Fischer indole synthesis over hydrous zirconia supported niobium oxide and boria zirconia catalysts

FISCHER INDOLE SYNTHESIS OVER HYDROUSZIRCONIA SUPPORTED NIOBIUM OIXDE AND BORIA-ZIRCONIA CATALYSTS ZHONG LIANG (B.Sci.(Hons.), NUS) A THESIS SUBMITTED FOR THE DEGREE OF MASTER OF SCIENCE DEPARTMENT OF CHEMISTRY NATIONAL UNIVERSITY OF SINGAPORE 2009 i ACKNOWLEDGEMENT First of all, I would like to express my gratitude to my supervisor Associate Professor Chuah Gaik Khuan for her great guidance and supervision in my research project. I also appreciate the help and support from Associate Professor Stephan Jaenicke. I would like to thank Dr. Nie Yuntong for her kind advices and assistance during my work as well as all other members of our research group for their help and encouragement. Special thanks must go to my parents for their understanding, encouragement and support. Lastly, financial support from National University of Singapore is gratefully acknowledged. TABLE OF CONTENTS ii PAGE Acknowledgement i Table of Contents ii Summary vi List of Tables viii List of Schemes x List of Figures xi Chapter I Introduction 1.1 Heterogeneous Catalysis 1 1.1.1 Introduction 1 1.1.2 Solid Acid/Base Catalysts 3 1.1.2.1 Solid Acid Catalysts 5 1.1.2.2 Solid Base Catalysts 8 1.1.2.3 Acid-Base Bifunctional Catalysts 10 1.1.3 Supported Oxide Catalysts 11 1.1.3.1 Introduction 11 1.1.3.2 Zirconia as Support Material 15 1.2 Synthesis of Indoles 18 1.2.1 Introduction 18 1.2.2 Fischer Indole Synthesis 19 1.2.3 Other Methods of Indole Synthesis 21 1.2.4 Catalysts Used in Fischer Indole Synthesis 23 1.3 Aims of Present Study 25 References 26 iii Chapter II Experimental 2.1 Powder X-ray Diffraction 31 2.2 Nitrogen Adsorption/Desorption Study 33 2.2.1 BET Surface Area Determination 33 2.2.2 Porosity by Gas Adsorption 35 2.3 Infrared Spectroscopy 36 2.4 Inductively Coupled Plasma Atomic Emission Spectroscopy 39 2.5 Thermal Analysis 41 2.5.1 Thermogravimetric Analysis 41 2.5.2 Simultaneous TGA/DTA 42 2.6 Determine Acidity Using Amine Titration Method with Indicators 43 2.6.1 Strength and Amount of Solid Acid 43 2.6.2 Determine the Acid Strength Using Hammett Indicator 44 2.6.3 Determine the Number of Acid Sites by Amine Titration 46 2.7 Temperature Programmed Desorption of Ammonia 47 2.8 Catalytic Activity Test 49 2.8.1 Fischer Indole Synthsis of 2-ethyl-3-propyl indole and 2-butyl-3-methyl indole 49 2.8.2 Analysis of Products by Gas Chromatography and GC-MS 51 2.8.3 Fischer Indole Synthesis of 1,2,3,4-tetrahydrocarbazole 53 References Chapter Ш 55 Study of Hydrous Zirconia Supported Niobium Oxide Catalysts 3.1 Introduction 57 iv 3.2 Preparation of Catalysts 3.2.1 Preparation of Hydrous Zirconia Support 60 60 3.2.2 Wet Impregnation of Niobium Oxide on Hydrous Zirconia Support 61 3.2.3 Synthesis of 25 wt.% Nb2O5/Zr(OH)4 by Coprecipitation Method 62 3.2.4 Synthesis of 25 wt.% Nb2O5/MCM-41 by Wet Impregnation 62 3.2.5 Recycling of Catalysts 63 3.3 Catalyst Characterization 64 3.3.1 Powder X-ray Diffraction 64 3.3.2 Textural Properties 66 3.3.3 Thermogravimetric Analysis 74 3.3.4 Acidity Test 77 3.4 Catalytic Activity 81 3.4.1 Effect of Niobium Oxide Loadings 81 3.4.2 Effect of Drying Temperature 86 3.4.3 Indole Reaction with Cyclohexanone 86 3.5 Recycling of Catalyst 89 3.5.1 Leaching of Catalysts 89 3.5.2 Regeneration of Catalysts 89 3.6 Conclusion 91 References 92 Chapter IV Study of Boria-Zirconia Catalysts 4.1 Introduction 95 4.2 Preparation of Catalysts 97 4.2.1 Preparation of boria-zirconia catalysts 97 v 4.2.2 Recycling of Catalysts 98 4.3 Catalyst Characterization 98 4.3.1 Powder X-ray Diffraction 98 4.3.2 Textural Properties 101 4.3.3 Thermal Analysis 104 4.3.4 Acidity Measurement 113 4.4 Catalytic Activity 118 4.4.1 Effect of Boria Loadings 118 4.4.2 Effect of Calcination Temperature 120 4.4.3 Recycling of Catalysts 121 4.5 Conclusion 122 References 125 Summary vi For the last 100 years, the synthesis of indoles and their derivatives have been a research topic of great interest. Every year, several thousands of indoles and indole derivatives have been synthesized in the research laboratories. The sustained interest in indoles is due to their wide range of biological activity. The Fischer indole reaction is an extremely useful and important method for the synthesis of a variety of biologically active indole-structure compounds for over a century due to its simplicity and efficiency. In recent years, zirconia has attracted much attention as both a catalyst and a catalyst support because of its high thermal stability and the amphoteric character of its surface hydroxyl groups. The application of zirconia as a catalyst support has been employed in many industrially important reactions. Zirconia is stable under oxidizing and reducing conditions and it possesses both acidic and basic properties. Also, unlike Al2O3, zirconia does not form solid solutions upon addition of a second metal oxide component. In this study, the Fischer indole reaction over two different solid acid catalysts, hydrous zirconia supported niobium oxide and boria-zirconia catalysts, were investigated. Hydrous zirconia supported niobium oxide catalysts with loadings of 10 to 40 wt.% Nb2O5 were prepared by wet impregnation with niobium chloride. The samples retained the high surface area and pore structure of the support, suggesting a uniform overlayer. Theoretical monolayers of 0.33 to 2.0 were achieved with these loadings. The Fischer indole synthesis of phenylhydrazine with 3-heptanone and with cyclohexanone was achieved with high conversions. The most active catalysts had vii loadings of 25 – 30 wt.% Nb2O5, which is close to a monolayer coverage. The activity of the catalysts was highest when dried at 100 – 200°C. Use of siliceous MCM-41 as a support also resulted in an active catalyst. The good activity of the supported niobium oxide catalysts makes this a green route to the synthesis of indoles. No leaching of cations was detected and the used catalysts were effectively regenerated by immersion in 30% H2O2 solution at 40°C. Boria-zirconia catalysts with boria loadings 3 to 40 wt.% were prepared by wet impregnation of boric acid on the hydrous zirconia support followed by calcination. At boria loadings below 25 wt.% , the catalysts had high surface areas. Aggregation of boron oxide occurred for boria loading of 25 wt.% with theoretically more than three overlayers on the surface of zirconia support. Infrared spectroscopic study showed that the first layer of boron oxide was composed of tetrahedral oxygencoordinated BO4 units whereas at boria loading higher than a monolayer, trigonal BO3 units were formed. The Fischer indole reaction of phenylhydrazine with 3-heptanone was achieved with high conversions. The most active catalysts had loadings of 5 – 10 wt.% B2O3, which is close to a monolayer coverage. This result is similar to that of Nb2O5/Zr(OH)4 samples where catalysts with niobium oxide approaching a monolayer showed the highest activity. The optimal calcination temperature for the catalysts was 500°C. The used catalyst could be simply regenerated by recalcining at 500°C, with full recovery of textural properties, acidity, and activity. LIST OF TABLES viii Page Table 1.1 E factors in the chemical industry 2 Table 1.2 Number of solid acid, base and acid-base bifunctional catalysts in industry processes 5 Infrared bands of pyridine adsorbed on solid acid catalysts in the 1700-1400 cm-1 region 39 Table 2.2 Strengths and weaknesses of ICP-AES 40 Table 2.3 Basic indicators used for measurement of acid strength 45 Table 3.1 Acidic properties of surface modified niobium oxide catalysts 60 Amount of reagents used in preparing the Nb2O5/Zr(OH)4 catalysts 62 Table 3.3 Textural properties of Nb2O5/Zr(OH)4 catalysts 68 Table 3.4 Niobium oxide loading and surface coverage 70 Table 3.5 Total amount of water lost (%) in TGA for Nb2O5/Zr(OH)4 catalysts 76 Calculation for amount of water lost H2O/oxides (mol/mol) for Nb2O5/Zr(OH)4 catalysts 76 Table 2.1 Table 3.2 Table 3.6 Table 3.7 Acid strength of Nb2O5/Zr(OH)4 catalysts measured using Hammett indicators 79 Table 3.8 Measuring number of acid sites of Nb2O5/Zr(OH)4 catalysts by n-butylamine titration 80 Table 3.9 Selectivity to 2-butyl-3-methyl indole (linear) and 2-ethyl3-propyl indole (bulky) 85 Table 3.10 Conversion to 1,2,3,4-tetrahydrocarbazole after 4 h 89 Table 4.1 Amount of reagents used in preparing the boria-zirconia catalysts 98 Table 4.2 Textural properties of boria-zirconia catalysts 102 Table 4.3 Boria loading and surface coverage 104 Table 4.4 Total amount of water lost (%) in TGA for boria-zirconia catalysts 108 ix Table 4.5 Table 4.6 Calculation for amount of water lost H2O/oxides (mol/mol) for boria-zirconia catalysts 108 Peak maxima of exothermic peaks and enthalpy of crystallization of hydrous zirconia and boria-zirconia catalysts 113 Table 4.7 Number of acid sites of boria-zirconia catalysts determined by NH3 TPD Table 4.8 Measuring acid strength of boria-zirconia catalysts using Hammett indicators 117 Number of acid sites of boria-zirconia catalysts measured by n-butylamine titration 116 Conversion and selectivity of 2-butyl-3-methyl indole (linear) and 2-ethyl-3-propyl indole (bulky) 119 Table 4.11 Textural properties of fresh and recycled boria-zirconia 122 Table 4.12 Acidity determined by NH3 TPD 123 Table 4.9 Table 4.10 115 LIST OF SCHEMES x Page Scheme 1.1 Zeolite-catalyzed vs. classical Friedel-Crafts acylation 7 Scheme 1.2 Sumitomo vs. manufacture 8 conventional process for caprolactam Scheme 1.3 Hydrotalcite-catalyzed condensation reactions 9 Scheme 1.4 Tethered organic bases as solid base catalysts 10 Scheme 1.5 Synthesis of ethyleneimine (EI) from monoethanolamine (MEA) 11 Scheme 1.6 Drugs containing indole structures 18 Scheme 1.7 Resonance structures of indole 19 Scheme 1.8 Mechanism of Fischer indole synthesis 20 Scheme 1.9 Fischer indole synthesis via Japp-Klingemann reaction 21 Scheme 1.10 Various routes for synthesis of substituted indole 22 Scheme 1.11 Indole synthesis via Sonogashira reaction 23 Scheme 2.1 Fischer indole reaction of phenylhydrazine with 3-heptanone 50 Scheme 2.2 Fischer indole synthesis of 1,2,3,4-tetrahydrocarbazole 54 LIST OF FIGURES xi Page Various arrangement of active oxide component on support material (a) portion of oxide component buried in the support (b) oxide component dispersed on the surface of support material (c) formation of monolayer (d) mixture of monolayer, aggregated particles and uncovered support surface. 13 Fig. 2.1 Reflection of X-rays from two planes 32 Fig. 2.2 The five types of adsorption isotherms 35 Fig. 2.3 Gas chromatogram after formation of E/Z phenylhydrazone 50 Fig. 2.4 Gas chromatogram of reaction mixture of Fischer indole reaction of phenylhydrazine with 3-heptaone 51 Gas chromatogram of reaction mixture of Fischer indole synthesis of 1,2,3,4-tetrahydrocarbazole 54 X-ray diffractograms of Nb2O5/Zr(OH)4 catalysts after drying at (a) 100°C (b) 200°C (c) 300°C 65 Fig. 3.2 X-ray diffractograms of MCM-41 and 25% Nb2O5/MCM-41 66 Fig. 3.3 Nitrogen adsorption/desorption Nb2O5/Zr(OH)4 catalysts 72 Fig. 1.1 Fig. 2.5 Fig. 3.1 Fig. 3.4 Fig. 3.5 Fig. 3.6 Fig. 3.7 Fig. 3.8 Fig. 3.9 curves and pore of Nitrogen adsorption/desorption curves and pore volume distribution of MCM-41 and 25% Nb2O5/MCM-41 73 Comparison of pore volume distribution (a) hydrous zirconia supported niobium oxide catalyst with Nb2O5 loadings from 0 – 40 wt.%, drying temperature 100°C (b) MCM-41 and 25% Nb2O5/MCM-41 73 TGA curves of hydrous zirconia , niobium oxide and 10 – 40 wt.% hydrous zirconia supported niobium oxide catalysts 75 Pyridine IR of 25% Nb2O5/Zr(OH)4 after evacuation at (a) 25°C (b) 100°C and (c) 200°C. 80 The Fischer indole reaction of phenylhydrazine with 3heptanone over (+) niobium oxide, (■) hydrous zirconia, and (×) 10% (□) 20% (▲) 25% (♦) 30% (○) 40% Nb2O5/Zr(OH)4 82 The Fischer indole reaction of phenylhydrazine with 3heptanone over (■) 25% Nb2O5/Zr(OH)4 coppt (×) H-beta (Si/Al = 12.5) (▲) 25% Nb2O5/MCM-41 (♦) 25% xii Nb2O5/Zr(OH)4 82 Fig. 3.10 Linear and bulky indole products with molecular dimensions in Å 84 Fig. 3.11 Conversion to indole products over 10 – 40 wt. % Nb2O5/Zr(OH)4 catalysts , Nb2O5 and Zr(OH)4 after drying at (…) 100°C, ( ) 200°C, and („) 300°C 88 The Fischer indole synthesis of 1,2,3,4-tetrahydrocarbazole over (♦) hydrous zirconia, and (□) 10% (■) 20% (×) 30% (▲) 40% Nb2O5/Zr(OH)4 87 Fig. 3.13 Catalyst leaching test 90 Fig. 3.14 Regeneration of 25 wt.% Nb2O5/Zr(OH)4 91 Fig. 4.1 X-ray diffractograms of boria-zirconia catalysts with boria loading from 3 to 40 wt.%, calcination temperature 500°C 100 X-ray diffractograms of 25 wt.% boria zirconia at calcination temperature 100°C, 500°C, and 650°C 100 Nitrogen adsorption/desorption curves and pore volume distribution of boria-zirconia catalysts with different boria loadings 105 Nitrogen adsorption/desorption curves and pore volume distribution of 25 wt.% boria-zirconia sample calcined at 100°C, 500°C, and 650°C 106 Infrared spectra of (a) boric acid (b) 3% (c) 5% (d) 10% (e) 20% (f) 25% (g) 30% and (h) 40% boria-zirconia 106 TG–DTA diagrams of (a) hydrous zirconia (b) 3% (c) 5% (d) 10% (e) 20% (f) 25% (g) 30% and (h) 40% boria-zirconia 112 Ammonia TPD profiles of (a) 3% (b) 5% (c) 10% (d) 20% (e) 25% (f) 30% (g) 40% boria-zirconia (calcination temperature 500°C) and (h) 25% boria-zirconia (calcination temperature 650°C) 114 Conversion of indoles over (♦) 3% (■) 5% (▲) 10% (×) 20% (◊) 25% (□) 30% (+) 40% boria-zirconia 119 Conversion of indoles over 25 wt.% boria-zirconia (♦) uncalcined (■) calcined at 500°C and (▲) calcined at 650°C 120 Fig. 3.12 Fig. 4.2 Fig. 4.3 Fig. 4.4 Fig. 4.5 Fig. 4.6 Fig. 4.7 Fig. 4.8 Fig. 4.9 xiii Fig. 4.10 Fig. 4.11 X-ray diffratograms of (a) fresh 5% boria-zirconia (b) 5% boria-zirconia recycle 2 Conversion of indoles over (♦) 5% boria-zirconia fresh (▲) 5% boria-zirconia recycle 1 (■) 5% boria-zirconia recycle 2 122 123 xiv Chapter I Introduction 1.1 Heterogeneous Catalysis 1.1.1 Introduction As the chemistry industry develops nowadays, it is recognized that environmentally friendly processes become an urgent need and the whole world reaches a consensus that the developing of green chemistry processes should be one of the major concerns in modern chemistry industry. In the last one or two decades, traditional concepts of process efficiency which focus exclusively on chemical yield is not the only origin of the chemistry industry. Instead, concepts of sustainable technology are also involved in order to eliminate waste and avoid the use of toxic or hazardous substances. From the green chemistry point of view, the primary pollution in the chemistry industry is at the source, and hence eliminating waste at the source is recommended rather than waste remediation after reaction processes. The magnitude of the waste problem in the manufacture of chemicals is measured by “E factor” in the amount of waste produced per kg product.[1,2] E factors in different segments of the chemistry industry is listed (Table 1.1). These tremendous quantities of waste consist primarily of inorganic salts, such as sodium chloride, sodium sulfate, and ammonium sulfate, which are formed during the reaction or in subsequent neutralization steps. The E factor increases dramatically on going downstream from bulk to fine chemicals and pharmaceuticals because of two possible reasons. One reason is that multi-step synthesis is involved in the production of the latter chemicals, and the other could be the use of stoichiometric (inorganic) reagents rather than catalytic methodologies. 1 Table 1.1 E factors in the chemical industry[3] Industry segment Product tonnagea E (kg waste/kg product) Bulk Chemicals < 104 - 106 50 Pharmaceuticals 10 - 103 25 → > 100 a Depending on the product this could be the capacity of a single plant or the worldwide production. Three catalytic methodologies, homogeneous, heterogeneous and enzymatic catalysis, are commonly used in organic synthesis in the chemistry industry. All three approaches have their advantages and limitations. Homogeneous catalysis by organometallic complexes[4] has wide applications in both bulk and fine chemical industry and is the method chosen for reactions such as carbonylations and hydrofomyation. Biocatalysis[5] has the advantage of mild reaction conditions and high chemo-, region-, and enantioselectivity, and it is now increasingly used in fine chemical manufacture. Solid, heterogeneous catalysts have the advantages of easy recovery and easy recycling and are most suitable for continuous processing. Moreover, heterogeneous catalysis has already been widely applied in oil refining and bulk chemicals manufacture.[6] The utility of solid catalysts in organic synthesis is feasible according to the experience and understanding from their industry applications. In the preparation of solid catalysts that can be employed in the chemistry industry, the demands for the structures and chemical compositions of the catalysts must be carefully considered. Three characteristics of solid catalysts are decisive:[3] 2 (1) The catalytically active surface ― a sufficiently large active surface area per unit weight or unit volume of catalyst is required; ― the catalytically active surface must be stable at the temperatures of pretreatment and the intended catalytic reaction; ― the catalytically active surface must have the desired structure and chemical composition. (2) The transport properties ― the transport of reactant molecules to the catalytically active surface, and of the reaction products from the surface, must proceed sufficiently rapidly; ― the transport of thermal energy to and from the catalyst particles must occur smoothly. (3) The mechanical strength ― for technical applications the mechanical strength of catalyst particles is most important. 1.1.2 Solid Acid/Base Catalysts Acid- and base-catalyzed processes are fundamentals in the oil refining and petrochemical industries as well as in the manufacture of a wide variety of specialty chemicals such as pharmaceuticals, agrochemicals and flavors and fragrances. Traditional BrØnsted acids (H2SO4, HF, HCl, p-toluene-sulfonic acid) or Lewis acids (AlCl3, ZnCl2, BF3) and typical bases (NaOH, KOH, NaOMe, KOBu+) have been used in many of these industry processes in liquid-phase homogeneous systems or on inorganic supports in vapor phase systems. The subsequent neutralization step of the 3 used acids and bases causes the generation of inorganic salts which at last are disposed into aqueous waste streams. Replacement of traditional BrØnsted and Lewis acids and bases with recyclable solid catalysts provides an effective solution to this salt generation problem.[7-10] The manufacturing costs and environmental problems associated with the neutralization and disposal of traditional acids and bases are reduced as the need for hydrolytic work-up no longer exists. The use of solid acids and bases as catalysts provides the following advantages:[11] (1) the catalysts can be designed to give high catalytic activity and selectivity, and also longer catalytic life; (2) non-corrosive, safer and easier to handle (3) repeated use is possible; (4) separation of the catalysts is simple, contamination of the product by trace amounts of catalyst is avoided; (5) disposal of used solid acid and base catalysts costs much less money as the neutralization step is eliminated The number of solid acid, base, and acid-base bifunctional catalysts used in industry processes are shown in Table 1.2. The number of solid acid catalysts is the largest due to its demand in the great progress of petroleum and petrochemical industry for the last 40 years. There are only ten processes for solid base catalysis at present because the study of solid base catalysts starts much later than that of solid acid catalysts. However, this study is now becoming more active. For acid-base bifunctional catalysts, the number is estimated to be 14, and the definition of solid 4 acid-base bifunctional catalyst is strictly limited to those having evidence for the bifunctional catalysis. Table 1.2 Number of solid acid, base and acid-base bifunctional catalysts in industry processes[12] Solid acid catalysts 103 Solid base catalysts 10 Sold acid-base bifucntional catalysts 14 Total 127 1.1.2.1 Solid Acid Catalysts A solid acid may be generally defined as a solid which changes the color of a basic indicator or a solid on which a base is chemically adsorbed. From more strictly BrØnsted and Lewis definitions, a solid acid shows a tendency to donate a proton or to accept an electron pair. Solid acids can be summarized in different gourps: (1) natural clay minerals (montmorillonite); (2) zeolites (H-ZSM-5, H-Beta, H-Y); (3) cation exchange resins (amberlyst-15); (4) heteropoly acids; (5) metal oxides (Al2O3, Nb2O5, WO3); (6) mixed oxides (SiO2-Al2O3, SiO2-ZrO2, TiO2-WO3). Solid acid catalysts can be effectively applied to most acid-promoted organic synthesis.[12] Typcial reactions are various electrophilic aromatic substitutions including nitrations, and Friedel-Craft alkylations and acylations, and numerous rearrangement reactions such as the Beckmann and Fries rearrangments. Friedel-Crafts acylation, which is widely applied in the fine chemicals industry, is one of the most representative examples where solid acid catalysts have 5 been used. In contrast to the corresponding alkylations, Friedel-Crafts acylations require more than one equivalent of Lewis acid (AlCl3 or BF3) because of the strong complexation of the Lewis acid by the ketone product. Zeolite beta is employed as a catalyst and latter commercialized in Friedel-Crafts acylation by Rhodia.[13] The discovery of the first zeolite-catalyzed Friedel-Crafts acylation may be considered as a benchmark in this area. The reaction was carried out in fixed-bed operation, where the acetylation of anisole with acetic anhydride gave p-methoxyacetophenone (Scheme 1.1). In the original process, acetyl chloride in combination with 1.1 equivalents of AlCl3 in a chlorinated hydrocarbon solvent generated 4.5 kg of aqueous effluent, containing AlCl3, HCl, solvent residues and acetic acid per kg of product. However, when zeolite is employed as an alternative catalyst, the production of HCl in both the acylation and in the synthesis of acetyl chloride are completely eliminated. In this solid acid catalyzed process, only 0.035 kg of aqueous effluent is generated, which is more than 100 times less, and consisted of 99% water, 0.8% acetic acid and 95% vs. 85-95%) is obtained. Also, the number of unit operations is reduced from twelve to two and the zeolite catalyst is easily recyclable. Hence, the Rhodia process is not only environmentally superior to the traditional process, but also more economically favorable. Therefore, a conclusion can be drawn that application of the appropriate heterogeneous catalyst, in addition to its obvious environmental advantages, also shows economically benefits. Another case comes from the manufacture of the bulk chemical caprolactam which is the raw material for Nylon 6. The conventional process (Scheme 1.2) involves the reaction of cyclohexanone with hydroxylamine sulfate (or another salt) to 6 produce cyclohexanone oxime, which then undergoes the Beckmann rearrangement in the presence of stoichiometric amounts of sulfuric acid or oleum. The overall two-step process generates ca. 4.5 kg of ammonium sulfate per kg of caprolactam in total and roughly equal amount in each step. O + CH3COCl OMe AlCl3 Solvent + HCl + CH3COOH OMe O + (CH3CO)2O H-beta OMe OMe Scheme 1.1 Zeolite-catalyzed vs. classical Friedel-Crafts acylation Ichihashi and coworkers at Sumitomo[14,15] developed a catalytic vapor phase Beckmann rearrangement over a high-silica MFI zeolite. When this is combined with another heterogeneous catalysis technology, developed by Enichem,[16] for the ammoxidation of cyclohexanone with NH3/H2O2 over the titanium silicalite catalyst (TS-1), a yield of caprolactam >98% (based on cyclohexanone) is obtained. The overall process generates caprolactam and two molecules of water from cyclohexanone, NH3 and H2O2 and is essentially salt-free. This process is currently being commercialized by Sumitomo in Japan. Another widely used reaction in fine chemicals manufacture is the acidcatalyzed rearrangement of epoxides to carbonyl compounds. In conventional methods, Lewis acids such as ZnCl2 or BF3·OEt2 are generally used in stoichiometric 7 amounts. But again, zeolites are found to be effectively solid, recyclable catalysts. Two commercially relevant examples are the rearrangements of α-pinene oxide[17,18] and isophorone oxide[19]. The products of these rearrangements are fragrance intermediates. The rearrangements of α-pinene oxide to campholenic aldehyde was catalyzed by H-USY zeolite[17] and titanium-substituted zeolite beta[18]. With the latter, selectivities up to 89% in the liquid phase and 94% in the vapor phase were obtained, which are greater than the best results obtained with homogeneous Lewis acids. NH3 + O2 NO H2 dil. H2SO4 (NH3OH)2SO4 Current Process H2SO4 (>1 eq) O NOH NH O TS-1 Sumitomo Process H2 + O2 H2O2 NH3 High Si MFI vapor phase Scheme 1.2 Sumitomo vs. conventional process for caprolactam manufacture 1.1.2.2 Solid Base Catalysts Similar to solid acids, a solid base is defined to be a solid which tends to accept a proton or to donate an electron pair. Solid base catalysts consist of primarily basic metal oxides and mixed oxides (MgO, CaO, ZnO, SiO2-MgO, Al2O3-MgO, TiO2-MgO). 8 The replacement of conventional bases, such as NaOH, KOH and NaOMe, by recyclable solid bases for a variety of organic reaction, becomes more active recently. One of the examples is synthetic hydrotalcite clays, or known as layered double hydroxide (LDHs), which have the general formula Mg8-xAlx(OH)16(CO3)x/2·nH2O. They are basically hydrated aluminum-magnesium hydroxides possessing a lamellar structure in which the excess positive charge is compensated by carbonate anions in the interlamellar space.[20, 21] Dehydroxylation and decarbonation occur during calcination of hydrotaciles, which led to the formation of strongly basic mixed magnesium-aluminum oxides. Reactions involving these strong solid base catalysts include inter alia, aldol,[22] Knoevenagel[23, 24] and Claisen-schmidt[24] condensations (Scheme 1.3). CN CHO + H2C CO3Et hydrotalcite CN CO2Et R2 CHO OH R1 OH hydrotalcite + COCH3 R1 R2 O O O + H2O2 hydrotalcite MeOH; 20°C O Scheme 1.3 Hydrotalcite-catalyzed condensation reactions Another approach for synthesis of recyclable solid bases is to precipitate organic bases onto the surface of materials like mesoporous silicas.[25, 26] For example, aminopropyl-silica, which are formed from reaction of 3-aminopropyl(trimethoxy) 9 silane with pendant silanol groups, was reported to be an active catalyst for Knoeveagel condensations.[27] A stronger solid base called MCM-TBD was obtained by functionalisation of mesoporous MCM-41 with the guanidine base, 1,5,7-triazabicyclo-[4,4,0]dec-5-ene (TBD), using a surface glycidylation technique followed by reaction with TBD (Scheme 1.4). The resulting material was active for Knoevenagel condensation, Michael addition and Robinson annulation.[28] N O O OH 1. (H3CO)3Si N N O MCM OH 2. N N H N TBD MCM Scheme 1.4 Tethered organic bases as solid base catalysts 1.1.2.3 Acid-Base Bifunctional Catalysts An acid-base bifunctioaal catalyst contains a weak acid site as well as a weak base site on a solid surface. The simultaneous cooperation of the acid and base sites exhibits surprisingly high catalytic activity and selectivity towards some organic synthesis requiring both acid and base. The number of these examples is increasing in the industrial application of the bifunctional catalysis. Ethyleneimine derivatives are commercially important chemicals which are used for the production of pharmaceuticals and various other amines and for the production of amine type functional polymer for coatings of paper and textile. Following the Wenker process, ethyleneimine was produced by intramolecular 10 dehydration of monoethanolamine in liquid phase using sulfuric acid and sodium hydroxide. However, problems such as low productivity and formation of large amounts of sodium sulfate occurred. In contrast, the vapor phase process using solid acid-base bifunctional catalyst is more advantageous than the liquid phase process, because the formation of undesirable by-products such as acetaldehyde, piperidine, ethylamine, and acetonitrile is minimized. A new efficient catalyst (Si-Ba-Cs-P-O) has been developed by Nippon Shokubai with the conversion of monoethanolamine and the selectivity for aziridine being 86% and 81% (Scheme 1.5).[29] The acid and base strengths of the catalyst are weaker than H0 = +4.8 and H_ = -9.4, respectively. This reaction is considered to be catalyzed by an acid-base bifunctional mechanism. OH OSO3H H2SO4 H2N H2N MEA (Liquid phase) NaOH N H EI Solid catalyst (Vapor phase) Scheme 1.5 Synthesis of ethyleneimine (EI) from monoethanolamine (MEA) 1.1.3 Supported Oxide Catalysts 1.1.3.1 Introduction Supported oxides have several advantages over unsupported materials. Firstly, the surface area of the catalyst can be increased by supporting the active oxide on a high surface area material. Also, supports are effective in improving the mechanical strength, the thermal stability and the lifetime of the catalyst. In addition, the 11 structural features and chemical compositions of supported oxides are usually different from the bulk oxide. The active oxide interacts with the surface of the supporting material in several ways (Fig. 1.1). In the first case, the active component and the support are intimately mixed throughout the catalyst particle. In this type of arrangement, the maximum interaction between the support and the active oxide can be achieved. However, a large portion of the active sites is hindered in the support. The second way of interaction is through catalytic dispersion. The active oxide molecules are bound to the surface of the support as a localized binary oxide. The third possibility is formation of monolayer on the support. Monolayer coverage is considered the best way of interaction for a number of oxides.[30] The concept of the oxide monolayer was suggested by Russell and Stokes.[31] By analogy with dispersion of small particles of active oxide on a support, an oxide is most efficient when present as a layer as thinly as possible over the surface of the support. The formation of a monolayer or a monomolecular dispersion maximizes the influence of the support to the active sites. However, in practice, formation of a monolayer of active oxide on the surface may not be possible. More frequently, the surface coverage will be the combination of areas of monolayer accompanied by some larger supported oxide particles and some uncovered surface. There are several methods commonly used to prepare supported oxide catalysts.[32] These methods differ mainly in the manner in which the active component is placed to contact with the support. 12 Support Support (a) (b) Support Support (c) (d) Fig. 1.1 Various arrangement of active oxide component on support material (a) portion of oxide component buried in the support (b) oxide component dispersed on the surface of support material (c) formation of monolayer (d) mixture of monolayer, aggregated particles and uncovered support surface. Coprecipitation is achieved by treatment of a solution containing both the support precursor and the catalytic precursor with a suitable precipitating agent. The resulting precipitate contains both the active component and the support material. The main disadvantage of using this method is that a large portion of the active sites is hindered in the support, and hence unavailable for catalysis. Another method for preparing supported oxide catalysts involves the dispersion of the active species or its precursor onto the surface of a support oxide. This can be achieved by equilibrium adsorption, incipient wetness or dry impregnation. 13 The incipient wetness impregnation method makes use of a solution of the active species to fill the pores of the support material. This is normally achieved by slowly addition of the solution to the solid support material with continuous stirring. The concentration of the solution can be adjusted to give the desired catalytic loading. Once the incipient wetness point is reached, the solvent is filtered or evaporated and the solid residue is calcined to give the catalyst. Depending on the solubility of the catalytic precursor, aqueous or organic solutions are prepared. The disruption of pores in the support may occur when high surface tension solvents such as water are evaporated resulting in a decrease of the surface area of the catalyst. The concentration of the precursor solution affects the average size of the catalytic particles. Normally, high catalytic loadings tend to give larger particles than low loadings. The drying step also has effect on the size of catalytic particles. A slow drying step makes the catalytic particles migrate towards the external surface of the support. Therefore, in order to obtain a uniform distribution of active sites throughout the support, a quick drying step is necessary. By mixing two or more different oxides together, materials with acidic properties different from that of bulk oxides can be generated. Three different proposals have been formulated to predict the acidity of mixed oxides. (i) In the case where the mixed oxide is an actual solid solution, acidity is created if the charge of the guest cation is lower than the charge of the host cation. Normally, BrØnsted acidity can be created as charge balance. Zeolite is a well-known example of this type of acidity generation. BrØnsted acidity is created by the substitution of tetrahedral Si4+ by A13+. (ii) According to Tanabe et al.,[33,34] Lewis or BrØnsted acidity may also be formed if the guest cation preserves its coordination number with a surrounding of 14 oxide ions in host oxide. Significant examples are mixed oxides e.g. TiO2-SiO2, ZrO2-SiO2 and Nb2O5-SiO2, or Nb2O5-TiO2, in which higher acidity is observed than their individual oxide components. (iii) Finally, Seiyama[35] proposed that acidity appears at the boundary where two oxides contact. In this case, each cation preserves its own environment. Only the oxide ion at the boundary is coordinated to two different cations. This model may be applied in the case of the dispersion of an oxide on a support. 1.1.3.2 Zirconia as Support Material In recent years, zirconia has attracted much attention as both a catalyst and a catalyst support because of its high thermal stability and the amphoteric character of its surface hydroxyl groups. Zirconia has the following advantages when used as a catalyst support:[36] (i) it interacts strongly with the active phase; (ii) it possesses high thermal stability and is more chemically inert than the classical supported oxides; (iii) it is the only metal oxide which may possess all four chemical properties: acidity, basicity, reducing ability and oxidizing ability; (iv) it exhibits super-acidic properties when modified with small quantities of sulfate ions. The extreme hardness and high specific mass of zirconia can also be an advantage for its potential use as catalyst support. The application of zirconia as a catalyst support has been employed in many industrially important reactions such as hydroprocessing,[37] oxidation of alcohols,[38,39] and synthesis of methanol and higher alcohols.[40] 15 Structure of Zirconia Zirconia goes through three stages of crystalline phase transformations. The monoclinic phase is stable up to 1200°C, the tetragonal phase is stable up to 1900°C, and the cubic phase is stable above 1900°C. In addition, a metastable tetragonal form is also found and is stable up to 650°C. The metastable tetragonal, the monoclinic, and the tetragonal zirconia are often used as catalysts. Hydrous zirconia is used as catalyst for organic synthesis as well. Compared with SiO2 and Al2O3 which are often used as catalyst supports, the surface area of zirconia is not very large. The surface area depends much on the calcination temperature. After calcined at 600°C, the surface area is in the range of 40 - 100 m2/g. By addition of another metal oxide component, a higher surface area may be obtained. An alternative for increase of surface area is by dispersing zirconia on the high surface area supports. Calcination of hydrous zirconia which is formed by hydrolysis of zirconium salts gives zirconium oxide. The preparation method affects the crystal forms of zirconia. Formation of monoclinic phase upon calcination is preferred if the zirconia precursor is aged for a long period while the tetragonal phase is dominant if aging is omitted. Surface Properties of Zirconia The surface of the metal oxides can exhibit acidic, basic, oxidizing, and reducing properties. Most metal oxides show only one of the properties. However, for zirconia, both acidic and basic properties are present on the surface though their 16 strength is rather weak. Meanwhile, zirconia also exhibits oxidizing and reducing properties. In solutions, acid and base are neutralized immediately. However, they may exist independently at the surface of solid materials. The acidic and basic sites on the surface of oxides work both independently and cooperatively. Therefore, zircoinia can be considered as an acid-base bifunctional oxide. By measuring the amount of chemically adsorbed ammonia and CO2, the numbers of acid sites and base sites can be estimated, respectively. Zirconia calcined at 600°C exhibits 0.6 pmol/m2 of acidic sites and 4 pmol/m2 of base sites. Pyridine absorption infrared spectra show the presence of only Lewis acid sites, but not BrØnsted acid sites.[41] Zirconia calcined at 500°C exhibits two types of OH groups at 3780 and 3680 cm-1.[42, 43] These two bands at 3780 and 3680 cm-1 are assigned to be the terminal and bridged OH groups, respectively.[43] Application as Catalyst Support Zirconia is stable under oxidizing and reducing conditions and it possesses both acidic and basic properties. Also, unlike Al2O3, zirconia does not form solid solutions upon addition of a second metal oxide component. Therefore, the use of zirconia as a support may be promising in despite of its relative low surface area. When zirconia is used as catalyst support, pronounced catalytic activities can be obtained. One example is the supported perovskite (La-Sr-Co-O) catalyst for the propane oxidation.[44] Both of the high surface area supports, Al2O3 and SiO2, were 17 not effective in promoting the activity of perovskite, but when perovskite was dispersed on zirconia surface, this supported catalyst exhibited activity 10 times higher than that of the original perovskite. Zirconia was also found to be a superior support for Re and Rh in the hydrogenation of CO2.[45-47] 1.2 Synthesis of Indoles 1.2.1 Introduction For the last 100 years, the synthesis of indoles and their derivatives have been a research topic of great interest. The first synthesis of indole was in 1866 and later in 1883, the Fischer indole synthesis was discovered, which is still the most versatile method for preparing indoles nowadays.[48] Every year, several thousands of indoles and indole derivatives have been synthesized in the research laboratories. The sustained interest in indoles is due to their wide range of biological activity.[49] The indole ring structures in the amino acid tryptophan and metabolites of tryptophan are important in the biological function of both plants and animals. The indole ring structure is also found in many natural products such as the indole alkaloids, fungal metabolites and marine natural products. Many indole derivatives have been used as drugs. Indomethacin is one of the first non-steroidal anti-inflammatory agents; sumatriptan is used in the treatment of migraine headache and pindolol is one of the β-adrenergic blockers (Scheme 1.6). CH2CO2H H3CO CH2CH2NH2 H3CHNO2SH2C CH3 OH OCH2CHCH2NHCH(CH3)2 N N H O Cl indomethacin sumatriptan N H pindolol Scheme 1.6 Drugs containing indole structures[50] 18 Indole is classified as a π-excessive aromatic compound. It is isoelectronic with naphthalene, with two of the ten π-electrons from the heterocyclic nitrogen atom (Scheme 1.7). The aromaticity of the ring is fundamental to the success of many synthetic methods. N H N H N H Scheme 1.7 Resonance structures of indole[50] 1.2.2 Fischer Indole Synthesis For over a century, the Fischer indole reaction has remained as an extremely useful and important method for the synthesis of a variety of biologically active indole-structure compounds.[51] The Fischer indole reaction is one of the simplest and most efficient methods by which enolizable N-arylhydrazones are easily transformed into indoles. In the Fischer indole synthesis, the ketone or aldehyde and the arylhydrazine are generally mixed and heated to form the hydrazone intermediate. Without isolation of the intermediate, indoles are formed by simply adding the appropriate acid or acid catalyst. Advantages of the Fischer indole reaction include the acceptance of a wide range of compatible functional groups around the aromatic ring and lack of a requirement for a functional group to form the new C-C and C-N bonds. The mechanism of the Fischer indole synthesis involves a complex series of acid-catalyzed reactions and rearrangements (Scheme 1.8). The generally key steps are (i) the condensation of arylhydrazines with ketones to form arylhydrazones, (ii) rearrangement of the arylhydrazones to form ene-hydrazines (iii) acid-catalyzed 19 [3,3]-sigmatropic rearrangement followed by intramolecular displacement of NH3 to produce indoles.[52,53] When an unsymmetrical ketone is used, the rearrangement of arylhydrazones results in the formation of a mixture of ene-hydrazine isomers. R1 R1 H+ N H N R2 R1 N N H H H R1 R2 NH2 NH R2 R1 R2 N H -NH2 NH2 R2 N Scheme 1.8 Mechanism of Fischer indole synthesis The arylhydrazones are frequently prepared via condensation of an arylhydrazine with a ketone. However, as few arylhydrazines are commercially available, an alternative way can be used to prepare the hydrazones by reduction of aryl diazonium salts, which are obtained from the appropriate aniline. Aryl diazonium salts can be converted directly to hydrazones via the Japp-Klingemann reaction (Scheme 1.9).[54] In the Japp-Klingemann reaction, the aryl diazonium salts are firstly treated with an active methylenyl or methinyl compound in acidic or basic environment to generate azo intermediates. The azo intermediates then react with β-ketoesters or β-ketoacids to form hydrazones. With β-ketoesters, deacylation occurs, but if β-ketoacids are used, decarboxylation happens with the carboxylic acid group being expulsed. The Japp-Klingemann procedure avoids the formation and use of arylhydrazines, which can be difficult to prepare and handle in some cases. 20 R2 O CO2R1 + R N2 N H R2 N CO2R1 CH2R2 CO2R1 N H R2 O + R N2 CO2H N H R2 N CO2R CH2R2 CO2R N H Scheme 1.9 Fischer indole synthesis via Japp-Klingemann reaction . 1.2.3 Other Methods of Indole Synthesis Other than Fischer indole synthesis, several classical methods have been reported (Scheme 1.10), including the Batcho–Limgruber synthesis from o-nitrotoluenes and dimethylformamide acetals,[55] the Gassman synthesis from N-haloanilines,[56,57] the Madelung cyclisation of Nacyl- o-toluidines[58,59] and the reductive cyclization of o-nitrobenzyl ketones.[60] One of the famous indole synthesis methods is the heteroannulation of 2-iodoanilines with alkynes via the Sonogashira cross-coupling reaction catalyzed by the palladium-copper catalysts.[62 - 65] This approach also presents many advantages:[66] (1) The Sonogashira reaction has been proved to be a reliable reaction for the synthesis of arylethylenes in high yields. (2) This indole synthesis method is widely employed as a key step for the synthesis of natural and pharmaceutical products. (3) It tolerates a broad variety of functional groups on the aromatic rings. (4) It is completely regiospecific at 2,3-positions in the indole ring as the 2,3-regioselectivity is directed by the choice of the alkyne used in the Sonogashira cross-coupling. 21 O R2 N NH R1 R2 N O R3 + R3 h R R3 N N R R1 R2 O R3 b f + R1 a R2 R3 N2 NHNH2 g R1 R2 Z Z N2 O + R3 R3 + OH/R1 N N R2 R e H R3 R c H d NO2 R3 + H2/Rh/C + CO/H2 N O NO2 O P RO N (OEt)2 R H N R1 O O R R1 + O P RO (OEt)2 N2 O Scheme 1.10 Various routes for synthesis of substituted indole[61] Generally, the indole synthesis via the Sonogashira reaction is carried out in two steps (Scheme 1.11). (1) The palladium-copper catalyzed Sonogashira crosscoupling between a 2-aminoaryl halide and an alkyne. (2) A copper(I) or palladium(II) catalyzed ring-closure by hydroamination/cyclization to give the indole product. A two-step procedure is generally required since the Sonogashira cross-coupling is the most efficient under mild reaction conditions, while the intramolecular heteroannulation requires stronger reaction conditions in order to obtain high conversion, such as the use of strong bases and high temperature. However, when the aniline is activated, for example, a presence of electron withdrawing substitution 22 groups like sulfonyl, the heteroannulation can occur under relatively mild reaction conditions. However, for industrial applications this methodology remains too expensive, mainly due to the high costs of catalysts and costs in multiple steps of handling, separation and refining procedures. No general simple and economical procedure has been described via the Sonogashira cross-coupling reaction with heterogeneous catalysts. Ph H I + NH2 Ph [Pd-Cu] [Pd-Cu] NH2 Ph N H Scheme 1.11 Indole synthesis via Sonogashira reaction 1.2.4 Catalysts Used in Fischer Indole Synthesis The Fischer indole synthesis is an acid-catalyzed reaction. Both BrØnsted and Lewis acids are effective. Catalysts employed include mineral acids (HCl, H2SO4, H3PO4), organic acids (acetic acid), and metal-halide Lewis acids (ZnCl2). A neutralization procedure is required after reaction if the traditional homogeneous BrØnsted and Lewis acids are employed in the synthesis. The disposal of salts or metal-containing waste generated is environmental unfriendly and highly costly. Applying heterogeneous solid acid catalysts for the Fischer synthesis avoids waste formation and has the advantages of simple separation, recovery, and regeneration of the catalysts. The possible solid catalysts include the acidic forms of ion-exchange resins, oxides such as silica, alumina and silica-alumina, and zeolites. 23 Suvorov et al.[67] investigated the cyclization of isolated aldehyde and ketone phenylhydrazones using γ-alumina. The reaction was conducted in the vapor phase, at atmospheric pressure and temperatures around 300°C. A maximum yield of 60% was obtained from acetaldehyde phenylhydrazone accompanied with thermal decomposition of the hydrazone and the formation of benzene and aniline by-products. Zeolites are microporous materials with pore diameters comparable with molecular dimensions. The shape selectivity of zeolites has wide applications in oil refining and bulk chemicals catalysis. The spatially restricted reaction environment of zeolite pores can be used in the Fischer indole reaction to control the selectivity towards the less bulky isomeric product. Thus, in addition to the advantages of easy recovery and regeneration, zeolites provide possibility of regioselective formation of a single indole isomer. The large-pore zeolites, such as beta, mordenite and faujasites X and Y are most applied in fine-chemical catalysis. Prochatska et al.[68] studied the Fischer indole synthesis of phenylhydrazine with five different unsymmetrical ketones using fourteen different zeolites in one-pot reactions. Yields of 80 - 90% were obtained from the phenylhydrazones of 3-hexanone, 3-undecanone, 5-methyl-3-heptanone, and 1-phenyl-2-butanone using large-pore zeolites. As the most striking example of shape selectivity, Prochazka et al.[68] reported that in the synthesis of indole from phenylhydrazine and 1-phenyl-2butanone, the selectivity of the least bulky isomer, 2-benzyl-3-methylindole was 93% where mordenite is used as catalyst, whereas only 2-ethyl-3-phenylindole was formed if the reaction was performed in acetic acid. 24 1.3 Aims of Present Study Fischer indole synthesis is important and widely applied in industrial production of indoles and their derivatives. The reaction involves acid-catalyzed condensation and cyclization in which water is liberated. Niobium oxide is a water-tolerant catalyst. Deposition of niobium oxide at the surface of hydrous zirconia support generates a novel catalyst with high acidity and water tolerance. Boria-zirconia is considered as a superacid catalyst. It is used for various solid acid-catalyzed reactions. The properties and activity of boria-zirconia catalysts can be improved by proper control of variable factors such as loadings and dispersion of boria. The recycling of boria-zirconia is convenient. The aims of this study include: (1) Synthesis of hydrous zirconia supported niobium oxide catalysts and boriazirconia catalysts. The factors of niobium oxide/boria loadings and the calcination temperatures are varied. (2) Characterization of the catalysts using techniques including powder X-ray diffraction (XRD), inductively coupled plasma atomic emission spectroscopy (ICP-AES), infrared spectroscopy (IR), thermogravimetry (TG), differential thermal analysis (DTA), amine titration, and temperature programmed desorption of ammonia (NH3 TPD). (3) Catalytic testing of the hydrous zirconia supported niobium oxide catalysts and boria-zirconia catalysts in the Fischer indole synthesis of phenylhydrazine with 3-heptanone and cyclohexanone. (4) Regeneration of the catalysts. 25 References [1] R. A. Sheldon, Chem. tech. 1994, 38 [2] R. A. Sheldon, J. Chem. Tech. Biotechnol. 1997, 68, 381 [3] R. A. Sheldon, H. van Bekkum, in Fine Chemicals through Heterogeneous Catalysis (Eds. R. A. Sheldon, H. van Bekkum) 2001 (Wiley-VCH: Weinheim, New York) [4] B. Cornils, W. A. Herrmann, in Applied Homogeneous Catalysis with Organometallic Compounds (Eds. B. Cornils, W. A. Herrmann) 1996 (VCH: Weinheim) [5] K. Drauz, H. Waldmann, in Enzyme Catalysis in Orgnic Syntheis (Eds. K. Drauz, H. Ealdmann) 1995 (VCH: Weinheim) [6] G. Ertl, H. Knozinger, J. Weitkamp, in Handbook of Heterogeneous Catalysis (Eds. G. Ertl, H. Knozinger, J. Weitkamp) 1997 (Wiley-VCH: Weinheim) [7] W. HÖlderich, Catal. Today 2000, 62, 115 [8] A. Corma, H. Garcia, Chem, Rev. 2003, 103, 4307 [9] R. A. Sheldon, R. S. Downing, Appl. Catal., A 1999, 189, 163 [10] A. Mitsutani, Catal. Today 2002, 73, 57 [11] K. Tanabe, in Catalysis: Science and Technology (Eds. J. R. Anderson, M. Boudart) 1981 (Springer-Verlag: Berlin, Heidelber, New York) [12] K. Tanabe, W. F. HÖlderich, Appl. Catal., A 1999, 181, 399 [13] M. Spagnol, L. Gilbert, D. Alby, in The Roots of Organic Development (Eds. J. R. Desmurs, S. Ratton) 1996 (Elsevier: Amsterdam) [14] H. Ichihashi, M. Kitamura, Catal. Today 2002, 73, 23 [15] H. Ichihashi, H. Sato, Appl. Catal., A 2001, 221, 359 26 [16] P. Roffia, G. Leofanti, A. Cesana, M. Mantegazza, M. Padovan, G. Petrini, S. Tonti, P. Gervasutti, in Stud. Surf. Sci. Catal. (Eds. G. Centi, F. Trifiro) 1990, vol. 55 (Elsevier: Amsterdam, New York) [17] W. F. HÖlderich, J. RÖseler, G. Heitmann, A. T. Liebens, Catal. Today 1997, 37, 353 [18] P. J. Kunkeler, J. C. van der Waal, J. Bremmer, B. J. Zuurdeeg, R. S. Downing, H. van Bekkum, Catal. Lett. 1998, 53, 135 [19] J. A. Elings, H. E. B. lempers, R. A. Sheldon, in Stud. Surf. Sci. Catal. (Eds. H. Chon, S. –K. Ihm, Y. S. Uh) 1997, vol. 105 (Elsevier Science B. V.: New York) [20] B. F. Sels, D. E. De vos, P. A. Jacobs, Cat. Rev. Sci. Eng. 2001, 43, 443 [21] A. Vaccari, Catal. Today 1998, 41, 53 [22] F. Figueras, D. Tichit, M. Bennani Naciri, R. Ruiz, in Catalyis of Organic Reactions (Ed. F. E. Herkes) 1998 (Marcel Dekker: New York) [23] A. Corma, R. M. Martin-Aranda, Appl. Catal., A 1993, 105, 271 [24] M. J. Climent, A. Corma, S. Iborra, J. Primo, J. Catal. 1995, 151, 60 [25] D. Brunel, A. C. Blanc, A. Galarnean, F. Fajula, Catal. Today 2002, 73, 139 [26] A. C. Blanc, S. Valle, G. Renard, D. Brunel, D. Macquarrie, C. R. Quinn, Green Chem. 2000, 2, 283 [27] D. Brunel, Micropor. Mesopor. Mater. 1999, 27, 329 [28] Y. V. Subba Rao, D. E. De Vos, P. A. Jacobs, Angew. Chem. Int. Ed. 1997, 36, 2661 [29] K. Tanabe, H. Hattori, T. Yamaguchi, T. Tanaka, in Acid-Base Catalysis (Eds. K. Tanabe, H. Hattori, T.Yamaguchi, T. Tanaka) 1989 (VCH: Weinheim) [30] G. C. Bond, Appl. Catal. 1991, 71, 1 27 [31] A. S. Russell, J. J. Stokes, Ind. Eng. Chem. 1946, 38, 1071 [32] G. C. Bond, in Heterogeneous Catalysts: principles and Applications (Eds. P. W. Atkins, J. S. E. Holker, A.K. Holliday) 1987 (Clarendon Press: Oxford) [33] K. Tanabe, T. Sumiyoshi, K. Shibata, T. Kiyoura, J. Kitagawa, Bull. Chem. Soc. Jpn. 1974, 47, 1064 [34] K. Tanabe, M. Misono, Y. Ono, H. Hattori, in Stud. Surf. Sci. Catal. (Eds. K. Tanabe, M. Misono, Y. Ono, H. Hattori) 1989, vol. 51 (Kodansha: Tokyo) [35] T. Seiyama, in Metal Oxides and Their Catalytic Actions (Ed. T. Seiyama) 1978 (Kodansha: Tokyo) [36] T. Bhaskar, K. R. Reddy, C. P. Kumar, M. R. V. S. Murthy, K. V. R. Chary, Appl. Catal., A 2001, 211, 189 [37] E. D. Whitney, Trans. Faraday Soc. 1965, 61, 1991 [38] R. C. Garvie, J. Phys. Chem. 1965, 69, 1238 [39] K. S. Mazdiyasni, A. T. Lynch, J. S. Smith, J. Am. Ceram. Soc. 1965, 48, 372 [40] K. S. Mazdiyasni, A. T. Lynch, J. S. Smith, J. Am. Ceram. Soc. 1966, 49, 286 [41] Y. Nakano, T. Iizuka, H. Hattori, K. Tanabe, J. Catal. 1978, 57, 1 [42] P. A. Agron, E. L. Fuller, H. F. Homes, J. Colloid. Interface. Sci. 1975, 52, 553 [43] T. Yamaguchi, Y. Nakano, K. Tanabe, Bull. Chem. Soc. Jpn., 1978, 51, 2482 [44] H. Fuji, N. Mizuno and M. Misono, Chem. Lett. 1987, 2147 [45] T. Itzuka, Y. Tanaka, K. Tanabe, J. Catal. 1982, 76, 1 [46] T. Iizuka, Y. Tanaka, K. Tanabe. J. Mol. Catal. 1982, 17, 381 [47] T. Itzuka, M. KoJima, K. Tanabe, J Chem. Soc. Chem. Commun. 1983, 638. [48] W. C. Sumpter, F. M. Miller, Chem. Heterocycl. Cmpds. 1954, 8, 1 [49] K. C. Joshi, P. Chand, Pharmazie 1982, 37, 1 28 [50] R. J. Sundber, in Indoles (Ed. O. Meth-Cohn) 1996 (Academic Press: London) [51] Robinson, B. in The Fischer Indole Synthesis (Ed. B. Robinson) 1982 (WileyInterscience: New York) [52] D. L. Hughes, Org. Prep. Proced. Int. 1993, 25, 607 [53] L. Przheval’ski, Y. Kostromina, I. I. Grandberg, Chem. Heterocycl. Compd. 1988, 24, 709 [54] R. R. Phillips, Org. React. 1959, 10, 1143 [55] R. D. Clark, D. B. Repke, Heterocycles 1984, 22, 195 [56] P. D. Johnson, P. A. Aristoff, J. Org. Chem. 1990, 55, 1374 [57] S. W. Wright, L. D. McClure, D. L. Hageman, Tetrahedron Lett. 1996, 37, 4631 [58] W. J. Houlihan, Y. Uike, V. A. Parrino, J. Org. Chem. 1981, 46, 4515 [59] D. A. Wacker, P. Kasireddy, Tetrahedron Lett. 2002, 43, 5189 [60] R. J. Sundberg, in Comprehensive Heterocyclic Chemistry (Eds. A. R. Katritzky, C. W. Rees) 1984, vol. 4 (Pergamon: Oxford) [61] S. Patil, R. Patil, Curr. Org. Synth. 2007, 4, 201 [62] J. Ezquerra, C. Pedregal, C. Lamas, J. Barluenga, M. Pérez, M. A. GarciaMartin, J. M. González, J. Org. Chem. 1996, 61, 5804 [63] M. C. Fagnola, I. Candiani, G. Visentin, W. Cabri, F. Zarini, N. Mongelli, A. Bedeschi, Tetrahedron Lett. 1997, 38, 2307 [64] H.-C. Zhang, K. K. Brumfield, B. E. Maryanoff, Tetrahedron Lett. 1997, 38, 2439 [65] H. -C. Zhang, H. Ye, A. F. Moretto, K. K. Brumfield, B. E. Maryanoff, Org. Lett. 2000, 2, 89 29 [66] S. Chouzier, M. Gruber, L. Djakovitch, J. Mol. Catal. A: Chem. 2004, 212, 43 [67] N. N. Suvorov. A. D. Pozdnyakov, M. Ya. Bykhovskii, V. V. Sadovnikov, J. Org. Chem. 1975, 11, 2684 [68] M. P. Prochazka, L. Eklund, R. Carlson, Acta Chem. Scand. 1990, 44, 610 30 Chapter II Experimental 2.1 Powder X-ray Diffraction X-ray diffraction is one of the tools for the investigation of the structures of materials.[1] X-ray diffraction was initially applied only for the determination of crystal structure. Later on, this technology was applied in other areas as well. For example, it is applied to the study of phase equilibria and the measurement of particle size nowadays. The powder X-ray diffraction method was first invented independently by German scientists Debye and Scherrer in 1916 and by Hull from the United States in 1917. This method generally makes use of the diffraction of monochromatic x-rays by a powder specimen (Fig. 2.1). Diffraction occurs whenever the Bragg law (equation 2.1) is satisfied: nλ = 2dsinθ (2.1) where n is an integer normally taken as 1, d is interplanar spacing, θ is the angle of diffraction, and λ is the wavelength of X-ray. Knowing the wavelength of the X-ray λ and the diffraction angle θ, the interplanar spacing d of various planes can be determined. The interplanar spacing d, measured at certain diffraction angles to the planes, is a function both of the plane indices (hkl) and the lattice constants (a, b, c, α, β, γ). For cubic system, the spacing equation is in a relatively simple form (equation 2.2): 31 d= a (2.2) h + k 2 + l2 2 In the tetragonal system the equation contains both a and c as they are not equal (equation 2.3): d= a (2.3) h 2 + k 2 + l 2 (a 2 / c 2 ) Using these spacing equations, the plane indices corresponding to the right diffraction angle can be determined. θ d x Fig. 2.1 Reflection of X-rays from two planes The particle size of very small crystals can be estimated by the Scherrer formula (equation 2.4) from the width of their diffraction curves: 32 τ= Kλ β cos θ (2.4) where τ is the mean crystallite dimension, K is the shape factor, λ is the X-ray wavelength, θ is the Bragg angle, and β is the line broadening at half the maximum intensity (FWHM) in radians In this study, the crystalline phases of the samples were determined by the powder X-ray diffraction using a Siemens D5005 diffractometer. The Cu anode was used and operated at 40 kV and 40 mA. The measured area was set to be 20 × 20 mm. The range in a typical scan started from 2θ = 1.5° to 70°. A step size of 0.02°, and a measurement time of 1 second/step were used. By comparing with the JCPDS files, the crystalline phases of the samples can be verified except for the samples with amorphous phase. 2.2 Nitrogen Adsorption/Desorption Study 2.2.1 BET Surface Area Determination The measurement of adsorbed vapors by a dispersed powder material gives valuable information on its surface area and its pore structure. It is possible to use the adsorption data to determine the specific surface area of the solid. Also in the case of a porous material, the pore size distribution can be estimated. Three main types of method are developed for the determination of adsorption isotherms, which are volumetric, gravimetric, and dynamic methods. Among these three methods, the most commonly used one is the volumetric method. Nitrogen is often chosen to be one of the most suitable adsorbate to determine the surface area for nonporous, macroporous, or mesoporous solids. The BET 33 monolayer of nitrogen is usually assumed in a closed packed form. To measure the surface area, first the quantity of gas adsorbed as a single layer of molecules on a sample should be determined. This adsorption happens at temperature near the boiling point of the adsorbate gas, and for nitrogen, the temperature is 77 K. The area occupied by one molecule of nitrogen has been estimated to be 0.162 nm2/mol[2]. The surface area is then calculated from the number of adsorbed molecules multiplied by the area occupied by each molecule. In 1938, Stephen Brunauer, Paul Hugh Emmett, and Edward Teller together published the BET theory and this theory is later on named with the first initials of their family names. BET theory is an extension of the Langmuir theory. It develops the isotherm of monolayer molecular adsorption to multilayer adsorption with the following hypotheses. Hypothesis 1: gas molecules physically adsorbed on a solid in layers infinitely. Hypothesis 2: there is no interaction between each adsorption layer. Hypothesis 3: the Langmuir theory can be applied to each layer. The BET equation is written in the form (equation 2.5): 1 c −1 ⎛ P ⎞ 1 = ⎜ ⎟ + v [( P0 / P) − 1] vmc ⎝ P0 ⎠ vm c (2.5) where P and P0 are the equilibrium and saturation pressure of adsorbates at the temperature of adsorption, v is the adsorbed gas quantity, vm is the monolayer adsorbed gas quantity, and c is the BET constant. The monolayer capacity vm, is defined as the quantity of the adsorbate that can be accommodated in a completely filled, single layer of molecules on the surface of 34 the solid. The specific surface area S (m2g-1) is directly proportional to the monolayer capacity. These two values are related by the simple equation (equation 2.6):[3] S= vm · N · Am × 10-20 M (2.6) where vm is the monolayer capacity in grams of adsorbate per gram of solid, M is the molecular weight of the adsorbate, Am (Å2) is the area occupied per molecule of adsorbate in the completed monolayer and the factor 10-20 is introduced in order to give S in square metres per gram. Volume Adsorbed/Desorbed (cm3/g) 2.2.2 Porosity by Gas Adsorption Relative Pressure (P/P0) Fig. 2.2 The five types of adsorption isotherms 35 In 1945, Brunauer proposed the classification of the different kinds of adsorption isotherms into five general forms (Fig. 2.2). Type I or Langmuir physisorption isotherms are characteristic of monolayer adsorption, and often exhibited by microporous solids having very small pores that all filled up at once. Type II isotherms are characteristic of multilayer adsorption, which is normally obtained from nonporous or macroporous adsorbent. Occasionally type III adsorption isotherms are observed where initially there is very little adsorption, but additional adsorption occurs more easily due to strong adsorbate-adsorbate interactions. Type IV and type V isotherms usually represent multilayers of gas adsorbed onto the surface of the pores in a mesoporous solid. Initially, the type IV or type V adsorption resembles a type II or a type III adsorption respectively. The nitrogen adsorption–desorption isotherms for the solid samples were measured using a Micromeritics Tristar porosimeter. Prior to measurements, the samples were degassed for 4 h under a flow of nitrogen at temperatures of 100°C – 300°C. 30 – 50 mg of sample was usually measured and liquid nitrogen bath was used to cool down the sample cells to the boiling temperature of nitrogen, 77 K. 2.3 Infrared Spectroscopy Infrared spectroscopy is one of the most widely used methods in both research and industry. It is simple and reliable, and large applications are found in works as quality control and dynamic measurement. Infrared spectroscopy has been highly successfully applied in both organic and inorganic chemistry. In heterogeneous catalysis, infrared spectroscopy can be used for 36 study of solid catalysts.[4] Such characterization can be very useful in determining the nature of the support, and surface groups such as hydroxyls on the suface of oxides. Bands arising from the bulk structure of catalysts or supports can readily be observed. From the infrared spectra of supported metals oxides, the information of the bonding modes of metal atoms with oxygen atoms can be determined, and hence the coordination structures of the supported and bulk metal oxides are determined. A small amount of sample was diluted with KBr salt (ca. 1:100), and after thorough mixing and grinding, wafers were prepared by pressing the fine powder in a 13 mm die under vacuum at ~15,000 psi. FT-IR spectra of samples in the wavelength region 4000-400 cm-1 were recorded at room temperature. A Bio-Rad FTS 165 FT-IR spectrometer was used at a resolution of 2 cm-1, and 32 scans were measured and averaged. Identifying the nature of the acid sites on acidic oxides and zeolites using a base molecule probe is another popular technique. Studies of ammonia and pyridine adsorbed on solid acid surfaces by infrared spectroscopy are useful to distinguish between BrØnsted and Lewis acid sites and to independently evaluate the amounts of the two different acid sites. Ammonia is one of the widely used probes for the characterization of the acid properties of solid acids. However, when ammonia is used as a detector for surface acidity, several complications may arise. Ammonia is such a strong base (pKa ~5) that it will react with extremely weak acid site. Therefore, it is not a very specific probe molecule which provides ideal properties for studies of surface acidity. Moreover, 37 dissociation of adsorbed ammonia to form NH2 and NH surface groups has been observed at temperature 500 K on partially dehydroxylated oxide surfaces.[5, 6] While pyridine is a relatively strong base (pKa ~9), it is significantly weaker than ammonia. The pyridine molecule can be coordinately bonded to aprotic sites, and it can also be protonated to form the pyridinium ion PyH+ on acidic OH groups or form hydrogen bonded pyridine with less acidic groups. The infrared spectra of pyridine chemisorbed on various solid in the region 1700 cm-1 to 1400 cm-1 contain clearly distinct characteristic bands of pyridinium ion PyH+ and hydrogen bonded pyridine,[7] so the corresponding types of surface acid sites can be easily distinguished. Infrared spectra measured for pyridine adsorption on solid acids in the region 17001400 cm-1 are presented in Table 2.1. Pyridinium ion from BrØnsted acid sites gave characteristic IR absorption bands at about 1485-1500 cm-1, 1540 cm-1, 1620 cm-1 and 1640 cm-1. The characteristic bands of pyridine coordinately bonded to Lewis acid sites appeared at 1447-1460 cm-1, 1448-1503 cm-1, ~1580 cm-1 and 1600-1633 cm-1. Hydrogen bonded pyridine showed adsorption bands around 1440-1470 cm-1, 14851490 cm-1, and 1580-1600 cm-1. The sample was pressed into an 8-10 mg thin self-supporting disk of 13 mm diameter. It was placed in a Pyrex cell with NaCl windows and degassed at 300°C for 2 h under vacuum. After cooling down to room temperature, an IR spectrum of the pellet was measured and recorded as the background. The sample was then dosed with pyridine at 22 mbar for 15 minutes and evacuated at room temperature for 1 h before the first pyridine adsorption IR spectrum was measured. The spectrum of the pyridine adsorbed on the sample is obtained by subtracting the pyridine adsorption IR 38 spectrum with the background. Subsequently, the sample was heated in vacuum at 100°C and 200°C for 1 h at each temperature, and cooled to room temperature for measuring the IR spectra. A Bio-rad FTS 165 FT-IR spectrometer was used for recording the spectra at a resolution of 2 cm-1 and 64 scans were averaged. Table 2.1 Infrared bands of pyridine adsorbed on solid acid catalysts in the 17001400 cm-1 regiona[7] H-bonded pyridine Coordinately bonded Pyridinium ion pyridine 1400-1447 (v.s.) 1447-1460 (v.s.) 1485-1490 (w) 1488-1503 (v) 1485-1500 (v.s.) 1580-1600 (s) ~1580 (v) 1540 (s) 1600-1633 (s) ~1620 (s) ~1640 (s) a Band intensities: v.s.―very strong; s―strong; m―medium; w―weak; v―variable 2.4 Inductively Coupled Plasma Atomic Emission Spectroscopy Inductively coupled plasma atomic emission spectroscopy (ICP-AES) is an analytical technique used for the detection of trace metals. It is composed of two parts. The inductively coupled plasma produce excited atoms and ions and the electromagnetic radiation emitted gives an emission spectroscopy at wavelengths of particular elements. The intensity of the emission is corresponding to the concentration of the element in the sample. 39 ICP-AES has a lot of strengths (Table 2.2). One of the most important features is that ICP-AES produces an extremely line-rich spectrum, and therefore spectral lines selected can be ideally free from spectral overlap with those from other species. Table 2.2 Strengths and weaknesses of ICP-AES[8] Strengthes Weaknesses ● 108 counts s-1 per atom ● Rich spectrum choice of spectral lines ● Large number of resolution elements ― Spectra clutter (“rich” spectra) ― Spectral interferences ● Atom, ion lines ● High background (continuum, bands) ● Straightforward spatial averaging ● Detector noise (low work function) ● Convenient diagnostics ● Cannot “see” some oxides, multiply ― Tg, Texc, ne, M(II)/M(I), MO, etc ● Simple instrumentation charged ions ● Cannot resolve most isotopes ― Alignment (visual), familiarity However, despite all the strengths, ICP-AES also has its shortcomings (Table 2.2). For example, the line-rich spectra produce tremendous spectral clutter as well. Spectral interferences thus can not be avoided and this increases the difficulty in the identification of chosen spectral lines. Futher, the desired spectral lines and also the spectral clutter are superimposed with a high background, which comes from the argon ion-electron recombination continuum. This high background compromises detection limits in ICP-AES, so the detectors used in AES are relatively inefficient and noisy compared with those in ICP-MS (mass spectrometry). Lastly, some metal oxides are unfortunately invisible to ICP-AES, and also dectection for the existence of multiply charged ions is diffcult. Most importantly, isotope-induced spectral shifts for 40 elements in the center of the Periodic Table are too small to be resolved. Isotopeanalysis capability is very limited in ICP-AES. A Thermo Jarrell Ash Duo Iris ICP-AES machine was used for ICP-AES analysis. For sample preparation, 1 – 5 mg of solid sample was weighed using microbalance and dissolved with two drops of 40% HF. The prepared solution was diluted with ultrapure water and finally neutralized with excess boric acid. 2.5 Thermal Analysis The analytical experimental techniques used to investigate the behavior of a sample during heating or cooling are named as thermal analysis (TA). Conventional TA techniques include differential scanning calorimetry (DSC), differential thermal analysis (DTA), thermogravimetry (TG), thermomechanical analysis (TMA), and dynamic mechanical analysis (DMA) etc. 2.5.1 Thermogravimetric Analysis Thermogravimetric analysis (TG) is an experimental technique in which the weight or the mass of a sample is measured as a function of temperature. In a typical TGA process, the sample is heated at a constant heating rate (dynamic measurement) or held at a certain temperature (isothermal measurement). However, in some other cases, such as in sample controlled TGA (SCTA) experiments, non-linear temperature programs are used. Based on different information required about the sample, the temperature programs used are of different types. Additionally, the atmosphere in the TGA experiment is crucial to the results. Reactive, oxidizing or inert atmosphere can 41 be used, and in certain experiments changing in the atmosphere during a measurement is also possible if necessary. A TGA curve is used to represent the results of a TGA measurement. In a TGA curve, usually mass or percent mass is plotted as y-axis and temperature is plotted as x-axis. Mass changes occur when the sample loses material in several different ways or reacts with the surrounding atmosphere. Steps in the TGA curve are created according to the mass change. The different effects that causes a sample to lose, or even gain mass include the following:[9] (1) Evaporation of volatile constituents; drying; desorption and adsorption of gases, moisture and other volatile substances; loss of water of crystallization. (2) Oxidation of metals in air or oxygen. (3) Oxidative decomposition of organic substances in air or oxygen. (4) Thermal decomposition in an inert atmosphere with the formation of gaseous products. (5) Heterogeneous chemical reactions in which a starting material is taken up from the atmosphere or reactions in which a product is evolved. (6) Ferromagnetic materials. (7) Uptake or loss of water in a humidity controlled experiment. 2.5.2 Simultaneous TGA/DTA At the same time with the actual thermogravimetric measurement, the DTA signal can be recorded with the equipped thermobalances. The DTA curve shows the energetic nature during weight loss as well as those thermal effects that are not accompanied by a change in mass, for example, melting, crystallization or a glass 42 transition. Endothermic DTA peaks usually correspond to the weight loss processes, while other processes can produce exothermic peaks. These peaks are often used to determine the onset temperatures of particular thermal effects. A TA SDT 2960 Double Beam Simultaneous DTA-TGA was used in this study. The samples were kept isothermally at 100°C for 30 min to remove any physically adsorbed water before increasing the temperature. The thermal analysis was then carried out by heating the samples to 900°C with a heating ramp 20 °C/min. The atmosphere used is flowing air at velocity 100 mL/min. 2.6 Determine Acidity Using Amine Titration Method with Indicators 2.6.1 Strength and Amount of Solid Acid The acid strength of a solid is defined as the ability to convert a neutral base absorbed on the surface into its conjugate acid form. When proton transfer occurs from the surface to the adsorbate, the acid strength is expressed by the Hammett acidity function H0 (equation 2.7):[10] H0 = pKa + log [B] [BH + ] (2.7) where [B] and [BH+] are, respectively, the concentrations of the neutral base (basic indicator) and its conjugate acid, and pKa equals to pK BH + . In the other case, instead of proton transfer, an electron pair is transferred, from the adsobate to the surface. H0 is expressed by (equation 2.8):[10] 43 H0 = pKa + log [B] [AB] (2.8) where [AB] is the concentration of the neutral base which reacted with the Lewis acid or electron pair acceptor, A. The amount of acid on a solid is usually expressed as the number or mmol of acid sites per unit weight or per unit surface area of the solid. The amount of a base used to react with the solid acid is measured as the amount of acidity. For the determination of strength and amount of a solid acid, two main methods are generally applied: an amine titration method using indicators and a gaseous base adsorption method. 2.6.2 Determine the Acid Strength using Hammett Indicator The acid strength of a solid can be determined from the color of suitable indicators adsorbed on its surface. If the color corresponds to the acid form of the indicator, the value of the H0 function of the solid surface is equal to or lower than the pKa of the conjugate acid of the indicator. Great acid strength is indicated by the low values of H0. Thus, for indicators undergoing color changes in this way, the lower the pKa, the greater the acid strength of the solid. The indicators used for the determination are listed in Table 2.3. 44 Table 2.3 Basic indicators used for measurement of acid strength[11] Indicators Color pKaa [H2SO4]b/% Base-form Acid-form Neutral red yellow red + 6.8 8 × 10-8 Methyl red yellow red + 4.8 ― Phenylazonaphthylamine yellow red + 4.0 5 × 10-5 p-Dimethylaminoazobenzene yellow red + 3.3 3 × 10-4 2-Amino-5-azotoluene yellow red + 2.0 5 × 10-3 Benzeneazodiphenylamine yellow purple + 1.5 2 × 10-2 Crystal violet blue yellow + 0.8 0.1 p-Nitrobenzeneazo- orange purple + 0.43 ― Dicinnamalacetone yellow red - 3.0 48 Benzalacetophenone colorless yellow - 5.6 71 Anthraquinone colorless yellow - 8.2 90 2,4,6-Trinitroaniline colorless yellow - 10.10 98 p-Nitrotoluene colorless yellow - 11.35 *c m-Nitrotoluene colorless yellow - 11.99 * p-Nitrofluorobenzene colorless yellow - 12.44 * p-Nitrochlorobenzene colorless yellow - 12.70 * m-Nitrochlorobenzene colorless yellow - 13.16 * 2,4-Dinitrotoluene colorless yellow - 13.75 * 2,4-Dinitrofluorobenzne colorless yellow - 14.52 * 1,3,5-Trinitrotoluene colorless yellow - 16.04 * (p’-nitro-diphenylamine) a b c pKa of the conjugate acid of indicator wt. percentage of H2SO4 in sulfuric acid solution which has the acid strength corresponding to the respective pKa The indicator is liquid at room temperature: the acid strength corresponding to the indicator is higher than the acid strength of 100 percent H2SO4 The acid strength of a solid superacid is very sensitive to moisture. Thus, to determine the acid strength, the vapor of an indicator is adsorbed on a solid sample through a breakable seal in a vacuum system at room temperature. The acid strength is determined from the color change of the indicator.[12] 45 2.6.3 Determine the Number of Acid Sites by Amine Titration The number of acid sites on a solid surface can be measured by amine titration following determination of acid strength by the above method. This method is performed by titrating a solid acid suspended in benzene with n-butylamine using a selected indicator. The number of acid sites at various acid strengths can be determined by amine titration using various indicators with different pKa values. Both proton donors and electron pair acceptors on the surface will react with either the electron pair of the indicator or that of the amine to form a coordination bond. Thus, the amine titration method gives only the total amounts of both BrØnsted and Lewis acid. This method is not suitable for colored or dark samples because it is difficult to observe the usual color change. However, several other methods can be applied for this case. By mixing the colored sample with a white substance of known acidity, the number of acid sites of the colored sample is calculated by subtracting the known acidity of the white substance from the total amount of acidity determined from the titration results. The spectrophotometric method and calorimetric titration of a solid acid with amine have also been used for the estimation of the acid amount of a colored or dark sample.[13-15] Recently, Hashimoto et al.[16] developed a method to measure the acid strength and distribution on a solid surface. By utilizing the chemisorption isotherms of a series of Hammett indicators on a solid suspended in a nonpolar solvent such as benzene or cyclohexane, the fraction of acid sites covered by the indicator was found and expressed by a Langmuir type equation involving both acid strength and the indicator concentration. A cumulative distribution curve of acid strength can be derived from the chemisorption isotherms of Hammett indicators. 46 Samples were freshly dried at 100°C in test tubes before carrying out the indicator tests, and were subjected to color tests immediately after drying. Removal of adsorbed water is necessary as the effect of water adsorption changed the color intensity of the adsorbed indicators or caused a shift to lower acid strengths. Color tests were carried out by transferring a few milligram of dried, powdered solid to a watch glass, and adding a few drops of 0.1% solution of indicator. From the results of the tests, it was decided whether the solid was basic or acidic to all indicators, or had an H0 value lying between two adjacent indicators’ pKa values. The amount of acid sites on the samples was measured by amine titration immediately after determination of acid strength by above method. The 0.04N n-butylamine solution was prepared by dissolving 0.1 mL of n-butylamine in a 25 mL volumetric flask and making up the volume using dried benzene. 0.05 g solid sample was weighed and transferred to a 50 mL flask containing 5 mL benzene. Three drops of the 0.1% methyl red solution was added to the sample suspension. The titration was conducted using a burette with small steps. 2.7 Temperature Programmed Desorption of Ammonia Gaseous bases adsorbed on acid sites give information about the acidity of the solid surface. A base adsorbed on strong acid site is more stable and more difficult to desorb than one adsorbed on a weak acid site. As temperature is increased, evacuation of the adsorbed bases from acid sites occurs. Bases adsorbed at weaker acid sites will be evacuated at lower temperature than those at stronger acid sites. Thus, the evacuation temperatures indicate the strength of various acid sites. 47 The amount of a gaseous base adsorbed chemically by a solid is proportional to the number of acid sites on the solid surface. Ammonia, n-butylamine, and pyridine are gaseous bases commonly used for the determination of strength and amount of a solid acid. The advantages of the gaseous base adsorption and desorption method are: (1) It can be used to determine the amount of acid sites for a solid at high temperature or under its actual working conditions. (2) It can also be applied even to colored samples. However, the disadvantage is the BrØnsted and Lewis acid sites can not be differentiated by this method. Temperature programmed desorption is one of the most useful characterization methods in the study of heterogeneous catalysts and was developed in 1963 by Canadian researchers Amenomiya and Cvetanovic.[17] Compared with other analytical techniques, TPD provides information which is closely related to the catalytic properties and the reactions in real work. In a TPD measurement, a sample is first heated in a vacuum or helium atmosphere to remove contamination. Then, a gaseous base is introduced to be adsorbed by the sample. With a programmed temperature rise, desorption of the adsorbed gas starts and the amount of gas desorbed is continuously monitored by a mass analyzer. The gaseous bases adsorbed at acid sites with different acid strength gain enough energy to escape at different temperatures. The desorbed gas is carried by helium and detected by the mass spectrometer. The temperature of the peak maximum provides information on the binding energy of the bound species. A home-built TPD system was used in this study. For pretreatment, the sample was first heated at 500°C under the flow of helium for 2 h. After preheating, the 48 sample was cooled to room temperature before introducing ammonia. After flushing with ammonia for 15 minutes, helium was passed over the sample for 2 h to flush away any physically adsorbed ammonia at 150°C. The sample was then heated at a constant heating rate of 10 °C/min and the ammonia desorbed over a temperature range of 150°C to 500°C was measured. The amount of ammonia desorbed corresponds to the number of acid sites on the surface. Calibration of the ammonia signal was carried out by injecting 250 μL of ammonia into the system and measuring the area of signal peak. A value of 3.61 × 10-8 As/mmol was obtained for measurement conditions of pre-chamber pressure 5 mbar and high vacuum pressure of 5 × 10-7 mbar. 2.8 Catalytic Activity Test 2.8.1 Fischer Indole Synthsis of 2-ethyl-3-propyl indole and 2-butyl-3-methyl indole The Fischer indole reaction of phenylhydrazine and 3-heptanone (Scheme 2.1) was carried out in a two-necked 25 mL round-bottomed flask equipped with a condenser. 0.202 mL (2 mmol) of phenylhydrazine (Aldrich, 97%) and 0.279 mL (2 mmol) of 3-heptanone (Fluka) were mixed in 5 mL solvent, p-xylene (Fluka). 1 mmol dodecane was added as internal standard. The flask was flushed with nitrogen and the reaction mixture was heated up to 140°C with stirring under nitrogen. After 30 minutes, the complete condensation of phenylhydrazine with 3-heptanone to form phenylhydrazone was confirmed by gas chromatography (Fig. 2.3). Then 0.25 g of the catalyst was introduced to catalyze the [3,3]-sigmatropic rearrangement followed by intramolecular cyclisation. Aliquots were removed at regular time intervals and analyzed by GC (Fig. 2.4). 49 H3CHC C2H5 NH C4H9 C2H5 O C4H9 + C4H9 NH 2 N H 4 C4H9 N NH 1 C2H5 NHNH2 CHC3H7 NH NH 3 C3H7 N H 5 C2H5 Scheme 2.1 Fischer indole reaction of phenylhydrazine with 3-heptanone Fig. 2.3 Gas chromatogram after formation of E/Z phenylhydrazone It has long been known that phenylhydrazones, although more stable than the aliphatic hydrazines, are quite sensitive to air oxidation, usually taking on a yellow to red colouration due to partial or complete conversion to the appropriate azo derivative.[18] In this study, the formation of phenylhydrazones and the synthesis of indole products were protected by inert nitrogen gas. Oxidation occurred very fast when phenylhydrazones were exposed to air at the high reaction temperature of 50 140°C, accompanied by an obviously yellow to red colour change. It is therefore necessary to protect the synthesis from air and other oxidizing compounds. Also, aliquots removed from the reaction vessel should be analyzed immediately using GC to minimize the possible oxidation of phenylhydrazones. Fig. 2.4 Gas chromatogram of reaction mixture of Fischer indole reaction of phenylhydrazine with 3-heptaone 2.8.2 Analysis of Products by Gas Chromatography and GC-MS Gas chromatography (GC) is a technique for separating and detecting volatile substances.[19] This separation technique is basically done by percolating a mixture gas substances over a stationary phase and depends upon the adsorptive properties of the column packing. Nowadays, the capillary column is used mostly in gas chromatography. It consists of long tube that can be made of metal, glass or quartz. The outside of the tube is coated with a polymer resin to provide mechanical strength and prevent atmospheric erosion. The external coating can remain stable up to about 250°C. The stationary phase is coated as a thin film on the internal surface of the tube. 51 The most important characteristic of the capillary column is its small flow impedance relative to that of the packed column. This permits very short columns operated at very high mobile phase velocities to provide fast separations. The column is the heart of any chromatograph. When a mixture gas passes though the column, the individual components in the mixture are separated from each other due to their different moving speed, and finally detected as individual sample bands.[20] As the sample components move through the column, they will dissolve in the stationery phase without any movement. The mobile phase (carrier gas) flows continuously through the column to elute the individual components of the sample. Components with a low solubility move faster than components with a high solubility. Therefore, individual components are separated. Nitrogen, helium, and hydrogen are the most common carrier gases. The components in the sample will equilibrate between the moving gas phase and the stationary liquid phase to form so-called theoretical plates.[19] An Agilent GC 6890 equipped with a HP-5 capillary column and an FID detector was used to analyze the samples. The temperature program used for separation of the products was as follow: Initial temperature: 80°C Holding time: 2 min Rate 1: 10 °C/min Temperature 1: 230 °C Rate 2: 20 °C/min Temperature 2: 300 °C The identity of the products was verified by comparison with the retention time of authentic samples and by GC–MS (Shimadzu GCMS QP5000, DB5MS column). 52 The conversion and selectivity of indole products are calculated by the following equations: %conversion = ([P1 ] + [P2 ]) / 12.15 ([P1 ] + [P2 ]) / 12.15 + ([R 1 ] + [R 2 ]) / 11.4 × 100 (2.9) %selectivity 1 = [P1 ] × 100 [P1 ] + [P2 ] (2.10) %selectivity 2 = [P2 ] × 100 [P1 ] + [P2 ] (2.11) where [P1], [P2] represent the peak areas from gas chromatogram for linear and bulky indole products respectively; [R1], [R2] represent the peak areas for corresponding E/Z phenylhydrazones; ECN for indole products is 12.15, and for phenylhydrazone reactants is 11.4. 2.8.3 Fischer Indole Synthesis of 1,2,3,4-tetrahydrocarbazole The Fischer indole reaction of phenylhydrazine with cyclohexanone to 1,2,3,4-hydrocarbazole was carried out in a similar way using 2mmol cyclohexanone instead of 3-heptanone (Fig. 2.5). Only a single hydrazone intermediate was formed, and therefore 1,2,3,4-tetrahydrocarbazole was the only product (Scheme 2.2). 53 NHNH2 + O N H N N H 8 Scheme 2.2 Fischer indole synthesis of 1,2,3,4-tetrahydrocarbazole Fig. 2.5 Gas chromatogram of reaction mixture of Fischer indole synthesis of 1,2,3,4-tetrahydrocarbazole 54 References [1] B. D. Cullity, in Elements of X-ray diffraction (Ed. B.D. Cullity) 1956 (Addison-Wesley Pub. Co.) [2] M. K. Ismail, Langmuir 1992, 8, 360 [3] S. J. Gregg, K. S. W. Sing, in Asdorption, Surface Area and Porosity (Eds. S. J. Gregg, K. S. W. Sing) 1967 (Academic Press: London, New York) [4] A. J. Lecloux, in Catalysis, Science and Technology (Eds. J. R. Aderson, M. Boudart) 1984, vol. 5 (Springer-Verlag: Berlin, Heidelber, New York) [5] J. B. Peri, J. Phys. Chem. 1965, 69, 211 [6] P. Fink, J. Datka, J. Chem. Soc. Faraday Trans. 1 1989, 85, 309 [7] E. P. Parry, J. Cata. 1963, 2, 371 [8] S. J. Hill, in Inductively Coupled Plasma Spectrometry and its Applications, second edition (Ed. S. J. Hill) 2007 (Blackwell Pub.) [9] P. Gabbott, in Principles and Applications of Thermal Analysis (Ed. P. Gabbott) 2008 (Blackwell Pub.) [10] L. P. Hammett, A. J. Deyrup, J. Am. Chem. Soc. 1932, 54, 2721 [11] K. Tanabe, in Catalysis: Science and Technology (Eds. J. R. Anderson, M. Boudart) 1981, vol. 2 (Springer-Verlag: Berlin, Heidelber, New York) [12] K. Tanabe, H. Hattori, Chem. Lett. 1976, 625 [13] L. Forni, Advan. Catal. 1974, 8, 65 [14] K. Tanabe, T. Yamaguchi, J. Res. Inst. Catal. 1966, 14, 93 [15] S. P. Walvekar, A. B. Halgeri, S. Ramanna, T. N. Srinivasan, Fertilizer Tech. 1976, 13, 241 [16] K. Hashimoto, T. Masuda, H. Motoyama, H. Yakushiji, M. Ono, I & EC Prod. Res. & Develop. 1986, 25, 243 55 [17] Y. Amernomiya, R. J. Cventanovic, J. Phys. Chem. 1963, 63, 144 [18] S. Patai; in The chemistry of the hydrazo, azo and azoxy groups (Ed. S. Patai) 1975 (Wiley: New York) [19] H. M. McNair, in Basic Gas Chromatography (Ed. H. M. McNair) 1988 (American Chemical Society: Washington, D. C.) [20] R. P. W. Scott, in Introduction to Analytical Gas Chromatography, Second Edition, Revised and Expanded (Ed. R. P. W. Scott) 1998 (Marcel Dekker, Inc.) 56 Chapter Ш Study of Hydrous Zirconia Supported Niobium Oxide Catalysts 3.1 Introduction Niobium materials have been receiving increasing interest in heterogeneous catalysis where they play a role as catalyst components or promoter species. Many niobium containing compounds have been studied for different purposes. Niobium oxides go through three steps of phase transitions on heating, changing from amorphous to crystalline T form at ca. 773 K, T to M form at ca. 1103 K, and M to H form at around 1473 K.[1,2] In between these temperatures, some other phases, for example, less crystalline TT form can be found as well.[3] Niobium oxide-based compounds generally comprise of octahedrally coordinated NbO6 structure. Based on whether the structure is corner- or edge-shared, the octahedral NbO6 is distorted to different extents.[4] The highly distorted NbO6 structure contains Nb=O bonds which are responsible for Lewis acid sites.[5] In contrast, the slightly distorted NbO6 structure together with the NbO7 and NbO8 groups only possess Nb-O bonds and are responsible for Brønsted acid sites. Lewis acid sites are present in all supported niobium oxide systems, but only in the Nb2O5/Al2O3 and Nb2O5/SiO2 systems, Brønsted acid sites are generated.[6] Hydrated niobium pentoxide (Nb2O5·nH2O), also known as niobic acid, has a high acid strength with H0 ≤ -5.6.[7] The acidity decreases when heated above 300°C and becomes almost neutral after calcination at 500°C. The presence of water is essential for the acidity in niobium oxide, making it a useful catalyst in reactions where there is water participation or liberation.[8] 57 On phosphating niobium oxide with H3PO4, niobium phosphate is formed.[9] Commercial amorphous niobium phosphate shows strong surface acidity that is retained at temperatures higher than that for niobium oxide. Like niobium oxide, niobium phosphate also promotes some acid-catalyzed reactions.[10] Thus, the acidic and catalytic properties of niobium phosphate are similar to that of niobium oxide, but the advantage is these properties are conserved at higher pretreatment temperatures.[11] Various functions of niobium-containing compounds are applied in heterogeneous catalysis. As a promoter or active phase, a small amount of niobia added enhance the catalytic activity and the selectivity of the known catalysts and prolong catalyst life.[10,12] Three factors have been found to affect the surface properties of supported niobia catalysts. Firstly, a study of support materials, selected from silica, alumina, magnesia, titania, zirconia, and zeolites showed that Lewis acidity is found in all of the supported niobium oxide systems, while Brønsted acid sites are only detected in niobia supported on alumina and silica. The choice of niobium precursors is another crucial factor. Potential niobium precursors include niobium chloride, niobium oxalate, niobium ethoxide, niobium acid, niobates, and niobium complexes with hydrocarbon ligands. The niobium oxide species are usually deposited at the outer surface of the catalysts as an overlayer. The formation and location of these surface niobium oxide species depend on the surface hydroxyl properties of the catalysts. Preparation methods affect the promotion effect of surface niobium oxide species as well. A simple impregnation method helps in a uniform dispersion of niobium oxides on the supports. 58 Niobium oxide has also been used as an oxide support for metals such as Ru, Rh, Pt, Re, Ni, Cr, W, Co, V, P, Ge, Mo, Sb, Pb, Bi and Fe. It exhibits a pronounced effect when used in this manner.[10,12] The properties of the niobia are promoted by the surface interaction with other metals, while its high selectivity is still maintained. For example, Jehng et al.[13] studied the properties of various surface modified niobium oxide catalysts. Metal oxides (Re2O7, CrO3, WO3, MoO3, and V2O5) were loaded on the niobium oxide surface by the incipient-wetness impregnation method. Acidity and BET studies revealed that the presence of the surface metal oxide species retained the surface acidic properties and surface area of niobium oxide during high calcination temperature (500 °C). At low metal oxide loadings (less than 10 wt.% ), the surface areas of the resultant supported metal oxide catalysts has the following order: V2O5/Nb2O5 > CrO3/Nb2O5 > WO3/Nb2O5 > MoO3/Nb2O5 > ReO3/Nb2O5 > Nb2O5. For 1% V2O5/Nb2O5, the surface area was very large, ca. 100 m2/g. Also, the acidity of 1% V2O5/Nb2O5 was similar to that of hydrous niobium oxide. Acidic properties of surface modified niobium oxide catalysts studied by pyridine absorption infrared spectroscopy were shown in Table 3.1. In the catalysis studies of methanol oxidation, 1% WO3/Nb2O5 exhibited selectivity towards CH3OCH3 similar to that of pure niobium oxide, but the catalytic activity of 1% WO3/Nb2O5 was doubled. Niobium compounds are also used directly as solid acid catalysts. A typical example is niobic acid. As previously stated, niobic acid has strong acid strength corresponding to 70% sulfuric acid. The existence of water molecular in the catalyst provides both BrØnsted and Lewis acid sites. Therefore, it can be used in many acidcatalyzed reactions such as dehydration, esterification, and hydrolysis where water molecules involve.[14-18] 59 Table 3.1 Acidic properties of surface modified niobium oxide catalystsa Amount of LASb Amount of BASb (μmol/g) (μmol/g) Niobium oxide (act. 200°C) 152 57 Niobium oxide (cal. 500°C) 0 0 1% Re2O7/ Nb2O5 24 8 1% CrO3/ Nb2O5 39 5 1% WO3/ Nb2O5 56 26 1% MoO3/ Nb2O5 12 0 1% V2O5/ Nb2O5 106 46 Catalyst a b all surface modified niobium oxide catalysts were calcined at 500°C LAS: Lewis acid site; BAS: BrØnsted acid site In this work, we explore the use of supported niobium oxide as possible catalyst in the Fischer indole synthesis. The molecular structures and reactivity of twodimensional niobium oxide overlayers on zirconia have been well investigated.[19-21] Surface hydroxyl groups play an important role in the loading and alumina, titania and zirconia tend to form a close-packed monolayer of surface niobium oxide. In contrast, the lower surface hydroxyl density on silica does not allow the formation of a closepacked monolayer. Hence, use of hydrous zirconia and silica as supports allows a comparison of the activity of the supported niobium oxide. 3.2 Preparation of Catalysts 3.2.1 Preparation of Hydrous Zirconia Support 20 g of ZrCl4 powder (Merck) was dissolved in 200 mL deionized water to make a 10% zirconium chloride solution. The acidic zirconium chloride solution was transferred via a peristaltic pump into excess 5 M ammonium hydroxide solution 60 under vigorous stirring using a mechanic stirrer. The pH of the mixture was adjusted to 9.5 using 5 M ammonium hydroxide. The gel solution was then transferred into a 400 mL Teflon round bottom flask and digested for 24 h at 100°C. The solid product was collected by filtering, washed with dilute ammonia nitrate solution until free of chloride, rinsed with deionized water and dried at 100°C overnight. 3.2.2 Wet Impregnation of Niobium Oxide on Hydrous Zirconia Support To synthesize 2 g of 10 wt.% Nb2O5/Zr(OH)4, 0.41 g of NbCl5 (Riedel-deHaen) was weighed and dissolved in 10 mL of ethanol. A solvolysis reaction occurs when NbCl5 and an alcohol are mixed together leading to the formation of a trichloroalkoxide.[22,23] NbCl5 + ROH(excess) Æ NbCl2 (OR)3 + 3HC1 About 1.8 g of hydrous zirconia support was added to the clear niobium chloride solution and the mixture was stirred. In order to precipitate the niobium oxide, 5 M ammonium hydroxide solution was added dropwise to the solution. The completion of precipitation was attained when no more gaseous ammonium chloride was evolved. Finally, the pH of the gel solution was adjusted to 10 using 5 M ammonium hydroxide. The gel solution was transferred into a 150 mL Teflon round bottom flask and digested for 24 h at 100°C. The product was filtered, washed and dried as for hydrous zirconia. Samples with various loadings of niobium oxide (20 wt.%, 25 wt.%, 30 wt.%, 40 wt.%) were synthesized in a similar manner (Table 3.2). To investigate the effect of drying temperature on the textural properties and catalytic activities of 61 the catalysts, all samples were dried at three different temperatures, 100°C, 200°C and 300°C. Table 3.2 Amount of reagents used in preparing the Nb2O5/Zr(OH)4 catalysts NbCl5 (g) Sample Name Zr(OH)4 (g) 10% Nb2O5/Zr(OH)4 1.8 0.41 20% Nb2O5/Zr(OH)4 1.6 0.81 25% Nb2O5/Zr(OH)4 1.5 1.02 30% Nb2O5/Zr(OH)4 1.4 1.22 40% Nb2O5/Zr(OH)4 1.2 1.63 3.2.3 Synthesis of 25 wt.% Nb2O5/Zr(OH)4 by Coprecipitation Method To prepare 2 g of catalyst, 1.02 g of NbCl5 (Riedel-de-Haen) and 2.84 g of ZrCl4 (Merck) were first dissolved in 10 mL of ethanol and 30 mL of deionized water separately. The two precursor solutions were then mixed to form an acidic mixture solution with pH ≈ 1. Then 5 M ammonium hydroxide solution was used to precipitate the zirconium and niobium cations in a similar way as described previously for the preparation of hydrous zirconia. The final pH was adjusted to 10 using 5 M ammonium hydroxide. The gel solution was transferred into a 150 mL Teflon round bottom flask and digested for 24 h at 100°C. The product was filtered, washed and dried as for hydrous zirconia. 3.2.4 Synthesis of 25 wt.% Nb2O5/MCM-41 by Wet Impregnation To compare the effect of the support, 25 wt.% Nb2O5/MCM-41 was synthesized following the procedure above. The mesoporous material MCM-41 was synthesized following the method by Cheng et al.[24] 62 TEAOH and CTABr were added to deionized water with stirring at 30°C until the solution became clear. Fumed silica was added and the solution was stirred for 2 h before allowing the gel to age for 24 h at room temperature. The molar composition of final gel mixture was 1.0 SiO2: 0.19 TEAOH: 0.27 CTABr: 40 H2O. The mixture was reacted for 48 h at 150°C in a Teflon-lined stainless-steel autoclave. The solid product was filtered, washed with deionized water, dried in air at 100°C overnight and finally calcined at 550°C for 8 h. 3.2.5 Recycling of Catalysts Three methods were used to regenerate the used 25 wt.% Nb2O5/Zr(OH)4 catalyst. Method 1: After the reaction, the catalyst was collected by centrifugation, washed with the solvent p-xylene for several times, and then dried in a 100°C oven overnight. Method 2: The catalyst was first washed with solvent, dried, and then immersed in 30% H2O2 solution with vigorously stirring. The color of the catalyst finally changed to pale yellow and the catalyst was again collected by centrifugation of the suspension. The dried catalyst was suspended in 50 mL of 5 M ammonia solution in a Teflon round bottom flask and digested for 24 h to regenerate the hydroxyl groups. After rehydration, the catalyst was filtered, washed with deionized water and dried before use. Method 3: As in method 2, the catalyst was washed with solvent and dried before immersion in 30% H2O2 solution. The mixture solution was stirred and heated 63 up to 40°C. The catalyst recovered its white color after a few hours and it was filtered and dried overnight. 3.3 Catalyst Characterization 3.3.1 Powder X-ray Diffraction Fig. 3.1 shows the X-ray diffractograms of the hydrous zirconia support and the Nb2O5/Zr(OH)4 catalysts after drying at 100°C, 200°C and 300°C. The hydrous zirconia support was X-ray amorphous after drying at 300°C. This result is expected as the reported transformation from the amorphous hydrous zirconia to crystalline zirconium oxide usually occurs after calcination above 450°C.[25] None of the Nb2O5/Zr(OH)4 samples show any observable diffraction peak indicating the lack of crystallinity in the niobium oxide overlayer. This may be due to the fact that the transition of the amorphous niobium oxide to the crystalline γ-form occurs only at 435°C.[4] The X-ray diffractogram of the MCM-41 support shows one intense peak at 2θ ~ 1.9°, as well as four weaker peaks at 3.3°, 3.8°, 5.1° and 5.8° (Fig. 3.2). These reflexes correspond to the diffraction planes (100), (110), (200), (210) and (300) respectively. The presence of these reflexes indicates that the MCM-41 formed has a very well-ordered hexagonal structure.[24] However, after impregnation with 25 wt.% niobium oxide, all the diffraction peaks could no longer be seen. The overall peak intensity is also decreased, which may be partly due to (i) a decrease in the diffraction contrast of the lattice when the pores are filled with niobium oxide, as well as (ii) a destruction of the well-structured pores as a result of impregnation. 64 Intensity (cps) 1600 1400 25% Nb2O5/Zr(OH)4 Coppt 1200 40% Nb2O5/Zr(OH)4 1000 30% Nb2O5/Zr(OH)4 800 25% Nb2O5/Zr(OH)4 600 20% Nb2O5/Zr(OH)4 400 10% Nb2O5/Zr(OH)4 Zirconia support 200 0 0 10 20 30 40 50 60 70 2θ (°) (a) 1400 Intensity (cps) 1200 40% Nb2O5/Zr(OH)4 1000 30% Nb2O5/Zr(OH)4 800 25% Nb2O5/Zr(OH)4 600 20% Nb2O5/Zr(OH)4 400 10% Nb2O5/Zr(OH)4 Zirconia support 200 0 0 10 20 30 40 50 60 70 2θ (°) (b) 1400 Intensity (cps) 1200 40% Nb2O5/Zr(OH)4 1000 30% Nb2O5/Zr(OH)4 800 25% Nb2O5/Zr(OH)4 600 20% Nb2O5/Zr(OH)4 10% Nb2O5/Zr(OH)4 400 Zirconia support 200 0 0 10 20 30 40 50 60 70 2θ (°) (c) Fig. 3.1 X-ray diffractograms of Nb2O5/Zr(OH)4 catalysts after drying at (a) 100°C (b) 200°C (c) 300°C 65 8000 100 7000 Intensity (cps) 6000 5000 4000 110 200 3000 210 2000 25% Nb2O5/MCM-41 300 1000 MCM-41 support 0 1.6 3.6 5.6 7.6 9.6 2θ (°) Fig. 3.2 X-ray diffractograms of MCM-41 and 25 wt.% Nb2O5/MCM-41 3.3.2 Textural Properties The hydrous zirconia dried at 100°C had the highest surface area of 377 m2/g (Table 3.3), but the surface area of niobium oxide was measured to be only 4.58 m2/g. After impregnation with niobium oxide, the surface area of the catalysts decreased to 171 - 314 m2/g. This is mostly likely caused by a reduction in the porosity of the hydrous zirconia support. When the drying temperature was increased to 200°C and 300°C, the surface area further decreased to 155 - 212 m2/g. The retention of the surface area despite the higher temperature suggests that the hydrous amorphous nature of the catalysts is still present. In comparison, a 25 wt.% Nb2O5-Zr(OH)4 sample prepared by coprecipitation method had a lower surface area, 163 m2/g, than that prepared by impregnation, 201 m2/g. The nature of the support also plays a role as supporting niobium oxide onto the mesoporous siliceous material MCM-41 results a drastic decrease of surface area. The surface area of the MCM-41 support dropped from 1009 m2/g to only 240 m2/g in 25 wt.% Nb2O5/MCM-41. 66 From nitrogen porosimetry, all the Nb2O5/Zr(OH)4 catalysts were found to have a wide range of pores ranging from 2.9 - 10 nm, with a bimodal distribution of pores centred around 3.3 nm and 5.6 nm (Fig. 3.3). Hydrous zirconia dried at 100°C had the largest pore volume, 0.47 cm3/g. After impregnation with niobium oxide, the pore volume decreased and the pore size distribution became narrower as shown in Fig. 3.5a. This suggests that the niobium oxide is uniformly deposited on the surface of hydrous zirconia support with little pore blockage. The 25 wt.% Nb2O5/Zr(OH)4 sample was prepared following activity studies, and a different batch of hydrous zirconia was used as support. As with the other samples, the pore size distribution of the supported sample is similar to that of the support, suggesting an even distribution of niobium oxide on the surface of hydrous zirconia. In contrast, for the MCM-41 supported samples, the mesopores centred around 2.9 nm in MCM-41 no longer existed after impregnation with niobium oxide indicating the blockage of the pores by the promoter (Fig. 3.5b). Instead, the remaining pores are shifted to bigger pore diameters, centred at ~ 3.5 nm. These results show that the sizes of niobium oxide clusters are at least 2.9 nm in diameter so that the bulk of the mesopores in MCM-41is no longer accessible. The loadings of niobium oxide in the catalysts were determined by inductively coupled plasma-atomic emission spectroscopy (ICP-AES). The results show that the actual niobium oxide loadings were slightly higher than the expected value for all samples (Table 3.4). For the MCM-41 supported sample, the actual and expected niobium oxide contents were very close, about 25%. Based on the crystalline structure of Nb2O5, the cross-sectional area of a Nb2O5 unit[21] is estimated to be 67 Table 3.3 Textural properties of Nb2O5/Zr(OH)4 catalysts Sample Drying SBET (m2/g) Temperature (°C) 10% Nb2O5/Zr(OH)4 Pore volume (cm3/g) 100 314 0.41 200 245 0.33 300 212 0.33 100 252 0.33 200 246 0.32 300 211 0.30 100 201 0.20 200 199 0.20 300 153 0.19 100 217 0.28 200 227 0.30 300 180 0.27 100 171 0.22 200 180 0.24 300 155 0.22 100 377 0.47 200 268 0.35 300 227 0.34 Nb2O5 ― 4.58 0.01 25% Nb2O5-Zr(OH)4 coppt 100 163 0.12 25% Nb2O5/MCM-41 100 240 0.69 MCM-41 500 1009 0.68 20% Nb2O5/Zr(OH)4 25% Nb2O5/Zr(OH)4 30% Nb2O5/Zr(OH)4 40% Nb2O5/Zr(OH)4 Zr(OH)4 support 68 0.32 nm2. Using this value, the theoretical monolayer coverage of the supported samples can be calculated as follows (equations 3.1 – 3.3). Calculation of theoretical monolayer coverage n(Nb2O5) = actual Nb2O5 content molecular weight of Nb2O5 (3.1) where n(Nb2O5) = number of moles of Nb2O5 in catalyst S(Nb2O5) = n(Nb2O5) × NA × 0.32 nm2 (3.2) where S(Nb2O5) = total surface area of Nb2O5 overlayer Monolayer coverage = S(Nb2O5) SBET of catalyst (3.3) The overlayer coverage of 10 to 40 wt.% Nb2O5/Zr(OH)4 catalysts was calculated to be between 0.22 to 2.0, where at 25% to 30% loadings, the overlayer of niobium oxide approached a monolayer. Jehng and Wachs[19,20] measured the surface coverage and surface density of niobium oxide overlayers on different oxide supports and found that supports such as Al2O3, TiO2, and ZrO2 with a high density of reactive surface hydroxyls could form a close-packed niobium oxide monolayer. In contrast, SiO2 support had a lower density of reactive surface hydroxyls and thus the surface density of niobium oxide formed was only ~ 1/20 compared with that of ZrO2 supported catalyst. Therefore, they concluded that a necessary condition for the formation of surface metal oxide overlayers is the presence of reactive surface hydroxyls on the oxide support. However, in this study the niobium oxide overlayer on the surface of MCM-41 support was formed using deposition impregnation method 69 which did not rely on the number of reactive surface hydroxyl groups. Therefore, the monolayer coverage in the 25% Nb2O5/MCM-41 was found to be 0.74. This value was much higher than that reported by Jehng and Wachs. Table 3.4 Niobium oxide loading and surface coverage Sample Drying Nb2O5 contenta temperature (wt.%) (°C) 10% Nb2O5/Zr(OH)4 20% Nb2O5/Zr(OH)4 25% Nb2O5/Zr(OH)4 30% Nb2O5/Zr(OH)4 40% Nb2O5/Zr(OH)4 25% Nb2O5/MCM-41 a b 100 11.7 Monolayerb 0.33 200 0.32 300 0.40 100 25.5 0.73 200 0.75 300 0.88 100 27.4 0.90 200 0.91 300 1.09 100 31.4 1.05 200 1.01 300 1.26 100 47.1 2.00 200 1.90 300 2.20 100 24.6 0.74 from ICP-AES measurement based on 0.32 nm2/Nb2O5 70 400 0.09 100 200 350 Volume adsorbed/desorbed (cm /g) 300 300 3 Volume adsorbed/desorbed(cm3/g) 100 200 300 0.08 250 200 150 100 50 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0 0 0.2 0.4 0.6 Relative Pressure (P/P0) 0.8 0 1 2 4 6 8 Average pore diameter (nm) 10 (a) hydrous zirconia 350 0.08 100 200 100 200 300 0.07 300 Volume adsorbed/desorbed (cm3/g) Volume adsorbed/desorbed(cm3/g) 300 250 200 150 100 50 0.06 0.05 0.04 0.03 0.02 0.01 0 0 0 0.2 0.4 0.6 Relative Pressure (P/P0) 0.8 0 1 2 4 6 8 Average pore diameter (nm) 10 (b) 10% Nb2O5/Zr(OH)4 300 0.06 100 100 200 300 200 300 0.05 Volume adsorbed/desorbed (cm3/g) Volume adsorbed/desorbed(cm3/g) 250 200 150 100 0.04 0.03 0.02 0.01 50 0 0 0 0.2 0.4 0.6 Relative Pressure (P/P 0) 0.8 1 0 2 4 6 8 Average pore diameter (nm) 10 (c) 20% Nb2O5/Zr(OH)4 71 250 0.06 100 100 200 300 200 300 0.05 Volume adsorbed/desorbed (cm3/g) Volume adsorbed/desorbed(cm3/g) 200 150 100 50 0.04 0.03 0.02 0.01 0 0 0 0.2 0.4 0.6 Relative Pressure (P/P0) 0.8 0 1 2 4 6 8 Average pore diameter (nm) 10 (d) 30% Nb2O5/Zr(OH)4 200 0.045 100 200 300 150 Volume adsorbed/desorbed (cm3/g) Volume adsorbed/desorbed(cm3/g) 100 200 300 0.04 100 50 0.035 0.03 0.025 0.02 0.015 0.01 0.005 0 0 0 0.2 0.4 0.6 Relative Pressure (P/P0) 0.8 0 1 2 4 6 8 Average pore diameter (nm) 10 (e) 40% Nb2O5/Zr(OH)4 0.09 200 100 200 300 100 0.08 200 Volume adsorbed/desorbed (cm3/g) Volume adsorbed/desorbed(cm3/g) 300 150 100 50 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0 0 0.2 0.4 0.6 Relative Pressure (P/P0) 0.8 1 0 2 4 6 8 Average pore diameter (nm) 10 (f) 25% Nb2O5/Zr(OH)4 Fig. 3.3 Nitrogen adsorption/desorption curves and pore volume distributions of Nb2O5/Zr(OH)4 catalysts (a) – (f) 72 500 2 450 1.5 400 1 MCM-41 350 Pore volume (cm3/g) volume adsorbed/desorbed (cm 3 /g) 2.9 MCM-41 300 250 200 150 0.5 0 0 5 10 15 20 (b) 0.06 25% Nb2O5/MCM-41 0.04 100 25% Nb2O5/MCM-41 50 0.02 0 0 0.2 0.4 0.6 0.8 1 0 Relative Pressure (P/P0) 0 5 (a) Average10diameter(nm) 15 20 (c) Fig. 3.4 (a) Nitrogen adsorption/desorption curves; pore volume distribution of (b) MCM-41 and (c) 25% Nb2O5/MCM-41 0.09 2 0.08 1.5 p o re v o lu m e (c m 3 /g ) pore volum e(cm 3/g) 0.07 0.06 Zr(OH)4 0.05 10% 0.04 20% 30% 0.03 1 0.5 40% 0.02 0.01 0 0 0 1 2 pore diameter(nm) 4 5 6 7 8 pore diameter(nm) (a) (b) 3 4 5 6 7 8 9 10 1 2 3 9 10 Fig. 3.5 Comparison of pore volume distribution (a) hydrous zirconia supported niobium oxide catalyst with Nb2O5 loadings from 0 – 40 wt.%, drying temperature 100°C (b) MCM-41 and 25% Nb2O5/MCM-41 73 3.3.3 Thermogravimetric Analysis Thermogravimetric analysis (Fig. 3.6) show that hydrous zirconia lost weight in a number of steps. Prior to the measurement, all samples were kept at 100°C for 30 min in order to remove physically adsorbed water. According to the IR investigations of Orio et al.[26,27] coordinated water can be detected on a zirconia oxide surface up to 200°C. Therefore, the first weight loss from 100°C to 200°C can be assigned as coordinated water (strongly bound chemisorbed water). Beyond this temperature, subsequence loss of water was ascribed to removal of surface hydroxyl groups to form zirconium oxide.[28] The total amount of water loss (Table 3.5 and 3.6) was calculated in the form of ratio H2O/oxides (mol/mol) as follows (equations 3.4 – 3.8). Calculation for amount of water lost H2O/oxides (mol/mol) Weight of oxides (mg) = total weight of sample (mg) × percentage weight 2 (%) (3.4) where percentage weight 2 (%) is the final weight percentage of sample after running TGA program Moles of oxides (mmol) = weight of oxides (mg) M (g/mol) (3.5) where M is the molecular weight of oxides, e.g. Molecular weight of 10% Nb2O5/ZrO2a = 0.1 × molecular weight of Nb2O5 + 0.9 × molecular weight of ZrO2 a catalyst after running TGA program is in oxide form Nb2O5/ZrO2 Weight of water lost (mg) = total weight of sample (mg) × percentage H2O (%) (3.6) 74 where percentage H2O (%) is the weight percentage of water lost Moles of H2O lost (mmol) = H2O/oxides (mol/mol) = weight of water (mg) molecular weight of H2O (18 g/mol) (3.7) Moles of H2O lost (mmol) Moles of oxides (mmol) (3.8) The amount of water lost for hydrous zirconia was 2.53 H2O/ZrO2 (mol/mol). The weight loss for bulk niobium oxide occurred below 400°C, and no further change in weight was observed for higher temperatures. The amount of water lost, 1.27 H2O/Nb2O5 (mol/mol), was smaller than that for hydrous zirconia. The weight loss patterns of the Nb2O5/Zr(OH)4 samples were similar to that for hydrous zirconia where constant weight was observed only after 500°C. However, the amount of water lost was comparable to that of bulk niobium oxide, 0.98 – 1.29 H2O/oxides (mol/mol). The smaller amount of water lost as compared to the support suggests that the niobium oxide overlayer stabilized the hydrated state of the hydrous zirconia support. 100 Nb2O5 90 weight (%) 10% Nb2O5/Zr(OH)4 20% Nb2O5/Zr(OH)4 25% Nb2O5/Zr(OH)4 80 30% Nb2O5/Zr(OH)4 40% Nb2O5/Zr(OH)4 70 Zr(OH)4 60 100 200 300 400 500 600 700 800 900 temperature(°C) Fig. 3.6 TGA curves of hydrous zirconia , niobium oxide and 10 – 40 wt.% hydrous zirconia supported niobium oxide catalysts 75 Table 3.5 Total amount of water lost (%) in TGA for Nb2O5/Zr(OH)4 catalysts Sample Sample Percentage Percentage Percentage weight weight 1a weight 2b H2Oc (%) (mg) (%) (%) Nb2O5 8.59 94.03 86.58 7.45 10% Nb2O5/Zr(OH)4 18.70 89.27 79.15 10.12 20% Nb2O5/Zr(OH)4 10.68 88.49 76.83 11.66 25% Nb2O5/Zr(OH)4 25.41 92.69 80.88 11.81 30% Nb2O5/Zr(OH)4 17.70 90.87 80.95 9.92 40% Nb2O5/Zr(OH)4 24.94 88.86 78.88 9.98 Zr(OH)4 27.05 93.96 68.62 25.34 a initial weight percentage of sample after isotherm at 100 °C for 30 min final weight percentage of sample after running TGA program c amount of water lost (%), equal to weight 1 (%) minus weight 2 (%) b Table 3.6 Calculation for amount of water lost H2O/oxides (mol/mol) for Nb2O5/Zr(OH)4 catalysts Sample Weight of Moles of Weight of Moles of H2O/oxides (mol/mol) oxides oxides H2O lost H2O lost (mg) (mmol) (mg) (mmol) Nb2O5 7.44 0.03 0.64 0.04 1.27 10% Nb2O5/Zr(OH)4 14.80 0.11 1.89 0.11 0.98 20% Nb2O5/Zr(OH)4 8.21 0.05 1.25 0.07 1.28 25% Nb2O5/Zr(OH)4 20.55 0.13 3.00 0.17 1.29 30% Nb2O5/Zr(OH)4 14.33 0.09 1.76 0.10 1.13 40% Nb2O5/Zr(OH)4 19.68 0.11 2.49 0.14 1.27 Zr(OH)4 18.56 0.15 6.86 0.38 2.53 76 3.3.4 Acidity Test Hammett indicators were used to determine the acid strength of the solid acid catalysts (Table 3.7). The bulk niobium oxide showed only weak acid sites (pKa = +3.3) at the surface. In contrast, the acid strength of the hydrous zirconia support was much stronger (pKa = -3.0), which was equivalent to 48 wt.% sulfuric acid solution. This can be attributed to the presence of surface hydroxyl groups on hydrous zirconia. Also, it is known that the digestion step in sample preparation procedure can help to increase the acidity of the catalysts through formation of BrØnsted acid sites and increase of Lewis acid sites.[29,30] This may explain the strong acidity of 25 wt.% Nb2O5-Zr(OH)4 sample prepared by coprecipitation method. On the other hand, after 10 wt.% niobium oxide was impregnated onto the hydrous zirconia, the acidity of the resulting sample dropped drastically. From Table 3.8, it was found that the density of acid sites decreased from 0.611 mmol/g in hydrous zirconia to 0.038 mmol/g in 10 wt.% Nb2O5/Zr(OH)4. Similar observations had been found where the addition of low loadings of niobium oxide onto titania and zirconia slightly decreased the number of Lewis acid sites and also the strength of these sites.[19,20] However, as the Nb2O5 loadings increased, the acidity of the corresponding impregnated catalysts increased as well. The acid strength of 20 to 25 wt.% niobium oxide impregnated sample was comparable to that of the bulk niobium oxide (pKa = +3.3). At even higher niobium oxide loading 30 to 40 wt.%, stronger acidity (pKa = -3.0) was exhibited. Onfroy et al.[31] reported that a threshold of niobium oxide loading was necessary for the formation of BrØnsted acid sites. Only when niobium oxide loading was higher than 1.2 Nb atoms/nm2, was BrØnsted acid sites formed. The number of BrØnsted acid sites increased with increasing niobium loadings. Raman and infrared results showed that at low loadings, monooxoniobate species were formed at the surface but at higher 77 loadings, dimeric and polymeric niobate species were formed. The density of acidic sites with pKa of at least +4.8 was determined by titration with n-butyalmine. It was found that the density of acidic sites increased with the niobium oxide loading. The acid sites per unit area are calculated as follows (equation 3.9): Acid site (m-2) = Acid site (mol/g) × NA SBET (3.9) where NA is the Avogadro constant, 6.023 × 1023, and SBET is the surface area of the catalysts The initial decrease in acid sites with the 10 wt.% Nb2O5/Zr(OH)4 sample suggests that the NH4OH used to precipitate the niobium oxide can also react with the acidic hydroxyl groups on hydrous zirconia. The consumption of these hydroxyl groups reduces the overall acidity of the sample. As the loading of niobium oxide increased, more hydroxyl groups on the support are utilized for anchoring these species through formation of new Nb-OH-Zr bridging hydroxyls.[32] However, the loss of these hydroxyl groups can be compensated by the acidic hydroxyl groups of niobium oxide, thus increasing the acid density and strength of the sample. Pyridine infrared spectroscopy was measured on 25% Nb2O5/Zr(OH)4 (Fig. 3.7). Both Lewis as well as Brønsted acid sites were found.[33] The band at ~1442 cm-1, indicative of H-bonded and Lewis-acid bonded pyridine, remained even after the sample was heated to 200°C. However, the absorption band due to Brønsted acidity at 1540 cm-1 was only observed up to 100°C. This can be attributed to removal of [H(H2O)n]+ species at the higher temperatures. 78 Table 3.7 Acid strength of Nb2O5/Zr(OH)4 catalysts measured using Hammett indicators Sample pKa a Methyl red Methyl yellow 2-aminoazotoluene 4-phenyldiphenylamine Dicinnamalacetone Benzalacetophenone +4.8 +3.3 +2.0 +1.5 -3.0 -5.6 10% Nb2O5/Zr(OH)4 +a - - - - - 20% Nb2O5/Zr(OH)4 + +a - - - - 25% Nb2O5/Zr(OH)4 + +a - - - - 30% Nb2O5/Zr(OH)4 + + +a +a +a - 40% Nb2O5/Zr(OH)4 + + + + + - Zr(OH)4 + + + + + - Nb2O5 + +a - - - - 25% Nb2O5-Zr(OH)4 coppt + + + + + - 25% Nb2O5/MCM-41 + + + + + - slightly color change 79 Table 3.8 Measuring number of acid sites by n-butylamine titrationa Sample Acid sites /m2 Zr(OH)4 0.611 9.76 x 1017 Nb2O5 0.260 9.91 x 1017 10% Nb2O5/Zr(OH)4 0.038 6.29 x 1016 20% Nb2O5/Zr(OH)4 0.226 5.40 x 1017 25% Nb2O5/Zr(OH)4 0.244 7.35 x 1017 30% Nb2O5/Zr(OH)4 0.282 7.83 x 1017 40% Nb2O5/Zr(OH)4 0.394 1.39 x 1018 25% Nb2O5-Zr(OH)4 coppt 0.495 1.24 x 1018 25% Nb2O5/MCM-41 1.035 3.82 x 1018 indicator used: methyl red (pKa = +4.8) 0.5 (c) 0.4 (b) 0.3 0.2 (a) 1700 absorbance a mmol/g 0.1 1650 1600 1550 1500 1450 0 1400 wavenumber(cm-1) Fig. 3.7 Pyridine IR of 25% Nb2O5/Zr(OH)4 after evacuation at (a) 25°C (b) 100°C and (c) 200°C. 80 3.4 Catalytic Activity 3.4.1 Effect of Niobium Oxide Loadings The Fischer indole reaction of phenylhydrazine with 3-heptanone to 2-butyl-3methyl indole and 2-ethyl-3-propyl indole was tested (Fig. 3.8). Over bulk niobium oxide, a conversion of only 32.6% was reached after 8 h reaction. The activity of the hydrous zirconia support was higher than niobium oxide, and a conversion of 58.6% was obtained in a similar reaction time. The increased activity of the hydrous zirconia should be due to the presence of surface hydroxyl groups which promote the surface acidity. When 10 wt.% niobium oxide was deposited on the hydrous zirconia, no significant change in conversion was observed. However, at higher loading range 20 to 30 wt.%, the catalysts were much more active. A maximum conversion of 91.1% was achieved for 25 wt.% Nb2O5/Zr(OH)4 sample on which the niobium oxide coverage was close to a monolayer. Further increase of the loading to 40 wt.% niobium oxide resulted in a lower conversion of only 68%. It is known that when two or more metal oxides are mixed together, materials with new acidic properties different from the original acid sites may be formed. Seiyama[34] proposed that acidity appears at the boundary where two oxides contact. The chemical environment of each of the cations does not change, but the the oxide anion at the boundary coordinate to two different cations. This model is usually applied in the case where an oxide is dispersed on a support. It is likely that the formation of Nb-OH-Zr bridging hydroxyl groups has significant impact to the activity of the catalysts in the Fischer indole reaction. As the niobium oxide loading increased from 10 to 25 wt.%, the niobium oxide overlayer gradually reached a monolayer, and the catalytic activity increased to a maximum. 81 100 90 80 conversion (%) 70 60 50 40 30 20 10 0 0 1 2 3 4 5 6 7 8 9 10 tim e(h) Fig. 3.8 The Fischer indole reaction of phenylhydrazine with 3-heptanone over (+) niobium oxide, (■) hydrous zirconia, and (×) 10% (□) 20% (▲) 25% (♦) 30% (○) 40% Nb2O5/Zr(OH)4 100.00 90.00 80.00 conversion (%) 70.00 60.00 50.00 40.00 30.00 20.00 10.00 0.00 0 2 4 6 reaction time (h) 8 10 Fig. 3.9 The Fischer indole reaction of phenylhydrazine with 3-heptanone over (■) 25% Nb2O5/Zr(OH)4 coppt (×) H-beta (Si/Al =12.5) (▲) 25% Nb2O5/MCM-41 (♦) 25% Nb2O5/Zr(OH)4 82 Further increase of niobium oxide loading led to a decrease in indole conversion because when the overlayer of niobium oxide exceeded a monolayer, bulk niobium oxide was formed on the surface of catalysts which blocked the active sites and hence reduced the activity. Similar observations were reported in V2O5/ZrO2 catalysts[35] and zirconia supported heteropoly acids.[36] A sample prepared by coprecipitation of hydrous zirconia and niobium oxide with 25 wt.% niobium oxide loading was less active than 25% Nb2O5/Zr(OH)4 as the conversion after 6 h was only 64% (Fig. 3.9). This is despite a higher density of acid sites for the coprecipitated sample (0.495 g/mmol). The results suggest that an overlayer of niobium oxide facilitated the catalytic reaction, rather than dispersed niobium sites, although using both preparation methods new acid sites could be created.[37] Indeed, the MCM-41 supported sample with 25 wt.% niobium oxide also showed better activity, 83% conversion, than the coprecipitated sample. The niobium oxide surface coverage on the MCM-41 sample was 0.74 monolayer. For comparison, microporous H-beta (Si/Al = 12.5) was tested and found to be less active than the supported niobium oxide samples. The conversion after 6 h was 72% for H-beta. In the Fischer indole reaction of phenylhydrazine with 3-heptanone, the selectivity to 2-butyl-3-methyl-indole (linear product) and 2-ethyl-3-propyl-indole (bulky product) was ca. 60:40 over all Nb2O5/Zr(OH)4 samples (Table. 3.9). Besides indoles, no other products were detected. Fig. 3.10 shows the two different indole isomers. Based on their molecular dimensions, the linear indole product is less bulky in steric conformation. Therefore, the selective synthesis of indole isomers, in principle, could be achieved using catalysts with spatially restricted reaction 83 environment. Zeolites were studied because these microporous aluminosilicates had micropores of diameters comparable with molecular dimensions.[38,39] When H-beta was used as catalyst in our experiment, the ratio of linear to bulky product was found to be 71:29. This is close to the ratio 74:26 reported by Kunkeler et al.[38] Zeolite Hbeta has a three-dimensional channel system with sizes 7.6 × 7.4 × 11 (Å). The pore structure is feasible for intraporous catalysis to occur whereas the use of molecular sieves with one-dimensional channel systems results in non-shape-selective catalytic activity because the reaction occurs at the outer surface of the crystallites.[40] Calculation based on thermodynamic equilibrium showed that the ratio of linear to bulky product should be 31:69. This is because the bulky indole isomer was calculated to be more stable by ca. 2.7 kJ/mol.[38] Based on the large discrepancy between the calculated value and experimentally obtained data, it was concluded that this reaction was kinetically controlled. The restricted transition state selectivity was suggested by Rigutto et al.[40] to be the most likely mechanism for determining regioselectivity. By this assumption, the selectivity should depend on the relative rates of formation of the intermediates enehydrazine isomers, and this step is assigned as the rate determining step of the catalytic indole reaction. N H 2-ethyl-3-propyl-indole N H 2-butyl-3-methyl-indole "bulky" "linear" 9.8 * 8.3 * 3.7 12.5 * 6.5 * 3.7 Fig. 3.10 Linear and bulky indole products with molecular dimensions in Å 84 Table 3.9 Selectivity a to 2-butyl-3-methyl indole (linear) and 2-ethyl-3-propyl indole (bulky) Drying Selectivity Selectivity Sample Temperature (linear) (bulky) (°C) 10% Nb2O5/Zr(OH)4 20% Nb2O5/Zr(OH)4 25% Nb2O5/Zr(OH)4 30% Nb2O5/Zr(OH)4 40% Nb2O5/Zr(OH)4 Zr(OH)4 Nb2O5 25% Nb2O5- Zr(OH)4 coppt 25% Nb2O5/ MCM-41 H-beta (Si/Al 12.5) a reaction time 8 h 100 200 300 100 200 300 100 200 60.3 59.6 59.8 60.4 59.8 59.8 59.8 59.5 39.7 40.4 40.2 39.6 40.2 40.2 40.2 40.5 300 100 200 59.1 60.5 60.1 40.9 39.5 39.9 300 100 200 300 100 200 59.9 60.7 60.3 59.9 60.2 59.8 40.1 39.3 39.7 40.1 39.8 40.2 300 100 300 100 100 500 60.0 60.9 61.6 60.7 63.1 71.0 40.0 39.1 38.4 39.3 36.9 29.0 In this study, although the supported niobium oxide catalysts have larger pore sizes than H-beta, preferential formation of linear 2-butyl-3-methyl indole was still observed. This suggests that the pores of the solid samples may play a role in the selectivity, although to a lesser extent than that of microporous zeolite. Indeed, this hypothesis is supported by the slightly higher selectivity of 63% to the linear indole 85 isomer over 25% Nb2O5/MCM-41 where the mean pore diameter is in the MCM-41supported sample is narrowly distributed around ~ 3.5 nm. 3.4.2 Effect of Drying Temperature Drying the catalysts at 200°C had no significant effect on the activity. For catalysts with niobium oxide loadings of 10 wt.% and 20 wt.%, the activity was even slightly increased (Fig. 3.11). However, after drying at 300°C, most of the samples except for 40 wt.% Nb2O5/Zr(OH)4 were less active. As discussed earlier, when drying the catalysts up to 200°C, only the coordinated water (strongly bound chemisorbed water) was lost. Beyond 200°C, the surface hydroxyl groups were removed which may led to a reduction of the surface acidity and active sites of the catalysts. Furthermore, the decrease of surface area could be another factor that accounted for the lower activity. 3.4.3 Indole Reaction with Cyclohexanone The synthesis of 1,2,3,4-tetrahydrocarbazole from phenylhydrazine and cyclohexanone was carried out by Nb2O5/Zr(OH)4 catalysts (Fig. 3.12). The reaction rate was about two times faster than that of phenylhydrazine with 3-heptanone. Similarly, the highest catalytic activity was obtained for catalyst with niobium oxide loadings close to a monolayer. In particular, 30 wt.% Nb2O5/Zr(OH)4 showed the best activity and the reaction almost completed within 4 h. The rate of reaction for 25 wt.% Nb2O5/MCM-41 was 86%, and this was similar to that of 25% Nb2O5/Zr(OH)4, 88% (Table 3.10). However, the coprecipitated 25 wt.% Nb2O5-Zr(OH)4 sample was less active and the conversion of carbazole after 4 h was only 54%. The reaction using microporous zeolite beta (Si/Al = 12.5) also led to a high reaction rate with a 86 conversion 84% in 4 h reaction time. Previously, Bhattacharya et al.[41] synthesized 1,2,3,4-tetrahydrocarbazole using a number of different zeolites. They reported the best yields of slightly higher than 60% were achieved by H-Y and H-beta. The reaction conditions were 2.7 mmol phenylhydrazine and 5.5 mmol cyclohexanone in 10 mL methanol with 1 g catalyst at 333 K and the reaction time was 18 h. The high conversion from Nb2O5/Zr(OH)4 catalysts showed that these catalysts are efficient for Fischer indole reaction of phenylhydrazine with ketones. 100.00 90.00 80.00 conversion (% ) 70.00 60.00 50.00 40.00 30.00 20.00 10.00 0.00 0 1 2 3 4 5 6 7 8 reaction time (h) Fig. 3.12 The Fischer indole synthesis of 1,2,3,4-tetrahydrocarbazole over (♦) hydrous zirconia, and (□) 10% (■) 20% (×) 30% (▲) 40% Nb2O5/Zr(OH)4 87 100 100C 90 200C 80 300C 70 60 50 40 30 20 10 0 Zr(OH)4 10% 20% 25% 30% 40% Nb2O5 Fig. 3.11 Conversion to indole products over 10 – 40 wt. % Nb2O5/Zr(OH)4 catalysts , Nb2O5 and Zr(OH)4 after drying at (…) 100°C, ( 200°C, and („) 300°C ) 88 Table 3.10 Conversion to 1,2,3,4-tetrahydrocarbazole after 4 h Sample Conversion (%) 10% Nb2O5/Zr(OH)4 67.1 20% Nb2O5/Zr(OH)4 86.9 25% Nb2O5/Zr(OH)4 88.3 30% Nb2O5/Zr(OH)4 99.6 40% Nb2O5/Zr(OH)4 82.5 25% Nb2O5- Zr(OH)4 coppt 53.8 25% Nb2O5/ MCM-41 86.2 Zr(OH)4 38.7 H-beta (Si/Al 12.5) 83.7 3.5 Recycling of Catalysts 3.5.1 Leaching of Catalysts The possibility of leaching was checked by filtering out the 30 wt.% Nb2O5/Zr(OH)4 from the reaction mixture of phenylhydrazine and cyclohexanone after 1 h. No further increase in conversion was observed in the catalyst-free system (Figure 3.13). 3.5.2 Regeneration of Catalysts Three methods are used to regenerate 25 wt.% Nb2O5/Zr(OH)4 catalysts. By washing with p-xylene solvent, the organic compounds adsorbed on the surface of catalyst were not removed as the catalyst still remained dark brown color. The activity after recycling was only ~ 50% of the activity for the fresh sample (Fig. 3.14). 89 60.00 removal of catalyst Conversion (%) 50.00 40.00 30.00 20.00 10.00 0.00 0 1 2 3 4 Reaction time (h) 5 6 Fig. 3.13 Catalyst leaching test Using method 2, washing the catalyst with H2O2 partially removed the adsorbed organic compounds. The color of the catalyst became pale yellow after washing. Rehydration of the catalyst in 5 M ammonia solution indeed resumed the acidity. With Hammett indicator tests and n-butylamine titration, the acid strength and number of acid sites of the rehydrated catalyst were determined to be even greater than the fresh sample, with pKa ≤ - 3.0 and number of acid sites = 0.592 mmol/g. However, the catalyst only regained ~ 80% activity in the Fischer indole reaction. In method 3, instead of washing catalyst in H2O2 at room temperature, the mixed solution was heated up to 40 °C. The adsorbed organic compounds were completely removed as the color of the catalyst changed back to white. The acid strength was the same as the fresh sample, but the number of acid sites decreased to 0.113 mmol/g. Surprisingly, the activity of the catalyst was fully recovered, and was 90 even higher than that of the fresh sample. This might be due to the generation of active hydroxyl groups during the regeneration procedures. 100 fresh 90 recycled 80 Conversion (%) 70 60 50 40 30 20 10 0 method 1 method 2 method 3 Fig. 3.14 Regeneration of 25 wt.% Nb2O5/Zr(OH)4 3.6 Conclusion Hydrous zirconia supported niobium oxide catalysts with loadings of 10 to 40 wt.% Nb2O5 were prepared by wet impregnation with niobium chloride. The samples retained the high surface area and pore structure of the support, suggesting a uniform overlayer. Theoretical monolayers of 0.33 to 2.0 were achieved with these loadings. The Fischer indole synthesis of phenylhydrazine with 3-heptanone and with cyclohexanone was achieved with high conversions. The most active catalysts had loadings of 25 – 30 wt.% Nb2O5, which is close to a monolayer coverage. The activity of the catalysts was the highest when dried at 100 – 200 °C. Use of siliceous MCM-41 as a support also resulted in an active catalyst. The good activity of the supported niobium oxide catalysts makes this a green route to the synthesis of indoles. 91 References [1] F. Fairbrother, in The Chemistry of Niobium and Tantalum (Ed. F. Fairbrother) 1967 (Elsevier: Amsterdam) [2] F. Holtzberg, A. Reisman, M. Berry, M. Berkenblit, J. Am. Chem. Soc. 1957, 79, 2039 [3] E. I. Ko, J. G. Weissman, Catal. Today 1990, 8, 27 [4] I. Nowak, M. Ziolek, Chem. Rev. 1999, 99, 3603 [5] S. M. Maurer, E. I. Ko, J. Catal. 1992, 135, 125 [6] J.-M. Jehng, I. E. Wachs, Catal. Today 1990, 8, 37 [7] K. Tanabe, M. Misono, Y. Ono, in New solid acids and bases : their catalytic properties, Stud. Surf. Sci. Catal. (Eds. K. Tanabe, M. Misono, Y. Ono, H. Hattori) 1989, vol. 51 (Kodansha: Tokyo) [8] M. Moraes, W. de S. F. Pinto, W. A. Gonzalez, L. M. P. M. Carmo, N. M. R. Pastura, E. R. Lachter, Appl. Catal. A 1996, 138, L7 [9] S. Okazaki, M. Kurimata, T. Iizuka, K. Tanabe, Bull. Chem. Soc. Jpn. 1987, 60, 37 [10] S. Okazaki, N. Wada, Catal. Today 1993, 16, 349 [11] A. Florentino, P. Cartraud, P. Magnoux, M. Guisnet, Appl. Catal. A 1992, 89, 143 [12] A. Kurosaki, T. Okuyama, S. Okazaki, Bull. Chem. Soc. Jpn. 1987, 60, 3541 [13] J. –M. Jehng, A. M. Turek, I. E. Wachs, Appl. Catal. A 1992, 83, 179 [14] K. Tanabe, Catal. Today 2003, 78, 65 [15] I. Nowak, M. Ziolek, Chem. Rev. 1999, 99, 3603 [16] Y. Izumi, Catal. Today 1997, 33, 390 [17] T. Okuhara, Chem. Rev. 2002, 102, 3641 92 [18] T. Ushikubo, Catal. Today 2000, 57, 331 [19] J.-M. Jehng, I. E. Wachs, Catal. Today 1993, 16, 417 [20] J.-M. Jehng, I. E. Wachs, J. Phys. Chem. 1991, 95, 7373 [21] R. M. Pittman, A. T. Bell, J. Phys. Chem. 1993, 97, 12178 [22] D. Brown, in Comprehensive Inorganic Chemistry (Ed. J.C. Bailar, Jr.) 1973, vol. 3 (Pergamon: London) [23] P. N. Kapoor, R. C. Mehrotra, J. Less Common Met. 1965, 8, 339 [24] C. Cheng, D. H. Park, J. Klinowski, J. Chem. Soc., Faraday Trans. 1997, 93, 193 [25] G. K. Chuah, S. Jaenicke, B. K. Pong, J. Catal. 1998, 175, 80 [26] L. Orio, V. Bolis, B. Fubini, C. Mortera, in Ceramics Today-Tomorrow’s Ceramics (Ed. A. Vincencini) 1991 (Elsevier: Amsterdam) [27] V. Bolis, C. Mortema, M. Volante, L. Orio, B. Fubini, Langmuir 1990, 6, 695. [28] J. Nawrocki, P. W. Carr, M. J. Annen, S. Froelicher, Anal. Chim. Acta 1996, 327, 261 [29] G. K. Chuah, S. H. Liu, S. Jaenicke, L. J. Harrison, J. Catal. 2001, 200, 352 [30] G. K. Chuah, S. Jaenicke, T. H. Xu, Surf. Interface Anal. 1999, 28, 131 [31] T. Onfroy, G. Clet, M. Houalla, J. Phys. Chem. B 2005, 109, 14588 [32] L. J. Burcham, J. Datka, I. E. Wachs, J. Phys. Chem. B 1999, 103, 6015 [33] E. P. Parry, J. Catal., 1963, 2, 371 [34] T. Seiyama, in Metal Oxides and Their Catalytic Actions (Ed. T. Seiyama) 1978 (Kodansha: Tokyo) [35] K. V. R. Chary, K. Ramesh, D. Naresh, P. V. R. Rao, A. R. Rao, V. V. Rao, Catal. Today 2009, 141, 187 [36] B. M. Devassy, S. B. Halligudi, J. Catal. 2005, 236, 313 93 [37] J.C. Védrine, G. Coudurier, A. Ouqour, P. G. Pries de Oliveira, J. C. Volta, Catal. Today 1996, 28, 3 [38] P. J. Kunkeler, M. S. Rigutto, R. S. Downing, H. J. A. de Vries, H. van Bekkum, in Progress in Zeolites and Microporous Materials, Stud. Surf. Sci. Catal. (Eds. H. Chon, S.–K. Ihm, Y. S. Uh) 1997, vol. 105 (Elsevier Science: Amsterdam) [39] R. S. Downing, P. J. Kunkeler, in Fine Chemicals through Heterogeneous Catalysis (Eds. R. A. Sheldon, H. van Bekkum) 2001 (Wiley-VCH: Weinheim) [40] M. S. Rigutto, H. J. A. de Vries, S. R. Magill, A. J. Hoefnagel, H. van Bekkum, in Heterogeneous Catalysis and Fine Chemicals III, Stud. Surf. Sci. Catal. (Eds. M. Guisnet,.J. Barbier, J. Barrault, C. Bouchoule, D. Duprez, G. Pérot, C. Montassier) 1993, vol. 78 (Elsevier Science: Amsterdam) [41] D. Bhattachary, D. W. Gammon, E. van Steen, Catal. Lett. 1999, 61, 93 94 Chapter IV Study of Boria-Zirconia Catalysts 4.1 Introduction Supported boria catalysts are widely used for various solid catalyzed reactions such as vapor phase Beckmann rearrangement of cyclohexanone oxime,[1] isomerization of 1-butene,[2] and dehydration of alcohol.[3] These catalysts are usually prepared using impregnation method, by depositing boria on metal oxide supports, such as SiO2, Al2O3, TiO2 and ZrO2 from an aqueous solution of boric acid. It is well established in the literature that the nature of support is crucial and responsible for the catalytic performance. By choosing an appropriate oxide support, the activity, selectivity and stability of the catalyst for a typical reaction can be greatly improved.[4] Zirconia is often used as the supporting material because for many catalytic reactions, zirconia supported catalysts produce better catalytic activity than those using conventional oxide supports.[5,6] Zirconia supported boron oxide containing 30 mol% of boron has been reported by Arata and coworkers[7,8] as a superacid catalyst with acidic strength H0 = − 13.0. The properties and activity of supported boria catalysts can be improved by proper control of variable factors such as loadings and dispersion of boria.[9–12] Xu et al.[4] studied the gas phase Beckmann rearrangement cyclohexanone oxime over zirconia supported boria catalysts. They compared the activity and selectivity of the zirconia supported boria catalyst with those supported on Al2O3, TiO2, SiO2, MgO and HZSM-5 at a boria loading level of 10 wt.%. It was found the zirconia supported 95 catalyst showed the best catalytic performance for lactam synthesis. The effect of boria loadings for the zirconia supported catalysts was also investigated. In the range of 0 – 20 wt.% boria loading, the selectivity to lactam increased with increasing boria loading, but as the conversion of cyclohexanone oxime decreased, the yield of lactam exhibited a maximum at a loading level of 10 wt.%. Malshe et al.[13] applied the boriazirconia solid catalyst in another acid catalyzed reaction, namely, the selective Cmethylation of phenol with methanol. A series of boria-zirconia catalysts with 5 -30 mol% boria loadings were tested. The maximum conversion was observed for 5 mol% boria-zirconia sample where a highly uniform dispersion of boria on the zirconia support was achieved. It is well established from the literature that the performance of supported boria catalyst is strongly dependent on the calcination temperature.[14-16] Mao et al.[17-20] did a systematic study on the activity of B2O3/ZrO2-TiO2 catalysts in the vapor phase Beckmann rearrangement of cyclohexanone oxime. They investigated the influence of the preparation method, boria loading, calcination temperature and also the deactivation and regeneration of the catalysts. It was found that the crystallization of ZrO2-TiO2 started at 600°C. At this temperature, the surface area, the acidity and the catalytic activity of the catalyst began to decrease. However, at a lower temperature of 500 °C, the acid strength was weak due to the incomplete interaction of B2O3 with ZrO2-TiO2. Deactivation of the B2O3/ZrO2-TiO2 catalyst occurred and was mainly caused by the deposition of coke. By calcining the deactivated catalyst at 600°C in air for 8 h, the activity was fully recovered. 96 In this work, the possible application of boria-zirconia solid acid catalysts in the Fisher indole reaction of phenylhydrazine with 3-heptanone was explored. The morphology, textural properties, and acidity of the catalysts are characterized. The effects of boria loading, calcination temperature and recycling of the catalysts are investigated. The results are compared with that of Nb2O5/Zr(OH)4 catalysts. 4.2 Preparation of Catalysts 4.2.1 Preparation of Boria-zirconia Catalysts Catalysts were prepared following a similar method described by Malshe et al.[21] In a typical synthesis procedure for 2 g of 25 wt.% boria zirconia, 0.888 g boric acid (Sigma-Aldrich) was dissolved in 50 mL deionized water. The zirconia support was prepared following the procedure described in chapter 3 and was added to the boric acid solution with constant stirring to obtain slurry. After heating to dryness, it was further dried overnight in an oven at 100°C and calcined at 500°C for 5 h. A series of catalysts containing 3, 5, 10, 20, 30 and 40 wt.% of boria were prepared similarly using the corresponding amount of boric acid and zirconia support (Table 4.1). All catalysts were characterized by X-ray diffraction and nitrogen adsorption/desorption measurement to study the crystal phases and textural properties. ICP-AES was used to determine the actual boria loading. Thermal analysis was performed to determine the phase transition enthalpy. Infrared spectroscopy revealed information about the bonding structures of boria overlayer. Acidity of the catalysts was measured by NH3 TPD and by base titration using Hammett indicators. 97 Table 4.1 Amount of reagents used in preparing the boria-zirconia catalysts H3BO3 (g) Sample Name Zr(OH)4 (g) 3% 1.94 0.107 5% 1.90 0.178 10% 1.80 0.355 20% 1.60 0.710 25% 1.50 0.888 30% 1.40 1.07 40% 1.20 1.42 4.2.2 Recycling of Catalysts The used catalyst was separated by centrifugation, dried and regenerated by calcining at 500°C for 5 h. The morphology, textural properties and acidity of the recycled catalysts were determined. The catalytic activity was tested for up to three rounds of experiments. The catalysts used in these three rounds are named “fresh”, “recycle 1”, and “recycle 2”, respectively. 4.3 Catalyst Characterization 4.3.1 Powder X-ray Diffraction The X-ray diffractogram of boria-zirconia catalysts with different boria loadings are shown in Fig. 4.1. For samples with boria loadings from 3 to 20 wt.%, the XRD results showed that only the amorphous phase was present despite the calcination at 500°C. The transformation of amorphous hydrous zirconia to crystalline zirconium oxide usually starts from 450°C.[22] The formation of metastable tetragonal phase in the boria-zirconia catalysts was postponed to higher temperatures, indicating that the deposition of boria overlayer stabilized amorphous phase of the zirconia 98 support. This is similar to previous observations for Nb2O5/Zr(OH)4 and also zirconia catalysts doped with other metals.[23] At boria loadings of 25 wt.% and higher, two diffraction peaks at 2θ ~ 15° and 28° were observed. These two peaks are consistent with the most intense peaks of boric acid showing that aggregation of boron oxide molecules occurs at high boria loadings.[24] Three 25 wt.% boria-zirconia catalysts were prepared at different calcination temperatures of 100°C, 500°C and 650°C, and their X-ray diffractograms are shown in Fig. 4.2. At temperatures of 100°C and 500°C, only the boric acid peaks at 2θ ~ 15° and 28° were detected. After calcination at 650°C, crystalline zirconia was formed. The peaks at 30.2°, 35.3°, 50.4°, and 59.7° are due to the diffractions of (111), (200), (220) and (131) planes of tetragonal ZrO2, whereas the peak around 28.2° can be assigned to be the diffraction of (11 1 ) of monoclinic ZrO2.[25] The peak at ~15° from boric acid could no longer be seen. The Debye-Scherrer equation was used to estimate the crystallite sizes of the boria-zirconia samples with boria loadings of 25 wt.% and higher from the measured line width of the peak ~28°. The average crystallite sizes are 17.8, 18.6 and 18.0 nm for 25, 30 and 40 wt.% boria zirconia, respectively. The three similar values indicate that although aggregation of boron oxide molecules starts at boria loading of 25 wt.%, the degree of aggregation does not aggravate with higher boria loadings. 99 1400 1200 40% Intensity (cps) 1000 30% 25% 800 20% 600 10% 400 5% 3% 200 0 0 10 20 30 40 50 60 70 2θ (°) Fig. 4.1 X-ray diffractograms of boria-zirconia catalysts with boria loading from 3 to 40 wt.%, calcination temperature 500°C 1200 1000 Intensity (cps) 800 650 °C 600 500 °C 400 100 °C 200 0 0 10 20 30 40 50 60 70 2θ (°) Fig. 4.2 X-ray diffractograms of 25 wt.% boria zirconia at calcination temperature 100°C, 500°C, and 650°C 100 4.3.2 Textural Properties Table 4.2 shows the surface area and pore volume of boria-zirconia catalysts. Impregnation of small amounts of boria (3 to 10 wt.%) increased the surface area from 179 to 226 m2/g. This promotion effect is also seen in zirconia catalysts with other metal dopants.[23] However, at higher loadings, 20 to 40 wt.%, the surface area dropped drastically, with only 23 m2/g for the 40 wt.% boria zirconia sample. This can be attributed to the aggregation of boron oxide to form bulk boron oxide at the surface of the zirconia support. This result is supported by the detection of bulk boric acid peaks in the XRD measurement. The pore volume of boria-zirconia catalysts exhibited a similar trend as that of the surface area. At boria loadings of 3 - 10 wt.%, the pore volume was relatively large, in the range of 0.24 – 0.27 cm3/g. Beyond 20 wt.%, a significant decrease of pore volume was observed. The nitrogen adsorption/desorption isotherms of all boria-zirconia samples calcined at 500°C were similar, with less volume of nitrogen adsorbed/desorbed for catalyst at high boria loadings (Fig. 4.3). The size of pores ranged from 2.9 – 10 nm with a narrow distribution centered at ~3.5nm and bigger pores from 4 to 10 nm. As the boria loading increased, the distribution of the bigger pores narrowed, suggesting that the overlayer is uniformly coating the pore channels. For loadings up to 30 wt.%, the small pores centered at ~3.5 nm are unaffected. Only for the 40 wt.% sample is there a significant loss in porosity. The isotherms of 25 wt.% boria-zirconia samples calcined at 100°C, 500°C, and 650°C were compared (Fig. 4.4). The sample dried at 100°C had a smaller area than that calcined at 500°C. This may be due to the incomplete removal of water from the sample, hence hindering the absorption of nitrogen. On the other hand, 650°C calcined sample exhibited a different isotherm 101 with the hysteresis loop at around P/P0 = 0.9. This shows that the smaller pores had collapsed after calcination at this temperature, resulting in only mesopores of 10 – 40 nm. Following calcination at 650°C the surface area decreased drastically from 115 to 20 m2/g. Table 4.2 Textural properties of boria-zirconia catalysts Sample Calcination SBET temperature Pore volume 2 (m /g) (cm3/g) (°C) 3% boria-zirconia 500 179 0.27 5% boria-zirconia 500 195 0.24 10% boria-zirconia 500 226 0.25 20% boria-zirconia 500 190 0.21 25% boria-zirconia 500 115 0.15 30% boria-zirconia 500 70 0.11 40% boria-zirconia 500 23 0.05 25% boria-zirconia 100 29 0.06 25% boria-zirconia 650 20 0.22 The amount of boria impregnated on the zirconia support was measured by inductively coupled plasma-atomic emission spectroscopy (ICP-AES). The actual boria loadings were slightly higher than the expected value for all boria-zirconia samples (Table 4.3). Based on the physical measurement results done by Xu et al.,[26] the monolayer coverage of boria on zirconia surface was estimated to be about 102 12 μmol B2O3/m2-ZrO2. Using this value, the monolayer coverage of the boriazicronia catalysts was calculated as follows (equations 4.1 – 4.3): Calculation of theoretical monolayer coverage n(B2O3) = actual B2O3 loading molecular weight of B2O3 (4.1) where n(B2O3) = number of moles of B2O3 in catalyst, Average number of boron atoms = n(B2O3) × 2 × NA SBET (4.2) where NA is the Avogadro Constant, Monolayer coverage = Average number of boron atoms 7.23 × 1018 atoms / m 2 (4.3) If the boria stayed at the surface, then monolayer coverage should occur for samples with loadings between 5 and 10 wt.%. At loadings of 20 wt.% and higher, the boria overlayer was more than three layers thick. These results agree with the observation of diffraction peaks of boria in the sample. It was reported[4] that at boria loadings below a monolayer, the first overlayer of boria on the zirconia surface should be composed of the tetrahedral oxygen-coordinated boron (BO4) units. Trigonal BO3 units are formed at higher boria loadings and are presumable stacked onto the interconnected BO4 units of the monolayer. Infrared studies show that samples with 20 – 40 wt. % boria exhibited absorption bands at 650 cm-1, 885 cm-1, and 1200 cm-1 and their intensity grew with boria content (Fig. 4.5). Comparing with the spectra of boric acid and borax (not 103 shown),[6] the absorption bands at 650 cm-1, 885 cm-1, and 1200 cm-1 can be assigned as the trigonal BO3 structures present in boric acid.[27,28] Samples with low loadings (3 – 10 wt.%) showed two broad absorption bands at 1250 – 1450 cm-1, and 1550 1700 cm-1, indicative of the presence of tetragonal BO4 units. As the boria loading increased, the band at 1250 – 1450 cm-1 shifted up to 1580 cm-1 while the band at 1550 - 1700 cm-1 diminished. Table 4.3 Boria loading and surface coverage Average number Sample Boria loadings (wt.%) a a of boron atoms Monolayer coverage -2 (m ) 3% boria-zirconia 3.79 3.66 ×1018 0.51 5% boria-zirconia 6.70 5.95×1018 0.82 10% boria-zirconia 13.57 1.07×1019 1.48 20% boria-zirconia 25.84 2.43×1019 3.36 25% boria-zirconia 32.16 5.00×1019 6.92 30% boria-zirconia 37.31 9.54×1019 13.2 40% boria-zirconia 44.67 3.47×1020 48 from ICP measurement 4.3.3 Thermal Analysis The TG-DTA diagrams of hydrous zirconia and boria-zirconia catalysts are plotted in Fig. 4.6. Below 200°C, the initial weight loss was attributed to the loss of adsorbed moisture on the surface of the catalysts. The total amount of water loss (Table 4.4 and 4.5) was calculated in the form of ratio H2O/oxides (mol/mol) as follows (equations 4.4 – 4.8). 104 250 0.06 3% 0.05 200 0.04 150 0.03 100 0.02 50 0.01 0 0 0 0.2 0.4 0.6 0.8 1 200 0 5 10 15 20 0 5 10 15 20 0 5 10 15 20 0.09 5% 0.08 150 0.07 0.06 0.05 100 0.04 0.03 50 0.02 0.01 0 0 0.2 0.4 0.6 0.8 200 1 10% 150 100 50 0 0 0.2 0.4 0.6 0.8 200 1 20% 150 100 50 0 0 0.2 0.4 0.6 0.8 150 1 Volume adsorbed/desorbed (cm3/g) Volume adsorbed/desorbed (cm3/g) 0 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0.1 0.09 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0 0 5 10 15 20 0.09 25% 0.08 0.07 100 0.06 0.05 0.04 0.03 50 0.02 0.01 0 0 0 0.2 0.4 0.6 0.8 100 0 5 10 15 20 0 5 10 15 20 15 20 1 30% 0.07 0.06 0.05 0.04 50 0.03 0.02 0.01 0 0 0 0.2 0.4 0.6 0.8 50 1 0.014 40% 0.012 0.01 0.008 0.006 0.004 0.002 0 0 0 0.2 0.4 0.6 0.8 Relative Pressure (P/P0) 1 0 5 10 Average diameter (nm) Fig. 4.3 Nitrogen adsorption/desorption curves and pore volume distribution of boriazirconia catalysts with different boria loadings 105 160 0.04 100°C 140 0.02 650˚C 0.01 Volume adsorbed/desorbed (cm3/g) N 2 volum e adsorbed/desorbed (cm 3 /g) 0.03 120 100 500˚C 80 60 40 100˚C 20 0 0 5 10 15 0.09 20 500°C 0.06 0.03 0 0 5 10 15 20 0.01 650°C 0.005 0 0 0 0 0.2 0.4 0.6 0.8 20 40 60 80 1 Average diameter (nm) Relative Pressure (P/P0) Fig. 4.4 Nitrogen adsorption/desorption curves and pore volume distribution of 25wt.% boria-zirconia sample calcined at 100°C, 500°C, and 650°C 350 (h) (g) 300 (f) 200 (d) 150 (c) (b) 100 50 (a) 1800 tansmittance(%) 250 (e) 1600 1400 1200 1000 800 0 600 wavenumber(cm-1) Fig. 4.5 Infrared spectra of (a) boric acid (b) 3% (c) 5% (d) 10% (e) 20% (f) 25% (g) 30% and (h) 40% boria-zirconia 106 Calculation for amount of water lost H2O/oxides (mol/mol) Weight of oxides (mg) = total weight of sample (mg) × percentage weight 2 (%) (4.4) where percentage weight 2 (%) is the final weight percentage of sample after running TGA program Moles of oxides (mmol) = weight of oxides (mg) M (g/mol) (4.5) where M is the molecular weight of oxides, e.g. Molecular weight of 5% B2O3/ZrO2a = 0.05 × molecular weight of B2O3 + 0.95 × molecular weight of ZrO2 a catalyst after running TGA program is in oxide form B2O3/ZrO2 Weight of water lost (mg) = total weight of sample (mg) × percentage H2O (%) (4.6) where percentage H2O (%) is the weight percentage of water lost Moles of H2O lost (mmol) = H2O/oxides (mol/mol) = weight of water (mg) molecular weight of H2O (18 g/mol) Moles of H2O lost (mmol) Moles of oxides (mmol) (4.7) (4.8) 107 Table 4.4 Total amount of water lost (%) in TGA for boria-zirconia catalysts H2Oc (%) Sample Weight (mg) Weight 1a Weight 2b (%) (%) 3% boria-zirconia 19.37 91.03 87.94 3.09 5% boria-zirconia 28.78 89.07 85.60 3.47 10% boria-zirconia 30.76 89.90 86.20 3.7 20% boria-zirconia 26.01 89.29 84.53 4.76 25% boria-zirconia 10.25 87.53 82.79 5.34 30% boria-zirconia 15.27 85.60 79.26 6.34 40% boria-zirconia 13.47 83.46 75.45 8.01 a initial weight percentage of sample after isotherm at 100°C for 30 min final weight percentage of sample after running TGA program c amount of H2O = weight change b Table 4.5 Calculation for amount of water lost H2O/oxides (mol/mol) for boriazirconia catalysts Sample Weight of Moles of Weight Moles of H2O/oxides oxides oxides of H2O H2O (mg) (mmol) (mg) (mmol) 3% boria-zirconia 17.03 0.140 0.60 0.033 0.24 5% boria-zirconia 24.64 0.204 1.00 0.055 0.27 10% boria-zirconia 26.52 0.225 1.14 0.063 0.28 20% boria-zirconia 22.00 0.196 1.24 0.069 0.35 25% boria-zirconia 8.49 0.077 0.55 0.031 0.40 30% boria-zirconia 12.10 0.113 0.97 0.054 0.48 40% boria-zirconia 10.16 0.100 1.08 0.060 0.60 (mol/mol) 108 The amount of water lost for boria-zirconia catalysts was 0.24 - 0.60 H2O/oxides (mol/mol), which was about half of the amount of water lost for Nb2O5/Zr(OH)4 samples. This is because the calcination temperature of boria-zriconia catalysts was 500°C and at this temperature, most of the surface hydroxyl groups were removed. Constant weight was obtained above 500°C. As the boria loading increased, the ratio H2O/oxides (mol/mol) increased as well, indicating that the hydroxyl groups remained in boria-zirconia catalysts were mainly from boria component, not zirconia. An endothermic peak with maximum at 130°C to 170°C was observed in the DTA profile for all boria-zirconia samples. This peak is attributed to the desorption of physically adsorbed water and the dehydration of H3BO3.[29] Because the water physically adsorbed on catalysts is more easily removed than that produced by dehydration of boric acid, a shift of the maximum of the peak towards higher temperatures with increasing of boria loading was observed, from about 130°C for 3 wt.% boria-zirconia to about 170°C for 40 wt% boria-zirconia. At the same time, the intensity of this endothermic peak increased with the increase of boria loading. Moreover, an exothermic band was found in all TGA profiles without a corresponding weight loss in TG diagram. From this observation, it was concluded the exothermic peak represented the crystallization of amorphous zirconia support. For pure hydrous zirconia, this phase transition happened at about 450°C. However, after impregnation with boria, the phase transition was postponed to 630 – 720°C. Therefore, impregnation of boria onto the surface of zirconia stabilized the amorphous phase of zirconia component. The area of the exothermal DTA peaks decreased as 109 increase of boria loadings. The relationship between the peak area (A) of the DTA curve and an enthalpy change (∆H) for a mass (m) of sample is as equation 4.9:[30] A = + ∆HmK (4.9) where K is the calibration constant, Using the crystallization enthalpy of pure zirconia 28.2 kJmol-1[31] (229 J/g), and exothermic peak area of zirconia integrated from the DTA curve, the calibration constant K can be found. Applying the value of K, the enthalpy of crystallinzation for boria-zirconia catalysts were calculated (Table 4.6). The glow phenomenon is believed to be a visible manifestation of the coalescence of primary colloidal particles to larger masses with consequence release of surface energy.[32] The exotherm could be a characteristic feature of both the crystallization and glow phenomenon occurring simultaneously. The enthalpy of crystallization of zirconia was lowered by deposition of boria on its surface. Also, except for 3 wt.% boria-zirconia sample, the enthalpy of crystallization for all other boria-zirconia decreased with increasing boria loadings, from 208 J/g for 5 wt.% boria-zirconia to 9.67 J/g for 40 wt.% sample. A decrease in the enthalpy of crystallization with increase in boria loading indicates that the glow phenomenon slowly decreases with increasing boria content. 110 0.08 Weight (%) 0.04 80 0.02 0 70 -0.02 0.03 91 0.025 exo Weight (%) 0.06 Temperature difference (°C/mg) (b) exo 90 0.035 92 (a) 0.02 90 0.015 0.01 89 0.005 0 88 Temperature difference (°C/mg) 100 -0.005 200 300 400 500 600 700 800 -0.04 900 87 100 200 300 Temperature (°C) 400 500 600 700 -0.01 900 800 Temperature (°C) 0.025 90 0.07 90 (c) (d) 0.015 0.01 88 0.005 87 0 -0.005 86 -0.01 85 100 200 300 400 500 600 Temperature (°C) 700 800 -0.015 900 89 exo 0.06 0.05 0.04 Weight (%) Weight (%) exo Temperature difference (°C/mg) 0.02 89 88 0.03 0.02 87 0.01 0 86 Temperature difference (°C/mg) 60 100 -0.01 85 100 200 300 400 500 600 700 800 -0.02 900 Temperature (°C) 111 88 0.025 0.08 (f) (e) 88 0.01 87 0.005 86 0 -0.005 85 -0.01 84 83 100 87 0.06 86 0.04 85 84 0 83 -0.02 -0.015 200 300 400 500 600 700 800 82 100 -0.02 900 200 300 400 500 600 700 0.05 86 83 0.01 82 exo 0 81 -0.01 80 -0.02 79 -0.03 400 500 600 Temperature (°C) 700 800 -0.04 900 0.06 81 Weight (%) 0.02 83 82 Temperature difference (°C/mg) 0.03 84 Weight (%) (h) 0.04 85 300 -0.04 900 0.08 (g) 200 800 Temperature (°C) Temperature (°C) 78 100 0.02 exo 0.04 80 79 0.02 exo 78 0 77 -0.02 76 75 100 200 300 400 500 600 700 800 Temperature difference (°C/mg) Weight (%) exo Weight (%) 0.015 Temperature difference (°C/mg) 0.02 89 Temperature difference (°C/mg) 90 -0.04 900 Temperature (°C) Fig. 4.6 TG–DTA diagrams of (a) hydrous zirconia (b) 3% (c) 5% (d) 10% (e) 20% (f) 25% (g) 30% and (h) 40% boria-zirconia TG curve DTA curve 112 Table 4.6 Peak maxima of exothermic peaks and enthalpy of crystallization of hydrous zirconia and boria-zirconia catalysts Sample Peak maxima Enthalpy (°C) (J/g) Hydrous zirconia 447 229 3% boria-zirconia 630 148 5% boria-zirconia 681 208 10% boria-zirconia 718 151 20% boria-zirconia 696 67.7 25% boria-zirconia 678 42.8 30% boria-zirconia 664 25.7 40% boria-zirconia 655 9.67 4.3.4 Acidity Measurement The surface acidity of boria-zirconia catalysts was measured by temperature programmed desorption using ammonia as a probe molecule. From the ammonia TPD profiles (Fig. 4.7), it can be seen that all desorption occurred below 400°C. For 3% boria-zirconia, the maximum of the desorption peak occurred between 150 – 270 °C, corresponding to medium acid sites. As boria loadings reached up to 20 wt.%, a slightly shift of the desorption peak to lower temperatures was observed. However, beyond 20 wt.%, the maximum shifted towards higher temperature, and for 30 wt.% and 40 wt.% samples, both strong (270 -320°C) and medium (150 – 270°C) acid sites were observed. The number of acid sites per unit area generally increased with boria loadings (Table 4.7). Malshe et al.[21] also found that the 30 mol% B2O3/ZrO2 had higher acid site density than the 5 mol% sample. 113 3.00E-11 2.50E-11 (h) (g) Ion current (A) 2.00E-11 (f) 1.50E-11 1.00E-11 (e) (d) (c) 5.00E-12 (b) (a) 0.00E+00 160 240 320 400 480 Temperature (°C) Fig. 4.7 Ammonia TPD profiles of (a) 3% (b) 5% (c) 10% (d) 20% (e) 25% (f) 30% (g) 40% boria-zirconia (calcination temperature 500 °C) and (h) 25% boria-zirconia (calcination temperature 650°C) The 25 wt.% boria-zirconia sample showed higher acid strength when calcined at 650°C than at 500°C. A pronounced desorption peak was observed at higher temperatures. A similar shift was found for B2O3/TiO2-ZrO2.[19] According to the postulate of Mao et al.[19], the phase transition of amorphous to crystalline form ZrO2 happened during 500 – 650°C, and therefore, before the transformation, the reaction of B2O3 with the zirconia support was not complete, so only weaker acid sites were produced. This phenomenon was also observed in the preparation of WO3/ZrO2, where tungsten oxide combines with ZrO2 to create superacid sites when tetragonal crystalline phase of ZrO2 was formed.[33] 114 Table 4.7 Number of acid sites of boria-zirconia catalysts determined by NH3 TPD Sample Name Weight Total acid sites Average acid sites (g) (mmol) (mmol/g) (m-2) 3% boria-zirconia 500°C 0.247 0.071 0.29 9.59 × 1017 5% boria-zirconia 500°C 0.250 0.054 0.22 6.49 × 1017 10% boria-zirconia 500°C 0.263 0.078 0.30 7.86 × 1017 20% boria-zirconia 500°C 0.253 0.114 0.45 1.42 × 1018 25% boria-zirconia 500°C 0.152 0.047 0.31 1.62 × 1018 25% boria-zirconia 650°C 0.153 0.056 0.37 1.11 × 1019 30% boria-zirconia 500°C 0.229 0.081 0.36 3.06 × 1018 40% boria-zirconia 500°C 0.237 0.042 0.18 4.66 × 1018 To compare the acidity of boria-zirconia catalysts with that of Nb2O5/Zr(OH)4 samples, color tests using Hammett indicator and amine titration were also performed to determine the acid strength and number of acid sites (Table 4.8 and Table 4.9). 3 wt.% and 5 wt.% boria-zirconia exhibited no color change for any of the indicators, indicating that the two catalysts had acid strength lower than pKa of +4.8. As the loading of boria increased, the acid strength of the catalysts increased as well. From 20 wt.% boria loading onwards, the catalysts showed acid strength with pKa greater than +1.5. However, 25 wt.% boria-zirconia with drying/calcination temperature 100 °C and 650 °C showed low acid strength. For 30 wt.% and 40 wt.% 115 Nb2O5/Zr(OH)4 samples, the acid strength were higher than all boria-zirconia catalysts, with pKa greater than -3.0. The number of acid sites showed a similar trend as that determined by NH3 TPD. The maximum number of acid sites was obtained at boria loading of 20 wt.%, and decreased for further deposition of boria. The number of acid sites determined by n-butylamine titration was generally larger than that determined by NH3 TPD except for 25 wt.% boria-zirconia calcined at 650°C. It is probably because the use of different base molecules as well as different operation temperatures. Compared with Nb2O5/Zr(OH)4 samples, at the same loading level, boria-zirconia catalysts contained more acid sites although their acid strength might be lower. Table 4.9 Number of acid sites of boria-zirconia catalysts measured by n-butylamine titrationa Sample Acid sites a b mmol/g /m2 10% boria-zirconia 0.315 8.39 x 1017 20% boria-zirconia 0.899 2.85 x 1018 25% boria-zirconia 0.808 4.23 x 1018 30% boria-zirconia 0.797 6.86 x 1018 40% boria-zirconia 0.477 1.25 x 1019 25% boria-zirconia (100°C)b 0.183 3.80 x 1018 52% boria-zirconia (100°C) 0.179 5.39 x 1018 indicator used: methyl red (pKa = +4.8) calcination/drying temperature; the other catalysts were calcined at 500°C 116 Table 4.8 Measuring acid strength of boria-zirconia catalysts using Hammett indicators Sample pKa a b Methyl red Methyl yellow 2-aminoazotoluene 4-phenyldiphenylamine Dicinnamalacetone Benzalacetophenone +4.8 +3.3 +2.0 +1.5 -3.0 -5.6 3% boria-zirconia - - - - - - 5% boria-zirconia - - - - - - 10% boria-zirconia + - - - - - 20% boria-zirconia + + + + - - 25% boria-zirconia + + + + - - 30% boria-zirconia + + + + - - 40% boria-zirconia + + + + - - 25% boria-zirconia (100 °C)b + +a - - - - 25% boria-zirconia (650 °C)b + +a - - - - slightly color change calcination/drying temperature; the other boria-zirconia catalysts were calcined at 500°C 117 4.4 Catalytic Activity 4.4.1 Effect of Boria Loadings The Fischer indole reaction of phenylhydrazine with 3-heptanone to 2-butyl-3methyl indole and 2-ethyl-3-propyl indole was tested (Fig. 4.8). At 5 wt.% boria loading, the maximum conversion of 85.1% was reached after 8 h. This is slightly lower than the maximum conversion of 91.1% obtained from 25 wt.% Nb2O5/Zr(OH)4. At higher boria loadings, the conversion of indole products decreased and the conversion was only 23.6% for 40 wt.% boria-zirconia catalyst. A similar trend was found in selective the C-methylation of phenol with methanol over 5 – 30 wt.% boriazirconia catalysts.[21] At 5 wt.% boron oxide loading, the overlayer of boria approached a monolayer coverage but at higher loadings, multiple layers of boria are formed. The results showed that a uniform dispersion of boria on zirconia leads to enhanced activity. When the loading decreased to 3 wt.%, a lower activity was obtained possibly due to a lower density of active sites on the zirconia support. Besides indoles, no other product was detected. The selectivity to linear and bulky indole product was ca. 60:40 (Table 4.10). This ratio is the same as that of Nb2O5/Zr(OH)4 samples and hydrous zirconia. Therefore, the ratio of indole selectivity is probably kinetically controlled. 118 100.00 90.00 80.00 Conversion (%) 70.00 60.00 50.00 40.00 30.00 20.00 10.00 0.00 0 2 4 6 Reaction Time (h) 8 10 Fig. 4.8 Conversion of indoles over (♦) 3% (■) 5% (▲) 10% (×) 20% (◊) 25% (□)30% (+) 40% boria-zirconia Table 4.10 Conversion and selectivitya of 2-butyl-3-methyl indole (linear) and 2ethyl-3-propyl indole (bulky) Conversion Selectivity Selectivity Sampleb (%) (linear) (bulky) (%) (%) a b 3% boria-zirconia 64.3 59.9 40.1 5% boria-zirconia 85.1 59.4 40.6 10% boria-zirconia 78.6 60.7 39.3 20% boria-zirconia 65.4 61.1 38.9 25% boria-zirconia 56.7 60.7 39.3 30% boria-zirconia 29.8 60.4 39.6 40% boria-zirconia 23.6 60.2 39.8 reaction after 8 h reaction; reaction conditions: 140°C, nitrogen protected all samples calcined to 500°C 119 4.4.2 Effect of Calcination Temperature The effect of calcination temperature on the catalytic activity of the Fisher indole synthesis was tested over a 25 wt.% boria-zirconia sample (Fig. 4.9). The uncalcined sample gave a low indole conversion of ~ 32% after 6 h. After calcination at 500°C, the activity increased and a conversion of 56.7% was obtained after 8 h. The increase of catalytic activity is most likely due to the increase of surface area from 29 to 115 m2/g and pore volume from 0.06 to 0.15 cm3/g. Further increase of the calcination temperature to 650°C led to a drastic decrease of activity. After 6 h, the conversion was only 7.1% for 650°C calcined sample. The formation of the crystalline structure as well as the decrease of surface area could be the reasons for the low activity. This is in contrast to the findings that crystalline boria-zirconia formed after calcination at 650 °C was the most active catalyst in the selective C-methylation of phenol with methanol.[21] 60.00 50.00 Conversion (%) 40.00 30.00 20.00 10.00 0.00 0 1 2 3 4 5 6 7 8 Reaction Time (h) Fig. 4.9 Conversion of indoles over 25 wt.% boria-zirconia (♦) uncalcined (■) calcined at 500°C and (▲) calcined at 650°C 120 4.4.3 Recycling of Catalysts Catalyst deactivation is a common problem in catalysis. The main reason for catalyst deactivation is blocking of active sites by the deposition of organic products on the surface of catalysts or poisoning of the acid sites by adsorption of basic products. Besides that, some other deactivation mechanisms such as the loss of boron oxide,[34] and the melting and agglomeration of boron oxide[35] have been reported. Regeneration of catalyst is always a major concern in heterogeneous catalysis. One of the most effective methods is recalcining the used catalysts up to 500°C to remove the adsorbed organic compounds. This method is used in the recycling of the boria-zirconia samples. From the X-ray diffractograms (Fig. 4.10), both the fresh and the recycled sample were X-ray amorphous. No zirconia or boria phase could be seen, indicating no agglomeration of boron oxide occurred. The surface area, pore volume and pore size of the fresh and recycled samples remained quite constant (Table 4.11). The average acid sites determined by ammonia TPD also showed no significant change (Table 4.12). Therefore, any organic compounds or basic products which were possibly adsorbed at the surface of the catalysts had been completely removed by recalcination of the catalysts at 500°C. From the catalytic activity, the indole conversions obtained after 8 h reaction for the fresh and recycled samples were similar (Fig. 4.11). Thus, recalcination of the 121 used boria-zirconia to 500°C effectively removed any adsorbed organic products, allowing the textural properties, acidity, and catalytic activity to be retained. Compared with Nb2O5/Zr(OH)4 catalysts, the recycling of boria-zirconia catalysts is much easier. Only a single and simple procedure is required to fully recover the properties and activities of the catalysts. 600 500 Intensity (cps) 400 300 (b) 200 100 (a) 0 0 10 20 30 40 50 60 70 2θ (°) Fig. 4.10 X-ray diffratograms of (a) fresh 5% boria-zirconia (b) 5% boria-zirconia recycle 2 Table 4.11 Textural properties of fresh and recycled boria-zirconia Pore volume Average pore Sample SBET (m2/g) (cm3/g) diameter (nm) 5% boria-zirconia (500°C) 195 0.24 3.5, 4.7 5% boria-zirconia (recycle 1) 187 0.25 3.5, 4.7 5% boria-zirconia (recycle 2) 186 0.24 3.5, 4.7 122 Table 4.12 Acidity determined by NH3 TPD Sample Name Weight Total acid sites (g) (mmol) Average acid sites (mmol/g) (m-2) 5% boria-zirconia (500°C) 0.250 0.054 0.22 6.49 × 1017 5% boria-zirconia (recycle 1) 0.224 0.044 0.20 6.31 × 1017 5% boria-zirconia (recycle 2) 0.298 0.068 0.23 7.35 × 1017 40.00 35.00 Conversion (%) 30.00 25.00 20.00 15.00 10.00 5.00 0.00 0 1 2 3 4 5 6 7 8 Reaction time (h) Fig. 4.11 Conversion of indoles over (♦) 5% boria-zirconia fresh (▲) 5% boria-zirconia recycle 1 (■) 5% boria-zirconia recycle 2 4.5 Conclusion Boria-zirconia catalysts with boria loadings 3 to 40 wt.% were prepared by wet impregnation of boric acid on the hydrous zirconia support followed by calcination. At boria loadings below 25 wt.% , the catalysts had high surface areas. Aggregation of boron oxide started from boria loading of 25 wt.%, with more than three overlayers formed on the surface of zirconia support. The first layer of boron oxide was composed of tetrahedral oxygen-coordinated BO4 units whereas at boria 123 loading higher than a monolayer, trigonal BO3 units were formed. The Fischer indole reaction of phenylhydrazine with 3-heptanone was achieved with high conversions. The most active catalysts had loadings of 5 – 10 wt.% B2O3, which is close to a monolayer coverage. This result is similar to that of Nb2O5/Zr(OH)4 samples where catalysts with niobium oxide approaching a monolayer showed the highest activity. The optimal calcination temperature for the catalysts was 500°C. The used catalyst can be simply regenerated by recalcining at 500°C, with textural properties, acidity, and activity fully recovered. 124 References [1] L. Forni, G. Fornasari, C. Tosi, F. Trifirό, A. Vaccari, F. Dumeignil, J. Grimblot, Appl. Catal. A 2003, 248, 47 [2] S. Sato, M. Kuroki, T. Sodesawa, F. Nozaki, G.E. Maciel, J. Mol. Catal. A 1995, 104, 171 [3] A. Delmastro, G. Gozzelino, D. Mazza, M. Vallino, G. Busca, V. Lorenzelli, J. Chem. Soc., Faraday Trans. 1992, 88, 2065 [4] B. –Q. Xu, S. B. Cheng, S. Jiang, Q. M. Zhu, Appl. Catal. A 1999, 188, 361 [5] K. Tanabe, M. Misono, Y. Ono, in New solid acids and bases : their catalytic properties, Stud. Surf. Sci. Catal. (Eds. K. Tanabe, M. Misono, Y. Ono, H. Hattori) 1989 (Kodansha: Tokyo) [6] T. Yamaguchi, Catal. Today 1994, 20, 199 [7] K. Arata, Appl. Catal. A 1996, 146, 3 [8] M. Hinto, K. Kato, K. Arata, in Acid-base catalysis II : proceedings of the International Symposium on Acid-base Catalysis II, Stud. Surf. Sci. Catal. (Eds. H. Hattori, M. Misono, Y. Ono) 1994, vol. 90 (Kodansha: Tokyo; Elsevier: Amsterdam, New York) [9] H. Sato, K. Urabe, Y. Izumi, J. Catal. 1986, 102, 99 [10] H. Sato, S. Hasebe, H. Sakurai, K. Urabe, Y. Izumi, Appl. Catal. 1987, 29, 107 [11] T. Curtin, J. B. McMonagle, B. K. Hodnett, Appl. Catal. 1992, 93, 75 [12] T. Curtin, J. B. McMonagle, B. K. Hodnett, Appl. Catal. 1992, 93, 91 [13] K. M. Malshe, P. T. Patil, S. B. Umbarkar, M. K. Dongare, J. Mol. Catal. A: Chem. 2004, 212, 337 [14] B. –Q. Xu, S. –B. Cheng, X. Zhang, S. –F. Ying, Q. –M. Zhu, Chem. Commun. 2000, 13, 1121 125 [15] B. –Q. Xu, S. –B. Cheng, X. Zhang, Q. –M. Zhu, Catal. Today 2000, 63, 275 [16] T. Curtin, J. B. McMonagle, M. Ruwet, B. K. Hodnett, J. Catal. 1993, 142, 172 [17] D. –S. Mao, G. –Z. Lu, Q. –L. Chen, Z. –K. Xie, Y. –X. Zhang, Catal. Lett. 2001, 77, 119 [18] D. –S. Mao, G. –Z. Lu, Q. –L. Chen, Appl. Catal., A 2004, 263, 83 [19] D. –S. Mao, G. –Z. Lu, Q. –L. Chen, Appl. Catal., A 2005, 279, 145 [20] D. –S. Mao, G. –Z. Lu, Q. –L. Chen, J. Mol. Catal. A 2005, 240, 164 [21] K. M. Malshe, P. T. Patil, S. B. Umbarkar, M. K. Dongare, J. Mol. Catal. A 2004, 212, 337 [22] G. K. Chuah, S. Jaenicke, B. K. Pong, J. Catal. 1998, 175, 80 [23] S. Jaenicke, G. K. Chuah, V. Raju, Y. T. Nie, Catal. Surv. Asia 2008, 12, 153 [24] M. W. Simon, S. S. Nam, W. Q. Xu, S. L. Suib, J. C. Edwards, C. L. O'Young, J. Phys. Chem. 1992, 96, 6381 [25] J. Zhao, Y. Yue, W. Hua, Z. Gao, Catal. Lett. 2007, 116, 27 [26] B. –Q. Xu, Acta Phys. Chim. Sin. 2001, 17, 443 [27] C. E. Weier, J. Res. Natn. Bur. Stand. 1966, 70A, 153 [28] J. L. Parsons, M. E. Milberg, J. Am. Chem. Soc. 1960, 43, 326 [29] D. –S. Mao, G. –Z. Lu, J. Solid State Chem. 2007, 180, 484 [30] T. Hatakeyama, Z. –H. Liu, in Handbook of Thermal Analysis (Eds. T. Hatakeyama, Z. –H. Liu) 1998 (Wiley: New York) [31] J. Livage, K. Dio, C. Mazires, J. Am. Ceram, Soc. 1968, 51, 349 [32] M. Sorrentino, L. Steinbrecher, F. Hazel, J. Colloid Interface Sci. 1969, 31, 307 [33] M. Hino, K. Arata, J. Chem. Soc., Chem. Commun. 1988, 1259 126 [34] T. Ushikubo, K. Wada, J. Catal. 1994, 148, 138 [35] S. Sato, K. Urabe, Y. Izumi, J. Catal. 1986, 102 127 [...]... distribution (a) hydrous zirconia supported niobium oxide catalyst with Nb2O5 loadings from 0 – 40 wt.%, drying temperature 100°C (b) MCM-41 and 25% Nb2O5/MCM-41 73 TGA curves of hydrous zirconia , niobium oxide and 10 – 40 wt.% hydrous zirconia supported niobium oxide catalysts 75 Pyridine IR of 25% Nb2O5/Zr(OH)4 after evacuation at (a) 25°C (b) 100°C and (c) 200°C 80 The Fischer indole reaction of... 30% (+) 40% boria -zirconia 119 Conversion of indoles over 25 wt.% boria -zirconia (♦) uncalcined (■) calcined at 500°C and (▲) calcined at 650°C 120 Fig 3.12 Fig 4.2 Fig 4.3 Fig 4.4 Fig 4.5 Fig 4.6 Fig 4.7 Fig 4.8 Fig 4.9 xiii Fig 4.10 Fig 4.11 X-ray diffratograms of (a) fresh 5% boria -zirconia (b) 5% boria -zirconia recycle 2 Conversion of indoles over (♦) 5% boria -zirconia fresh (▲) 5% boria -zirconia. .. of indole[ 50] 1.2.2 Fischer Indole Synthesis For over a century, the Fischer indole reaction has remained as an extremely useful and important method for the synthesis of a variety of biologically active indole- structure compounds.[51] The Fischer indole reaction is one of the simplest and most efficient methods by which enolizable N-arylhydrazones are easily transformed into indoles In the Fischer indole. .. (e) 20% (f) 25% (g) 30% and (h) 40% boria -zirconia 106 TG–DTA diagrams of (a) hydrous zirconia (b) 3% (c) 5% (d) 10% (e) 20% (f) 25% (g) 30% and (h) 40% boria -zirconia 112 Ammonia TPD profiles of (a) 3% (b) 5% (c) 10% (d) 20% (e) 25% (f) 30% (g) 40% boria -zirconia (calcination temperature 500°C) and (h) 25% boria -zirconia (calcination temperature 650°C) 114 Conversion of indoles over (♦) 3% (■) 5% (▲)... Nb2O5/Zr(OH)4 catalysts , Nb2O5 and Zr(OH)4 after drying at (…) 100°C, ( ) 200°C, and („) 300°C 88 The Fischer indole synthesis of 1,2,3,4-tetrahydrocarbazole over (♦) hydrous zirconia, and (□) 10% (■) 20% (×) 30% (▲) 40% Nb2O5/Zr(OH)4 87 Fig 3.13 Catalyst leaching test 90 Fig 3.14 Regeneration of 25 wt.% Nb2O5/Zr(OH)4 91 Fig 4.1 X-ray diffractograms of boria -zirconia catalysts with boria loading from... solid base catalysts 10 Scheme 1.5 Synthesis of ethyleneimine (EI) from monoethanolamine (MEA) 11 Scheme 1.6 Drugs containing indole structures 18 Scheme 1.7 Resonance structures of indole 19 Scheme 1.8 Mechanism of Fischer indole synthesis 20 Scheme 1.9 Fischer indole synthesis via Japp-Klingemann reaction 21 Scheme 1.10 Various routes for synthesis of substituted indole 22 Scheme 1.11 Indole synthesis. .. 3heptanone over (+) niobium oxide, (■) hydrous zirconia, and (×) 10% (□) 20% (▲) 25% (♦) 30% (○) 40% Nb2O5/Zr(OH)4 82 The Fischer indole reaction of phenylhydrazine with 3heptanone over (■) 25% Nb2O5/Zr(OH)4 coppt (×) H-beta (Si/Al = 12.5) (▲) 25% Nb2O5/MCM-41 (♦) 25% xii Nb2O5/Zr(OH)4 82 Fig 3.10 Linear and bulky indole products with molecular dimensions in Å 84 Fig 3.11 Conversion to indole products over. .. synthesis of indole was in 1866 and later in 1883, the Fischer indole synthesis was discovered, which is still the most versatile method for preparing indoles nowadays.[48] Every year, several thousands of indoles and indole derivatives have been synthesized in the research laboratories The sustained interest in indoles is due to their wide range of biological activity.[49] The indole ring structures in... 500°C 100 X-ray diffractograms of 25 wt.% boria zirconia at calcination temperature 100°C, 500°C, and 650°C 100 Nitrogen adsorption/desorption curves and pore volume distribution of boria -zirconia catalysts with different boria loadings 105 Nitrogen adsorption/desorption curves and pore volume distribution of 25 wt.% boria -zirconia sample calcined at 100°C, 500°C, and 650°C 106 Infrared spectra of (a)... on zirconia surface, this supported catalyst exhibited activity 10 times higher than that of the original perovskite Zirconia was also found to be a superior support for Re and Rh in the hydrogenation of CO2.[45-47] 1.2 Synthesis of Indoles 1.2.1 Introduction For the last 100 years, the synthesis of indoles and their derivatives have been a research topic of great interest The first synthesis of indole ... the Fischer indole reaction over two different solid acid catalysts, hydrous zirconia supported niobium oxide and boria-zirconia catalysts, were investigated Hydrous zirconia supported niobium oxide. .. Material 15 1.2 Synthesis of Indoles 18 1.2.1 Introduction 18 1.2.2 Fischer Indole Synthesis 19 1.2.3 Other Methods of Indole Synthesis 21 1.2.4 Catalysts Used in Fischer Indole Synthesis 23 1.3... and temperature programmed desorption of ammonia (NH3 TPD) (3) Catalytic testing of the hydrous zirconia supported niobium oxide catalysts and boria-zirconia catalysts in the Fischer indole synthesis