine surfactants, is much larger (from 1.7 to 3.0 nm) than that exhibited by the electrostatic counterparts. 52 The much thicker pore walls of our HMS materials could be attributed to the absence of electrostatic repulsions between the neutral intermediate silicate species and to the lack of charge matching interactions between the neutral surfactant and the neutral inorganic precursor. 52:56 As illustrated by Fajula and co - workers 57 thicker pore walls are highly desired as improving the thermal and hydrothermal stability of the mesoporous framework. In this sense neutral templating may provide for the preparation of mesoporous molecular sieves with much improved thermal stability. 26 Since the S+ 1' electrostatic templating pathway is based on charge matching between a cationic surfactant and an anionic inorganic reagent, the template is strongly bonded to the charged framework and difficult to recover. In the originally disclosed approach 45 the template was not recovered, but simply burned off by calcination at elevated temperatures. Recently, the scientists at Mobil came up with a better idea, namely, to ion - exchange the ionic surfactant with an acidic cation donor solution. 53 Also, Stucky and co - authors have demonstrated that much of the template - halide ion pairs in MCM-41 prepared by the acidic 8*“ X‘ 1+ method can be displaced by ethanol extraction. 51 Thus, ionic template recovery is possible, provided that exchange ions or ion pairs are present in the extraction process. In contrast, our 3° 1° templating approach to mesoporous molecular sieves allowed for the facile and environmentally benign recovery of the neutral template by simple solvent extraction. 52 Owing to the absence of counterions the recovered neutral template can be easily recycled and reused by simple evaporation of the solvent. Recently, scientists at Mobil reported the 29Si MAS NMR of as - synthesized M418 samples prepared by the electrostatic S+ 1' pathway. 59 Using deconvolution techniques, they resolved three broad and overlapping peaks in the spectra of MCM-41, MCM-48 and two peaks in the spectrum of the thermally unstable lamellar phase. The spectra of both hexagonal MCM-41 and cubic MCM—48 were essentially the same as for amorphous silica. This is not surprising giving the amorphous character of the framework walls and the expected wide range of Si-O—Si bond angles. Interestingly, all 8* 1' phases exhibited more non - condensed Si(SiO)x(OH)4.x framework units (where X: 2 or 3, respectively, for Q2 and Q3) than fully condensed (X: 4) units as evidenced by the low ratios of Q4/Q2+Q3 (on average around 0.67). This result is extremely significant and implies a relatively low degree of cross - linking of the "electrostatic” molecular sieves. The low degree of cross - linking (a ratio of 0.77) is also evident from the data of Davis et al. 5055 who showed that cationic template removal by ion - exchange with HCl - ethanol solutions, as well as by calcination resulted in further cross - linking of the Q2 and Q3 units into Q4 units. The fact that the HCl - ethanol treatment causes some condensation of the silanol groups is not surprising since the acid - catalyzed condensation of silanols is well known. However, HCl - ethanol treatment causes significant lowering of the d100 values, which translates into thinner !&. 27 pore walls or smaller mesopore sizes. In this connection Davis and co - workers made a very important observation, namely, that 100% Q4 state or full cross - linking of the electrostatic M415 framework can not be reached because SiO' groups are needed for charge compensation of the quaternary ammonium cationic template (5*). This fact suggests that the ionic template in a typical M415 preparation is strongly bound to the negatively charged framework and difficult to recover. The 29Si MAS NMR spectra of the as - synthesized M415 phases prepared by the acidic S+ X' I+ pathway generally also exhibit low Q4/Q3 ratios < 1.0. 5156 However, the positive charge of the surfactant is balanced in part by the excess of halide counterions which allowed for displacement of 85 wt °/o of the template - halide ion pairs by solvent extraction. In contrast to the electrostatic templating pathways our neutral 5° 1° pathway affords much more cross - linked HMS frameworks, as judged by the 29Si MAS NMR spectra. 56 This is evidenced by the very high Q4 / Q3 ratios of our as - synthesized and solvent extracted samples (2.70 and 3.12, respectively). This observation is not surprising due to the lack of electrostatic repulsions and charge matching and the corresponding thicker pore walls of our HMS materials. The adsorption pr0perties of electrostatically templated M415 materials have been subject of intensive studies from_ the first days of their discovery. Beck et al. demonstrated 46 that both benzene and N2 adsorption isotherms of their MCM-41 samples exhibit a sharp adsorption uptake in the low relative pressure region indicative of capillary condensation in framework - confined mesopores. The BET specific surface area 2 was estimated to be around 1000 m2/ g and the total adsorbed volume in the range of 0.7 to 1.2 cm3/ g. The mesopore size distribution of S+ 1‘ MCM-41 samples was calculated by the method of Horvath - Kawazoe. 50 This method assumes a slit - shaped pore model and was originally designed for the determination of micropore size distribution. Surprisingly, the mesopore sizes determined by this method was found to be in a good agreement with these determined from transmission Electron micrographs. 'Ihe mesopore size of MCM-41 samples prepared by 8*” I‘templating was found to be in the range from 1.8 to 3.7 nm, depending on the surfactant chain length. 46 Figure 10 A shows typical N2 adsorption - desorption isotherms for amorphous silica, zeolite NaY and MCM-41 prepared by the $4“ I‘ pathway. It is obvious that the isotherm for MCM-41 differs quite "v 28 dramatically from these for the amorphous silica and zeolite NaY. The presence of a large hysteresis loop at Pi/Po > 0.5 in the isotherm of the amorphous silica is indicative of non - uniform textural or interparticle mesopores. The isotherm of zeolite NaY (pore size ~0.74 nm) does not exhibit any hysteresis loop and is characterized by strong adsorption uptake at very low relative pressures owing to adsorption in micropores. 2 Due to the absence of framework - confined mesopores NaY and amorphous silica lack the sharp adsorption feature observed on the isotherm of MCM-41. Davis and co - workers measured cyclohexane and water adsorption isotherms for a S+ 1' pure - silica MCM-41 and aluminosilicate MCM-41 samples. 55 Both samples exhibited very high cyclohexane adsorption capacities (0.4 g/ g dry solid) and very low water adsorption capacities (0.05 g / g). However, the aluminosilicate showed slightly higher adsorption affinity toward water. This result shows that pure - silica MCM-41 samples are hydrophobic and that the Al substitution could introduce partial hydrophilicity. This feature of the pure - silica MCM-41 samples makes them very attractive as potential adsorbents for large organic pollutants from waste or drinking waters. The N2 adsorption - desorption isotherms of MCM-41 samples prepared by the acidic 5“ X' 1+ templating route are very similar to those exhibited by the 5* I' counterparts. 5256 However, both electrostatic templating routes afford MGM-41 samples that lack appreciable textural mesoporosity. This is evidenced by the absence of a significant hysteresis loop in their N2 adsorption - desorption isotherms in the region of Pi/Po > 0.4. The lack of textural mesoporosity for the electrostatically templated MCM-41 materials could impose serious limitations on their use in diffusion controlled processes. Recently, we have demonstrated that 8° 1° templating affords HMS materials with very small scattering domain sizes (less than 17.0 nm) and complementary framework - confined and textural mesoporosity (see Figure 10 B). 5256 The small crystallite size and substantial textural mesoporosity are Very desirable for accessing the framework - confined mesopores and for imProving the performance as adsorbents and catalysts. 61:52 In summary, the electrostatic templating pathways to ordered mesostructures used charged surfactants ions (8*) as templates in order to assemble an inorganic framework from charged inorganic oxide precursors (I' or 1+). These charged templates are usually expensive, strongly bonded to the litre 1E Zitljfe N "ileum 29 i- .(A) ., l g. E 0 VA I'll/l ('7) 8 // / U3 " // E / amorphous / Z P" / silica , Q / .0 /' // ‘5 / / g 31 / NaY / // (II “—1) 0) E 2 O > a? Vqume adsorbed (cc/9. STP) °( i T? '3‘ .7: "E "'."6"*T7"'_'.‘§" ".’§ Pifpo '1 Figure 10. N2 adsorption - desorption isotherms for (A) amorphous silica, Zeolite NaY, MCM-41 prepared by the S+ 1' pathway, and (B) HMS prepared by the neutral 5° 1° templating route. 30 charged inorganic framework and difficult to recover. In addition, some charged templates, such as quaternary ammonium ions are highly toxic and, therefore, potential health hazards. In general the electrostatically bonded templates are removed from the framework of 5+ I'MCM-41 materials by either a combustion process or by an ion - exchange reaction with an ion donor solution. Also, ion pairs are necessary in order to displace the template from the framework of 5+ X' 1+ MCM-41 materials. Electrostatic templating affords as - synthesized MCM-41 materials with a relatively low degree of framework cross - linking and small framework wall thickness (from 0.8 to 1.2 nm). This could seriously influence the thermal and hydrothermal stability of these molecular sieves especially in applications that require severe regeneration conditions. In addition, MGM-41 samples prepared by electrostatic templating seem to lack appreciable textural mesoporosity. The lack of textural mesoporosity could lead to serious diffusion limitations in many potential applications. Neutral templating (8° 1°) allows for the preparation of mesoporous molecular sieves with more completely cross - linked framework, large framework wall thickness, small particle sizes and complementary framework - confined and textural mesoporosity. In addition the 8° 1° approach allows for cost reduction by employing less expensive reagents and mild reaction conditions while at the same time providing for the effective and environmentally benign recovery and recyclability of the template. 31 D. Catalytic Applications of Mesoporous Molecular Sieves Involving Large Organic Molecules 1. Important Catalytic Applications of Microporous Zeolites and Molecular Sieves The isomorphous substitution of synthetic zeolites and molecular sieves with metal atoms capable of performing different chemical (mostly catalytic) tasks is quickly emerging as an important aspect of today’s approach to the design of heterogeneous catalysts. For the purpose of our discussion we will briefly review the catalytic applications of some well known and relatively new industrially important microporous zeolites TS—l, zeolite X, Y and the high silica ZSM-S. Titanium silicalite-1 (denoted TS-1) with a ZSM-5 framework and pore size of ~ 0.6 nm is emerging as a valuable industrial catalyst due to its ability to oxidize organic molecules under mild reaction conditions. The hydrothermal synthesis of TS-l was first accomplished by Taramasso et al. in 1983. 63 The 3-D framework of TS-l, shown in Figure 11 A, confines a micropore network of intersecting 10 - membered, parallel, elliptical (0.51 x 0.57 nm) channels along [100] and zig - zag, nearly circular (0.54 i 0.02 nm) channels along [010]. 64 Therefore, the size of the framework - confined micropores of TS-l is equal to the size of these intersecting channels (~ 0.6 nm). Titanium silicalite was found to be an effective liquid phase oxidation catalyst for a variety of organic molecules in the presence of H202 as oxidant. The broad spectrum of TS-l catalyzed reactions includes oxidation of alkanes, 65 oxidation of primary alcohols to aldehydes and secondary alcohols to ketones, 55 epoxidation of olefins, 67 hydroxylation of aromatic compounds 68 and oxidation of aniline. 59 The production of catechol and hydroquinone from phenol over TS-l is now an industrially established process. 70 There is a wide spread notion that the exceptional catalytic activity of TS—l is due to the presence of site - isolated titanium atoms in the micropores of the silicalite host (see Figure 11 B). 32 (B) 055 + H20. +H202 sao—n—osa : m: - H20. +120, osa Figure 11. (A) Schematic representation of the TS-l solid framework (left) and corresponding pore network (right). 54 (B) A formation of an active Ti - peroxocomplex upon addition of aqueous H202 to site - isolated tetrahedral Ti atoms in the framework of TS—l. In addition, the ability of these Ti sites to easily undergo coordination change in the presence of H20 and H202 and to form a very active titanium peroxocomplex is believed to be of primary importance for the observed activity. However, because of the small pore size of the inorganic framework the number of organic compounds that can be oxidized by TS-l is strongly limited to molecules having kinetic diameters equal to or less than about 0.6 nm. Another titanium silicalite, TS-2, with MEL structure was recently reported to exhibit similar oxidation properties. 71 The similar catalytic behavior of TS-Z is not surprising in view of the nearly identical size of the silicalite-2 framework - confined micropore channels (~ 0.53 nm). Very recently, a Ti - substituted analog of yet another zeolite (zeolite B) with slightly larger micropore size has been reported by Corma et al. 72 The main incentive for preparing Ti - substituted analog of zeolite Bwas to be able to take advantage of its slightly larger micropore size network composed by intersecting 12 - membered ring (0.76 x 0.64 nm) channels along [001] and 12 - 33 membered channels (0.55 nm) along [100]. However, the catalytic oxidation chemistry of Ti - substituted zeolite B, with the exception of the slightly higher conversion of cyclododecane relative to TS-l, was again confined to the well known small substrates subjectable to catalytic oxidation over T84 and TS—2 molecular sieves. In addition, the presence of Al3+ in zeolite B affords a hydrophilic framework exhibiting much higher acidity than TS-1 and TS—2. This precludes the possibility for catalytic oxidations of bulky alkyl substituted aromatics or phenols without dealkylation of the alkyl groups. The small micropore size of two recently discovered Ti - substituted molecular sieves, namely ETS-10 73 and Ti-ZSM-48, 74 would most likely confine their catalytic oxidation chemistry again to substrates with small kinetic diameters. Vanadium - substituted silicalite-1 and 2 (denoted VS-l and VS-2) were also recently reported 75 but due to the embedding of V in the same silicalite microporous framework, the catalytic oxidation activity of these molecular sieves is again limited to small organic substrates with kinetic diameters of less than 0.6 nm. The industrially important aluminum containing zeolites X, Y and high silica ZSM-S are used on a very large scale as cracking, alkylation and isomerization catalysts. Because of their uniform pore size and shape - selective properties they are much more active and specific than the amorphous alumina - silica catalysts. The substitution of Al in the frameworks of these zeolites requires protons for balancing the negative charge. These protons are localized at one of the four oxygen atoms of the A104 tetrahedra. The AlOHSi Bronsted acidity sites are responsible for the high activity of the above zeolites exhibited in a broad range of catalytic conversions of alkanes, alkenes and alcohols. There is increasing demand in recent years for treating heavier feeds and for shape - selective catalytic alkylation or isomerization involving large organic molecules. However, the small uniform pore size of these industrially important catalysts (S 0.74 nm) severely limits their catalytic potential to molecules of small kinetic diameters. Therefore, there is a need for a new metal - substituted mesoporous molecular sieves capable of transforming organic species with kinetic diameters > 0.6 nm, especially bulky aromatics. Such metal - substituted mesoporous molecular sieves would greatly complement and extend the catalytic chemistry of the above microporous zeolites toward much larger organic molecules. ear 3.4.x a l CJTI 2. Catalytic Applications of Metal - substituted Mesoporous Molecular Sieves Both mesoporous molecular sieves MCM-41 and HMS offer exciting opportunity for the preparation of large pore analogs of the above industrially important catalysts. The first preparation of Ti - substituted MCM-41 was demonstrated simultaneously by us 61 and Corma et al. 76 However, these latter workers used the Mobil’s S+ 1‘ templating route (Pathway 1) and prolonged hydrothermal synthesis conditions to prepare their Ti-MCM-41 analog. The crystallinity of this particular material was very poor and the claim for a hexagonal Ti-MCM-41 material is doubtful giving the reported d - spacing of 2.8 nm. Thus, the XRD pattern of Corma’s Ti-MCM-41 most likely corresponds to a lamellar rather than a hexagonal phase. The Ti - site isolation in this framework was studied by diffuse reflectance UV visible spectroscopy and IR spectroscopy. The sample exhibited a UV absorbance at 210 - 230 nm and IR band at 960 cm'l, which were assigned to site - isolated Ti atoms in tetrahedral (210 nm) and octahedral (230 nm) coordination and to Si- 05'----Ti5+ group stretching vibrations, respectively. However, we 77 and others 73 have observed this IR band for Ti free HMS and MCM-41 and therefore it can not be considered as an evidence for Ti site isolation. Finally, the catalytic activity of this Ti-MCM-41 sample was illustrated by the epoxidation of rather small organic molecules such as hex-l-ene and norbornene in the presence of H202 and tertbutylhydroperoxide (THP) as oxidants. Simultaneously with the report of Corma et al., we reported 61 the preparation of a hexagonal mesoporous silica (HMS) molecular sieve and a Ti - substituted analog (Ti-HMS) by the acid catalyzed hydrolysis of inorganic alkoxide precursors in the presence of a partially protonated primary amine surfactants (S°/S+). In the same article we demonstrated the first ambient temperature preparation of Ti-MCM-41 molecular sieve using the acidic 8+ X' 1+ templating route (Pathway 3). We also reported that both Ti-HMS and Ti- MCM—41 exhibit remarkable catalytic activity for peroxide oxidation of very large aromatic substrates such as 2,6-di-tert-butylphenol to the corresponding quinone. 61 Due to its complementary textural and framework - confined 35 mesoporosity Ti-HMS showed superior catalytic activity for the oxidation of this large organic substrate. Recently, Sayari and colleagues reported 79 the preparation and catalytic activity of V-MCM-41 for peroxide oxidation of 1-naphthol and cyclododecane. However, the preparation of this molecular sieve again involved an electrostatic S+ 1' templating pathway and hydrothermal treatment at 100°C for 6 days. In addition, the reaction mixture contained significant amounts of N a+ ions which are known to be an unwanted impurity, significantly lowering the catalytic activity of the microporous TS-l. 30 Corma et al. have more recently compared 81 the catalytic activity of Ti- MCM-41 and Ti-zeolite Bfor epoxidation of a-terpineol and norbornene. Due to its large uniform mesopore size Ti-MCM-41 exhibited superior catalytic activity than Ti-zeolite B for oxidation of these bulky olefins. Very recently, we have also reported a neutral 8° 1° templating pathway to mesoporous molecular sieves. 52 We have used this neutral templating route to prepare Ti-HMS molecular sieves with different Ti loadings. 32 The 8° 1° templating strategy allowed for the effective and environmentally benign recovery and recycling of the neutral primary amine template form Ti-HMS by simple solvent extraction. Ti-HMS showed superior catalytic activity for oxidation of the bulky 2,6-DTBP relative to the Ti-MCM-41 counterpart at all nominal Ti loadings in the range from 1 to 10 mol °/o. This has been attributed to the complementary textural mesoporosity of Ti-HMS which facilitates access to the framework - confined mesopores. In another development Sayari et al. compared the catalytic activity of Ti- HMS and Ti-MCM-41 for peroxide oxidation of 2,6-DTBP, 1-naphthol and norbomylene. 33 Ti-HMS was found to exhibit better H202 selectivities as compared to Ti-MCM-41. The decomposition of H202 over Ti-HMS was found to be very limited, whereas in the case of Ti-MCM-41 all remaining H202 was decomposed. Thus, these authors confirmed our observations that both Ti- HMS and Ti-MCM-41 are very promising catalysts for the oxidation of large organic substrates. Scheme I summarizes most of the reported to date catalytic reactions performed over metal - substituted MCM-41 and HMS materials. 36 CATALYST REACTION REF. H202 CH "CH(CH CH 76 CH2=CH(CH2)3CH3 ——» 2 2>3 3 , Tt-HMS “my“ Tl-MCM-‘l TH? v-MCM-n Lb“! —’ LQ 76 catalyst 0 ’0 H202 83 catalyst TBl-IP we” # —.u E catalyst 0" OH H202 _. 61 catalyst l O 0‘ 0 H202 GO —> 73.79.33 catalyst O Al-MCM-dl 20H Ph 86 catalyst Al-MCM-ll catalyst ROH + —> 87 R0 0 Scheme 1. Reactions involving large organic molecules catalyzed by metal - Shbstituted MCM-41 and HMS as reported to date. 37 The catalytic expectations regarding A1 - substituted MCM-41 derivative are very high. Perhaps, a convincing evidence for that is the large number of patents on MCM-41 applications granted to Mobil Oil during 1991 - 92. 84 Most of these applications deal with oligomerisation and isomerization of C3 to C10 olefins over Al-MCM-41. There are also patents describing catalytic cracking and dealkylation of branched aromatics. 35 Very recently, Corma and co - workers, reported that acidic Al-MCM-41 can catalyze the alkylation of electron reach bulky aromatics, such as 2,4-DTBP with cinnamyl alcohol. 85 This reaction did not proceed with any significant conversion over microporous zeolite Y but mesoporous Al-MCM-41 afforded a good yield of the corresponding Friedel - Crafts products. In another development Kloetstra and van Bekuum 37 demonstrated that MGM-41 aluminosilicate could catalyze the tetrahydropyranylation of alcohols and phenols. The large mesopore size of MCM-41 allowed for the accommodation and the selective catalytic transformation of cholesterol to the corresponding tetrahydropyranyl ether (see Scheme 1). In conclusion, it is obvious that the invention of mesoporous molecular sieves will have an enormous impact on the industrial heterogeneous catalysis. Myriads of new and exiting opportunities for shape - selective catalytic transformations of large organic molecules are now open for exploration. We have no doubt that the day is not far off when their industrial application will be considered a common practice. 38 References 1. 2. 3. I. C. P. Broekhoff, B. G. Linsen, in Physical and Chemical Aspects of Adsorbents and Catalysts, B. G. Linsen, Ed., (Academic Press, London, 1970), p. 1. K. S. W. Sing, D. H. Everett, R. A. W. Haul, L. Moscou, R. A. Pierrotti, I. Rouquérol, T. Siemieniewska, Pure Appl. Chem, 57, 603 (1985). M. E. Davis, Chem. Ind., 1992, 137. 4. C. N. Satterfield. Heterogeneous Catalysis in Practice. (McGraw - Hill, New 5. 6. 10- 11- 12- 13- 14- 15- 16- 17- 18- 19, 20- York, 1980). The Colloid Chemistry of Silica, H. E. Bergna, Ed., (Adv. Chem. Ber. v. 234, ACS, Washington, DC, 1994). K. Wefers, C. Misra. Oxides and Hydroxides of Aluminum, (Alcoa Technical Paper No. 19, Alcoa Laboratories, 1987) p. 52. E. M. Flanigen, in Proc. 5th Intl. Conf. Zeolites, L. V. C. Rees, Ed., (Heyden, London, 1980) p. 760. D. E. W. Vaughan, Chem. Eng. Prog., 1988, 25. R. M. Barrer. Zeolites and Clay Minerals as Sorbents and Molecular Sieves, (Academic Press, London, 1978) p. 34. D. W. Breck. Zeolite Molecular Sieves: Structure, Chemistry and Use, (Wiley, London, 1974). R. M. Barrer, I. Chem. Soc., 1948, 127. R. M. Barrer, P. I. Denny, I. Chem. Soc., 1961, 971. W. M. Meier, D. H. Olson, Atlas of Zeolite Structure Types, 3rd edn. (Butterworth - Heinemann, London, 1992). R. M. Barrer, Zeolites, 1, 130 (1981). B. M. Lok, T. R. Cannan, C. A. Messina, Zeolites, 3, 282 (1983). M. E. Davis, R. F. Lobo, Chem. Mater., 4, 756 (1992). H. Gies, B. Marler, Zeolites, 12, 42 (1992). P. A. Jacobs, I. A. Martens. Synthesis of high - silica aluminosilicate zeolites, Stud. Surf. Sci. Catal. vol. 33, B. Delmon and I. T. Yates, Eds., (Elsevier, Amsterdam, 1987). I(. R. Franklin, B. M. Lowe, in Stud. Surf. Sci. Catal., vol. 49, 174 (1988). S. I. Zones, M. M. Olrnstead, D. S. Santilli, I. Am. Chem. Soc., 114, 4195 (1992). 21. 24. 26. 27. 28. 29. 30. 31. 32. 33. 35- 36- 37- 38- 39- 40- 41 42- 39 S. L. Burkett, M. E. Davis, I. Phys. Chem, 98, 4647(1994). R. von Balmos, I. B. Higgins. Collection of Simulated XRD Powder Patterns for Zeolites. A special issue of Zeolites, 10, 313 (1990). M. K. Rubin, E. I. Rosinski, C. I. Plank, 11.5. Patent 4,151,189 (1979). L. D. Rollmann, E. W. Valyocsik, US. Patent 4,108,881 (1978). R. A. Hearmon, A. Stewart, Zeolites, 10, 608 (1990). j R. M. Milton, 1.1.5. Patent 2,882,243 (1959). R. M. Milton, 11.8. Patent 2,882,244 (1959). C. Ho, C. B. Ching, D. M. Ruthven, Ind. Eng. Chem. Res., 26,1407 (1987). P. K. Shrivastava, R. Prakosh, I. Sci. Res., 11, 13 (1989). E. Roland, in Zeolites as Catalysts, Sorbents and Detergents Builders, Stud. Surf. Sci. Catal. vol. 46, H. G. Karge and I. Weitkamp, Eds., (Elsevier Sci. Pub., Amsterdam, 1989) pp. 645-659. A. Corma, A. Martinez, Adv. Mater., 7, 137 (1995). P. B. Weisz, V. I. Frillette, R. W. Maatrnan, E. B. Mower, I. Catal., 1, 307 (1962). ' M. E. Davis, Acc. Chem. Res., 26, 111 (1993). S. T. Wilson, B. M. Lok, C. A. Messina, T. R. Cannon, E. M. Flanigen, I. Amer. Chem. Soc., 104, 1146 (1982). M. E. Davis, C. Saldarriaga, C. Montes, I. M. Garces, C. A. Crowder, Nature, 331, 698 (1988). . M. Estermann, L. B. McCusker, Ch. Baerlocher, A. Merrouche, H. Kessler, Nature, 352, 320 (1991). Q. Huo, R. Xu, S. Li, Z. Ma, I. M. Thomas, R. H. Jones, A. M. Chippindale, I. Chem. Soc., Chem. Commun., 875 (1992). V. Soghmonian, Ch. Qin, R. Haushalter, I. Zubieta, Angew. Chem, Int. Ed. Engl., 32, 610 (1993). M. I. Annen, D. Young, M. E. Davis, 0. B. Cavin, C. R. Hubbard, I. Phys. Chem, 95, 1380 (1991). A. Merrouche, I. Patarin, H. Kessler, M. Soulard, L. Delrnotte, I. L. Guth, Zeolites, 12, 226 (1992). K. Beneke, G. Lagaly, Am. Mineral, 68, 818 (1983); G. Lagaly, K. Beneke, A. Weiss, Am. Mineral, 60, 642 (1975). T. Yanagisawa, T. Shirnizu, K. Kuroda, C. Kato, Bull. Chem. Soc. Ipn., 63, 988 (1990). . 43.} ill 0. .7 - a 43. 45. 46. 47. 48. 49. 50- 51- 52- 53 - 54- 55- 56- 57, 40 I. 5. Beck, C. T.-W. Chu, I. D. Johnson, C. T. Kresge, M. E. Leonowicz, W. J. Roth, J. C. Vartuli, W0 Patent 91/ 11390 (1991). J. S. Beck, II. 5. Patent 5,057,296 (1991). J. L. Casci, in Advanced Zeolite Science and Applications, Stud. Surf. Sci. Catal. vol. 85, I. C. Jansen, M. Stacker, H. G. Karge and J. Weitkamp, Eds., (Elsevier, Amsterdam, 1994) pp. 329-355. 1 J. S. Beck, J. C. Vartuli, W. J. Roth, M. E. Leonowitz, C. T. Kresge, K. D. Schmitt, C. T-W. Chu, D. H. Olson, E. W. Sheppard, S. B. McCullen, J. B. Higgins, J. L. Schlenker, I. Am. Chem. Soc., 114, 10834 (1992); C. T. Kresge, M. E. Leonowicz, W. J. Roth, J. C. Vartuli, J. S. Beck, Nature, 359, 710 (1992). V. Luzzati, in Biological Membranes, D. Chapman, Ed., (Academic, New York, 1968) pp. 71-123. S. Inagaki, Y. Fukushima, K. Kuroda, J. Chem. Soc. Chem. Commun., 8, 680 (1993). A. Monnier, F. Schiith, Q. Huo, D. Kumar, D. Margolese, R. S. Maxwell, G. D. Stucky, M. Krishnamurty, P. Petroff, A. Firouzi, M. Janicke, B. F. Chmelka, Science, 261, 1299 (1993). C.-Y. Chen, S. L. Burkett, H.-X. Li, M. E. Davis, Microporous Mater., 2, 27 (1993); M. E. Davis, C.-Y. Chen, S. L. Burkett, R. F. Lobo, in Better Ceramics Through Chemistry VI, MRS Symp. Proc., v. 346, A. K. Cheetham, Ed., (Pittsburgh, MRS, 1994) p. 831. Q. Huo, D. I. Margolese, U. Ciesla, P. Feng, T. Gier, P. Sieger, R. Leon, P. M. Petroff, F. Schiith, G. D. Stucky, Nature, 368, 317 (1994). P. T. Tanev, T. J. Pinnavaia, Science, 267, 865 (1995). M. Froba, P. Behrens, J. Wong, G. Engelhardt, Ch. Haggen-Miiller, G. van de Goor, M. Rowen, T. Tanaka, W. Schwieger, in Advances in Porous Materials, Sh. Komarneni, J. S. Beck, D. M. Smith, Eds., MRS Symp. Proc. Ser., vol. 371, (Pittsburgh, MRS, 1995), pp. 99-104. R. Schmidt, D. Akporiaye, M. Stocker, 0. H. Ellestad, in Zeolites and Related Microporous Materials, State of the Art 1994, Stud. Surf. Sci. Catal. vol. 84, J. Weitkamp, H. G. Karge, H. Pfeifer and W. Holderich, Eds., (Elsevier, Amsterdam, 1994), pp. 61-68. C.-Y. Chen, H.-X. Li, M. E. Davis, Microporous Mater., 2, 17 (1993). P. T. Tanev, T. J. Pinnavaia (manuscript in preparation). N. Coustel, F. Di Renzo, F. Fajula, J. Chem. Soc. Chem. Commun., 1994, 967. .nl/ MN. .3.-- Mix FM. 69. ,"J menNhnw.../mwnnw 58. 59. 61. 62. 815981 69. 70- 71- 73- 74- 7S- 76- 78- 41 D. D. Whitehurst, 11.5. Patent 5,143,879 (1992). J. C. Vartuli, K. D. Schmitt, C. T. Kresge, W. J. Roth, M. E. Leonowitz, S. B. McCullen, S. D. Hellring, J. S. Beck, J. L. Schlenker, D. H. Olson, E. W. Sheppard, Chem. Mater., 6, 2317 (1994). G. Horvath, K. J. Kawazoe, I. Chem. Eng. Ipn., 16, 470 (1983). P. T. Tanev, M. Chibwe, T. J. Pinnavaia, Nature, 368, 321 (1994). N. S. Gnepr et al., C. R. Acad. Sci. Ser. 2, 309, 1743 (1989); B. Chauvin, F. Fajula, F. Figueras, C. Gueguen, J. Bousquet, J. Catal., 111, 94 (1988). M. Taramasso, G. Perego, B. Notari, 1.1.5. Patent 4,410,501(1983). G. T. Kokotailo, S. L. Lawton, D. H. Olson, Nature, 272, 437 (1978). . D. R. C. Huybrechts, L. De Bruycker, P. A. Jacobs, Nature, 345, 240 (1990). 66. A. Esposito, C. Neri, F. Buonomo, 11.5. Patent 4,480,135 (1984). 67. 68. A. Esposito, M. Taramasso, C. Neri, F. Buonomo, Brit. Patent 2,116,974 C. Neri, A. Esposito, B. Anfossi, F. Buonomo, Eur. Patent 100,119 (1984). (1985); A. Tangaraj, A. Kumar, P. Ratnasamy, Appl. Catal., 57 L1 (1990). H. R. Sonawane, A.V. P01, P. P. Moghe, S. S. Biswas, A. Sudalai, I. Chem. Soc., Chem. Commun., 1994, 1215; S. Gontier, A. Tuel, Appl. Catal. A, 118, 173 (1994). G. Bellussi, M. Clerici, F. Buonomo, U. Romano, A. Esposito, B. Notari, Eur. Patent 200663 (1986). J. S. Reddy, R. Kumar, J. Catal., 130, 440 (1991). , M. A. Camblor, A. Corma, A. Martinez, J. Perez-Pariente, J. Chem. Soc. Chem. Commun., 1992, 589. M. W. Anderson, 0. Terasaki, T. 0hsuna, A. Phillipou, S. P. MacKay, A. Ferreira, J. Rocha, S. Lidin, Nature, 367, 347 (1994). D. P. Serrano, H.-X. Li, M. E. Davis, I. Chem. Soc. Chem. Commun., 1992, 745. M. S. Rigutto, H. van Bekkum, Appl. Catal., 68, L1 (1991); P. R. H. Rao, A. V. Ramaswamy, P. Ratnasamy, J. Catal., 137, 225 (1992). A. Corma, M. T. Navarro, J. Perez Pariente, J. Chem. Soc. Chem. Commun., 1994, 147. P. T. Tanev, T. J. Pinnavaia, unpublished results. A. Sayari, V. R. Karra, J. S. Reddy, 1. L. Moudrakovski, in Advances in Porous Materials, Sh. Komarneni, J. S. Beck and D. M. Smith, Eds., MRS Symp. Proc. Sen, vol. 371, (Pittsburgh, MRS, 1995), (in press). 42 79. K. M. Reddy, 1. Moudrakovski, A. Sayari, I. Chem. Soc. Chem. Commun., 1994, 1059. 80. C. B. Khouw, M. E. Davis, I. Catal., 151, 77 (1995). 81. A. Corma, M. T. Navarro, J. P. Pariente, F. Sanchez, in Zeolites and Related Microporous Materials, State of the Art 1994, Stud. Surf. Sci. Catal., vol. 84, J. Weitkamp, H. G. Karge, H. Pfeifer and W. Holderich, Eds., (Elsevier, Amsterdam, 1994), pp. 69-75. 82. T. J. Pinnavaia, P. T. Tanev, W. Jialiang, W. Zhang, in Advances in Porous Materials, Komarneni, Sh., Beck, J. S., Smith, D. M., Eds., MRS Symp. Proc. Ser., vol. 371, (Pittsburgh, MRS, 1995), pp. 53-62. 83. J. S. Reddy, A. Dicko, A. Sayari, ACS Symp. Ser., Proceedings of the Division of Petroleum Chemistry, Anheim, CA, 1995 (in press). 84. N. A. Bhore, Q. N. Le, G. H. Yokomizo, 11.3. Patent 5,134,243 (1992); Q. N. Le, R. T. Thomson, US. Patent 5,191,144 (1993). 85. Q. N. Le, R. T. Thomson, LLS. Patent 5,232,580 (1993). 86. E. Armengol, M. L. Cano, A. Corma, H. Garcia, M. T. Navarro, I. Chem. Soc. Chem. Commun., 1995, 519. 87. K. R. Kloetstra, H. van Bekkum, J. Chem. Res. (S), 1995, 26. CHAPTER TWO A NEUTRAL TEMPLATING ROUTE T0 MESOPOROUS MOLECULAR SIEVES 43 Abstract A neutral templating route to mesoporous molecular sieves is demonstrated based on hydrogen bonding interactions and self - assembly between neutral primary amine micelles (8°) and neutral inorganic precursors (1°). The S° 1° templating pathway affords ordered mesoporous materials with thicker framework walls, smaller X - ray scattering domain sizes, and substantially improved textural mesoporosities in comparison to M41S materials templated by quaternary ammonium cations of equivalent chain length. This synthetic strategy also allows for the facile, environmentally benign recovery of the cost - intensive template by simple solvent extraction methods. The S° I° templating route provides for the synthesis of other oxide mesostructures, (such as aluminas), that may be less readily accessible by electrostatic templating pathways. A T—_— 45 A. Introduction The discovery by Mobil researchers of the M415 family of mesoporous molecular sieves was accomplished by a self - assembly process involving electrostatic interactions between positively charged quaternary ammonium micelles and inorganic anions as framework precursors.1 Recently, Schiith, Stucky, and their co - workers 2 extended the electrostatic assembly approach by proposing four complementary synthesis pathways. Pathway 1 involved the direct co - condensation of a cationic surfactant (St) with anionic inorganic species (1') to give assembled ion pairs (8+ 1'), the original synthesis of MCM-41 silicates being a prime example. 1 In the charge - reversed situation (pathway 2), an anionic template (5‘) was used to direct the self - assembly of cationic inorganic species (1*) through 8' 1+ ion pairs. Pathways 3 and 4 involved counterion (X' or M+) - mediated assemblies of surfactants and inorganic species of similar charge. These counterion - mediated pathways afforded assembled solution species of the type S+ X' I+ (where X' = C1' or Br ') or, 8' Mt I' (where M+= Na+ or K+), respectively. The viability of pathway 3 was demonstrated by the synthesis of a hexagonal MCM-41 silica with quaternary ammonium cations under strongly acidic conditions (5 to 10 M HCl or HBr) in order to generate and assemble positively charged framework precursors. 2 Also, we have reported 3 the preparation of a mesoporous silica molecular sieve and a Ti - substituted analog by the acid - catalyzed hydrolysis of inorganic alkoxide precursors in the presence of primary ammonium ions. Because all of the above pathways are based on charge matching between ionic surfactants and ionic inorganic reagents, the template is strongly bonded to the charged framework and difficult to recover. In the original Mobil approach, 1 the template was not recovered, but simply burned - off by calcination at elevated temperatures. Recently, it has been demonstrated that the ionic surfactant in pathway 1 materials can be removed by ion - exchange in cation donor solutions. 4 Also, the template - halide ion pairs in the framework of acidic pathway 3 materials are displaced by ethanol extraction. 2 Thus, ionic template recovery is possible, provided that exchange ions or ion Pairs are present in the extraction process. We report a neutral templating route to mesoporous molecular sieves that is complementary to pathways 1 through 4. Our approach is based on 46 hydrogen bonding and self - assembly between neutral primary amine surfactants (5°) and neutral inorganic precursors (1°). This neutral S° I° templating route, which we denote pathway 5, affords mesostructures with larger wall thicknesses, small scattering domain sizes, and complementary textural mesoporosities relative to pathway 1 and 3 materials. The thicker pore walls improve the thermal and hydrothermal stability 5 of the mesopore framework, and the small crystallite domain size introduces textural mesoporosity, which facilitates accessing the framework - confined mesopores. 3'5 The 8° 1° pathway also allows for the facile recovery of the template by simple solvent extraction. B. Experimental We have prepared ordered mesoporous materials in the presence of C3 to C13 primary amines in water with ethanol as a co - solvent. The use of a co - solvent improved template solubility. The synthesis of a hexagonal mesoporous silica with dodecylamine (DDA) as the template illustrates the novelty of our synthetic strategy. In a typical preparation, tetraethyl orthosilicate (TEOS) (1.0 mol) was added under vigorous stirring to a solution of amine (0.27 mol) in ethanol (9.09 mol) and deionized water (29.6 mol). The reaction mixture was aged at ambient temperature for 18 hours, and the resulting hexagonal mesoporous silica (denoted HMS) was air - dried on a glass plate. Template removal was achieved by mixing 1 g of the air - dried HMS with 150 ml of hot ethanol for 1 hour. The product was then filtered and washed with a second 100 - ml portion of ethanol. This extraction procedure was repeated twice, and the crystalline product (denoted HMS-EB) was air - dried at 80°C. C. Results and Discussion The powder X - ray diffraction (XRD) patterns of HMS and HMS-BE are shown in Figure 12. Included for comparison is the pattern for an HMS sample calcined in air at 630°C for 4 hours (designated HMS-C). All patterns are similar and exhibit a single diffraction peak corresponding to d - spacing of 3.8, 4.0, and 3.5 nm, respectively. Higher order Bragg reflections of the hexagonal structure are not resolved. However, we and others have 47 >.. [- — m Z [.‘E‘. E (C) (B) k (A) I 1 L l l l l 1 l L i l 1 l 1 l 1 l l 0 4 8 12 16 20 DEGREES (29) Figure 12. Powder XRD patterns of (A) HMS, (B) HMS-C, and (C) HMS-EB. The patterns were measured on a Rigaku Rotaflex diffractometer equipped with a rotating anode and Cu - K; radiation (A. = 0.15418 nm). demonstrated 3:7 that similar "single reflection" MCM-41 - type products, still possess short - range hexagonal symmetry. The diffuse scattering at ~ 5° is attributable tohkO reflections that are broadened due to small crystallite domain effects. We believe that strong electrostatic interactions and charge matching are essential for the formation of the long - range - ordered hexagonal phase. The single XRD reflections and small x - ray scattering domain sizes of our hexagonal materials (see below) suggests that their formation is governed by weak, non - ionic interactions. We note that the intensity of the d100 reflection of HMS-C is twice, and that of HMS-EB four times that of the as - synthesized HMS. Thus, the removal of template by solvent extraction tends to preserve the crystallinity of the product whereas local heating during calcination might cause some degradation of the mesoporous framework. 48 Important trends are revealed by comparing in Figure 13 the N2 adsorption - desorption isotherms and the corresponding Horvath - Kawazoe 3 pore size distribution curves for HMS-C and HMS-BE. 0.25 > r 0.2 0.1 0 b dWIdR b 0.1 p b D p _, 0.05 ? p 011.5 2 2.5 3 3.5 4 4.5 5 m -4 mm VOLUME ADSORBED (cm3/g) oilililhltliluimlili o 0.2 0.4 0.6 0.8 1 PIIPO Figure 13. Nitrogen adsorption (—) and desorption (- -) isotherms for (A) HMS-C and (B) HMS-BE. 9 Insert: The corresponding Horvath - Kawazoe pore size distribution curves; d W/dR is the derivative of the normalized nitrogen volume adsorbed with respect to the pore diameter of the adsorbent. Both samples exhibit complementary textural and framework - confined mesoporosity, as evidenced by the presence of two separate, well - expressed hysteresis loops. The isotherm for HMS-EB is similar to that for HMS-C. The specific surface areas of both samples are also similar (1000 and 1150 m2/ g, respectively). These results imply that the template is efficiently removed by solvent extraction from the pore network of our molecular sieve. The ethanol - extracted product is thermally stable, as evidenced by the retention of the 4.1 - nm d spacing after calcination in air at 450°C for 7 hours. In accord with the preperties of MCM-41 materials, both samples exhibit a hysteresis loop in the 49 Pi/Po = 0.15 - 0.4 region 9 indicative of framework - confined mesopores (Figure 13). The size of these pores, as determined from the pore size distribution curves (see Fig. 13 insert), is 2.4 and 2.7 nm for HMS-C and HMS- EE, respectively. The smaller pore size of HMS-C could be attributed to the partial collapse of the mesoporous framework upon calcination. Table 2 summarizes the properties of the calcined. mesoporous silicas obtained by the 5° 1° and the electrostatic templating routes. The following features are evident: (i) all mesoporous products templated with neutral amines of C3 to C13 alkyl - chain length exhibit single XRD reflections with dioo and a0 values that are larger than those of the corresponding MGM-41 samples prepared with ionic surfactants; (ii) the scattering domain sizes of the materials prepared by neutral templating are much smaller (5 17 nm) than those of the corresponding MCM-41 materials (2 93 nm); (iii) the HK framework - confined mesopore sizes are similar for both classes of materials prepared with surfactants of equivalent alkyl - chain length; and (iv) the framework wall thicknesses of our 8° 1° mesostructures (2 1.7 nm) are consistently larger than those of MGM-41 materials prepared from charged templates of equivalent alkyl - chain length. The much thicker framework walls, together with the small scattering domain size and substantial textural mesoporosity of our materials, provide strong evidence in support of the weak templating forces governing the 5° 1° self - assembly mechanism. » The ability to completely displace the template from the mesoporous framework by ethanol extraction, without the need for exchange ions or ion pairs, also supports the neutral templating mechanism. The efficient removal of the template was confirmed by the absence of C-H vibrations in the infrared spectrum of the ethanol - extracted product and by the absence of a TGA weight loss over the temperature range where the amine normally undergoes thermal decomposition. 10 50 Table 2. Properties of calcined mesoporous silicas assembled by different templating routes. Templating Template (1100 Scattering a0 HK pore Wall route alkyl-chain (nm) domain (nm) size thickness length size (nm) (nm) (It!!!) 8° 1° C3 3.6 11.0 4.2 1.6 2.6 5° 1° C10 3.4 15.7 3.9 2.0 1.9 S° 1° C12 3.6 17.0 4.2 2.4 1.8 8° 1° C14 3.7 15.4 4.3 2.2 2.1 5° 1° C16 4.8 11.4 5.5 2.5 3.0 8° 1° C13 4.2 14.5 4.8 3.1 1.7 St 1' C3 ' 2.7 - 3.1 1.8 1.3 S+ 1' C10 ' 2.9 - 3.3 2.2 1.1 S+ 1' C12 I 3.1 240.3 § 3.6 2.5 1.1 S+ 1' C14 ' 3.3 - 3.8 3.0 0.8 5+ X‘ I+ C15 3.3 93.6 3.8 2.6 ‘ 1.2 5+ 1' C15 1 3.6 93.3 § 4.2 3.0 1.2 8" X‘ I+ C13 3.4 137.6 3.9 2.7 1.2 ' From ref. 1. I From ref. 13. 1 From ref. 14. § Calculated from the published XRD patterns. The scattering domain size was determined from the line width of the dloo x - ray reflection. The repeat distance (a0) between pore centers of the hexagonal structure is calculated from the XRD data with the formula a0 = 2d100/‘13. The framework - confined mesopore size (HK pore size) was determined by Horvath - Kawazoe (HK) analysis 8 of the N2 adsorption isotherms. The framework wall thickness is determined by subtracting the HK mesopore size from the repeat distance between pore centers. 51 Further evidence in support of the postulated mechanism was obtained by 14N NMR spectroscopy of neutral and intentionally protonated primary amine template solutions and a wet HMS product. 11 In agreement with the reported 14N NMR spectra of aqueous cetyltrimethylammonium cation solutions, 12 the spectrum of a partially protonated DDA solution (0.27 DDA : 0.054 HCl : 9.09 EtOH : 29.6 H20) exhibits a single isotropic resonance due to the 14N nucleus in a tetrahedral environment (one - fifth of the DDA protonated). This implies that significant concentrations of protonated amine should readily be detected by the applied NMR technique. In the absence of tetrahedral symmetry the signal produced by the 14N nucleus is broad and can not be detected. 12 Because the 14N NMR spectrum of our neutral DDA templating solution is featureless, the fraction of nitrogen centers adopting a tetrahedral environment is insignificant. In addition, the 14N NMR spectrum of the wet HMS product also lacks the isotropic resonance of a protonated amine. This result further verifies that the templating of our ordered mesoporous materials occurs primarily by the assembly of neutral amine molecules. We propose that the formation of our silica mesostructures occurs through the organization of the surfactant molecules into neutral rod - like micelles. Upon hydrolysis of TEOS, the resultant Si(OC2H5)4.,((OH)x species most likely participate in hydrogen - bonding interactions with the surfactant head groups. Further hydrolysis and condensation of the silanol groups results in short - range hexagonal packing of the micelles and framework wall formation. The versatility of this neutral templating approach is confirmed by our preparation of a non - layered templated alumina and a Ti - substituted analog. These ordered materials are obtained by a procedure similar to that for HMS with neutral octyl or dodecylamine as the template and neutral aluminum and titanium - aluminum alkoxide precursors. Both products exhibit strong dloo reflections, accompanied with more or less pronounced diffuse scattering centered at diffraction angles were the remaining hkO reflections of the hexagonal phase are expected. In addition, we have prepared a templated lamellar Si02 with neutral LIZ-diaminododecane. This product did not collapse upon removal of template by calcination in air at 630°C for 4 hours. 52 D. Conclusion The neutral 8° 1° templating route to regular mesoporous materials offers distinct advantages over the electrostatic templating methods of pathways 1 to 4. The ordered mesostructures obtained by this templating strategy have larger framework wall thicknesses, small crystallite domain sizes, and complementary textural mesoporosities in comparison to M418 materials templated by quaternary ammonium cations of equivalent chain length. In addition, the use of a non - ionic templating approach allows for the environmentally benign recovery of the cost - intensive template by simple solvent extraction. This circumvents the need for exchange ions or ion pairs to displace the ionic template. The neutral 8° 1° route complements electrostatic templating pathways to mesostructured materials and promises to facilitate the synthesis of ordered compositions (such as aluminas) that may be less readily accessible by ionic templating. 53 References and notes 1. 9‘2“?!" 9° C. T. Kresge, M. E. Leonowicz, W. J. Roth, J. C. Vartuli,]. 8. Beck, Nature 359, 710 (1992); J. S. Becket al I. Am. Chem. Soc. 114,10834 (1992). Q. Huo etal., Nature 368, 317 (1994). P. T. Tanev, M. Chibwe, T. J. Pinnavaia, ibid., p. 321. D. D. Whitehurst, US. Patent 5,143,879 (1992). S. B. McCullen and J. C. Vartuli, US. Patent 5, 156, 829 (1992); N. Coustel, F. Di Renzo, F. Fajula, I. Chem. Soc. Chem. Commun. 1994, 967 (1994). N. S. Gnepr et al., C. R. Acad. Sci. Ser. 2, 309, 1743 (1989); B. Chauvin, F. Fajula, F. Figueras, C. Gueguen, J. Bousquet, I. Catal. 111, 94 (1988). J. C. Beck et al., W0 91/113090 (1991); R. Schmidt, D. Akporiaye, M. Stocker, O. H. Ellestad, in Zeolites and Related Microporous Materials, State of the Art 1994, Studies in Surface Science and Catalysis, J. Weitkamp, H. G. Karge, H. Pfeifer and W. Holderich, Ed. (Elsevier Science B. V., Amsterdam, 1994), vol. 84, pp. 61—68. G. Horvath and K. J. Kawazoe, J. Chem. Eng. Ipn. 16, 470 (1983). . Pi is the equilibrium pressure of the adsorbate and Po is the saturation pressure of the adsorbate at the temperature of the adsorbent; adsorbed volume is at standard temperature and pressure. 10. The thermogravimetric analysis (TGA) was performed on a Cahn system TG analyzer with a heating rate of 5°C/m1n. The analysis of the as - synthesized HMS reveal ~ 47 °/o total weight loss upon heating to 800°C. Three distinguishing endothermic features centered at approximately 148, 238, and 460°C were observed. The first effect is attributed to the release of adsorbed water, the second to the desorption and decomposition of template, and the third to dehydroxylation of the surface. In contrast, the TGA of HMS-EB gave 11 % total weight loss with ~ 9 % corresponding to water desorption and surface dehydroxylation. 11. The 14N NMR spectra were recorded at 28.88 MHz on a Varian 400 VXR solid state spectrometer. Chemical shifts were referenced to an aqueous ammonium chloride solution. A quadrupole echo sequence with quadrature phase cycling was used with 90° pulse length of 4 us and interpulse delay (1:) of 115 to 120 .“3' Depending on the sample, 30 to 20000 54 transients were collected with a spectral width of 200 kHz and a recycle time of 140 ms. 12. C-Y. Chen, H-X. Li, M. E. Davis, Microporous Mater. 2, 27 (1993). 13. K. M. Reddy, 1. Moudrakovski, A. Sayari, J. Chem. Soc. Chem. Commun. 1994, 1059 (1994). 14. C-Y. Chen, H-X. Li, M. E. Davis, Microporous Mater. 2, 17 (1993). CHAPTER THREE A COMPARISON OF MESOPOROUS MOLECULAR SIEVES PREPARED BY IONIC AND NEUTRAL SURFACT ANT TEMPLATIN G APPROACHES 55 56 Abstract The physical properties of MCM-41 silicas prepared by direct 8+ 1' and counterion - mediated 8+ X' 1+ templating pathways and HMS silicas prepared by neutral 8° 1° templating method have been compared. The framework wall thickness, degree of framework cross - linking, pore structure, ease of template removal and thermal stability of the samples as well as the properties of the templating solutions were studied by means of XRD, TEM, SEM, N2 adsorption - desorption, 298i MAS NMR, TGA, chemical analysis and 14N liquid NMR. HMS materials prepared by neutral templating exhibited a smaller crystallite size, larger framework wall thickness, more completely cross - linked framework, superior thermal stability, and complementary textural mesoporosity in comparison to MCM-41 materials prepared with quaternary ammonium surfactants with the same chain length. Due to strong electrostatic interactions and charge matching, the cationic template in 8+1" MCM-41 materials is strongly bound to the framework and difficult to recover. In contrast, the 8° 1° templating approach allows for the effective and environmentally benign recovery and recycling of the neutral template. 57 A. Introduction The preparation of mesoporous molecular sieves was brought a stage nearer by the work of Kuroda et al. 1 on the surfactant directed rearrangement of single - layered siliceous host kanemite (NaH8i205.3H20) into long range cross - linked product. The driving forces for this self - assembly process were electrostatic interactions between assemblies of intercalated quaternary ammonium surfactants (8+) and the negatively charged layered host (I'). In 1992 scientists at Mobil Oil Research and Development accomplished the preparation of the first broad family of mesoporous molecular sieves (denoted M418) using different anionic silica precursor species. 2:3 Three members of the M418 family of materials were distinguished: hexagonal MCM-41; cubic MCM-48; and a lamellar silica phase. The originally proposed mechanism for MCM-41 formation involved strong electrostatic interactions and charge matching between liquid crystals of quaternary ammonium cations (8*), as structure directing agents, and anionic silicate oligomer species (I'). 2 Recently, Stucky and his colleagues extended the electrostatic assembly approach by proposing four complementary synthesis routes. 4 Pathway 1 involved the direct co - condensation of anionic inorganic species (1') with a cationic surfactant (8*) affording assembled ion pairs of a type 8+ 1', the original synthesis of MCM-41 silicas being the prime example. 2:3 In Pathway 2, the charge reversed case, an anionic template (8') was used to direct the assembly of cationic inorganic species (1+) into S‘ I+ organic - inorganic biphase arrays. Pathways 3 and 4 involved counterion (X‘ or M+) - mediated assemblies of surfactants and inorganic species of similar charge. These counterion - mediated routes afforded assembled solution species of type 8* X' P” (where X' = C1' or Br ‘) or, 8' Mt I' (where M+= Na’r or K+), respectively. The viability of Pathway 3 was demonstrated by the synthesis of a hexagonal MCM- 41 using a quaternary ammonium cationic template and strongly acidic conditions (5 - 10 M HCl or HBr) in order to generate and assemble positively - charged framework precursors. 4 Very recently, we have proposed a neutral (8° 1°) templating route 5 to mesoporous molecular sieves that is based on hydrogen bonding and self - assembly between neutral primary amine surfactants (8°) and neutral inorganic precursors (1°). This neutral 8° 1° templating route, which we 58 denoted pathway 5, allowed the preparation of hexagonal mesoporous silicas (HMS) with larger wall thicknesses for improved thermal stability, while at the same time it provided for the facile recovery and recycling of the neutral template by simple solvent extraction. 5 This work compares the physical properties and points out some important differences between the mesoporous molecular sieves'obtained by ionic and neutral templating methods. B. Experimental 1. Materials The sources of silica were N brand sodium silicate solution, containing 29 wt °/o 8i02 and 9 wt % N aOH, and tetraethylorthosilicate (TEOS) obtained form P. Q. Corporation and Kodak, respectively. The surfactants were: (i) quaternary ammonium salts with formula CnH2n+1N(CH3)3Br (denoted as Cn't), and (ii) primary amines with formula CnHZMINHz (denoted as Cn°) where n = 8, 10, 12, 14, 16 and 18. The surfactants were obtained from Aldrich, Lancaster and TCI America. Sulfuric acid (96 wt %) was purchased from J. T. Baker Chemical Co. Hydrochloric acid (36 - 38 wt °/o) was obtained from EM Science. Absolute ethyl alcohol (EtOH) was used as a co - solvent during HMS synthesis and as a solvent during template removal by solvent extraction. - 2. Synthesis The following describes the specific synthesis procedures. MCM-41 obtained by Pathway 8+ I'. These samples were prepared as described previously 3 using CnH2n+1N(CH3)3Br surfactants with alkyl chain lengths n= 8, 10, 12, 14, 16 and 18. Sodium silicate solution was used as a source of silica and the pH of the reaction mixture was adjusted as described 3 to ~ 10 with diluted sulfuric acid. The synthesis was carried out in an autoclave at 100°C for 6 days. MGM-41 obtained by Pathway 8+ X' 1+. 4 For these modified preparations, the appropriate amount of CnH2n+1N(CH3)3Br template (n= 8, 10, 12, 14, 16, 18) was dissolved in the acid/ water mixture followed by the addition of TEOS. This afforded a reaction mixture of the following molar composition: 1 TEOS : 0.13 CnH2n+1N(CH3)3Br : 5.4 HCl : 150 H20. 59 The reaction mixture was exposed to the open atmosphere and aged under vigorous stirring at ambient temperature for 1 week in order to prepare the crystalline MCM-41 product. Small amounts of deionized water were added during the aging process in order to compensate for the evaporation. The product was recovered by filtration, washed with deionized water and air - dried. ' HMS obtained by Pathway 8° 1°. These preparations were accomplished by hydrolysis of TEOS in the presence of CnH2n+1NH2 primary amines (n= 8, 10, 12, 14, 16, 18), water and EtOH as a co - solvent. The use of a co - solvent improved template solubility. In a typical preparation, TEOS was added to vigorously stirred solution of amine in ethanol and deionized water affording a reaction mixture of the following molar composition: 1 TEOS : 0.27 CnH2n+1NH2 : 9.09 EtOH : 29.6 H20. The reaction mixture was aged at ambient temperature for 18 hours, and the resulting hexagonal mesoporous silica (denoted HMS) was recovered by filtration, washed with deionized water, and air dried. 3. Template removal Template removal was achieved either by calcination in air at 630°C for 4 hr or by solvent extraction. The solvent extraction was performed by stirring 1 g of the air - dried product in 150 mL of hot (75°C) EtOH for. 1 hour. The product was then filtered and washed with another 100 - mL portion of ethanol. This extraction procedure was repeated twice, and the crystalline product was air - dried in an oven at 80°C for 1 hr. In some cases, complete cross - linking of the structure of the solvent extracted materials was accomplished by subsequent calcination in air at 630°C for 4 hr. 4. Analytical techniques Powder X - ray diffraction patterns were measured on Rigaku Rotaflex diffractometer equipped with a rotating anode and using Cu - K, radiation. Samples for TEM observation were prepared by embedding in L. R. White (hard) acrylic resin and sectioned on an ultramicrotome. The thin sections (~80 nm) were supported on 300 mesh copper grids and subsequently carbon coated to improve stability and reduce charging. The transmission electron micrographs were taken on a JEOL 100 CX microscope equipped with lanthanum hexaboride (LaB5) gun using an accelerating voltage of 120 kV and 6O 20 um objective lens aperture. The scanning electron micrographs were taken on a JEOL JSM 6400V using an accelerating voltage of 8 kV and a 16 mm working distance. The N2 adsorption - desorption isotherms were measured at -196°C on a Coulter Omnisorp 360 CX Sorptometer using a continuous adsorption procedure. Before measurement, samples were evacuated overnight at 150°C and 10'6 torr. The BET surface area was calculated from the linear part of the BET plot according to IUPAC recommendations. 5 The pore size distribution was estimated from the adsorption branch of the isotherms by the method of Horvath and Kawazoe. 7 29Si MAS NMR spectra were obtained on a Varian 400 VRX solid state NMR spectrometer at 79.5 MHz using 7 mm zirconia rotors, a pulse delay of 800 s and a sample spinning frequency of 4 kHz. The pulse delay time was determined by measuring the T1 (relaxation time) value for each group of samples and selecting the larger measured value (120 s) as a base for the determination. Thus, all spectra were acquired with pulse delay of more than 6 times the maximum measured T1 value. The chemical shifts were referenced to TMS (tetramethylsilane). 14N NMR spectra were recorded at 28.88 MHz on a Varian 400 VXR solid state spectrometer. Chemical shifts were referenced to an aqueous ammonium chloride solution. A quadrupole echo sequence with quadrature phase cycling was used with 90° pulse length of 4 us and interpulse delay (1) of 115 to 120 us. Depending on the sample, 30 to 20000 transients were collected with a spectral width of 200 kHz and a recycle time of 140 ms. The TGA curves were recorded in air on a Cahn TG system 121 thermogravimeter using a heating rate of 5°C / min. Chemical analyses for C, H, N, and Na were carried out by Microanalysis Laboratory, School of Chemical Sciences, University of Illinois at Urbana - Champaign, IL. C. Results and Discussion 1. Structure and morphology Figure 14 A shows the powder X - ray diffraction (XRD) patterns of the calcined HMS and MCM-41 materials prepared using neutral and charged surfactants of different alkyl chain length. All HMS materials prepared by 8° 1° 61 templating exhibit single dloo reflections accompanied by more or less pronounced diffuse scattering centered at ~ 5°29. Higher order Bragg reflections are not resolved in the patterns of the 8° 1° HMS materials. However, we and others have demonstrated 5'39 that similar ”single reflection” materials still have short - range hexagonal symmetry. We attribute the single XRD reflections of HMS materials to small scattering domain sizes (see below) and to disorder resulting from the weak, nonionic forces that operate in the 8° 1° assembly process. As the surfactant chain length increases from C3° to C13° the position of the [100] reflection shifts toward lower 29 angles or toward higher dloo - spacings. This result is not unexpected and implies that surfactants with longer alkyl chain lengths would afford materials with larger uniform mesopore size. Similar XRD patterns are indicated in Figure 14 B for the calcined 8+ 1' MCM-41 samples prepared at 100°C. However, these 8+ 1‘ materials exhibit sharper £1100 reflections accompanied with diffuse scattering or, in some cases, additional 110, 200 and 210 reflections indicative of an hexagonal phase. The absence of reflections with non - zero 1 component has been attributed 2:3 to the lack of local order in the walls of the hexagonal MCM-41. The observed hkO reflections were assigned to scattering arising from the long range - ordered hexagonal uniform mesopores. It is interesting to note that for these 8* I' MCM-41 materials (prepared at 100 °C) surfactant chain lengths of at least C12+ units are necessary in order to observe the additional reflections indicative of hexagonal structure. 0n the other hand, the calcination of the 8* I‘ MCM—41 sample prepared with C16+ template significantly affects the 110, 200 and 210 reflections observed on the pattern of the as - synthesized material. The absence of these higher order reflections in the calcined product was unusual in view of the low heating rate of 1°C/ min. In this connection, our attempt to prepare a 8+ I'MCM-41 silica at 100°C with C18+ quaternary ammonium ions afforded an as - synthesized material with multiple reflections that could not be assigned to any of the known M418 phases (Figure 15 A). The calcination of this material afforded a completely amorphous product (pattern B). Perhaps, the combustion of large amounts of organic phase during calcination of the C16+ and C13+ templated samples leads to a significant thermal destruction of the framework walls. Figure 14 C shows the powder XRD patterns of the calcined MCM-41 samples prepared at ambient temperature by the acidic S+ X" 1+ pathway. All 62 $8555 L -x +m use. .5 3 3.3.3 .85. 5202 Q E. .38.. -_ t... .3 8.3.3 .85. ~36: av .85.... masses.- ._ .m .3 3.323 .85. m2: 9: a: we. .2 .3 .2 .2 .w "5 .535 5.6 5:. Ease.- ao ensued... 3.8 5828 van Aozuv 3.53m: its moan—9:2 8.56 52829: genome-m2: pogo—mu we 85on Dmx .8950.“ .3 0.3me 6N. mummomo 8% mummomo , 6N. mummomo ca 3. c o as mp or m o ow up or m o ple b» P h w. .J. D P PIL b F D P D - 55 P by p F P o... l. D b h r . O m .:o 020 N X 08:0 2 o _. m n x 0' X o 0 a 0: 0 MI W M m v— +30 w 451-4 0:0 0 0 3.5.05. 3.5.05. mi: +_ -x +m .— a.“ 0— cm (Ill 001 n: so A3 63 Intensity (A) W L (a) j I I I I I I I I I I I I I I I I I o 5 1 o 1 20 Degrees (26) Figure 15. Powder XRD diffraction patterns of (A) as - synthesized and (B) calcined (650°C) 5+ 1' MCM-41 samples prepared from C13+ surfactant. 64 attempts to prepare MCM-41 materials from Cn“ surfactants with n < 14 afforded X - ray amorphous or very poorly ordered products. In addition, it seems that surfactant chain lengths of at least C16+ units are necessary in order to observe the additional 110 and 200 reflections of the hexagonal phase. In contrast to the 5* I“ approach, the S+ X‘ I+ pathway allows for the preparation of stable and very well - ordered MCM-41 products With surfactants of alkyl chain lengths 2 C16+. The comparison of these S+ X' 1+ MCM-41 XRD patterns with the ones corresponding to the other electrostatic counterpart (compare Figures 14 C and 14 B) reveals another interesting fact. Different surfactant chain lengths are necessary in order to observe long range - ordered hexagonal phase in the two reaction systems. These differences could be attributed to the possibility of different conformations adopted by the quaternary ammonium surfactants under different pH conditions. In addition to pH, the surfactant and reagents concentration, the presence of different ions, the ionic strength, and the temperature of the reaction may all influence the conformations adopted by the quaternary ammonium surfactants of equivalent chain length. Figure 16 shows representative TEM lattice images of HMS prepared by neutral templating and MCM-41 prepared by S+ X' I+ pathway. In accord with the adsorption data (see below) both materials exhibit uniform hexagonal mesopores. Due to the high degree of order MCM-41 exhibits long range hexagonal ordering of the mesopores (see image B). On the other hand, the image for HMS (image A) shows that the mesopores are oriented in a short - range hexagonal fashion. Figure 17 shows representative SEM micrographs of calcined HMS (photo A) and MCM-41 samples (photos B and C). The materials exhibit very different morphologies. Our neutral HMS material consists of small non - uniform aggregates that are distinctly composed by particles with very small sizes. In contrast, MCM-41 sample prepared by 8* I'templating pathway exhibits larger, pitted and more compact aggregates (see photo B). Contrary to the findings of Mobil, 2 we did not observe any hexagonal prismatic aggregates for the MCM-41 samples prepared with Cu“ and C16“ surfactants. Well - ordered hexagonal prismatic aggregates were also reported by Stucky et al. 4 for the 8* X' 1* MCM-41 materials. However, the morphology of our MCM-41 sample prepared by 5* X' 1+ templating is different. Each aggregate consists of a multiple ”cotton - candy” - like particles. Therefore, the intensity of agitation 65 Figure 16. Representative TEM micrographs for calcined (A) 8° 1° HMS and (B) S+ X' 1"“ MCM-41 samples. 65A Figure 16. (Cont’d). 66 66B 67 Figure 17. Representative SEM micrographs of calcined (A) 8° 1° HMS, (B) 8* I' MCM-41, and (C) S+ X' 1* MCM-41 samples. 67A — ltJm F2 L01 X4; 588 15mm 68 used in the MCM-41 synthesis may be responsible for the preparation of samples with different particle morphology. 2. Sorption properties and framework wall thickness Figures 18 - 20 show the N2 adsorption - desorption isotherms for calcined 5° 1° HMS, 5+1" MCM-41, and 8“ X' 1+ MCM-41 samples prepared with neutral and cationic surfactants of different alkyl chain lengths. For the purpose of simplifying the following discussion we would like to define and differentiate the terms ”framework - confined” mesoporosity and ”textural” mesoporosity. The framework - confined mesoporosity is the mesoporosity corresponding to the total pore space of the uniform framework - confined mesopores. These mesopores are formed upon hydrolysis, condensation and polymerization of the inorganic precursor followed by crystallization of the silicate framework around the rod - like surfactant micelles. The size of these framework - confined mesopores is equal to the diameter of the framework channels or to the diameter of the parent rod - like micelles and can be varied by changing the surfactant alkyl chain length or by the addition of auxiliary organics. 2'5 The presence of framework - confined mesopores is indicated by the sharp adsorption step or well - expressed hysteresis loop in the corresponding N2 adsorption - desorption isotherm centered in the low relative pressure (Pi/Po) region from 0.1 to 0.5 (see for example Figure 18). The textural mesoporosity is the mesoporosity that can be attributed to the non - uniform voids and channels formed by interparticle contacts. Each of these particles in the case of HMS and MCM-41 molecular sieves is composed of certain number of framework unit cells or framework - confined uniform mesopores. The size of the textural mesopores is determined by the size, shape and the number of interfacial contacts of these particles or aggregates and usually is at least one order of magnitude larger than the size of framework - confined mesopores. The presence of textural mesoporosity is verified by the appearance of a well - defined hysteresis loop in the corresponding N 2 adsorption - desorption isotherms located at the Pi/Po region from 0.5 to 1.0. An quantitative indication of the extent of certain mesoporosity is provided by the sharpness of the sorption step or the area encapsulated by the hysteresis loop (see Figure 18). Thus, the sharper the adsorption step or the larger the area contained by a given hysteresis loop, the larger the amount of the corresponding mesoporosity. 69 8° l° HMS VOLUME ADSORBED (cc/g, STP) 1200 o .1 .2 .3 .4 .5 .6 .7 .a .9 1.0 PilPo Figure 18. N 2 adsorption - desorption isotherms for calcined 5° 1° HMS samples prepared with neutral primary amine surfactants (Cn°) of different chain lengths (n= 8, 10, 12, 14, 16, 18). 70 The N2 adsorption - desorption isotherms for our HMS samples prepared from neutral primary amine surfactants are shown in Figure 18. It is obvious that the shorter alkyl chain lengths (C3° - C10°) afford predominately textural mesoporosity, whereas longer alkyl chains afford mostly framework - confined mesoporosity. In addition, the surfactants with intermediate chain length (such as C12°) seems to give a balanced or complementary framework - confined and textural mesoporosity. The N2 adsorption - desorption isotherms of the 5* I'MCM—41 samples prepared with quaternary ammonium surfactants are shown in Figure 19. Again shorter alkyl chain lengths afford predominately textural mesoporosity, whereas longer alkyl chains give mostly framework - confined mesoporosity (with some exceptions). However, the textural mesoporosity exhibited by these samples is negligible in comparison to the HMS materials. The MCM-41 samples prepared by the acidic S+ X' 1* pathway exhibit N2 adsorption - desorption isotherms similar to those of the 8" I‘ electrostatic counterparts. They also lack appreciable textural mesoporosity (see Figure 20). The lack of textural mesoporosity of the MCM-41 samples prepared by both electrostatic pathways could impose serious diffusional limitations in adsorption and catalytic processes involving bulky organic molecules. Table 3 summarizes the properties of the calcined mesoporous silicas prepared by neutral and electrostatic templating routes. The following trends are evident: (i) all 8° 1° mesoporous materials templated with primary amines of alkyl chain length from C3° to C13° exhibit single XRD reflections with Lima and a0 values that are larger than those of the corresponding MCM-41 samples prepared with ionic surfactants; (ii) the scattering domain sizes of the materials prepared by neutral templating are smaller (5 17 nm) than those of the corresponding MCM-41 materials (2 34 nm); (iii) the HK framework - confined mesopore sizes are similar for both classes of materials prepared with surfactants of equivalent alkyl - chain length; and (iv) the framework wall thicknesses of our 8° 1° mesostructures (2 1.7 nm) are consistently larger than those of MCM—4l materials prepared from charged templates of equivalent alkyl - chain length. The small framework wall thickness of 8* I‘ MCM-41 materials has been attributed 10 to the limiting efi'ects of both the electrostatic repulsions between anionic silicate oligomers, and the charge compensation between cationic micelles and anionic inorganic species. 71 S“ l' MOM-41 VOLUME ADSORBED (cc/g, STP) 1200 Pi/Po Figure 19. N 2 adsorption - desorption isotherms for calcined S+ 1' MCM-41 samples prepared with quaternary ammonium cationic surfactants (Cn+) of different chain lengths (n= 8, 10, 12, 14, 16). 72 5+ X' l+ MOM-41 600 VOLUME ADSORBED (cc/g, STP) Pi/Po Figure 20. N2 adsorption - desorption isotherms for calcined 5* X‘ 1* MCM-41 samples prepared with quaternary ammonium cationic surfactants (Cn+) of different chain lengths (n= 14, 16, 18). 73 TABLE 3. Properties of calcined mesoporous silicas assembled by different templating routes. I Templating Template dloo Scattering a0 HK pore Wall : route alkyl (nm) domain (nm) size thickness 5 chain size (nm) (nm) ; len_gth (nm) ‘ 8° 1° C3 3.6 11.0 4.2 1.6 2.6 i 8° 1" C10 3.4 15.7 3.9 2.0 1.9 { 5° 1° C12 3.6 17.0 4.2 2.4 1.8 5 8° 1° C14 3.7 15.4 4.3 2.2 2.1 4 5° 1° C16 4.8 11.4 5.5 2.5 3.0 5° 1° C18 4.2 14.5 4.8 3.1 1.7 5+ 1- 68 2.8 - 3.2 1.7 1.5 5+ 1- C10 3.0 34.0 3.5 2.1 1.4 3+ 1- C12 3.2 52.0 3.7 2.4 1.3 . 5+ 1- C14 3.1 102.6 3.6 2.1 1.5 , 3+ 1- C16 3.1 90.0 3.6 2.2 1.4 . 5+ 1- C18 amorph - - - - 3+ x-1+ 514 2.7 - 3.1 - - 3+ x- 1+ C16 3.1 53.0 3.6 2.2 1.4 5+ x- 1+ C18 3.4 66.2 3.9 2.7 1.2 The scattering domain size was determined from the line width of the d100 X - ray reflection. The repeat distance (a0) between pore centers of the hexagonal structure is calculated from the XRD data using the formula a0 = 2d1oo/N/3. The framework - confined mesopore size (HK pore size) was determined by Horvath - Kawazoe (HK) analysis of the N2 adsorption isotherms. The framework wall thickness is determined by subtracting the HK mesopore size from the repeat distance between pore centers. 74 The small variation of the framework wall thickness of our MCM-41 materials (see Table 3) is supported by the findings of Stucky et al. (0.8 to 0.9 nm) 10 and Davis and co - workers (~1.0 nm). 11:12 The framework wall thickness of MCM—41 samples prepared by the acidic 8“ X‘ I+ pathway also seems to be similar, 1.0 - 1.2 nm. 4 The small framework wall thicknesses of the electrostatically templated 'MCM-41 materials could seriously limit their thermal and hydrothermal stability in adsorption and catalytic applications requiring severe regeneration. Recently, Fajula and co - workers demonstrated that MCM-41 samples with thicker pore walls exhibit higher thermal stability. 13 The much thicker framework walls of the HMS materials prepared by 8° 1° pathway could be expected to afford superior thermal and hydrothermal stability to that exhibited by MCM-41 materials prepared by electrostatic templating. Table 4 summarizes the sorption properties of the ordered mesoporous silicas prepared by neutral and electrostatic templating. Both classes of materials exhibit similar values of BET specific surface area (SBET) and framework - confined mesopore volume (Vfi), but differ quite significantly in total pore volume (Vt), textural mesopore volume (Vtex), and ratio of textural to framework - confined mesoporosity (Vtex/Vfr). Thus, the HMS materials exhibit much larger Vt, Vtex, and Vtex/ Vf, ratio relative to the electrostatically templated MCM-41 samples. In addition, neutral templating allows for a large variation of the ratio of Vtex/Vfr mesoporosity, whereas electrostatic templating affords uniformly low values for this ratio. The low textural mesoporosity could have serious limitations on the potential application of the electrostatic MCM-41 materials in adsorption and catalytic processes controlled by diffusion. 75 TABLE 4. Sorption properties of calcined mesoporous silicas assembled by different templating routes. ‘ Templating Template Route alkyl chain length 8° 1° C8 8° 1° C10 5° 1° C12 5° 1° C14 5° 1° C15 8° 1° C13 8"” 1' C8 8"“ I' 5"” 1' 5+ 1' 8"" 1' S+ 1' S+ X' 1+ 8"" X‘ 1+ 8" X' 1"” 76 3. 2‘o’Si MAS NMR Recently, Mobil scientists reported the 29Si MAS NMR of as - synthesized M418 samples prepared by the electrostatic S+ 1‘ pathway. 14 Using deconvolution techniques they resolved three broad and overlapping peaks in the spectra of MCM.41, MCM-48 and two peaks in the spectra of the thermally unstable lamellar phase. The spectra of both hexagonal MCM-41 and cubic MCM-48 were essentially the same as that for amorphous silica. This is not surprising giving the amorphous character of the framework walls and the expected wide range of Si-O-Si bond angles. Interestingly, all 8* 1' phases exhibited more non - condensed Si(SiO)x(OH)4.x framework units (where X: 2 or 3, respectively, for Q2 and Q3) than fully condensed (X: 4) units as evidenced by the low ratios of Q4/Q2+Q3 (typically, around 0.67). 14 This result is extremely significant and implies a relatively low degree of cross - linking of the ”electrostatic” S+ 1' molecular sieves. The low degree of cross - linking of the as - synthesized S+ 1' MCM-41 (a ratio of 0.77) is also evident from the data of Davis et 511.1142 These authors first suggested that a 100% Q4 state corresponding to full cross - linking of the electrostatic MCM—41 framework can not be reached because unlinked SiO‘ groups are needed for charge compensation of the quaternary ammonium cationic template (5*). The previously reported 29Si MAS NMR spectra of as - synthesized MCM- 41 phases prepared by the acidic 5* X' 1+ pathway also exhibited low 04/ Q3 ratios < 1.0. 4 The positive charge on the surfactant is partially balanced by an excess of halide counterions. This allowed for displacement of ~85 wt % of the template - halide ion pairs by solvent extraction. 4 Figure 21 shows the 29Si MAS NMR spectra of our as - synthesized and calcined mesoporous molecular sieves prepared by neutral and electrostatic templating. The spectrum of the as - synthesized HMS material (A) reveals a much larger 04/ Q3 ratio (2.68) or much more highly cross - linked framework than the electrostatic MCM-41 counterparts (spectra B and C). The low Q4/ Q3 ratios of 0.71 and 1.18 exhibited by our as - synthesized S+ 1' and S+ X' 1+ MCM- 41 silicas, respectively, are in good agreement with the previously reported data (see above). The comparison of the 29 Si MAS NMR spectra of the as - synthesized and calcined samples emphasizes yet another important point. AS-SYNTHESIZED CALCINED 04/03 04/03 (C) 1.18 IIIIII " fl r r - . v - . - v "'.' "r -- v 'v -- I I . - '1' ' 1' I 1""1" '1 "'1 " 1'": & -50 -60 -70 fl J) .100 4” 440 .160 «0 -50 -& ~70 N c” 4m -.l20 440 -|60 PPM Figure 21. 29Si MAS NMR spectra of as - synthesized (A) 5° 1° HMS, (B) 5* I‘ MCM-41, (C) 8* X' 1+ MCM-41, and calcined (D) 5° 1° HMS, (E) 8"” I' MCM-41, (F) 5* X' 1* MCM-41 mesoporous molecular sieves. 78 The ratio of Q4 / Q3 for our 5° 1° material does not change significantly upon calcination (compare spectra (A) and (D)). This observation is not surprising due to the lack of electrostatic repulsions and charge matching and the corresponding thicker pore walls of our HMS materials prepared by neutral templating. The fact that the amount of non-condensed Q3 units remains essentially the same could be attributed to the incorporation of some of the (SiO)3Si-OH groups into the thicker pore walls. 4. Template removal and thermal stability Figure 22 shows the powder XRD patterns of as - synthesized, ethanol - extracted, ethanol - extracted and calcined mesoporous molecular sieves prepared by neutral and electrostatic templating pathways. The XRD data for our HMS materials (see Figure 22 A) clearly indicate preservation of the structure upon template removal by solvent extraction. It is remarkable that subsequent calcination of this ethanol - extracted sample did not affect the crystallinity of the product. We note that the intensity of the d 100 reflections of the ethanol - extracted and ethanol - extracted and calcined HMS products are four times that of the as - synthesized HMS. This is probably due to the more completely cross - linked framework of the HMS materials and to efficient removal of the neutral template. This suggests that preferred method for template removal should be ethanol extraction followed by calcination. In contrast to the above results are the data for the electrostatic S+ 1' MCM- 41 materials (see Figure 22 B). Our attempt to remove the cationic template by ethanol extraction resulted in significant broadening of the corresponding XRD reflections. It is very important to note that subsequent calcination resulted in almost complete decomposition of the structure. These observations are in accord with the low 04/ Q3 ratio or insufficient degree of framework cross - linking of these materials as determined by 298i MAS NMR (see Figure 21). As illustrated by Figure 22 C, the crystallinity of the MCM-41 sample prepared by the acidic 5" X' 1+ pathway retained after template removal by ethanol extraction and subsequent calcination. This is probably caused by the successful displacement of most of the template - halide ion pairs (S+ X') by solvent extraction. 79 43.62 L -x +m Q 2.. also: .. +m é .mzm L .m g 82.x. 838.9: machomoflfi “5538 was pmaombxw - 3556 €688.68 - 3:23 .pmnmmmficzm - an we .56th Dmx 8225 .3 «gamma 5N. mummomo 65 mummomo 63 $553 cm 2. or m c cu m_. or m 9 ON m.. o_. m c ...._...._T..._._.. .Ilnhlllp.__ru__._.__.__ _...__..._._._b____ 63665593 j umfimofigmém it»); 3535588 x 62092340550 36858-6:sz bosomzxmnocwfio \ 3528 new “62082340556 ALISNELNI 850.3 can 6208.6-.0556 posofio pea C neomzxonocafio 3-202 3.20: ms...“ +— .x +w u— +m O O SC EC 2: 80 Figure 23 shows the TGA curves of as - synthesized, ethanol - extracted, and ethanol - extracted and calcined mesoporous molecular sieves prepared by neutral and electrostatic templating pathways. The TGA curves for S° I° HMS samples are depicted in Figure 23 A. Several weight loss features can be distinguished in the TGA curve for the as - syntheSized HMS material. The region from ~ 20 to 150°C could be attributed to desorption of water, the second region from ~ 150 to 300°C, to the decomposition and combustion of the organic template, and the third from ~ 300 to 520°C, to dehydroxylation of the surface. However, the TGA curve of the ethanol - extracted HMS, as well as the corresponding x 10 blow - up of the curve (shown in Figure 23 A as insert), clearly lack an inflection point indicative of template weight loss. This result suggests that the neutral template can be efficiently removed and recycled from the framework of the HMS materials by simple solvent extraction. The so extracted template could be easily reused by simple evaporation of the solvent. Further verification of the efficient extraction of the neutral template from HMS materials is obtained by comparing the N2 adsorption - desorption isotherms and corresponding HK pore size distribution curves presented in Figure 24. The isotherms and pore size distribution curves of both calcined HMS (curves A) and ethanol - extracted HMS (curves B) are very similar. The specific surface areas of both samples are also similar (1150 and 1000 m2/ g, respectively). These results again imply that the neutral template is effectively removed by solvent extraction from the pore network of our 5° 1° molecular sieve. The size of the framework - confined mesopores, as determined from the pore size distribution curves (see Figure 24 insert), is 2.4 and 2.7 nm for the calcined HMS and ethanol - extracted HMS, respectively. The smaller pore size of the calcined HMS could be attributed to the partial collapse of the mesoporous framework upon calcination. In contrast to these findings, the TGA curve for the ethanol - extracted MCM-41 sample prepared by S+ 1' templating (see Figure 23 B) indicates more than 25 % weight loss due to decomposition and combustion of remaining cationic template. This implies that the framework of the as - synthesized S+ 1' MCM-41 materials is negatively charged and that the cationic template is strongly bound to the framework and difficult to recover. 81 o..mm:»- \\\_ I: x5 (I) 2 E (a) .- ethanol-extracted z 8 001 2.2 - 002 1.3 L x5 8 as-synthesizad o 001 2.3 (A) £02 1.2 x5 I I I I I I l I I I I I I I—T 0 5 1 0 1 5 20 DEGREES (20) Figure 28. Powder X - ray diffraction patterns of (A) as - synthesized, (B) ethanol - extracted, and (C) ethanol - extracted and calcined TPLM. 101 Figure 29. Transmission electron micrographs of the as - synthesized TPLM sample clearly showing regions of (A and B) multilamellar vesicles and (B and C) lattice fringes of well - ordered lamellar TPLM phase. 101A 102 Figure 29. (Cont’d). 102A 103 Figure 29. (Cont’d). 103A 104 Figure 29. (Cont’d). 104A 105 Figure 30. Scanning electron micrographs of (A and B) the as - synthesized, and (C) calcined TPLM sample. 105A 106 200 subset rather than the expected lamellar one. However, the possibility of solvent extraction reorganizing our lamellar surfactant - inorganic arrays into hexagonal structure similar to that observed for kanemite 7 is precluded by the TEM images that are alike those shown in Figures 29 C and D. Also, corresponding electron diffraction patterns observed from the most ordered regions of the sample exhibited typical lamellar arrangement of diffraction maxirna. In addition, the N2 adsorption - desorption isotherm for the solvent - extracted material is typical for pillared lamellar solid (see below) and very different from that of a hexagonal MCM-41. In contrast to all previously reported templated lamellar phases our ethanol - extracted and ethanol - extracted and calcined sample clearly retain their crystallinity (see Figure 28 curves B and C). Both the as - synthesized and ethanol - extracted TPLM samples exhibit superior thermal stability relative to the conventional pillared lamellar materials. Our experiments indicate that the thermal stability of TPLM samples is comparable to the hexagonal MCM-41 and HMS molecular sieves. This is evidenced by the retention of characteristic X - ray diffraction patterns upon calcination in air up to 800°C for 4 hr. We postulate that the formation of TPLM occurs in a manner reminiscent to the natural biomineralization processes. As illustrated in Figure 31 our biomineralization process most likely occurs trough the assembly of neutral diamine surfactants (S°-5°) into multilamellar vesicles with the surfactant head groups pointing toward the solution (Figure 31). The multilamellar region of these vesicles is composed by closely - packed layers of surfactant molecules divided by water layers. The width of this multilamellar region judging from TEM data is ~ 32 nm. Therefore, approximately 10 to 15 silica layers or S°-S° I° units could be accommodated (in a sandwich - like manner) by replacing the water from the surfactant - water interfaces of this multilamellar region. The addition of TEOS to the above solution of multilamellar surfactant vesicles affords the hydrolyzed neutral Si(OC2H5)4- xOHx species. These species are most likely capable to penetrate the vesicle interface, diffuse into the multilamellar regions and to participate in H - bonding interactions with the lone pairs on the surfactant head groups. Further cross - linking and polymerization of adjacent silica species most likely causes simultaneous growth of the parallel layers and intergallery pillars. This is possible because of the close proximity of the layers and the lack of electrostatic repulsions and charge matching between the neutral silica 107 Multilamellar vesicle // \ \ :\ sex _\ /// \\.\.\\\ m \\\\\\ \ .__\ ~\ 3. \: \ \ §\\ \ \\ \\ L .‘r \ \~ / organic-inorganic complex Figure 31. Proposed biomimetic I° S°-S° I° templating mechanism of formation of TPLM. 108 oligomer species and the neutral surfactant. Thus, in contrast to the existing electrostatic templating approaches which afforded 2-D unstable inorganic lamellar phases our biomimetic neutral templating method allows for the preparation of highly cross - linked and stable porous lamellar materials. Further support for the biomimetic intravesicular assembly mechanism is obtained from the comparison of the 29Si MAS NMR spectra and the corresponding 04/ Q3 ratios presented in Figure 32. The spectrum of the as - synthesized TPLM exhibits two well - defined Q3 and Q4 peaks and a Q4/Q3 ratio of 1.16. This ratio differs substantially from the previously reported Q4 / Q3 ratios (from 0.71 to 0.92) 27'” for the lamellar M41S materials prepared by S“ I' and S+ X' 1* templating and indicates higher degree of cross - linking for - our as - synthesized TPLM. We attribute this difference to the possible presence of pillars in the gallery of our templated material. Ethanol extraction affords further polymerization of the structure as evidenced by the much higher Q4 / Q3 ratio of 1.63. The removal of template by calcination leads to an even more completely cross - linked structure. This is due to the transformation of a large portion of the SiO3OH groups into SiO4 units (Q4 / Q3 = 3.41). The 29Si MAS NMR data for as - synthesized TPLM are to be contrasted with the 04/03 ratios observed for the as - synthesized lamellar M418 phases (0.71 - 0.92), 27'” hexagonal MCM-41 (0.75), 27 hexagonal HMS (2.7), 30 and cubic MCM-48 (0.61 - 0.91) templated phases. 9'29 This suggests that our templated TPLM samples are distinguishable form the previously reported templated lamellar, hexagonal and cubic phases. We believe that this difference is due to: (i) the ability of the neutral or partially - hydrolyzed inorganic precursor to penetrate freely the hydrophilic and hydrophobic regions of the vesicles; (ii) the lack of electrostatic repulsions between the neutral silica precursor species; (iii) the absence of charge matching surfactant- inorganic phase interactions; and (iv) to the close proximity of the layers which may allow for the formation of intergallery pillars of condensed silica species. Important trends are revealed by the comparison in Figure 33 of the TGA curves of the above samples. Four effects of weight loss are distinguished in the TGA curve of the as - synthesized sample (curve A). The first effect of weight loss (at 25 to 130°C) is due to desorption of water and the second (130 to 225°C) to the partial loss of the diamine template. The third effect (225 to 400°C) is probably caused by the combustion of the intergallery template, 109 04/03 3.41 (C) ~40 ~50 ~60 ~70 ~80 ~90 ~100 ~120 ~140 pill Figure 32. 29Si MAS NMR spectra of (A) as - synthesized, (B) ethanol - extracted, and (C) calcined TPLM. 110 100 100 I ““3 r . (c) . 90 _ — 90 o . . °\ . _ - - B . I it”) " ( ) u so Q 80 P - I o - _| - I; i 1 0 70 L - 70 Tu - - 3 E Z 60 - - 60 C (A) 1 50 I I I I r I I I I I I I I I I I f 50 0 200 400 600 800 TEMPERATURE, °C Figure 33. TGA curves of (A) as - synthesized, (B) ethanol - extracted, and (C) ethanol - extracted and calcined TPLM. 111 whereas the forth effect (from 400 to 600°C) is most likely due to dehydroxylation of the surface. At this temperature almost all of the silanol groups are transformed into Q4 sites. This is evidenced by the very small (~3 - 4 %) weight loss above 600°C. The total weight loss for the as - synthesized sample is 48 % with only 28 % loss due to the neutral organic species. The large dehydroxylation weight loss of ~ 10 wt °/o fOr our TPLM sample is very characteristic for typical layered material but not observed on the TGA curves of the hexagonal MCM-4l and HMS samples. 30 This observation is very significant and further supports the assumption for a porous lamellar material. Most of the neutral template can be efficiently removed from the structure of our TPLM material by simple solvent extraction. This is evidenced by the lack of a significant weight loss due to template for the ethanol - extracted sample (see Figure 33 curve B). Furthermore, solvent extraction seems to retain the crystallinity and even enhance the cross - linking of the TPLM material (see Figures 28 and 32). The extracted neutral template could be easily recycled and reused by simple evaporation of the solvent. The TGA data for the ethanol - extracted and calcined TPLM material (curve C) is almost featureless. This fact is quite interesting and suggests that this treatment sequence could produce hydrophobic TPLM materials with very high adsorption affinity for organic molecules. The chemical analysis of the as - synthesized TPLM shows a molar composition of 1 SiOzz 0.32 DADD, whereas that of the ethanol - extracted material is 1SiOz: 0.057 DADD. The C / N ratio calculated from the DADD molecular formula is 5.15. The ratio determined from chemical analysis data for the as - synthesized TPLM sample is 5.44 whereas that for the ethanol - extracted sample is 123.0. The large deviation of the C/ N ratio for the solvent - extracted product could be attributed to the presence of residual ethanol in the gallery pores of the sample. A simple recalculation of the amount of template based on the N content and the theoretical C / N ratio of 5.15 allowed for a rough determination the actual molar silica : template ratio in the ethanol - extracted product: ISiOzz 0.003 DADD. Both chemical analysis and TGA results are in agreement suggesting that the neutral template could be efficiently extracted from the framework of our TPLM materials by simple solvent extraction. The N2 adsorption - desorption isotherms and Horvath - Kawazoe pore Size distribution curves (PSD) of the TPLM samples are presented in Figures 34 112 (E) 500 400 “ 300 200 Volume Adsorbed (cm3g'1) 100 0' ' '012' 1 .014. ' j016' ' 'ois' ' '1 Pi/Po Figure 34. N2 adsorption - desorption isotherms for (A) calcined Al - pillared montmorillonite, (B) calcined TPLM, (C) ethanol - extracted TPLM, (D) ethanol - extracted and calcined TPLM, and (E) calcined MCM-41. 113 2.5 0.5 MGM-41 F 04 _ (D) 50.0:- 2 '- in; : 0.1 P otfii------.----.- .— 1.0 2.0 3.0 4.0 5.0 PORESIZEUII‘II) 1.5 - g . ‘ ’ (C) B . u . 1 .. (B) 0.5 ~ ‘ (A) t A 0 WWII 0 0.5 1.0 1.5 2.0 PORE SIZE (nm) Figure 35. Horvath — Kawazoe pore size distribution curves for (A) calcined Al - pillared montrnorrilonite, (B) calcined TPLM, (C) ethanol - extracted TPLM, (D) ethanol - extracted and calcined TPLM. Insert: the Horvath - Kawazoe pore size distribution of calcined mesoporous MCM-41. 114 and 35, respectively. Included for comparison are the curves for typical Al - pillared montmorillonite 31 and MCM-41 prepared by the 5* X’ I+ pathway. The corresponding pore structure parameters are summarized in Table 6. Table 6. Properties of the TPLM samples prepared by neutral templating. Sample calcined TPLM I ethanol-extracted ethanol-extracted 1 & calcined TPLMl ‘ Al-Mont. 31 I MCM-41 1‘ The C351- constant was calculated from the linear part of the BET equation. 23 1 The total pore volume, Vtotal: was taken from the desorption branch of the isotherm at Pi/Po = 0.98 assuming complete surface saturation. § The volume of micropores, Vmicm was determined by the t - plot method. # The volume of mesopores, Vmeso, was estimated from the equation: Vmeso= Vtotal -Vmicm. 1' The average pore size was determined from the maxima of the corresponding Horvath - Kawazoe micropore size distribution curves (see Figure 35). All TPLM samples exhibit similar Type I isotherms with hysteresis loops of Type H4. 23 Our isotherms are very different from the isotherm of the hexagonal MCM-41 material (compare curves B, C, D and E in Figure 34). This is evidenced by the absence of sharp adsorption step or capillary condensation at low relative pressures (0.08 to 0.35) considered to be a MCM-4l fingerprint. In addition, the corresponding PSD curves are also very different (see Figure 35). Thus, the curve for MCM-41 shows maximum at framework - confined mesopore size of 2.2 nm, whereas our distribution curves are featureless in this region. The comparison of the data in Table 6 shows that both classes of samples possess similar specific surface areas. However, MCM-41 exhibits only framework - confined mesoporosity and little or no microporosity or textural 115 mesoporosity. In contrast, our samples exhibit broad micropore size distribution curves centered near 0.6 and 1.2 nm and well - developed micro- and mesoporosity (see Figure 35 and Table 6). In addition, presence of microporosity is also indicated by the large CBET constants exhibited by our materials. The N2 adsorption - desorption isotherms and pore size distribution curves of our TPLM materials are very similar to these for the A1 - pillared montmorillonite and typical for pillared lamellar solids. 31:32 We associate the microporosity of our samples with the intergallery region, whereas the mesoporosity is probably due to interparticle contacts (textural mesoporosity). It is interesting to note that our samples differ by the size of their hysteresis loops or by area contained in the hysteresis loops. The well - expressed hysteresis loops could be attributed to the textural mesopore blocking effect. The pore blocking effect is caused by the differences in the size of the textural cavities and the size of the textural windows that connect them to the surface. 33 Thus, large differences between these sizes will invoke more pronounced effect of pore blocking or larger hysteresis loops. The adsorption properties of our calcined and ethanol - extracted TPLM samples are very similar. This again implies that the neutral template has been efficiently removed by solvent extraction from the pore network of our lamellar material. The fact that the calcined TPLM exhibits much smaller hysteresis loop could be attributed to partial collapse of the textural mesopores during calcination. In contrast, ethanol extraction seems to preserve the textural mesoporosity of the product (see Figure 34 curve C). In addition, the subsequent calcination of ethanol - extracted TPLM seems to preserve not only the textural mesoporosity but also to enhance the microporosity of the sample. This is evidenced by the larger hysteresis loop and much sharper micropore size distribution exhibited by ethanol - extracted and calcined TPLM. Therefore, preferred method for template removal should be solvent extraction followed by calcination. This treatment sequence allows not only for the facile recycling of the neutral template but also enhances the porosity of our lamellar materials. The fact that our isotherms and corresponding pore size distribution curves are typical for pillared lamellar solids supports the assumption for pillared lamellar material and precludes the possibility for an analogy with the templated M418 phases. 116 D. Conclusions The preparation of porous lamellar silicas by novel I° S°-S° I° biomineralization approach involving nucleation and growth in the interlayer region of multilamellar vesicles of neutral diamine surfactants (S°- S°) is demonstrated. The driving force for this process is H - bonding interactions between the lone pairs on the surfactant head groups and the intermediate Si(OC2H5)4-xOHx species produced by the hydrolysis of the neutral TEOS inorganic precursor (1°). The close proximity of the layers, the lack of electrostatic repulsions between the above intermediate silica species and the absence of charge matching interactions are the cause for the biomimetic nucleation and growth of highly cross-linked porous lamellar silica. The presence of multilamellar vesicles and the growth of TPLM phases into the intravesicular region are verified by the TEM and SEM micrographs. The intergallery spacings measured from the TEM lattice images agree well with the 61001 spacings obtained from XRD data. The novel I° S°-S° I° approach allows for the preparation of TPLM materials with high 04/ Q3 ratios, exceptional thermal stability (up to 800°C) and specific surface area and pore volume of the magnitude of the hexagonal MCM-41 and HMS materials. In addition, the N2 adsorption - desorption isotherms and pore size distribution curves of the TPLM materials are very similar to these exhibited by typical pillared lamellar solids. The lack of electrostatic interactions allows for the efficient and environmentally benign recovery and recycling of the neutral diamine template by simple solvent extraction. This novel biomineralization templating approach complements the existing approaches to pillared lamellar solids and could provide for the convenient preparation of a myriad of stable metal - substituted porous lamellar silicates or porous lamellar metal oxide compositions that can not be accessed by the conventional electrostatic templating techniques. 117 References . W. W. O’Dell, U. 8. Patent 1,984,380 (1934). . M. E. Davis, Chem. Ind., 1992, 137-139 (1992). . T. I. Pinnavaia, Science, 220, 365-371 (1983). . R. Burch, Catal. Today, 2, 185-366 (1988). . A. Galameau, A. Barodawalla, T. J. Pinnavaia, Nature, 374, 529 (1995). M. E. Landis, B. A. Aufdembrink, P., Chu, I. D. Johnson, G. W. Kirker, M. Rubin, I. Amer. Chem. Soc., 113, 3189-3190 (1991); I. S. Dailey and T. I. Pinnavaia, Chem. Mater., 4, 855-863 (1992). 7. S. Inagaki, Y. Fukushima, K. Kuroda, I. Chem. Soc. Chem. Commun., 8, 680 (1993). 8. I. 5. Beck, J. C. Vartuli, W. I. Roth, M. E. Leonowitz, C. T. Kresge, K. D. Schmitt, C. T-W. Chu, D. H. Olson, E. W. Sheppard, S. B. McCullen, I. B. Higgins, I. L. Schlenker, I. Am. Chem. Soc., 114, 10834 (1992); C. T. Kresge, M. E. Leonowicz, W. I. Roth, I. C. Vartuli, J. S. Beck, Nature, 359, 710 (1992). 9. Q. Huo, D. I. Margolese, U. Ciesla, P. Feng, T. Gier, P. Sieger, R. Leon, P. M. Petroff, F. Schiith, G. D. Stucky, Nature, 368, 317 (1994). 10. Q. Huo, D. I. Margolese, U. Ciesla, D. G. Demuth, P. Feng, T. Gier, P. Sieger, A. Firouzi, B. F. Chmelka, F. Schiith, G. D. Stucky, Chem. Mater., 6, 1176 (1994). 11. P. T. Tanev, M. Chibwe, T. I. Pinnavaia, Nature, 368, 321 (1994). 12. A. Corma, M. T. Navarro, I. Perez Pariente, I. Chem. Soc. Chem Commun., 1994, 147-148 (1994). 13. E. Armengol, M. L. Cano, A. Corma, H. Garcia, M. T. Navarro, I. Chem. Soc. Chem. Commun., 1995, 519-520 (1995). 14. C.-G. Wu and T. I. Bein, Science, 264, 1757-1758 (1994). 15. M. Dubois, Th. Gulik-Krzywicki, B. Cabane, Langmuir, 9, 673 (1993). 16. M. Ogawa, I. Amer. Chem. Soc., 116, 7941 (1994). 17. T. Dabadie, A. Ayral, C. Guizard, L. Cot, C. Lurin, W. Nie, D. Rioult, Mater. Sci. Forum, (Soft Chemistry Routes to New Materials), 152-153, 267-70 (1994); A. Ayral, C. Guizard, L. Cot, I. Mater. Sci. Lett., 13, 1538-1539 (1994). 18. S. Mann, Chem. Ind., 93-96 (1995). 19. M. Meyer, C. Wallberg, K. Kurihara, I. H. Fendler, I. Chem. Soc. Chem. Commun., 90-91 (1984). pxmstht—I 1 118 20. I.-P. Roman et al., I. Coll. Int. Sci, 144, 324-338 (1991). 21. S. Mann, I. P. Hannington and R. I. P. Williams, Nature, 324, 565-567 (1986). 22. D. D. Archibald and 8. Mann, Nature, 364, 430-433 (1993); I. M. Schnur, Science, 262, 1669-1676 (1993). 23. K. S. W. Sing, D. H. Everett, R. A. W. Haul, L. Moscou, R. A. Pierrotti, J. Rouquérol, T. Siemieniewska, Pure Appl. Chem, 57, 603-619 (1985). 24. G. Horvath and K. I. Kawazoe, I. Chem. Eng. Ipn., 16, 470 (1983). 25. G. Scholzen, K. Beneke, G. Lagaly, Z. Anorg. Allg. Chem, 597, 183-196 (1991) 26. K. Beneke and G. Lagaly, Am. Mineral., 68, 818-826 (1983). 27. A. Monnier, F. Schiith, Q. Huo, D. Kumar, D. Margolese, R. S. Maxwell, G. D. Stucky, M. Krishnamurty, P. Petroff, A. Firouzi, M. Janicke, B. F. Chmelka, Science, 261, 1299 (1993). 28. M. Froba, P. Behrens, I. Wong, G. Engelhardt, Ch. Haggen-Miiller, G. Van de Goor, M. Rowen, T. Tanaka, W. Schwieger, in Advances in Porous Materials, Sh. Komameni, I. S. Beck, D. M. Smith, Eds., MRS Symp. Proc. Ser., vol. 371, Pittsburgh, MRS, 1995, pp. 99-104. 29. I. C. Vartuli, K. D. Schmitt, C. T. Kresge, W. I. Roth, M. E. Leonowitz, S. B. McCullen, S. D. Hellring, J. S. Beck, I. L. Schlenker, D. H. Olson, E. W. Sheppard, Chem. Mater., 6, 2317-2326 (1994). 30. P. T. Tanev and T. I. Pinnavaia (manuscript in preparation). 31. J.- R. Butruille, Ph. D. Thesis, Michigan State University, 1992. 32. C. Pesquera, F. Gonzales, 1. Benito, S. Mendioroz, in Characterization of Porous Solids II, Stud. Surf. Sci. Catal., vol. 62, F. Rodriguez-Reinoso, I. Roquerol, K. W. Sing and K. K. Unger, Eds., (Elsevier Science Publishers B. V., Amsterdam, 1991), pp. 625-633. 33. P. T. Tanev and L. T. Vlaev, I. Call. Inter. Sci., 160, 110-116 (1993). CHAPTER FIVE NEW Ti - SUBSTITUTED MESOPOROUS MOLECULAR SIEVES FOR CATALYTIC OXIDATION OF AROMATIC COMPOUNDS 119 120 Abstract Titanium silicalite is an especially effective molecular sieve catalyst for the selective oxidation of alkanes, the hydroxylation of phenol and the epoxidation of olefins in the presence of H202. 1'3 However, the number of organic compounds that can be oxidized is greatly limited by the relatively small pore size (~ 0.6 nm) of the host framework. 4 A mesoporous molecular sieve capable of catalyzing the oxidation of organic species with kinetic diameters > 0.6 nm, especially aromatics, would greatly complement the catalytic chemistry of titanium silicalite. But the synthesis of transition metal - substituted mesoporous molecular sieves for the oxidation of large organic substrates is unexplored. Here we report the synthesis of new mesoporous silica molecular sieves with site - isolated titanium and unique catalytic activities for the selective peroxide oxidation of 2,6- di-tert- butyl phenol to the corresponding quinone and the conversion of benzene to phenol. 121 A. Introduction The suitability of conventional zeolites and molecular sieves for adsorption, separation and catalysis often is limited by pore sizes that seldomly exceed 1.0 nm. Recently, two major breakthroughs have been reported for the preparation of extra large pore molecular sieves: (i) the discovery by Mobil researchers 5'7 of a broad family of mesoporous M41S silicas, in particular, MCM-41 with hexagonal symmetry and uniform pore sizes in the range 2.0 - 10.0 nm and (ii) the preparation by Kuroda and his co - workers 3 of an ordered mesoporous silica by surfactant - directed rearrangement of a layered precursor (kanemite). Both classes of compounds offer the possibility of designing large pore titanium - substituted catalysts analogous to titanium silicalite. Our efforts to prepare such materials by templated synthesis have led to new mesoporous silica molecular sieves with catalytically active titanium centers. B. Experimental A new hexagonal mesoporous silica (designated HMS) and its titanium - substituted derivative (Ti-HMS) were prepared by acid catalyzed hydrolysis of Si(OC2H5)4 or of Ti(i- OC3H7)4 : Si(OC2H5)4 mixtures in alcoholic solutions using dodecylamine (DDA) as a template. This preparative strategy differs from all previously reported mesoporous molecular sieve syntheses insofar as a primary rather than a quaternary ammonium ion surfactant is used for the first time as a templating reagent. The pure silica derivative was prepared by adding a clear solution of Si(OC2H5)4 (1.00 mol) in ethanol (6.54 mol) to a stirred solution of DDA (0.27 mol) and HCl (0.02 mol) in water (36.3 mol). Allowing the resulting gel to age 18 h at ambient temperature afforded the crystalline templated product. An analogous procedure was used to prepare Ti-HMS from Ti(i - OC3H7)4 and Si(OC2H5)4 at 1 z 100 molar ratio, except that a 6.5 : 1.0 molar mixture of CszOH : (CH3)2CHOH was used in place of ethanol. Also, the replacement of DDA in our ambient temperature procedure by the customary quaternary ammonium ion template [C16H33N(CH3)3]+, abbreviated CTMA, afforded a titanium - substituted derivative of an authentic MCM-41 silica with crystallographic and mesoporous properties in accord with those reported 12 previously for the pure silica analog. 7 Prior to the measurements to be described below, all samples were calcined in air at 650°C for 4 h to remove the structurally incorporated template. C. Results and Discussion As shown in Figure 36 the X - ray powder diffraction patterns for HMS and Ti-HMS are nearly identical, with both materials exhibiting a single diffraction peak near 3.8 nm. An intense low angle reflection also is a characteristic feature of Ti-MCM-41, but owing to the long range hexagonal order weaker 110, 200, and 210 reflections are observed in the 26 range from 4.0 to 70°. The uniform mesopore structure of Ti-MCM-41 is evident in the TEM lattice image shown in Figure 37 A. Higher order Bragg reflections are not resolved in the XRD patterns of HMS and Ti-HMS (cf., Figure 36). Similar "single reflection" MCM-41 - type products, obtained by quaternary ammonium template synthesis and possessing substantial amounts of a hexagonal mesopore phase, also have been reported in the patent literature. 5 Electron micrographs of HMS and Ti- HMS reveal a very fine particle morphology (< 40 nm). Thus, the diffuse scattering at ~ 5° for these two materials may arise from broadening of hk0 reflections due to finite size effects. However, the electron. diffraction patterns (cf., Figure 37 B) exhibit well - defined hexagonal arrangement of diffraction maxima, clearly establishing a crystallographic symmetry analogous to MCM-41 phases. The regular mesoporous structures of HMS, Ti-HMS and Ti-MCM-41 are manifested in their sorption properties. Figure 38 presents the N2 adsorption - desorption isotherms and the corresponding Horvath - Kawazoe 9 pore size distribution curves for all three materials. Included for comparison are the corresponding plots for microporous titanium silicalite, TS-1, as prepared by the method of Tangaraj et a1. 10 and nonporous TiOz. In accord with the N2 sorption properties of MCM-41 silicas, 7 the sharp adsorption steps in the P/Po = 0.20 - 0.35 region and the corresponding maxima in the pore distribution curves indicate the existence of framework - confined uniform mesopores in the range 2.6 - 3.0 nm. Although the template chain is shorter for DDA than CTMA, HMS and Ti-HMS exhibit slightly larger pore sizes than Ti-MCM-41, suggesting that 123 hkl d(nm} § 100 3.26 110 1.92 200 1.67 210 1.25 110 200 210 x 3 Ti-MCM-4l Intensity O 4 8 12 16 20 Degrees (2 6) Figure 36. X - ray powder diffraction patterns of mesoporous molecular sieves. The diffraction patterns were recorded on a Rigaku Rotaflex diffractometer equipped with a rotating anode and using Cu - K; radiation (A. = 0.15418 nm). 124 Figure 37 (A) Transmission electron micrograph of the hexagonal pore structure of Ti-MCM-41. (B) Electron diffraction pattern of Ti-HMS. The TEM image was obtained on JEOL 100CX using an accelerating voltage of 120 kV and 20 um objective lens aperture, and the electron diffraction pattern was taken at 100 kV using 20 um diffraction aperture. e strum image was md 20 1”“ 1 at 100]“, 124A’ 125 1400 gt Ti-MCM-41 8 v r -- S g C’ ‘m s ‘°” _r 2'3 HMS 81L 3 Ti-HMS °" ”'0‘ . TI" , Tl-HMS 1 2 3 4 Effective pore diameter (um) Volume adsorbed (cm-33" STP) r * 4—_=“'01J o 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Relative pressure (P/Po) Figure 38. Nitrogen adsorption - desorption isotherms for mesoporous molecular sieves (Ti-HMS, HMS, and Ti-MCM-41), microporous titanium silicalite (TS-1), and nonporous anatase (T iOz). Insert: The corresponding Horvath - Kawazoe pore size distribution curves. Prior to measurement each sample was evacuated overnight at 150°C and 10'6 Torr. The isotherms were measured at -196°C on a Coulter Omnisorp 360CX Sorptometer using standard continuous adsorption procedures. 126 the two surfactants adopt different conformations during synthesis. The framework - confined mesopore volumes for Ti-HMS and Ti-MCM—41 are very similar (0.68 and 0.69 mL/ g, respectively), but the two materials differ greatly in textural (interparticle) mesoporosity, as indicated by a comparison of the hysteresis loops beyond P/Po = 0.8. For Ti-HMS the textural mesopore volume in the 10 - 30 nm range is 1.11 mL/ g vs. 0.03 mL/ g for Ti-MCM-41. These differences in textural porosity reflect the crystallographic distinctions indicated in the XRD patterns of these materials and underscore the discriminating templating properties of primary and quaternary ammonium ion surfactants. Stucky and his co - workers 11 have proposed that the templating of hexagonal mesoporous structures proceeds through lamellar assemblies of surfactant molecules and not by direct templating around rod - shaped micelles, as originally suggested. 6: 7 It appears that the way such lamellar assemblies rearrange into hexagonal structures depends on the structure of the polar head group, as well as on the length of the surfactant alkyl chain. Today's environmental concerns demand the streamlining of the catalytic processes for the production of fine chemicals. Both Ti-HMS and Ti-MCM-41 offer great promise in contributing to this need by complementing the catalytic oxidation chemistry of TS-1. Two peroxide oxidation reactions, namely, the conversion of 2,6- di-tert- butyl phenol (2,6-DTBP) to the corresponding quinone and the hydroxylation of benzene to phenol, were used to demonstrate their exceptional catalytic efficiency. The catalytic experiments were conducted at 62°C for 2 h using 100 mg of catalyst, 0.5 mmol of 2,6-DTBP or 1.0 mmol of benzene as substrate, 0.13 mol acetone and 29 mmol of 30% aqueous H202. The catalytic results, along with N2 BET surface areas, 12 are compared in Table 7 with those for HMS, TS-l, an amorphous titanium silicate gel (TSG), and bulk TiOz (anatase). The TSG was obtained by preparative methods equivalent to a Ti- HMS synthesis, but in the absence of template. Efforts to prepare mesoporous titania by templated synthesis gave only the bulk anatase phase. As in the case of TS-l, Ti-HMS and Ti-MCM-41 are effective catalysts for the direct hydroxylation of benzene. All three materials circumvent the need for a time consuming pre - alkylation step 13 to achieve the oxidation of the aromatic ring. However, the results for 2,6-DTBP emphasize the importance of Ti-HMS and Ti-MCM-41 mesoporosity in achieving the oxidation of larger substrates. 2,6-DTBP cannot be accommodated by the micropore structure of TS-l. Owing 127 Table 7. Specific surface areas and catalytic oxidation pr0perties of Ti-HMS, Ti-MCM-41 and related materials “. Specific 2,6-DTBP Ox1dation Benzene Oxidation Sample surface area Conversionb Quinone Conversionb Phenol (mz/g) Selectivity Selectivity (°/o) (%) (%) (%l 83 >95 37 >95 20¢ >98 68 HMS 1120 6.8 50 - - TS-l 398 6.5 TSG 7 12 67 0 i 0 TiOz 20 14.5 >95 . Ti-HMS . Ti-MCM-41 1031 1345 ‘1 Reaction conditions are provided in the text. b Blank reactions in the absence of any catalyst gave conversions of 6.8 % and 0 °/o, respectively, for 2,6—DTBP and benzene oxidation. C A conversion of 96 % was observed after 18 h reaction time, whereas the corresponding blank reaction in the absence of catalyst gave a conversion of 16%. 128 to the absence of appreciable surface areas and / or Ti site isolation, amorphous TSG and bulk TiOz are equally inefficient for the oxidation of 2,6-DTBP, as well as for benzene. The general unsuitability of TSG - type materials for peroxide oxidations of aromatics also has been reported recently by other workers. 14 Thus, Ti-HMS and Ti-MCM-41 are important complements to TS-1 in extending peroxide oxidation chemistry to large aromatic substrates. Finally, we note that Ti-HMS is more active than Ti-MCM-41 for 2,6-DTBP oxidation, but the reverse is true for the oxidation of benzene. The hydroxylation of aromatic rings over both catalysts presumably occurs via a common mechanism, perhaps by a peroxotitanate pathway as postulated for TS-l.3 However, access to the framework titanium sites of Ti-HMS by the larger substrate most likely is facilitated by the exceptionally high textural mesoporosity of this catalyst. In contrast, the somewhat smaller framework pore structure of Ti-MCM-41, along with the lack of appreciable textural mesoporosity for this material, probably contributes to its lower reactivity toward 2,6-DTBP. Diffusional effects in these materials will be the subject of future studies. 129 References 1. Taramasso, M., Perego, G. & Notari, B. US Patent No. 4, 410, 501 (1983). 2. Huybrechts, D. R. C., De Bruycker, L. 8r Jacobs, P. A. Nature 345, 240-242 (1990). 3. Notari, B. Structure - Activity and Selectivity Relationships in Heterogeneous Catalysis (eds. Grasselli, R. K. & Sleight, A. W.) 243-256 (Elsevier, Amsterdam, 1991). 4. Flanigen, E. M. et al. Nature 271, 512-516 (1978). 5. Kresge, C. T., Leonowicz, M. E., Roth, W. I. 8: Vartuli, I. C. U. S. Patent No. 5, 098, 684 (1992). 6. Kresge, C. T., Leonovicz, M. E., Roth, W. 1., Vartuli, J. C. 8: Beck, I. S. Nature 359, 710-712 (1992). 7. Beck, I. S. et al. I. Am. Chem. Soc. 114, 10834-10843 (1992). 8. Inagaki, 8., Fukushima, Y. 81.- Kuroda, K. I. Chem. Soc. Chem. Commun. 8, 680- 682 (1993). 9. Horvath, G. & Kawazoe, K. I. I. Chem. Eng. Ipn. 16, 470-475 (1983). 10. Tangaraj, A., Kumar, R., Mirajkar, S. P. & Ratnasamy, P. I. Catal. 130, 1-8 (1991). 11. Monnier, A. et al. Science 261, 1299-1303 (1993). 12. Sing, K. S. W. et al. Pure Appl. Chem. 57, 603-619 (1985). 13. Sheldon, R. A., New Developments in Selective Oxidation (eds. Centi, G. & Trifiro, F.) 1-32 (Elsevier, Amsterdam, 1990). 14. Neumann, R., Chava, M. & Levin, M. I. Chem. Soc. Chem. Commun. 22, 1685- 1687 (1993). at?" CHAPTER SIX Ti - SUBSTITUTED MESOPOROUS MOLECULAR SIEVES FOR CATALYTIC OXIDATION OF LARGE AROMATIC COMPOUNDS PREPARED BY NEUTRAL TEMPLATING ROUTE 130 131 Abstract A new synthesis route to open framework mesostructures based on H - bonding and self - assembly between neutral primary amine surfactants (S°) and neutral inorganic precursors (1°) has been used to prepare hexagonal mesoporous silicas containing site isolated titanium centers. These new titanosilicates, designated Ti-HMS, exhibit exceptional catalytic reactivity for the oxidation of substrates too large to access the pore structure of conventional titanosilicates, such as titanium silicalite, TS-l. The catalytic properties of Ti-HMS materials for the peroxide oxidation of 2,6-di-tert-butylphenol are compared with those of microporous TS-l and a mesoporous Ti-MCM-41 analog prepared by an electrostatic templating mechanism using a liquid crystal quaternary ammonium surfactant. 132 A. Introduction The suitability of conventional zeolites and molecular sieves for adsorption, separation and catalysis often is limited by pore sizes that seldomly exceed 1.0 nm. A major breakthrough in the preparation of extra large pore zeolites has been provided by Mobil researchers in reporting 1 a broad family of liquid crystal templated M415 materials. One member of this family, namely, MCM-41 silicate exhibited hexagonal symmetry and uniform pore size in the mesopore range from 2.0 to 10.0 nm. Also, Kuroda and his co - workers 7- have prepared an ordered mesoporous silica by quaternary ammonium surfactant rearrangement of a single - layered silicic acid precursor (kanemite). More recently, Schiith, Stucky and their co - workers 3 elaborated upon the Mobil approach, which is based on the assembly of cationic surfactants (8+) and anionic inorganic precursors (I'), and designated it the S+ 1' pathway. Also, they demonstrated a new charge - reversed S‘ I+ surfactant - reagent assembly mechanism, as well as counterion - mediated S+ X' 1" and S‘ M+ I' pathways, where X' = C1' or Br" and M+ = N a+ or Ki“. Recently, we have reported 4 the preparation of a mesoporous silica molecular sieve and a Ti - substituted analog by acid catalyzed hydrolysis of inorganic alkoxide precursor in the presence of a partially protonated primary amine surfactants (S+/S°). We also have demonstrated 5 a new templating route to open framework mesostructures based on H - bonding and self - assembly between neutral primary amine surfactants (8°) and neutral inorganic precursors (1°). Our new 5° 1° templating approach is complementary to the above electrostatic templating pathways. When we applied the 5° 1° pathway to the synthesis of hexagonal mesoporous silicates (denoted HMS) using a neutral primary amine as the template and tetraethyl orthosilicate as the inorganic reagent we obtained derivatives with more cross - linked framework, thicker framework walls, superior thermal stability, smaller X - ray scattering domain size, and substantial textural mesoporosity, relative to MCM-41 analogs prepared by an electrostatic S+ 1' or 8" X‘ I+ templating pathway. 5 Titanium silicalite, TS-l, is an especially effective molecular sieve catalyst for the selective oxidation of alkanes, the hydroxylation of phenol and the epoxidation of olefins in the presence of H202.5'3 However, the number of organic compounds that can be oxidized is greatly limited by the relatively small 133 pore size (~ 0.6 nm) of the host framework. 9 A mesoporous molecular sieve capable of catalyzing the oxidation of organic species with kinetic diameters > 0.6 nm, especially aromatics, would greatly complement the catalytic chemistry of titanium silicalite. HMS and MCM-41 silicates offer exciting opportunities for the preparation of large pore analogs of the industrially important TS-1 catalyst. Indeed, we have reported 4 the preparation of Ti - substituted HMS and MCM-41 derivatives and have demonstrated the ability of these materials to catalyze the oxidation of aromatic derivatives that are too large to access the micropore structure of TS-l. Also, several other groups reported more recently that Ti - substituted MCM-41 silicas,10'12 as well as a V - substituted derivative,13 are promising oxidation catalysts. Here we report the synthesis of a new mesoporous Ti-HMS prepared by a 8" 1° pathway using dodecylamine as a template. Also, we report the catalytic activity of this Ti-HMS derivative for the selective peroxide oxidation of 2,6-di- tert- butylphenol to the corresponding quinone, and compare the results with those for TS-l and a Ti-MCM-41 analog prepared by a S+ X' 1+ templating route. B. Experimental 1. Materials Ii:I:IMS The Ti-HMS was prepared in the presence of primary amine in water with ethanol as a co - solvent. The use of a co - solvent improved template solubility and product crystallinity. For this particular experiment a tetraethyl orthosilicate (TEOS) (1.0 mol) was diluted in 6.54 mol of ethanol (EtOH). A solution of 0.01, 0.02, 0.05, 0.07 or 0.1 mol Ti(i-C3H7O)4 in 1.99 mol isopropyl alcohol (i- PrOH) was quickly added to the TEOS solution under vigorous stirring. The resulting clear mixture was then heated and stirred at 65 - 80°C for 3 h to obtain the polymerized -Ti-O-Si- species. A separate solution of 0.27 mol of dodecylamine (DDA) in 21.2 mol of water was prepared. The aged clear -Ti-O-Si- solution was added to the above template solution under vigorous stirring. The resulting reaction mixture was vigorously stirred and aged at ambient temperature for 18 h in order to prepare the Ti - substituted hexagonal mesoporous silicas (1 - 10 "/0 Ti-HMS). The hJ' 134 template removal was achieved by calcination in air at 650°C for 4 h or by EtOH extraction. The template extraction was carried out by mixing 1 g of the air - dried product with 100 ml of EtOH at 45 - 75°C for 30 min. The product was then filtered and washed with another portion of EtOH (100 ml). The above washing procedure was repeated twice and the product was air dried at 80°C. Ii]. TS-l was synthesized according to the procedure of Tangaraj et al 14 modified as follows: TEOS (1.0 mol) was diluted in 3.98 mol of i- PrOH and aqueous solution of TPAOH (20%) (0.36 mol) was added dropwise under vigorous stirring. A portion of Ti(i-C3H70)4 (0.01 or 0.08 mol) was measured in 1.0 mol of i- PrOH and added to the solution of TEOS and template. The reaction mixture was stirred for 1 h and additional amount of deionized water (20.6 mol) was added. The mixture was subjected to heating at 65 - 80°C for 3 h and another 63.5 mol - portion of water was added followed by homogenization for 30 min. The obtained reaction mixture was placed in a autoclave and heated at 170°C for 48 h. The TS-1 product was washed, filtered, and calcined in air at 650°C for 4 h. W1 Ti-MCM-41 was obtained by a procedure similar to that reported elsewere.4 In a typical preparation TEOS (1.0 mol) was added to 6.54 mol of EtOH under vigorous stirring. A solution of 0.01, 0.02 or 0.1 mol Ti(i-C3H7O)4 in 1.99 mol isopropyl alcohol (i- PrOH) was quickly added to the TEOS solution under vigorous stirring. The resulting clear mixture was then heated and stirred at 65 - 80°C for 3 h to obtain the -Ti-O-Si- polymerized species. The solution of the template was obtained by dissolving hexadecyltrimethyl ammonium bromide (0.2 mol) in deionized water (91.5 mol) and HCl (1.0 mol). The aged clear Ti-O-Si solution was then added to the above template solution under vigorous stirring. The resulting mixture was stirred and aged at ambient temperature for 24 h in order to prepare the Ti - substituted MCM-41 material. 2. Characterization The powder X - ray diffraction patterns were measured on Rigaku Rotaflex diffractometer equipped with a rotating anode and using Cu - K, radiation. The N2 adsorption - desorption isotherms were measured at -196°C on a Coulter 135 Omnisorp 360CX Sorptometer using a continuous adsorption procedure. Before measurement, samples were evacuated overnight at 150°C and 10‘6 torr. 3. Catalytic Evaluation The catalytic activity of Ti-HMS and Ti-MCM-‘41 silicates as well as TS-l was tested using the liquid phase oxidation of 2,6-di-tert-butylphenol (2,6-DTBP) in the presence of H202 as a probe reaction. The reaction was carried out by refluxing in a glass flask at 62°C for 2 h using 10 m1 of acetone as solvent, a substrate : catalyst ratio of 14 : 1 ( w / w ) and adequate amount of 30% H202 (substrate : H202 ratio of 1 : 1, 1 : 2, and 1 : 6 was used respectively). The reaction products were quantitatively analyzed by means of HP-5890 GC equipped with SPB-20 fused silica capillary column ( 30 m x 0.53 mm x 0.50 um ) and a FID detector using 1,4-di- tert-butylbenzene as internal standard. C. Results and Discussion The powder X - ray diffraction (XRD) pattern of the calcined 1 % Ti-HMS is shown in Figure 39. Intensity l I T I l I I T l l I I I I T o 5 10 15 20 Degrees (29) Figure 39. Powder XRD pattern of 1 % Ti-HMS sample calcined at 650°C for 4 h. 136 The pattern exhibits a strong relative intensity 4100 reflection at 4.0 nm d - spacing and a diffuse scattering centered at ~ 1.8 nm. Higher order Bragg reflections of the hexagonal structure are not resolved. However, we and others have demonstrated 4. 15 that similar "single reflection" MCM-41 - type products, still possess short - range hexagonal symmetry. The scattering domain size of 1 % Ti- I-IMS, calculated from the line width of the time reflection is only about 17 nm. Thus, the diffuse scattering at ~ 1.8 nm could be attributed to broadening of the remaining hk0 reflections due to the small domain size effects. Surprisingly, the X - ray diffraction pattern of the ethanol - extracted 1 % Ti-HMS product (not shown) exhibits four times stronger dmo reflection at 4.8 run than that of the calcined 1 % Ti-HMS analog. In addition, a well - expressed diffuse scattering centered at 2.2 nm is also observed. The XRD patterns of our 1 % TS-l and 1 % Ti-MCM-41 are essentially identical to those published previously. 4'14 The N2 adsorption - desorption isotherm for the calcined 1% Ti-HMS sample is shown in Figure 40 A. 1‘00 1. Volume adsorbed (cc/9. STP) - Figure 40. Nitrogen adsorption - desorption isotherms for (A) calcined 1 % Ti- HMS and (B) ethanol - extracted 1 % Ti-HMS sample. 137 Included for comparison is the isotherm for the as - synthesized 1 % Ti-HMS sample in which the template has been removed by ethanol extraction (Figure 40 B). Both N2 adsorption - desorption isotherms are similar exhibiting complementary framework - confined and textural mesoporosity as evidenced by the presence of two well - defined, separate hysteresis loops. The textural mesoporosity in Ti-MCM-41 and Ti-HMS materials is the porosity that can be attributed to non - uniform voids and channels between elementary particles or aggregates of such particles (grains). Each of these elementary particles is composed of certain number of framework unit cells or framework - confined uniform mesopores. The size of the textural mesopores is determined by the size, shape and the number of interfacial contacts of these particles or aggregates and usually is at least one order of magnitude larger than that of the framework - confined mesopores. The SBET of both samples are also similar - 1046 and 803 m2 / g, respectively. This indicates that the neutral template has been efficiently removed from the framework of our 8° 1° Ti-HMS samples by EtOH extraction. Due to the lack of ions the extracted organic template, in the form of EtOH solution, can be recycled and reused after simple concentration of the solution. The total pore volume of the calcined 1 % Ti-HMS is 1.75 cc/ g. The pore volume corresponding to the uniform framework mesopores is 0.71 cc/ g and the v01ume of textural mesopores is 1.04 cc / g. The ratio of textural to framework mesoporosity here is 1.46. The total pore volume of the ethanol - extracted 1 % Ti-HMS is 1.31 cc/ g, the volume of framework - confined mesopores is 0.69 cc/ g and that of the textural mesopores is 0.62 cc/ g. The corresponding ratio of textural to framework - confined mesoporosity for this sample is 0.9. This suggests that ethanol extraction completely removes the template from the framework - confined mesopores of Ti-HMS, while at the same time providing less textural mesoporosity. The much well-pronounced ”framework - confined” hysteresis loop for the ethanol - extracted 1 % Ti-HMS sample suggests that template removal by ethanol extraction preserves the framework-confined mesopores or the crystallinity of the product. The corresponding Horvath - Kawazoe pore size distribution curves for both samples are presented in Figure 41. The size of framework - confined uniform mesopores of the calcined 1 % Ti- HLMS product is 2.8 and that of the ethanol - extracted 1 % Ti-HMS is 3.6 nm. The larger uniform pore size of the ethanol - extracted sample could be attributed to 138 the preservation of the crystallinity upon EtOH extraction whereas the smaller pore size of the calcined material is probably due to partial collapse of the mesoporous framework during calcination. dW/dR g A . J L i ‘a A 3 3 Effective pore diameter (nm) Figure 41. Horvath - Kawazoe pore size distribution curves for (A) calcined 1 % Ti-HMS and (B) ethanol - extracted 1 % Ti-HMS sample. We reported 5 that our HMS samples, prepared by neutral templating with primary amines of different alkyl - chain length, exhibit constantly larger framework wall thickness16 (2 1.7 nm) than that of MCM-41 materials prepared by electrostatic templating (S 1.2 nm). The framework wall thickness of our calcined 1 % Ti-HMS is also larger (1.8 nm) relative to that for the calcined 1 % Ti-MCM-41 material (1.2 nm). The catalytic evaluation results are listed in Tables 8, 9, 10 respectively. 139 Table 8. Effect of the nominal Ti - loading on the Catalytic activity for peroxide oxidation of 2,6-DTBP over mesoporous molecular sieves. a Catalyst ' Conversion Mononuclear Mononuclear : Quinone Yield Quinone Selectivity (%) (%) (%) 26 14 54 Ti-HMS 55 32 58 (5° 1°) 50 24 48 65 31 48 . 98 46 47 : Ti-MCM-41 1 8 3 38 (8+ X' 1*) 2 16 6 37 39 46 3 . _ 24 36 0.3 — a the reactions were carried out at reflux temperature (62°C) in a glass flask for 2 h using substrate : catalyst ratio of 14 : 1 (w/w), substrate : H202 molar ratio of 1 : 6 and substrate to solvent (acetone) molar ratio of 1 : 27. b nominal Ti - loading. 140 Table 8 provides a comparison of the catalytic activities of Ti-HMS, Ti-MCM-41 and TS-l for the peroxide oxidation of 2,6-DTBP to the corresponding quinone. Ti-I-[MS exhibits superior catalytic activity and mononuclear quinone selectivity relative to the electrostatically - templated Ti-MCM-41 at any nominal Ti loading. This superior activity of the Ti-HMS samples could be attributed to the substantial amount of complementary textural mesoporosity that facilitates accessing of the framework - confined mesopores. Due to the small micropore size (~ 0.6 nm ) of the silicalite framework9 TS-l samples under equivalent conditions are substantially less reactive. However, access to Ti-MCM-41 relative to Ti-HMS seems to be related not only to the absence of complementary textural mesoporosity 4 but also to differences in titanium siting. It may be that the larger wall thickness of Ti-HMS results in titanium bond angles that enhance the reactivity of the peroxotitanyl intermediate. Future studies are planned which will relate wall thickness to the intrinsic reactivity of the titanium centers. The data in Table 8 also show that the activity of the catalysts increases as the titanium loading is increased to 10 mole %. Above 10 % titanium substitution, the conversion begins to decrease. We should note that these titanium loadings are nominal values and do not necessarily correspond to the framework content of titanium. ' Table 9. Effect of the substrate : H202 ratio on the catalytic activity of 10 % Ti- HMS. a Conversion Quinone Mononuclear ' ‘ SubstrateszOz I ratio yield Quinone (mol/ mol) (%) (%) selectivity (%) 1 : 1 14 5 36 1 :2 30 10 33 1 :6 98 47 a Reaction conditions are the same as in Table 8. 141 The results in Table 9 illustrate that the conversion of 2,6-DTBP over 10 % Ti- HMS increases as the substrate : H202 mole ratio is decreased from 1 : 1 to 1 : 6. Interestingly, the selectivity toward the mononuclear quinone remains in the 33 - 47 % range even though the conversion ranges from 14 - 98 %. Other reaction products formed in the reaction include the binuclear 3,3’, 5,5’- tetra-tert-butyl- 4,4’-diphenoquinone and small amounts of dealkylated quinone. Table 10. Effect of the method of template removal on the catalytic activity of 1 % Ti - substituted mesoporous molecular sieves.a Method of template Conversion Mononuclear removal Quinone Yield 1 Catalyst (%) (%) Ti-HMS ethanol - extracted 30 7 5° I° ethanol - extracted and calcined 42 11 calcined 26 3 E Ti-MCM-41 ethanol — extracted 24 5 8* X' I+ ethanol - extracted and calcined 28 7 21 calcined a Reaction conditions are the same as in Table 8. Since the neutral amine template used in the synthesis of Ti-HMS and the template - halide ion pairs in S+ X' 1+ Ti-MCM-41 could be displaced by solvent extraction we have an unique opportunity to compare the effects of the method of template removal on the intrinsic reactivities of the site isolated titanium centers. As illustrated by the data in Table 10, 1 % Ti-HMS shows superior catalytic activity relative to the 1 % Ti-MCM-41 despite of the method of template removal. This again could be attributed to the complementary textural mesoporosity of Ti-HMS and possible differences in Ti - siting. It is very interesting to note that the ethanol - extracted and non-calcined samples exhibit substantial catalytic activity. According to our knowledge this data illustrates for 142 first time that non - calcined transition metal - substituted silica molecular sieves could exhibit superior catalytic activity relative to the calcined counterparts for peroxide oxidation of large aromatic molecules. The ethanol - extracted and subsequently calcined Ti-HMS sample exhibits much higher catalytic activity relative to the electrostatic Ti-MCM-41 counterpart. Therefore preferred method for template removal from these mesoporous molecular sieves should be ethanol extraction followed by calcination. Further work is in progress to more fully address the differences in Ti - siting and catalytic activity of these mesoporous materials. 143 References and Notes 1. C. T. Kresge, M. E. Leonowicz, W. J. Roth, J. C. Vartuli and]. S. Beck, Nature 359, 710 (1992); J. S. Beck, M. C. Vartuli, W. J. Roth, M. E. Leonowitz, C. T. Kresge, K. O. Schmitt, C. T-W. Chu, D. H. Olson, E. W. Sheppard, S. B. McCullen, J. B. Higgins and J. L. Schlenker, I. Am. Chem. Soc. 114, 10834 (1992). 2. S. Inagaki, Y. Fukushima and K. Kuroda, I. Chem. Soc. Chem. Commun. 8, 680 (1993). Q. Huo, D. I. Margolese, U. Ciesla, P. Feng, T. Gier, P. Sieger, R. Leon, P. M. Petroff, F. Schiith and G. D. Stucky, Nature 368, 317 (1994). . P. T. Tanev, M. Chibwe and T. J. Pinnavaia, Nature 368, 321 (1994). . P. T. Tanev and T. J. Pinnavaia, Science 267, 865 (1995). . M. Taramasso, G. Perego, and B. Notari, US Patent 4 410 501 (1983). . D. R. C. Huybrechts, L. De Bruycker and P. A. Jacobs, Nature 345, 240 (1990). . B. Notari, in Structure - Activity and Selectivity Relationships in Heterogeneous Catalysis, edited by R. K. Grasselli and A. W. Sleight (Elsevier, Amsterdam, 1991) pp. 243-256. 9. E. M. Flanigen, J. M. Bennett, R. W. Grose, J. P. Chen, R. L. Patton, R M. Kircher and J. V. Smith, Nature 271, 512 (1978). 10. A. Corma, M. T. Navarro and J. P. Pariente, I. Chem. Soc. Chem. Commun. 1224, ' 147. 11. A. Corma, M. T. Navarro, J. P. Pariente and F. Sanchez, in Zeolites and Related Microporous Materials, State of the Art 1994, Studies in Surface Science and Catalysis, edited by J. Weitkamp, H. G. Karge, H. Pfeifer and W. Holderich (Elsevier Science B. V., Amsterdam, 1994), vol. 84, pp. 69-75. 12. O. Franke, J. Rathousky, G. Schulz-Ekloff, J. Starek and A. Zukal, in Zeolites and Related Microporous Materials, State of the Art 1994, Studies in Surface Science and Catalysis, edited by J. Weitkamp, H. G. Karge, H. Pfeifer and W. Holderich (Elsevier Science B. V., Amsterdam, 1994), vol. 84, pp. 77-84. 13. K. M. Reddy, 1. Moudrakovski and A. Sayari, I. Chem. Soc. Chem. Commun. 1294, 1059. 14. A. Tangaraj, R. Kumar, S. P. Mirajkar and P. Ratnasamy, I. Catal. 130, 1 (1991). 15. J. C. Beck, C. T-W. Chu, I. D. Johnson, C. T. Kresge, M. E. Leonowitz, W. J. Roth and J. C. Vartuli, WO Patent 91 / 113090 ( 8 August 1991 ); R. Schmidt, D. Akporiaye, M. Stocker and O. H. Ellestad, in Zeolites and Related Microporous 9’ QVGUIUF- H-H 144 Materials, State of the Art 1994, Studies in Surface Science and Catalysis, edited by J. Weitkamp, H. G. Karge, H. Pfeifer and W. Holderich (Elsevier Science B. V., Amsterdam, 1994), vol. 84, pp. 61-68. 16. The framework wall thickness was calculated as a difference between the unit cell parameter (a0 = 2d100/ ‘13) of the hexagonal structure and the framework - confined mesopore size taken from the corresponding Horvath - Kawazoe pore size distribution curve. t..-- l... "lllllllll"lll'llllll