UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Synthesis and characterization of thermally stable acidic forms of zeolite A Sawada, James Alexander 2000

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2000-566161.pdf [ 7.54MB ]
Metadata
JSON: 831-1.0059649.json
JSON-LD: 831-1.0059649-ld.json
RDF/XML (Pretty): 831-1.0059649-rdf.xml
RDF/JSON: 831-1.0059649-rdf.json
Turtle: 831-1.0059649-turtle.txt
N-Triples: 831-1.0059649-rdf-ntriples.txt
Original Record: 831-1.0059649-source.json
Full Text
831-1.0059649-fulltext.txt
Citation
831-1.0059649.ris

Full Text

Synthesis and Characterization of Thermally Stable Acidic Forms of Zeolite A by J A M E S A L E X A N D E R S A W A D A B.Sc. (Honours) The University of Waterloo, 1994 A THESIS S U B M I T T E D IN P A R T I A L F U L F I L L M E N T OF T H E R E Q U I R E M E N T S F O R T H E D E G R E E OF D O C T O R OF P H I L O S O P H Y in T H E F A C U L T Y OF G R A D U A T E S T U D I E S (Department of Chemistry) We accept this thesis as conforming To the required standard T H E U N I V E R S I T Y OF B R I T I S H C O L U M B I A July 2000 © James Alexander Sawada, 2000 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. The University of British Columbia Vancouver, Canada Date I T Ifr 7. find DE-e (2/88) Abstract Zeolite A has several attributes which make it attractive as a solid acid catalyst. The framework has pores on the order of 4-5 A which show a strong selectivity toward small molecules by excluding species with larger diameters than linear hydrocarbons. The pores intersect to form a 3-dimensional channel network which is advantageous as it allows for free diffusion throughout the entire structure. Furthermore, zeolite A is already commercially available and is inexpensive to produce. Acidic forms of zeolite A , however, have long been considered unachievable as a result of the reduced stability of the zeolite A framework compared to those of more siliceous zeolites. This thesis describes the preparation and characterization of a series of zeolite A materials which demonstrate both thermal stability and Bronsted acidity. Zeolite A materials were prepared whose frameworks contained both C a 2 + and N F L * ions. The three materials studied had C a 2 + : N H 4 + ratios of 70:30, 60:40, and 50:50. The materials were systematically examined using powder X-ray diffraction, electron microscopy, solid-state N M R , and gas physisorption to establish their homogeneities and thermal stabilities. Finally, catalytic testing of the materials for the conversion of methanol to olefins was carried out to establish their activity in a model acid-catalyzed reaction. n Table of Contents Abstract »' Table of Contents iii List of Figures vi List of Tables >* Symbols and Abbreviations x Acknowledgements xi Chapter 1 General Introduction 1 1.1 STRUCTURE A N D COMPOSITION OF ZEOLITES 1 /././ Zeolite A 4 1.1.2 Zeolites X and Y 6 1.2 SUMMARY 7 1.3 TECHNIQUES FOR ZEOLITE CHARACTERIZATION 8 1.3.1 Powder X-ray Diffraction 8 1.3.2 Nuclear Magnetic Resonance Spectroscopy of Solids 10 1.3.2.1 The Zeeman Interaction 12 1.3.2.2 Chemical Shift Interaction 13 1.3.2.3 Dipolar Interaction 16 1.3.2.4 Scalar Coupling Interaction 16 1.3.2.5 Quadrupolar Interaction 16 1.3.2.6 High Power Decoupling 17 1.3.2.7 Magic Angle Spinning.. 18 1.3.3 Physisorption 21 1.3.3.1 Surface Area Calculations 24 1.3.3.2 Pore Size Calculations 25 1.4 RESEARCH OBJECTIVE 27 1.5 REFERENCES 28 Chapter 2 Materials and Methods 31 2.1 SYNTHESIS OF ZEOLITE A 31 2.1.1 Preparation of NaA 34 2.2 ION E X C H A N G E OF ZEOLITE A 35 2.2.1 Calcium-Exchanged Zeolite A 37 2.2.2 Ammonium-Exchanged Zeolite A 37 iii 2.3 SOLID-STATE ION E X C H A N G E 38 2.3.1 Preparation of Mixed Ca2+/NH4+ Materials 39 2.4 CHARACTERIZATION 40 2.4.1 Powder X-ray Diffraction 40 2.4.2 Scanning Electron Microscopy 41 2.4.3 Room-Temperature Hexane Adsorption 41 2.4.4 High-Temperature Adsorption of Small Molecules 44 2.4.5 Physisorption 44 2.4.6 Magic-Angle Spinning Nuclear Magnetic Resonance Spectroscopy 47 2.4.7 Catalytic Activity Tests 48 2.5 SIMULATED POWDER DIFFRACTION PATTERNS 49 2.6 REFERENCES 51 Chapter 3 Characterization of the Zeolite Starting Materials 54 3.1 INTRODUCTION 54 3.1.1 SEM. 55 3.1.2 Powder XRD : 55 3.1.3 29Si MAS NMR 58 3.1.4 Nomenclature 59 3.2 AS-SYNTHESIZEDNaA 63 3.3 A S - E X C H A N G E D C a A A N D N H 4 A 66 3.4 CHARACTERIZATION OF THE M I X E D MATERIALS 66 3.5 REFERENCES 75 Chapter 4 Activation and Thermal Stability of the Zeolite A Materials 76 4.1 INTRODUCTION 76 4.2 T H E R M A L STABILITY OF N H 4 A 77 4.3 T H E R M A L STABILITY OF C a A 80 4.3.1 High Temperature Transformation of CaA at 825 °C 80 4.3.2 Thermal Stability at 500 °C 84 4.4 T H E R M A L STABILITY OF T H E M I X E D MATERIALS 85 4.4.1 Structural Characterization 87 4.4.2 Sorption Characteristics 102 4.5 SUMMARY 107 iv 4.6 REFERENCES 108 Chapter 5 Characterization of the Acidic Sites in the Activated Mixed Materials 110 5.1 INTRODUCTION 110 5.2 CHARACTERIZATION OF A C I D SITES IN C a 2 + / N H 4 + ZEOLITE A 111 5.3 SUMMARY 123 5.4 REFERENCES 124 Chapter 6 Exploratory Catalytic Testing of the Mixed Materials 125 6.1 INTRODUCTION 125 6.2 C A T A L Y T I C ACTIVITY STUDY FOR THE CRACKING OF « - H E X A N E 126 6.3 PRELIMINARY C A T A L Y T I C STUDY OF THE CONVERSION OF M E O H TO HYDROCARBONS BY T H E ZEOLITE A MATERIALS 131 6.4 CONCLUSIONS 138 6.5 REFERENCES 140 Chapter 7 Conclusions and Suggestions for Further Work 141 Appendix A: Physisorption Data 147 v List of Figures Figure 1.1 Three commercially important zeolite structure types 2 Figure 1.2 The framework structure of zeolite A 5 Figure 1.3 The effect of path length on the relative phase of two incoming X-rays 9 Figure 1.4 The geometry of a typical Bragg-Brentano diffractometer 10 Figure 1.5 Energy diagram for the Zeeman interaction for a spin =1/2 nucleus 13 Figure 1.6 Schematic representation of the 1 3 C N M R adsorption spectrum of a carbonyl functionality in a single crystal, a powder, and in solution 15 Figure 1.7 Effect of the quadrupolar interaction on Zeeman energy levels of a spin 5/2 nucleus 17 Figure 1.8 Definitions of the angles used to describe the orientation of the 1-S internuclear vector as it is rotated about an axis inclined at the magic angle 20 Figure 1.9 The five isotherm classifications according to B D D T 22 Figure 2.1 Experimental setup for the preparation of N a A 34 Figure 2.2 Schematic of T G A setup for vapour adsorption 42 Figure 2.3: Deviation of an isotherm due to slow outgassing 45 Figure 2.4 L o w pressure region of an isotherm collected using A r at 87 K 46 Figure 2.5 Reactor design for the catalytic testing of the zeolite A materials 50 Figure 3.1 Powder patterns of a sample of 100%CaA and an equal mixture of C a A and amorphous material 57 Figure 3.2 Possible Si connectivities to A l in an aluminosilicate zeolite framework 58 Figure 3.3 The five possible local environments of a Si atom in a zeolite framework and the corresponding chemical shift ranges 60 Figure 3.4 S E M image of as-synthesized N a A and commercial N a A 63 Figure 3.5 Powder X R D pattern of as-synthesized N a A 64 Figure 3.6 Powder X R D pattern of as-synthesized N a A and the simulated powder pattern of hydrated N a A 65 vi Figure 3.7 S E M images of as-exchanged C a A and as-exchanged N H 4 A 66 Figure 3.8 Powder X R D patterns for as-synthesized N a A , as-exchanged C a A , and as-exchanged N H 4 A 67 Figure 3.9 Overlaid powder X R D patterns of C a A and N H 4 A 68 Figure 3.10 Powder X R D pattern of a 50:50 mixture of C a A and N H 4 A after mixing 69 Figure 3.11 Equilibration of a 50:50 mixture of as exchanged C a A and as-exchanged N H 4 A at various times 70 Figure 3.12 Powder X R D patterns of as-mixed 50%Ca, 60%Ca, and 70%Ca zeolite A 72 Figure 3.13 Si N M R spectra of the as-synthesized, as-exchanged, and as-mixed zeolite A materials 74 Figure 4.1 Powder X R D patterns illustrating the thermal behaviour of N H 4 A 78 Figure 4.2 Powder X R D pattern of N H 4 A evacuated overnight at 100 °C 79 Figure 4.3 Surface area of C a A as a function of time at 825 °C 81 Figure 4.4 Powder X R D patterns of C a A exposed to 825 °C for various times 82 Figure 4.5 Room temperature powder X R D patterns of as-exchanged C a A and C a A activated at 500 °C and rehydrated for 1 day in air 84 Figure 4.6 Powder X R D patterns of the activated mixed materials rehydrated in air 88 Figure 4.7 Partial powder X R D patterns of as-mixed and activated/rehydrated samples of 50%Ca 90 Figure 4.8 2 9 S i and 2 7 A 1 solid-state M A S N M R spectra of the activated materials 94 Figure 4.9 Deconvolution results of the activated/rehydrated 50%Ca material 96 Figure 4.10 208 M H z 2 7 A 1 spectra of activated N H 4 A and 50%Ca material 100 Figure 4.11 Adsorption of «-hexane at 30 °C on the 50%Ca material 104 Figure 4.12 Pore volume differential plots for the activated materials 106 Figure 5.1 ' H chemical shift ranges for framework protons in zeolites 111 Figure 5.2 Room temperature ! H M A S N M R spectra of the as-mixed mixed materials .... 112 vi i Figure 5.3 Room temperature ' H M A S N M R spectra of the activated/rehydrated mixed materials 114 Figure 5.4 ' H M A S N M R spectrum of N H 4 A activated at 500 °C 115 Figure 5.5 Room temperature ' H M A S N M R spectra of the activated/re-ammoniated/ rehydrated materials 116 Figure 5.6 Powder X R D patterns of the as-mixed, activated, and re-ammoniated 50%Ca material , 118 Figure 5.7 Powder X R D patterns of the as-mixed, activated, and re-ammoniated 60%Ca material 119 Figure 5.8 Powder X R D patterns of the as-mixed, activated, and re-ammoniated 70%Ca material 120 Figure 5.9 Overlay of a section of the powder patterns of samples of the as-mixed, activated, and re-ammoniated 50%Ca material 122 Figure 6.1 n-Hexane uptake at various temperatures for the 70%Ca material 130 Figure 6.2 Methanol uptake at ca. 410 °C for the 70%Ca material 133 Figure 6.3 Chromatograms for the M T O reaction at 400 °C after 5 minutes on stream for the 70%Ca material and the methanol bypass sample 135 vi i i List of Tables 1 3 * Table 1.1 Typical magnitudes for the nuclear spin interactions for C in a 4.7 Tesla magnetic field 12 Table 1.2 Important N M R parameters of selected nuclei 14 Table 3.1 Electron microprobe results for zeolite starting materials 61 Table 3.2 Ideal and calculated C a 2 + contents for the as-exchanged and as-mixed materials 62 29 Table 4.1 Contributions of the various Si(nAl) sites and amorphous material to the Si spectra of the activated mixed materials 97 Table 4.2 Octahedral and tetrahedral contributions to the 2 7 A 1 spectra of the activated mixed materials 98 27 Table 4.3 Contributions of the various aluminum environments to the 208 M H z A l spectra of samples of activated N H 4 A and 50%Ca 101 Table 4.4 rc-Hexane adsorption on zeolite A materials at 1 atm and 30°C 103 Table 4.5 Surface areas of the activated materials from argon physisorption 105 Table 6.1 Retention times of the various product species from the cracking of «-hexane. . 127 Table 6.2 Catalytic activity results and cc-values for the cracking of rc-hexane over the various zeolite A materials 128 Table 6.3 Catalytic activity results and a-values for the cracking of rc-hexane over the various zeolite A materials, courtesy of Mob i l O i l Corporation 129 Table 6.4 Percent conversion of M e O H to hydrocarbons over the mixed zeolite A materials and reference materials 136 Table 6.5 Product distributions from the conversion of methanol to hydrocarbons over the zeolite A materials and reference materials 137 Table A . 1 Argon physisorption data at 87 K for the zeolite A materials 147 ix ( Symbols and Abbreviations A Angstrom B 0 Applied external magnetic field d Distance between two crystal planes 5 Chemical shift of a nucleus E D X Energy dispersive X-ray y Gyromagnetic ratio ( N M R ) , surface tension (physisorption) h Planck's constant IR Infrared X Wavelength M A S Magic angle spinning N M R Nuclear magnetic resonance P Pressure P /P 0 Partial pressure P 0 Saturation pressure ppm Parts per mil l ion Q Quadrupolar coupling constant 0 Angle between the incident and reflected X-ray beam and the crystal plane 9 m Magic angle R Retention time r.f. Radio frequency S E M Scanning electron microscopy T Tesla T G A Thermal gravimetric analysis T M S Tetramethyl silane u 0 Larmour frequency of a nucleus in an applied magnetic field cor Rotor frequency W Weight of the adsorbed gas phase W m Weight of a monolayer of adsorbed gas X R D X-ray diffraction Z Atomic number x Acknowledgements I would like to acknowledge my supervisor, Dr. Col in Fyfe, for his assistance and direction in this project and I extend my appreciation to the members of the Fyfe research group for their assistance and input over the years. I am indebted to the Electronic and Mechanical Engineering services as their generous support greatly facilitated many of my experiments. I would like to thank my wife, Shelley, for her patience, support, and understanding throughout this work and my parents for their commitment to my higher education. x i Chapter 1 General Introduction 1.1 Structure and Composition of Zeolites Historically, zeolites have been defined as crystalline aluminosilicate materials having open framework structures but, more recently, the classification of a zeolite has been expanded to include the vast array of materials that are structurally similar to zeolites but contain different component atoms. Any three-dimensional network structure composed of tetrahedrally coordinated atoms linked together through oxygen bridges having a framework density less than 21 tetrahedral atoms/1000 A 3 can be considered zeolite-like.[l] A s such, pure silicates, aluminophosphates (A1PO), silicoaluminophosphates (SAPOs), metal-containing aluminophosphates (MeAPO) , gallophosphates, and so on are all considered as zeolites. For a comprehensive compilation of zeolites and their properties the reader is directed to the texts by Barrer,P] Breck ,P] Dyer,[4] and Englehardt and Michel.[5] Zeolites have a unique combination of properties which has allowed them to gain widespread use in adsorbent and catalytic processes. Zeolites are distinct materials by nature of their uniform micropore-size distributions, high surface areas, ion-exchange capabilities with the potential for internal acidity (in some cases), and often high thermal stabilities. Because zeolite A is the subject of the present study, the following discussion of zeolite characteristics is restricted to aluminosilicate zeolites and their properties. Framework diagrams are often used to portray zeolite structure types because they offer a clear depiction of the various features of the zeolite. The framework structures of three important zeolites are shown in Figure 1.1. Each vertex in the illustrations is occupied by either a silicon or aluminum and the bridging framework oxygens have been omitted for clarity. Zeolite structures are composed entirely of corner sharing tetrahedra which join 1 together in such a way that a 3-dimensional framework structure is formed containing a network of interconnected pores and channels. The characteristic high surface areas of zeolites result from the open framework structures in which (in some cases) every tetrahedron is exposed as part of a surface. (a) (b) (c) Figure 1.1 Three commercially important zeolite structure types, (a) zeolite A, (b) ZSM-5 and (c) faujasite The large openings in the framework structures define the pores of the zeolite framework. The pores determine the largest molecule that can diffuse into the framework and are referred to as n-membered rings where n is the number of tetrahedra that delineate the opening. The array of identical pores allow the materials to act as very discriminating molecular sieves. For example, n-paraffins can pass through an 8-membered ring but iso-paraffins are excluded. It is apparent from the framework diagrams that 4-, 5-, and 6-membered rings are also present in zeolite structures. These rings are not significant because, when the framework oxygens are taken into account, the diameter of the 6-membered rings excludes all but the smallest molecules such as water and the 4-and 5-membered rings have no actual aperture at all. Though many distinct framework structures are known for aluminosilicate zeolites, each wi l l contain either 8-, 10-, or 12-membered rings and in some cases more than one type. 2 Typically, the pore sizes range from 3.5-4.5 A for 8-rings, 4.5-6 A for 10-rings, and 6-8 A for 12-rings.[6] The composition of zeolites can be represented by the general formula: M n + x / n (A10 2 )x (Si0 2 )yw(H 2 0) The water is contained within the void space of the structure and can typically be reversibly removed and re-adsorbed without damaging the framework. The silicon to aluminum ratio (y/x) is a significant aspect of the framework composition as it controls, in part, the thermal stability and catalytic activity of the material. The presence of aluminum in a silicate lattice imparts a net negative charge to the framework which is balanced by the cations that reside in the framework. The cations (typically alkali metal cations) are usually exchangeable and can be replaced by a wide variety of alkali, transition metal, inorganic, and organic cations. Ion exchange is also usually the first step toward generating acidic zeolites. When the resident cation is exchanged by the ammonium ion and the material heated to high temperatures, ammonia is liberated and protons remain to charge-balance the framework. The residual protons exhibit Bronsted acidity; a discovery which revolutionized solid-acid catalysis. A scheme for producing an acidic zeolite through ion exchange and activation (via thermal treatment) is shown below. NaZ + N H 4 + ^ ^ N H 4 Z + Na + NH 4Z — • HZ + NH 3(g) Zeolites are superior solid-acid catalysts not only because they are more acidic than conventional amorphous aluminosilicates but because the acid sites are contained inside the frameworks. This property allows the fixed dimensions of the frameworks' pores and channels to exert an influence on a reaction by limiting the maximum size of molecule that 3 can diffuse in or out of the system. This property has useful consequences in many acid-catalyzed reactions because it allows some control over the product distributions. U] 1.1.1 Zeolite A The synthesis and characterization of zeolite A was reported in 1956L&] by researchers at the Linde A i r Product division of Union Carbide. No naturally occurring analogue to the zeolite A structure exists but isostructural materials have since been prepared with different S i / A l ratios and different framework compositions (aluminogermanate, gallophospate, and silicoaluminophosphatet^ ]). The framework and channel structure of zeolite A is shown in Figure 1.2 with one portion of the framework highlighted. This portion is called a sodalite cage and may be referred to when discussing the locations of the various species inside the zeolite framework. The access to the sodalite cages is controlled by the 6-membered rings which, as previously mentioned, only very small molecules such as water can penetrate. A representation of the channel system of zeolite A formed by the 8-membered rings is shown beside the framework structure in Figure 1.2. The intersecting channel system of zeolite A is ideal because the three dimensional array of intersecting channels allows for diffusion of species through the entire structure even i f some of the passages are obstructed. A l l species, other than water, that are adsorbed into the framework wi l l be located in this channel system. The zeolite A framework from a conventional synthesis consists of strictly alternating silica and alumina tetrahedra yielding a S i / A l = 1, the maximum allowed by Lowenstein's rule[9] which forbids A l - O - A l linkages in aluminosilicates. Consequently, the framework of zeolite A also contains the largest possible amount of charge-balancing cations. This property 4 Figure 1.2 The framework structure of zeolite A with a sodalite cage highlighted and a representation of the channel system formed by the interconnecting 8-membered rings. makes zeolite A particularly attractive as an ion-exchange agent because of the quantity of exchangeable cations per mass of zeolite. Further, the pore volume of zeolite A , at 0.45 cm /cm [3] is among the highest of all the zeolites and gives the material a clear advantage where adsorption processes are concerned. Zeolite A materials can be prepared with a range of useful pore sizes through ion exchange. The as-synthesized material (NaA) contains N a + ions in the framework and admits molecules with a diameter of 4 A , the C a 2 + exchanged material (CaA) is capable of adsorbing molecules as large as 5 A in diameter, while the K + exchanged form of zeolite A ( K A ) cannot admit molecules larger than 3 A . These materials are sold commercially as type 4A, 5A, and 3A molecular sieves respectively. The framework aluminum, however, is a source of framework instability as the aluminum is subject to attack by acids and water vapour at high temperatures; conditions which are prevalent in commercial catalytic processes. The zeolite A structure contains the maximum amount of framework aluminum possible and the dealumination that the material suffers as a result of exposure to steam or acids destroys the framework The sensitivity of 5 the framework toward acids extends to internal acidity; the ammonium-exchanged form of zeolite A (NH4A) decomposes upon heating to only moderate temperatures. The thermal and hydrothermal stabilities of zeolite A are influenced by the resident charge-balancing cation and enhanced stabilities are found when the material is exchanged with Ca ions. [3] 1.1.2 Z e o l i t e s X a n d Y Naturally occurring, more siliceous zeolites such as mordenite (S i /A l = 5) were known in the 1950s to have superior thermal and acid stabilities which provided an impetus to develop synthetic zeolites with higher S i / A l ratios. [10] Zeolite X , developed shortly after zeolite A , has the framework topology of the rare mineral faujasite (see Figure 1.1) and can be prepared with a range of S i / A l ratios between 1.1 and 1.25.[11] The pore volume of the material is comparable with zeolite A but the 10-membered rings have the capacity to admit large hydrocarbons which allowed the material to be used as a cracking catalyst. The larger S i / A l ratio slightly enhances the thermal stability of zeolite X but the hydrothermal stability is still poor and the material decomposes when exposed to steam at 350 °C.[3] Zeolite Y , introduced in 1959, shares the same framework structure as zeolite X but can be prepared with a S i / A l ratio between 1.5 and 2.5. A s a result, zeolite Y shows vastly superior thermal and hydrothermal stabilities compared to the more aluminous zeolites and shows a greater activity in catalytic applications. It is possible to further enhance the stability of the zeolite Y framework by steaming the material (to remove framework aluminum) and increasing the framework S i / A l ratio producing so-called ultrastable Y (or USY)[12] catalyst current used in crude oil cracking. 6 1.2 S u m m a r y The preparation of zeolites with successively larger S i / A l ratios steered interest toward the large-pore, thermally robust zeolite Y and interest in the less siliceous zeolites waned. A s a result, the largest commercial uses of zeolite A have been as a desiccant, a detergent builder[13] (to soften water by exchanging the C a 2 + and M g 2 + in the water for Na + ) , and as a molecular sieve in such applications as the commercial ISOSIV process used to separate iso-and n-paraffinsl 1^] Zeolite A isostructures with higher S i / A l ratios can be synthesized using quaternary ammonium salts (ZK-4, N - A H ' 3]) and it has been demonstrated in this laboratory that a completely siliceous form of zeolite A can be generated by exposing the zeolite Z K - 4 to elevated temperatures in the presence of water vapour; a process commonly known as steaming. Highly siliceous forms of zeolite A would also be of interest as catalyst supports for transition metals. While it is possible to generate highly siliceous forms of zeolite A starting from exotic forms, the starting materials are prohibitively expensive. A s a result of its reduced hydrothermal stability, zeolite A is best suited to small molecule reactions at low temperatures. Various ion-exchanged forms of zeolite A have been tested in an array of catalytic reactions involving organic reactants such as the dehydration of alcohols, the conversion of alcohols to haloalkanes, and the formation of simple amines from alcohols. The oxidation of C O and H 2 S using 0 2 have been tried over N a A as has the conversion of ammonia to N 2 and H 2 and ozone to oxygen. [ 15] N o acid forms of zeolite A have been explored in detail. 7 1.3 Techniques for Zeolite Characterization 1.3.1 Powder X-ray Diffraction Owing to the microcrystalline nature of most synthetically prepared zeolites, powder X-ray diffraction is an essential characterization technique. While large crystals suitable for single-crystal X-ray diffraction studies can be synthesized for a few zeolites,[ 16-25] powder X-ray diffraction is a representative technique because applications using zeolite materials utilize the microcrystalline powders. X-ray diffraction is a scattering phenomenon where the crystals act as a diffraction grating toward an incoming beam of X-rays. The electrons surrounding the atoms in the crystals scatter the X-rays in all directions and only under specific conditions wi l l the scattered rays interfere constructively and produce a diffracted beam. Figure 1.3 illustrates the diffraction process by representing a crystal as a being composed of crystal planes separated by a distance, d, which are exposed to an incoming X-ray beam composed of perfectly parallel X-rays at an angle 9 to the crystal planes. It can be seen from the figure that a difference of 2a (equivalent to 2dsin9) exists between the path length of rays R and R' which causes the two reflected rays to exit the crystal with different relative phases. Unless special conditions are met, this causes the two rays to interfere deconstructively, producing a diffracted beam with diminished intensity. The net effect, summed over all the incident X -rays and the entire sample, is that no diffracted beam is observed because no phase coherence exists between the component rays. The condition for the incoming X-rays to produce a diffracted beam is given by the Bragg equationP6, 27] (Equation 1.1) where n represents the "order of reflection" which can be any integral value of 1 or greater, A, is the wavelength of the radiation, d is the inter-8 Figure 1.3 The effect of path length on the relative phase of two incoming X-rays atomic (or inter-layer) spacing, and 6 is the angle between the incident and reflected X-ray beam and the crystal plane. n^ = 2dsin8 (1.1) The Bragg equation indicates that for constructive interference to occur, the path difference between the two X-rays must be an integral number of wavelengths. When this condition is met, the component X-rays exist the crystal in phase and constructively interfere to produce a diffracted beam. Experimental powder X-ray diffraction patterns are collected as a function of 28 because of the geometry of the diffractometer. The detector must be positioned at an angle 28 from the incident X-ray beam in order to collect the diffracted beam reflected at an angle 6 from the sample surface. Figure 1.4 illustrates the configuration of a typical Bragg-Brentano diffractometer with a fixed X-ray source using sample and detector rotation. The detector is rotated at twice the rate of the sample to maintain the 6/28 arrangement. The collection and interpretation of powder diffraction patterns of zeolites are treated in subsequent chapters. 9 source Figure 1.4 The geometry of a typical Bragg-Brentano diffractometer 1.3.2 Nuclear Magnetic Resonance Spectroscopy of Solids High-resolution, solid-state N M R spectroscopy is a powerful technique for structure elucidation and is an invaluable tool for studying catalytic materials. The information about the local ordering gained through N M R spectroscopic investigations of a sample is complementary to the information regarding its long-range order obtained through powder X R D . The two techniques in combination provide a substantial amount of information on the structures and stabilities of zeolitic materials. The key difference between solution and solid-state N M R is that, in solution, the rapid, isotropic tumbling of the molecules averages the magnetic spin interactions. Certain interactions, such as the chemical shift and the scalar spin-spin coupling (J-coupling) are averaged to their isotropic values and are observed as sharp resonances in the spectrum. The isotropic values for the dipole-dipole and quadrupolar interactions are zero and are not directly observed in solution state N M R spectra. In solids, while limited motions may occur, the rapid isotropic tumbling necessary to average the various interactions does not occur and, 10 as such, the N M R spectra of solids are influenced to a much greater extent by the above interactions. In the solid state these interactions can produce severely broadened, relatively featureless spectra; obscuring the diagnostic chemical shift information. A nucleus in an applied magnetic field is involved in several different types of interactions: i) the Zeeman interaction of the nucleus with the applied magnetic field ii) the interaction of the nucleus with the surrounding electron density (chemical shielding) which affects the chemical shift of the nucleus iii) scalar coupling with other nuclei iv) dipolar coupling with other nuclei v) quadrupolar interactions for nuclei with spins greater than 1/2 The Hamiltonian that describes the total nuclear spin interaction is the sum of the individual interaction Hamiltonians and can be written as: [28] H = H z + Hcs + Hsc + Hoc + H Q (1.2) The relative importance of the different nuclear spin interactions which contribute to N M R spectroscopy in the solid state wi l l depend on the nucleus being observed, the applied magnetic field strength, and on the particular system being studied. Table 1.1 lists the various interactions and their magnitudes for C nuclei in solution and in the solid state. The approach of high-resolution solid state N M R spectroscopy has been to remove or average the characteristic solid-state interactions to simplify the spectrum and obtain isotropic values for structural investigations. For a more detailed treatment of the theory and practice of solid-state N M R spectroscopy, the reader is referred to the texts by Abragam,[29] 11 Engelhardt and Michel,[5] Ernst, Bodenhausen, and Wokaum,[30] FyfeJ^ l ] and Mehring.[32] Table 1.1 Typical magnitudes for the nuclear spin interactions for 1 3 C in a 4.7 Tesla magnetic field!3 3! Spin Interaction Magnitude in Solids Magnitude in Liquids Zeeman 50 M H z 50 M H z Chemical Shielding Up to 1 kHz Isotropic value Scalar Coupling ~ 15 kHz 0 Dipolar Coupling ~ 200 H z ~ 200 H z Quadrupolar Coupling Up to - 100 M H z 0 1.3.2.1 The Zeeman Interaction The Zeeman interaction is the basis of N M R spectroscopy and occurs for all nuclei possessing non-zero spin. It is usually far larger than the other interactions with the exception of some quadrupolar interactions. The interaction of the magnetic moment of the nucleus and the applied field, B 0 , yields 21 + 1 energy levels (where I is the nuclear spin quantum number) separated by an energy proportional to the gyromagnetic ratio of the nucleus and the applied magnetic field (Figure 1.5). The Zeeman interaction determines both the frequency at which the nucleus is observed in a specific magnetic field and, through the Boltzmann distribution, affects the detection sensitivity of the nucleus. The natural abundance and relaxation times of the nucleus also affect its sensitivity, i.e., how easily the nucleus can be observed experimentally. 12 m = -1/2 E A E = (h/2;t)YB0 m = +1/2 B 0 Figure 1.5 Energy diagram for the Zeeman interaction for a spin = 1/2 nucleus The interaction of a nuclear spin with an applied external magnetic field can be described by Equation 1.3 where Hz is the Zeeman Hamiltonian, y is the gyromagnetic ratio, h is Planck's constant, B0 is the applied magnetic field, and / is a quantum mechanical The Zeeman interaction is linear with the applied field and an increase in the applied magnetic field produces a larger separation of the quantized energy levels and, correspondingly, a larger population difference which serves to increase the sensitivity. The sensitivity of a nucleus is also a function of its gyromagnetic ratio and its natural abundance. Table 1.2 lists some important N M R parameters of selected nuclei. The nuclei of interest in the present study are in bold typeface. 1.3.2.2 Chemical Shift Interaction The chemical shift, a , of a nucleus result from the interaction of the nucleus with the surrounding electron density. The chemical shift is the most diagnostic feature in the spectrum because it is sensitive to the geometry, bonding, and nature of the atoms surrounding the nucleus. The chemical shift is a three-dimensional quantity and, as a result, its value wi l l depend on the orientation of the nucleus in the magnetic field. In solution, the operator with eigenvalues (1(1 + 1)) / 2 were I is the spin quantum number. (1.3) 13 anisotropy of the chemical shift is averaged and an isotropic value for the chemical shift of each nucleus is observed as a result of the rapid tumbling of the molecules though space and Table 1.2 Important NMR parameters of selected nucleil3^] Nuclear Spin Gyromagnetic Ratio, y NMR Frequency Natural Abundance (%) Isotope (JO7 rad T1 S') at 9.4 T 'H 1/2 26.7520 400.22 99.985 2 H 1 4.1066 61.40 0.015 13C 1/2 6.7283 100.58 1.108 l 5 N 1/2 -2.7120 40.54 0.37 1 9 F 1/2 25.1810 376.30 100 2 7 A 1 5/2 6.9760 104.22 100 2 9Si 1/2 -5.3188 79.46 4.7 3 1 p 1/2 10.8410 161.92 100 through all possible orientations. The chemical shift observed in solution is, therefore, an average of the three principal components of the chemical shift tensor (Equation 1.4) rjiso = 1 /3 ( o x x + rjyy + o z z ) (1.4) In the solid state, the rigidity of the samples prevents such motions. A powder consists of a large number of individual crystallites which are oriented at random to the magnetic field. A s a result, all possible orientations of the nucleus with respect to the magnetic field, B 0 , w i l l be present simultaneously and the summation of this distribution of chemical shifts produces a broad peak. Such a spectrum is not structurally diagnostic and the range of frequencies covered by the anisotropic chemical shift broadening of a single resonance in a solid can be 14 similar to, or larger than, the entire range of isotropic chemical shifts found in solution. This behaviour is illustrated in Figure 1.6 and is common for all of the solid-state interactions. O 5 (ppm) Figure 1.6 Schematic representation of the 1 3 C NMR adsorption spectrum of a carbonyl functionality in (a) a single crystal at two different orientation to the magnetic field vector B 0 , (b) a polycrystalline sample where there are contributions from the random distribution of orientations giving the chemical shift anisotropy pattern and (c) in solution, where the random motion of the molecules yields the isotropic average chemical shift. [31] 15 1.3.2.3 Dipolar Interaction The dipolar interaction is a through-space interaction between two nuclear spins. In liquids the rapid tumbling motion of the molecules averages the dipolar interactions to zero. In solids, strong dipolar interactions remain because the rigidity of the sample prevents motional averaging. The dipolar coupling constant between two nuclei, I and S, is defined by: Z> = ^ # (1.5) 1 6 ; r V 7 2 2 where /u0 is the permittivity of free space (4n«l 0" kg m s" A" ), 77 and ys are the gyromagnetic ratios of the nuclei, and r is the distance between them. The 1/r3 dependence of the dipolar coupling means that the magnitude of the dipole-dipole interaction falls off rapidly with internuclear separation. The dipolar coupling wi l l be most pronounced between neighbouring nuclei that have large gyromagnetic ratios, such as ' H . 1.3.2.4 Scalar Coupling Interaction Scalar coupling, or J-coupling, is the result of indirect coupling between nuclei through their bonding electrons. The interaction is not averaged by the rapid tumbling in solution and appears in the solution spectrum as a splitting of the peaks into multiplets. J-coupling is also present in the solid-state but the magnitude of the interaction is much smaller than those of competing interactions and is normally not observed. 1.3.2.5 Quadrupolar Interaction The quadrupolar interaction is of the same order of magnitude as the Zeeman interaction and for nuclei with spin >l/2, the effects of quadrupolar interactions can dominate the spectrum. The quadrupolar interaction causes a perturbation in the Zeeman energy levels 16 which is illustrated in Figure 1.7 for a spin 5/2 nucleus such as 2 7 A1. The shift in the energy levels scales as a factor of the quadrupolar coupling constant, Q. Zeeman Splitting Zeeman Splitting + Quadrupole Interaction m, 5/2 3/2 1/2 -1/2 -3/2 -5/2 1 k vD+2Q • J I t v 0 v 0+Q > V 0 1 r k V 0 ' i \ ' 1 Vo k v 0 - Q r v 0 v 0 -2Q ' Figure 1.7 Effect of the quadrupolar interaction on the Zeeman splitting of a spin 5/2 nucleus In the absence of the quadrupolar interaction, all the transitions between the different Zeeman levels are of the same energy. It would appear from the diagram that additional peaks should appear in the spectrum of the nucleus in the presence of the quadrupolar interaction because the different transitions are no longer all of the same energy. However, for non-integer spin quadrupolar nuclei, only the central (1/2 to -1/2) transition is observed because it is not affected by the first-order quadrupolar interaction but only by the second-order interaction which is much weaker. 1.3.2.6 High Power Decoupling For most isotopically dilute nuclei, the major source of line broadening in the solid state is due to heteronuclear dipolar coupling with abundant spins. In the case of zeolites, 17 broadening of the dilute Si nuclei could be caused by dipolar coupling with H nuclei associated with the water molecules and/or ammonium ions in the framework. It is possible to eliminate the heteronuclear dipolar interaction by continuously irradiating the protons at their Larmor frequency with a strong r.f. field. This causes the ' H nuclei to rapidly flip between their two spin states which averages the heteronuclear dipolar coupling to zero. To be effective, the frequency of the spin-flip must be equivalent to the magnitude of the dipolar interactions and the pulse must be applied throughout the entire pulse sequence. 1.3.2.7 Magic Angle Spinning Magic angle spinning is the primary method used to remove or average the broadening caused by these solid-state interactions. The method takes advantage of the first-order orientational dependence of the chemical shift anisotropy and the dipolar interactions whose Hamiltonian contains the term (3cos § - 1) where (j) is the angle between an internuclear vector, fixed in space by nature of the rigid solid, and the main magnetic field. Spinning the sample about an inclined axis imposes a time-dependent motion that mimics the isotropic motions found in liquids. The dipole-dipole interaction is a convenient example to use and can be represented by two nuclei, I and S which are fixed in space, separated by a distance r. The illustration in Figure 1.8 depicts the various angles that are used to describe the orientation of the internuclear vector as it is rotated about an axis inclined at an angle of 54.74° to the main magnetic field. The rotation of the sample causes the internuclear vector, r, to trace out a conical path (defined by the angles a and (3) which causes the angle between the internuclear vector and the main magnetic field, <j), to vary. The average value of (3cos § - 1), denoted by the < > brackets, through the course of rotation is described by[35, 36] : 18 < 3cos 2 <j) - 1 >=- (3cos 2 9 m -1)(3cos 2 p -1) + — (s in29 m sin2pcos(co rt + a) + sin Pcos(2cort + a)) where cor is the frequency of rotation of the sample about the spinning axis. The angle, p, between the internuclear vector and the magnetic field is constant for each pair of nuclei but a polycrystalline sample wi l l contain a distribution of nuclei pairs whose P angles collectively span from 0 to TZ. A sufficiently high value of cor (where the frequency of rotation is fast compared to the frequency of the interaction) wi l l average the time-dependent portions of Equation 1.6 to zero and the remaining term wi l l depend on the (3cos 8 m - 1). The angle, 9 m , is a physical parameter and can be set experimentally. By choosing the angle between the sample holder 2 2 and the magnetic field to be 54.74° the (3cos Om - 1) term goes to zero which causes <3cos <j) - 1> to average to zero for all values of p. This behaviour is unique for this angle and, accordingly, 54.74° is called the 'magic angle'. Rotating the sample at the magic angle, i.e., magic-angle spinning or M A S , serves to average the dipolar coupling and the chemical shift anisotropy and narrow the lines. The ' H -2 9 S i dipolar interactions can be averaged by spinning above 3 k H z while ' H - ' H dipolar couplings require spinning rates above 8 k H z to achieve narrow lines through M A S . Spinning the samples at frequencies less than the chemical shift anisotropy does not completely average the time-dependent portions of Equation 1.6 which produces a series of "spinning sidebands" in the spectrum which are separated from the isotropic chemical shift by integral multiples of the spinning rate. 19 t». Figure 1.8 Definitions of the angles used to describe the orientation of the I-S internuclear vector as it is rotated about an axis inclined at the magic angle to the main magnetic field 20 The quadrupolar interaction has a second angular dependence that is not removed by spinning at the magic angle and, as a result, the second-order quadrupolar interaction is only partially reduced by magic angle spinning. This interaction is inversely field dependent, however, and can be diminished by increasing the magnitude of the applied magnetic field; resulting in narrower lines at higher field strengths. For complete averaging of the interaction, other methods must be used. Dynamic Angle Spinning (DAS) 7] and Double-Angle Rotation (DOR)[38] spin the sample about two axes simultaneously while Multiple-Quantum Magic Angle Spinning ( M Q M A S ) P 9 ] averages the quadrupolar interaction using M A S and a sophisticated pulse sequence. Virtually all of the nuclei present in zeolites can be studied by M A S N M R . The S i and A l spectra provide information about the framework structure and stability and can provide quantitative information about the numbers and types of framework species present. The ' f l spectra can be used to identify non-acidic silanol groups, ammonium ions, and acid protons present in the frameworks. The analysis of the different spectra w i l l be discussed in subsequent chapters as the data is presented. 1.3.3 Physisorption Gas adsorption is widely used to characterize porous solids. The I U P A C classification separates porous materials into three categories: microporous (<20 A),, mesoprorous (20-500 A) and macroporous (>500 A). Zeolites and zeolite-like materials are considered ultramicroporous since their pore sizes range from 4 A to 12 A. The surface area and pore structure are among the most basic properties of a zeolite catalyst. The total surface area is often related to catalytic activity as it controls the quantity of reactant that can be 21 absorbed into the framework to react with the active sites while the pore architecture is responsible for the selectivity of the material toward both reactants and products. Physisorption is modelled on the physical adsorption of a non-selective gas at a temperature low enough to cause the condensation of the adsorbate on the surface of the sample. The adsorbate-adsorbate and adsorbate-adsorbent interactions are an important aspect of physisorption since they control how gas adsorbs on surface of the sample and, coupled with the characteristics of the sample, can produce a variety of isotherm shapes. Five general isotherm types have been established by Brunauer, Deming, Deming, and Teller based on an extensive survey of isotherms reported in the literaturet^O] and are shown in Figure 1.9 where W represents the weight of gas adsorbed, P is the adsorbate equilibrium o P/P 0 1 o P/P 0 1 o P/P„ 1 o P/P„ 1 o P/P 0 1 Figure 1.9 The five isotherm classifications according to BDDTHO] pressure, P 0 is the saturation pressure (beyond which the adsorbate liquefies), and P /P 0 is the partial pressure of the system. The surface area of zeolites is almost entirely internal; 22 contained within the microporous framework. Type I isotherms are typical for microporous materials such as zeolites where pore-filling prevents further condensation from occurring. Type II isotherms are typical non-porous powders and mesoporous materials where the inflection point in the isotherm occurs near the completion of a monolayer. After the first monolayer is completed, additional layers are adsorbed until, at the saturation pressure, the number of layers becomes infinite. Type III isotherms result from systems where adsorption is dominated by adsorbate-adsorbate interactions. In Type III systems, the interaction of the adsorbate with the adsorbed layer is stronger than the adsorbate interaction with the adsorbent surface. Type IV isotherms are characteristic of materials with pore diameters ranging from 30-2000 A . The increase in slope at higher partial pressures indicates the progressive filling of the pores. A Type V isotherm results when adsorption in a porous system is dominated by Type III adsorbate interactions. Nitrogen gas is typically used for surface area analysis because it is inexpensive and readily available. Nitrogen, however, has a quadrupole moment which can interact with the aluminum in zeolite frameworks. The additional attractive forces cause the adsorbate to be strongly adsorbed at very low partial pressures where micropore filling occurs. It has been found that nitrogen isotherms for zeolites[41-43] ^ aluminophosphates[44] do not clearly differentiate between materials with appreciably different pore sizes and, further, that the pore size distributions calculated from the data do not agree well with the expected pore diameters calculated from single-crystal X-ray structural analysis of the materials. Argon gas lacks a quadrupole moment and interacts less strongly with the zeolite framework and is thus used in applications where nitrogen gas cannot be used. The aforementioned studies!^ 1-44] further demonstrate that isotherms collected using argon gas 23 and liquid argon as the cold bath clearly differentiate between materials with even slightly different pore sizes. As a result, the high-resolution physisorption experiments carried out in the present work used argon as the adsorbate at 87 K . 1.3.3.1 Surface Area Calculations The B E T equation,[45, 46] derived by Brunauer, Emmett, and Teller, has gained wide acceptance and is used to calculate the specific surface areas for a variety of materials. The B E T method, however, assumes that the surface on which the adsorbate is condensing does not restrict the number layers of gas that can be adsorbed. [47] This is important when studying zeolites whose pore and channel dimensions are not substantially larger than the adsorbate. Clearly, the restrictive dimensions of the channels in zeolite A preclude multilayer adsorption from taking place. A s a result, the Langmuir equation should be used in place of the B E T equation for calculating the surface areas of small-pore zeolites. The Langmuir equation is more appropriate for calculating the surface areas for microporous materials because it is modelled on systems displaying Type I isotherms where adsorption is limited to, at most, a few molecular layers. Such is the case for zeolite A where, once a monolayer of adsorbate has formed, the channels are filled; leaving no room for additional adsorption. The Langmuir equation is given in Equation 1.7 where P is the adsorbate pressure, W is the weight of the gas adsorbed, A' is a constant describing the rates of adsorption and desorption P 1 P —=——+ — (1.7) W KWm Wm of the adsorbate, and Wm is the weight of the monolayer of gas. [45] The equation is arranged in such a way that a plot of PIW versus P w i l l yield a straight line with a slope of \IWm and 24 an intercept of \IKWM. With the weight of the monolayer known, the total surface area can be calculated from Equation 1.8 where 5*, is the surface area of the sample, WM is the weight of the monolayer, N is Avagadro's number, A is the cross-sectional area for the adsorbate, and M i s the molecular weight of the adsorbate. W NA M 1.3.3.2 Pore Size Calculations Pore size calculations for microporous materials using gas adsorption data is an area of continuing development and no single model seems to be successful for all types of microporous materials. The Kelv in equation is widely used for calculating pore sizes and relates the equilibrium vapour pressure of a curved surface, as is found in a capillary, with the equilibrium pressure of the same liquid on a planar surface. [45] For cylindrical pores the Kelv in equation can be expressed by: -2rV(\\ \n(P/PJ = — J — - (1.9) RT \r) where P /P 0 is the partial pressure of the system, ^and Fare, respectively, the surface tension and molar volume of the liquid condensed in the pore, and r is the pore radius. The Kelv in equation, however, only applies to pores larger than 20 A and, therefore, cannot be used to calculate micropore radii. The Horvath-Kawazoe (H-K) method was developed for measuring the pore sizes in ultramicroporous carbon materials where the pores are considered to be slit-like, i.e., much longer and deeper than they are wide.[42, 43, 48, 49] jfoe model is based on adsorption occurring on the surfaces of two parallel planes whose separation is not much greater than twice the diameter of the adsorbate molecule or atom. The model is appealing because it is 25 conceptually simple and directly correlates pore size as a function of the partial pressure in a Kelvin-like equation which depends on the physical properties of the adsorbate-adsorbent pair. RTln(P/P0) = N n A + n 2 A 2 r~\L-d) f R~4 R -10 R - 4 R -10 ~\ K2{L-dl2f ~ 9(L -dITf 3(d/2)3 + 9(d/2)9 , (1.10) The Horvath-Kawazoe correlation between the partial pressure (P/P 0) and the pore diameter (L) is shown in Equation 1.10 where L is the separation of the planes (or pore dimension), d is the arithmetic mean of the diameters of the adsorbate and adsorbent atoms, A / and A2 are the Lennard-Jones constants for the adsorbent and adsorbate, respectively, ni and ri2 are the number of atoms per unit area of adsorbent and adsorbate, respectively, N is Avagadro's number, and r is distance between the gas and the surface at zero interaction energy. Applying the Horvath-Kawazoe model to zeolites has been criticized because the original model does not account for curvature such as that present in the cylindrical channels in zeolites. [42, 48] A refinement to the Horvath-Kawazoe model has been developed that accounts for the oxide surfaces and cylindrical geometry of the channels of zeolites and seems to produce more accurate values for the pore dimensions of medium-pore zeolites (ca. 6-10 A ) based on argon isotherms collected at 87 K.[48, 50] jfoe refinements do not seem to be equally successful for all zeolites and the pore sizes of small-pore (< 5 A ) and large pore (>10 A ) zeolites are overestimated and underestimated, respectively, compared to the values expected from the X-ray structural data. Which parameters are used for the Horvath-Kawazoe calculation affects only the magnitude of the calculated pore size which could be a concern when characterizing unknown materials. The absolute size of the pores is less important in the present study because the pore structure of zeolite A is well known and the 26 materials examined in the present study are being compared against each other within a series. The main interest was whether the materials were measurably different from one another; a change in the calculation of the magnitude of the pore size wi l l not affect any trend within the series. 1.4 Research Objective The thermal stability and 5 A pores of C a A are desirable characteristics for potential zeolite A-based catalysts but the material is not acidic. A purely acidic form of zeolite A (HA) cannot be produced from the ammonium-exchanged form of zeolite A (NH4A) because the material decomposes upon mild thermal treatment. The basic idea behind the present work is that, by preparing a zeolite A material that contains both C a 2 + and N H 4 + ions, the resulting material, when activated, should retain the thermal stability of C a A and the acidic character of H A . The objective of the present study was to prepare mixed C a 2 + / N H 4 + forms of zeolite A using conventional synthesis and ion-exchange methods, activate them and fully characterize all the materials in terms of crystallinity and acidity. The final activated materials were then to be tested for catalytic activity with accepted test reactions and procedures. A systematic investigation was carried out to establish that the frameworks of the mixed-ionic materials contained a homogeneous distribution of C a 2 + and N H 4 + ions, that the materials were thermally stable, and that the activated C a 2 + / N H 4 + materials displayed Bronsted acidity. If these issues are addressed, any catalytic activity in the sample(s) could be attributed to a homogeneous, thermally stable, acidic sample. 27 1.5 References [I] W. M . Meier, D . H . Olson, C. Baerlocher, Atlas of Zeolite Structure Types, Elsevier, London 1996. [2] R. M . Barrer, Zeolites and Clay Minerals as Sorbents and Molecular Sieves, Academic Press, London 1978. [3] D . W. Breck, Zeolite Molecular Sieves, John Wiley and Sons, Inc, New York 1984. [4] A . Dyer, An Introduction to Zeolite Molecular Sieves, John Wiley and Sons, New York 1988. [5] G . Englehardt, D . Michel , High-Resolution Solid-State NMR of Silicates and Zeolites, John Wiley & Sons, New York 1987. [6] E . M . Flanigan, in H . van Bekkum, E . M . Flanigan (Eds.): Introduction to Zeolite Science and Practice, Elsevier, Amsterdam 1991. [7] P. A . Jacobs, J. A . Martens, in H . van Bekkum, E. M . Flanigan (Eds.): Introduction to Zeolite Science and Practice, Elsevier, Amsterdam 1991. [8] D. W. Breck, W. G . Eversole, R. M . Mil ton, T. B . Reed, T. L . Thomas, J. Am. Chem. Soc 75(1956) 5963. [9] W. Loewenstein, Amer. Mineralog. 39 (1954) 92. [10] E . M . Flanigan, Pure andAppl. Chem. 52 (1980) 2191. [II] W . O. Haag, R. M . Lago, P. B . Weisz, Nature 309 (1984) 589. [12] R. Szostak, Stud. Surf Sci. Catal, Vol. 58 1991, p. 153. [13] H . G. Karge, J. Weitkamp, Stud, in Surf. Sci. Catal. 46, Elsevier B . V , Amsterdam 1989. [14] I. E . Maxwel l , W. J. H . Stork, in H . van Bekkum, E . M . Flanigan, J. C. Jansen (Eds.): Introduction to Zeolite Science and Practice, Elsevier, Amsterdam 1991, p. 571. [15] P. B . Venuto, Micropor. Mater. 2 (1994) 297. [16] Y . Sun, T. song, S. Qiu, W. Pang, J. Shen, D . Jiang, Y . Yue, Zeolites 15 (1995) 745. [17] A . Kuperman, S. Nadimi, S. Oliver, G . A . Ozin, J. M . Garces, M . M . Olken, Nature 5(55(1993)239. 28 [18] J. E . Lewis, C. F. Clemens, M . E . Davis, J. Phys. Chem. 100 (1996) 5039. [19] S. J. Weigal, J. Gabriel, E . G . Puebla, A . M . Bravo, N . J. Henson, L . M . B u l l , A . K . Cheetham, J. Am. Chem. Soc 118 (1996) 2427. [20] J. L . Guth, H . Kessler, J. M . Higel, J. M . Lamblin, J. Patarin, A . Seive, J. M . Chezeau, R. Wey, ACS Symp. Ser. 398 (1989) 176. [21] S. Shumizu, H . Hamada, Angew. Chemie 38 (1999) 2725. [22] J. L . Guth, L . Delmotte, M . Soulard, N . Brunard, J. F. Joly, D. Espinat, Zeolites 12 (1992)929. [23] D. T. Hayhurst, P. J. Mell ing, W . J. K i n , W. Bibby, ACS Symp.Ser. 398 (1989) 233. [24] S. Qiu, J. Y u , G. Shu, O. Terasaki, Y . Nozue, W. Pang, R. X u , Micropor. and Mesopor. Mater. 21 (1998) 245. [25] J. F. Charnell, J. Crystal Growth 8 (1971) 291. [26] H . P. Klug , L. E . Alexander, X-Ray Diffraction Procedures For Poly crystalline and Amorphous Materials, Wiley-Interscience, New York 1974. [27] B . D. Cullity, Elements of X-Ray Diffraction, 2nd Ed., Addison-Wesley Publishing Company, Inc., Reading, Mass. 1978. [28] E . R. Andrew, in D. M . Grant, R. K . Harris (Eds.): Encyclopedia of Nuclear Magnetic Resonance, John Wiley and Sons, New York 1996, p. 2891. [29] A . Abragam, The Principles of Nuclear Magnetism, Oxford University Press, London 1961. [30] R. R. Ernst, G . Bodenhausen, A . Wokaum, Principles of Nuclear Magnetic Resonance in One and Two Dimensions, Clarendon Press, Oxford 1990. [31] C. A . Fyfe, Solid State NMR for Chemists, C .F .C . Press, Guelph, Ontario 1983. [32] M . Mehring, Principles of High Resolution NMR in Solids, Springer-Verlag, New York 1983. [33] L. W. Jelinski, M . T. Melchior, in C. Dybowski, R. L . Lichter (Eds.): NMR Spectroscopy Techniques, Marcel Dekker, New York 1987. [34] R. K . Harris, Nuclear Magnetic Resonance Spectroscopy, John Wiley and Sons, New York 1986. 29 [35] R. A . Wind, in A . I. Popov, K . Hallenga (Eds.): Modern NMR Techniques and their Applications in Chemistry, Marcel Dekker, New York 1991. [36] M . M . Maricq, J. S. Waugh, J. Chem. Phys. 70 (1979) 3300. [37] A . Llor, J. Virlet, J. Chem. Phys. Lett 152 (1988) 248. [38] A . Samoson, E . Lippmaa, A . Pines, Mol. Phys. 65 (1988) 1013. [39] L . Frydman, J. S. Harwood, J. Am. Chem. Soc. 7/7 (1995) 5367. [40] S. Brunauer, L . S. Deming, W. S. Deming, E . Teller, J. Am. Chem. Soc. 62 (1940) 1723. [41] M . J. G . Janssen, C. W. M . van Oorschot, in P. A . Jacobs, R. A . van Stanten (Eds.): , Elsevier Science Publishers, Amsterdam 1989, p. 633. [42] S. Storck, H . Bretinger, W. F. Maier, Appl. Catal. A: General 174 (1998) 137. [43] A . F. Venero, J. N . Chiou, Mat. Res. Soc. Symp. Proc. Ill (1988) 235. [44] P. E . Hathaway, M . E . Davis, Cat. Lett. 5 (1990) 333. [45] S. Lowel l , J. E . Shields, Powder Surface Area and Porosity, Chapman and Hal l , London 1991. [46] S. W. Gregg, K . S. W. Sing, Adsorption, Surface Area, and Porosity, Academic Press, New York 1982. [47] D . J. C. Yates, Can. J. Chem. 46 (1968) 1695. [48] A . Saito, H . C. Foley, J. AlChE 37 (1991) 429. [49] G . Horvath, K . Kawazoe, J. Chem. Eng. Jpn. 16 (1983) 470. [50] A . Saito, H . C. Foley, Micropor. Mater. 3 (1995) 531. 30 Chapter 2 Materials and Methods 2.1 Synthesis of Zeolite A Due to the commercial importance of zeolite A , there has been an ongoing interest in determining how the various synthesis parameters affect the crystallite size, morphology, and rate of formation. B y understanding the factors which influence the final product, the zeolite properties can be tailored for specific purposes. For instance, the use of zeolite A in detergent applications favours crystals without sharp edges to prevent damage to the textiles while at the same time relying on a narrow crystallite-size distribution to ensure uniform exchange kinetics. [1] The composition of the synthesis gel and the conditions under which it is crystallized combine to influence both the phase and the characteristics of the crystallizing zeolite. [2-5] Studies have been carried out to determine the effects of gel aging, [6-8] heating time and temperature,[25 7] alkalinity,[5> 7-10] a n d stirring rate[5> 11] on the synthesis of zeolite A . These factors are important considerations as they each may affect the crystallization rate, as well as the crystallite size, morphology, and distribution. Other factors that can influence the growth of zeolite A crystals are the addition of "stabilizing" agents[12-14] to the gel and the act of seeding the synthesis mixture. 7] The crystallization of zeolite A from the synthesis gel is preceded by an induction, or aging period which is connected to the partial dissolution of the silica and alumina components from the gel into the aqueous phase and the formation of crystal nuclei within the gel framework. The partial dissolution of the gel skeleton provides "active" silica and alumina species in the form of soluble hydroxides which subsequently react with the crystal 31 nuclei in the gel and initiate crystal growth. This nucleation period is influenced strongly by both the temperature at which the gel is aged and alkalinity of the synthesis mixture. The induction time can be shortened by raising the temperature or by increasing the alkalinity of the mix. The crystallization of the zeolite microcrystals proceeds rapidly after nucleation of the gel. A n interesting observation, common to several of the studies, is that the growth rate of the crystals is independent of crystallite size. This reinforces the importance of controlling the aging of the gel since it is the properties of the gel nuclei which determine the particle size and distribution of the final product. The crystallization time is also influenced by temperature and while crystallization wi l l occur i f the mixture is maintained at the induction temperature (typically 60-80 °C) for longer times, raising the temperature increases the rate of crystallization and produces a narrower particle-size distribution. [2, 12] Raising the alkalinity of the synthesis mixture by adding base shortens the nucleation time and in doing so, produces a greater quantity of smaller nuclei. The small gel nuclei w i l l , in turn, form small crystallites. Conversely, decreasing the alkalinity of the system wi l l produce larger crystals but longer times are required to crystallize the zeolite from the synthesis mix. Whether the system is continuously stirred or left static w i l l only effect the particle-size distribution of the product. The use of stirring to keep the gel homogenized throughout the nucleation and crystallization processes produces materials with a narrower particle-size distribution than those obtained from static or incompletely mixed systems. The use of vigorous stirring also tends to produce materials consisting of smaller crystallites than those obtained from static mixtures. Static or incompletely mixed systems tend to produce a wide particle-size distribution in materials crystallized from a typical synthesis gel. A wide 32 particle-size distribution can impart an unwanted heterogeneity in the sample in terms of the exchange kinetics and thermal behaviour of the product. A synthetic method to generate crystals large enough to be studied by single-crystal X -ray diffraction techniques was developed by Charnell[12] in 1971 and uses triethanolamine (TEA) in the synthesis mix. The role of T E A in the preparation of large N a A crystals is unclear. It has been proposed that the T E A acts as an aluminum complexing agent which delays the incorporation of aluminum into the crystals and allows larger crystals to form. Another proposal views the additive as a gel "stabilizer" which prevents the formation of a large number of nuclei and, subsequently, larger crystals are produced. While the role of T E A has not been clearly established, the net effect of adding T E A to the synthesis mixture is that the nucleation period of the gel and the crystallization time of the zeolite are lengthened and larger crystals are formed. The synthesis of large batches of material from such a synthesis was not practical for this work since the time required for synthesis is on the order of three weeks and batch sizes were only 40 g. The wide particle-size distribution obtained from the Charnall synthesis is also unfavourable for this work as a uniform particle-size distribution was desired. B y combining higher aging and crystallization temperatures with a highly alkaline solution and maintaining a vigorously stirred reaction mixture, it is possible to crystallize batches of N a A in a short period of time with a uniform distribution of small crystallites. The preparation of N a A in the present work was adapted from a preparation reported by RolandH ] and produces a material comparable with commercially available products. 33 2.1.1 Preparation of NaA The synthesis of zeolite A does not require high temperatures or pressures and can be done on a benchtop with standard glassware. The setup employed is shown in Figure 2.1. A n external stirring motor was employed to ensure the zeolite gel remained homogenous at all times during the reaction. A glass wool heating mantle, attached to a Variac, was used as a heat source. Variable speed Thermometer \ Glass connecting Heating mantle ^ v. L rod Teflon paddle \ (— i _ Y ) Figure 2.1 Experimental setup for the preparation of NaA Two separate aqueous solutions were prepared which, when mixed, formed the precursor aluminosilicate gel from which zeolite A was crystallized. The aluminate solution was prepared by dissolving 71.74 g of N a O H and 72.63 g sodium aluminate (Anachemia, 96% technical grade) in 266 m L of distilled water. Once all reactants had dissolved, the aluminate solution was filtered to remove insoluble, suspended particulates. The silicate 34 solution was prepared by diluting 125 m L of Ludox (30% Si02) with 141 m L of distilled water. The two solutions were mixed in a 1 L resin kettle at room temperature for ten minutes until the gel had homogenized. The overall molar composition of the zeolite gel was: N a O H : A 1 2 0 3 : S i 0 2 : H 2 0 5.6 : 1 :2.23 : 71.7 The white, opaque gel was stirred constantly, heated to 60 °C, and maintained between 60-65 °C for 1.5 hours. After the initial aging period, the zeolite was crystallized by heating the gel to 90 °C and holding it at this temperature for 30 minutes. If the reaction was maintained at 90 °C for longer than 30 minutes, an impurity sodalite phase began to form. The hot zeolite slurry was filtered and rinsed with 12 liters of distilled water. The resultant zeolite cake was allowed to dry at ambient conditions. This preparation of N a A yielded approximately 125 g of zeolite. 2.2 Ion Exchange of Zeolite A The exchange of N a + for an extensive number of alkali, alkali earth, and transition metal ions has been studied and the resulting equilibria measured.t 1^-!^] jhe studies establish the selectivity of N a A for different cations, the maximum degree of exchange that can be achieved for the incoming cation, and the physical properties of the resulting ion-exchanged systems. Current interest in the kinetics of the ion exchange process in zeolite A stems from the commercial importance of zeolite A as a detergent builder and as such they focus on N a + / C a 2 + and N a + / C a 2 + / M g 2 + systems. [19-22] 35 Complete exchange of the N a ions in N a A is complicated by several factors. It has been established that extraneous N a A l C h species can exist in the framework of N a A and that the impurity remains trapped inside the sodalite cages of the structure. H71 To penetrate the sodalite cage, the exchanging cation must pass through a 6-membered oxygen ring with an approximate diameter of 2.5 A . H 7 ] This dimension excludes most cations, including C a 2 + and N H 4 + . Not all preparations of N a A contain the sodium aluminate impurity and those preparations that do can vary in the amount of NaAlC>2 the material contains.H^] It is possible, therefore, to exchange all of the N a + ions responsible for charge-balancing framework aluminum atoms but still have residual N a + ions in the material. Another complication that has been encountered when trying to achieve complete exchange is the phenomenon of hydronium exchange. Some studies have recorded discrepancies between the total amount of aluminum in an exchanged sample and the total amount of ions present. Since each equivalent of aluminum in the sample must be offset by an equal equivalent of charge-balancing cations (any sodium aluminate impurity included), the deficit was attributed to exchange with H 3 O from the solution. [19-21,23-26] W h i l e hydronium exchange remains a reasonable explanation to help explain unexpected results, it remains to be established whether exchange of H 3 0 + is actually taking place and whether, in fact, it is the only process operating to alter the balance of ions in the zeolite. For practical purposes it is not necessary to achieve complete exchange of N a + for C a 2 + or NH4 + . B y employing concentrated exchanging solutions and exposing the N a A material to multiple exchanges, a degree of exchange for C a 2 + H 8 , 21, 27] a n ( j NH4 + [28-30]j n excess of 90% can be expected. A n exchange time of one hour is sufficient to allow the 36 materials to reach exchange equilibrium as crystallites on the order of 1 um exchange very rapidly. [19] The same external stirring motor and teflon stir paddle used in the preparation of N a A were employed in the aqueous exchanges to ensure constant and complete mixing of the powders in the exchanging solution. The Ca - and NFL, -exchanges were carried out simultaneously in separate vessels. This ensured that the temperatures at which the two exchanges were carried out did not vary and the relative humidity under which the exchanged products were left to dry was the same. 2.2.1 Calcium-Exchanged Zeolite A The calcium form of zeolite A (CaA) was generated via aqueous ion-exchange. A 40 g sample of the as-synthesized N a A material was crushed to a loose powder and added to a 1 L round-bottom flask containing 670 m L of 1.5 M CaCL.. The slurry was stirred for one hour at room temperature, filtered, and rinsed with 5 L of room-temperature distilled water. The exchange reaction and subsequent rinsing was repeated once and the filter cake was allowed to dry under ambient conditions. 2.2.2 Ammonium-Exchanged Zeolite A The ammonium form of zeolite A was prepared by the same process as CaA. A 40 g sample of the as-synthesized material was exchanged twice with 1 L measures of 1 M NH4CI. After each exchange, the material was filtered and rinsed with 5 L of room-temperature, distilled water. The resultant ammonium-exchanged zeolite was allowed to dry under ambient conditions. 37 2.3 Sol id-State Ion Exchange While the ion exchange of zeolites is typically carried out by mixing a sample of the parent zeolite in a solution of the exchanging cation it is possible to achieve cation exchange in the solid state. B y carrying out the ion exchange in the solid state, the sensitivity zeolite A shows to low pH exchanging solutions is removed and since the incoming cation is not enlarged by a hydration sphere, the barrier to diffusion into the lattice is reduced, resulting in potentially higher degrees of exchange than with conventional aqueous ion exchange. [31> 32] Solid-state exchange can take place between the zeolite crystallites and a metal salt or between zeolite crystallites of different compositions. The former process is typically carried out under anhydrous conditions with calcined zeolites and dry metal salts at elevated temperatures. The zeolite-zeolite exchanges are carried out with the hydrated materials at ambient conditions. The two solid-state ion exchanges are distinguished from each other based on the degree of hydration of the exchanging system. Exchanges involving hydrated zeolite crystallites are termed contact-induced ion exchange while the anhydrous, high-temperature exchanges are designated solid-state ion exchanges. The solid-state ion exchange process was observed in the 1970s with reactions between zeolite Y and metal chloride salts[33, 34] a n c j h a s been recently reviewed.[35] A n exchange of the alkali metal in the salt for the framework cation occurs when the materials are intimately mixed and heated to temperatures above 250 °C. If the reaction occurs between a metal salt and an alkali-balanced zeolite, the products of the reaction wi l l be the metal-exchanged zeolite and an alkali metal salt. For example, M C l n + N a Z - > nNaCl + M Z (2.1) 38 If the reaction takes place between a metal salt and an acidic zeolite the exchange product leaves as a volatile product. M C l n + H Z -> M Z + n H C l t (2.2) These reactions assume a stoichiometric amount of each material is present. In both cases, the products that form from the reactions are easily distinguishable from the starting materials. This aspect of the solid-state reaction permits facile study of the system by several different methods.[35] The process of contact-induced ion exchange was illustrated by Fyfe, et al. using a L i A - N a A system. P 6] The contact-induced exchange process provides a simple means of precisely tailoring the charge-balancing cation composition of a zeolite without requiring the use of external reagents. The exchange process takes place through inter-crystallite contact which causes ion migration at the crystal faces. By intimately mixing two zeolite powders with different cation composition and allowing the system to come to equilibrium, each crystallite should have a composition equal to that of the original mix. The rate of the exchange process is strongly affected by the degree of hydration of the zeolitet^7, 38] a n ( j by the mobility of each ion in the lattice. 2.3.1 Preparation of Mixed Ca 2 +/NH 4 + Materials Zeolite A samples containing different proportions of C a 2 + and N H } + ions were prepared by contact-induced ion exchange. Mixed C a 2 + / N H 4 + materials were prepared by mixing the 'dry' powders of the as-exchanged calcium and ammonium materials (as described above) in weight percentages that corresponded to 70%, 60%, and 50% C a A in the mix. M i l d hydrothermal treatment was selected over thermal treatment to ensure that the lattice remained fully hydrated at the slightly elevated temperature. The materials were 39 intimately mixed in uncapped polyethylene bottles and were sealed in a desiccator containing distilled water. The desiccator was then placed in an oven at 45 °C for three days. 2.4 Characterization The microcrystalline character of the zeolite powders necessitates the use of an array of complementary techniques to gain a more complete understanding of the structure and behaviour of the materials. 2.4.1 Powder X-ray Diffraction (XRD) X-ray powder diffraction patterns of all zeolite materials were obtained using a Rigaku Rotaflex diffractometer. The diffractometer was equipped with a copper (Cu K a X = 1.54184 A) rotating anode source. The filament was run at a voltage of 50 k V and a current of 150 mA. The radiation was not monochromated so the contribution from Cu K a 2 X-rays (X = 1.154439 A ) can be resolved at angles above 35° 26 in some of the powder patterns. A s all samples were uniformly microcrystalline, no grinding or pretreatment of the samples was necessary prior to packing the samples. It was necessary to pack the samples tightly into the aluminum samples holders as the diffractometer was equipped with a goniometer having a vertical sample mount. Some preferred orientation effects from the cubic crystals must be expected as a result of the packing process but since the powder patterns of the materials were very reproducible from batch to batch, the contribution was considered to be relatively constant. Powder patterns were collected at a rate of 5°/min from 4 to 50° 29 with a step size of 0.02°. A l l of the powder X R D patterns in the present work 40 were collected on samples at room temperature since the instrument was not equipped with a variable-temperature sample stage. 2.4.2 Scanning Electron Microscopy (SEM) S E M images and electron microprobe analyses were acquired with a Hitachi S-2300 coupled to a Quartz PCI® digital imaging system which allows high resolution S E M micrographs and E D X (energy dispersive X-ray) results to be captured directly to a computer and printed off on standard paper. The electron beam voltage was set to 20 k V and the microscope was optimized for high-resolution work at magnifications up to 15000 X . Samples for imaging were prepared by sonicating a small amount of the zeolite powder in alcohol and dropping the suspension onto a thin glass plate mounted to a standard aluminum S E M stub. Once the alcohol had evaporated, the samples were gold coated to a thickness of 10-50 nm to make them suitably conductive for imaging. Samples for electron probe microanalysis ( E P M A ) were prepared by compressing the zeolite powder into disks measured approximately 1 cm diameter and 3-4 mm thick. The disks were directly mounted with conductive adhesive to a standard S E M stub and graphite coated. 2.4.3 Room-Temperature Hexane Adsorption The hexane adsorption experiments were carried out on a T A Instruments Inc. TA2000 thermal analysis system equipped with a TG51 thermogravimetric analyzer. A procedure, adapted from a study of a sodium-ammonium zeolite X system,[26] w a s developed to study the uptake of «-hexane at 30 °C. The equipment was optimized to retain an internal anhydrous atmosphere by sealing the external data cable connection with silicone 41 caulking compound and lightly greasing all O-ring seals with a small amount of high-vacuum grease. A schematic of the experimental setup is shown in Figure 2.2. The requirement for an anhydrous atmosphere arises from the extremely hydrophilic nature of dehydrated, aluminosilicate zeolites. Any water vapour resident in the instrument or introduced into the sample chamber or balance housing is scavenged by the activated zeolite. It is known that the adsorption of small amounts of water can significantly decrease the amount of hexane the material can adsorb.H 5] For this reason, gases were dried through columns of P2O5 prior to being introduced into the system. The sample and balance chambers of the instrument are not separately sealed. It is thus possible for vapours introduced into the sidearm of the instrument to diffuse back into the balance housing. To avoid the complications of hexane condensing on the balance assembly, two separate gas streams (He and N2) were used in the process of activating the zeolite material and measuring the hexane uptake. N 2 3-way valve Gas washbottle Exhaust < Furnace Balance housing He Figure 2.2 Schematic of TGA setup for vapour adsorption 42 A stream (100-150 cc/min) of nitrogen gas (industrial grade, 99.995%) was used both as a purge gas and as a carrier gas for the hexane vapour. The nitrogen stream was saturated with rc-hexane by passing the gas through a porous glass frit in a gas washbottle containing dry /7-hexane. The dry nitrogen gas and the hexane-saturated nitrogen stream were connected to the T G A via a 3-way valve attached to the gas inlet port. The balance assembly was continuously kept under a stream (200 cc/min) of helium (prepurified 99.996%) by purging through the balance housing. The helium flow rate was set to exceed the nitrogen flow rate to prevent the diffusion of hexane vapour into the balance housing. During the activation of the materials, the two three-way valves were opened to allow only dry nitrogen into the sample chamber. During the measurement of the adsorption of hexane, the 3-way valves were switched to admit the hexane vapour-nitrogen stream into the sample chamber. This procedure was employed to eliminate any deviations in the microbalance reading that occur due to suddenly changing the flow rate of gases through the instrument. Sample sizes of approximately 10-15 mg were used for analysis. Samples were activated by heating from ambient conditions to 600 °C at 10°/min. To ensure that materials were free of water and ammonia, the samples were maintained at 600 °C for two hours before cooling to 30-33 °C. The adsorption of n-hexane on the zeolites was measured by maintaining the temperature at approximately 30 °C and passing a hexane-saturated nitrogen gas stream through the sample chamber. The gain in sample mass with respect to time was recorded. 43 2.4.4 High-Temperature Adsorption of Small Molecules The equipment in Figure 2.2 was also employed to study the uptake of small molecules at high temperatures. In particular, the uptakes of methanol and n-hexane at elevated temperatures were studied in the present work. Sample sizes and gas flow rates matched those of Section 2.4.3. Samples were activated by heating at 10°/min from ambient to 500 °C and held isothermally at 500 °C for 15 minutes. Sorption studies were carried out by holding the activated sample at the selected temperature and passing the vapour of the probe molecule over the zeolite. The sample mass with respect to time was recorded. 2.4.5 Physisorption Surface area and porosity measurements were carried out with a Micromeritics A S A P 2010 instrument with micropore capability. Argon was used as the analysis and saturation pressure gas and liquid argon was used for the cryogen bath. A n analysis protocol, specific for zeolite A , was developed to circumvent difficulties encountered in measuring gas adsorption at very low partial pressures in microporous materials. The complication manifested itself as a deviation in the isotherm toward less positive values at low partial pressures (ca. 0.0001 P/P 0 ) causing the isotherm, in essence, to bend back on itself. This behaviour prevents any meaningful data from being collected in the microporous region of the isotherm. Longer evacuation times reduced, but did not remove, the effect. It was discovered that the deviation was caused by residual nitrogen and helium gases slowly diffusing out of the zeolite lattice. A n example of such an isotherm is shown in Figure 2.3. To minimize the effects of outgassing, it is necessary for the samples to be free of water and to minimize their exposure to N 2 and He. Samples were calcined at 10°/min to 44 500 °C in a programmable muffle furnace. After four hours at 500 °C the samples were loaded straight from the furnace into degassed sample tubes. The sample tubes containing the dry zeolite were then degassed at 350 °C on the analysis port of the instrument. Carrying out the sample preparation in this way prevents the equipment from backfilling the sample tube with nitrogen gas. Nitrogen gas is strongly physisorbed in the zeolite A lattice and is time-consuming to remove completely prior to analysis. 160 140 120 100 ,—s CD 73 80 CO 60 < CD E 40 o > 20 0 1e-7 1e-6 1e-5 1e-4 P/Po 1e-3 1e-2 Figure 2.3: Deviation of an isotherm due to slow outgassing. Nitrogen gas was used as the adsorbate at 77 K. The introduction of helium into the system occurs when the instrument measures the free space, or void fraction, of the sample tube. Once the helium is introduced into a microporous system it can take many hours of evacuation before all of the gas has been removed. [39] To avoid this complication, free-space measurements were always carried out 45 in a separate analysis after the isotherm had been collected. Figure 2.4 shows the low pressure region of an isotherm collected using A r at 87 K using a sample where the aforementioned complications were removed Approximately 0.25 g of freshly calcined material was used for analysis. Adsorption isotherms were collected by repeatedly dosing the system with 3 cc volumes of argon and measuring the resulting partial pressure. This method produces isotherms with a resolution 1e-6 1e-5 1e-4 1e-3 P/P„ 1e-2 1e-1 1e+0 Figure 2.4 Low pressure region of an isotherm collected using Ar at 87 K where the sample was not exposed to any gasses prior to analysis. suitable for Horvath-Kawazoe pore size distribution analysis. Since the micropores are filled at low partial pressures, data points beyond a partial pressure of 0.03 P /P 0 were not used in the calculation of the surface area using the Langmuir equation. Omitting the data beyond the point at which the micropores are filled produces a more accurate value for the internal surface area of the sample 46 2.4.6 Magic-Angle Spinning Nuclear Magnetic Resonance Spectroscopy (MAS NMR) M A S N M R experiments were carried out on a Bruker M S L 400 spectrometer. Three nuclei were examined in the study of the zeolite A materials. For Si experiments the stator was aligned to the magic angle using K I while for 2 7 A 1 experiments K B r was employed. 2 9 S i (79.495 M H z ) spectra were externally referenced to hydrated N a A at -89.4 ppm from TMS,[4°] 2 7 A l (104.26 M H z ) spectra were externally referenced to A 1 ( H 2 0 ) 6 at 0 ppm and *H (400.13 M H z ) spectra were externally referenced to C H C 1 3 at 7.2-7 ppm from T M S l 4 1 ] Single pulse experiments without proton decoupling were employed for collecting spectra of each nuclei. 2 9 S i spectra were collected using a 14 mm stator mounted on a homebuilt probe. The samples were run at a spinning rate of 2-2.5 kHz. Sweepwidths of 6 kHz and 10 kHz were used for the as-synthesized and activated materials respectively. The nuclei were excited with a 45° pulse measuring 4.2 ps. 2 7 A 1 spectra were collected using a Doty 5 mm stator system mounted on a homebuilt M A S probe. The spinning rate of the samples was determined by measuring the frequency difference between the spinning side bands. The spinning rate for all samples was ca. 10 kHz. A spectral sweepwidth of 50 kHz was used, yielding a chemical shift span of 480 ppm. The nuclei were excited with a 0.6 ps pulse corresponding to a tip angle of 30°. The small tip angle allows for a shorter recycle delay and permits more transients to be collected in less time than i f a 90° pulse had been used. A recycle delay of 100 ms followed the collection of the FID to allow the system to relax to equilibrium. Approximately 3000 scans were collected for each spectrum. 47 Proton spectra were run on the same probe and stator combination that was used for the Si experiments. A spectral sweepwidth of 38.5 k H z (96 ppm) was employed. The FIDs were collected following a 7 \is excitation pulse. Data collection was followed by a 500 ms recycle delay. 2.4.7 Catalytic Activity Tests Activity tests were carried out in a quartz reactor system coupled to a Shimadzu 14A gas chromatograph fitted with flame ionization detector and an inline automatic sampling/ injection valve. A 1/8" stainless steel packed column containing Chromosorb 101 was used as the separating phase. A schematic of the reactor system is shown in Figure 2.5. After sample injection, the oven was heated from 30 °C to 175 °C at 10°/min and held at 175 °C for 5 minutes to allow the products to completely elute from the column. Samples of untreated material were added to a quartz reactor tube (I.D. 7.5 mm) into which a plug of quartz wool had been added to act as a support for the powder. The sample and quartz wool support were aligned to sit at the midpoint of the reactor tube. The samples were activated under an 80 cc/min stream of helium by heating from ambient conditions to 500 °C at a rate of 10°/min. Samples were held at 500 °C for 15 minutes to complete the activation step before the furnace was warmed or cooled to the reaction temperature. The thermocouple used to control the furnace temperature was located outside the quartz reactor to minimize reactions with the hot metal surface. The Mob i l O i l a-cracking test[42] w a s used to evaluate the C a A , 50, 60, and 70%Ca materials' activity toward the catalytic cracking of ^-hexane. Conditions were chosen to achieve a significant degree of rc-hexane conversion below 40% as outlined in the procedure. The hexane cracking activity studies were carried out at 540 °C with an rc-hexane (98% 48 Aldrich) vapour flow rate of 18 cc/min. Approximately 750 mg (unactivated weight) of zeolite was loaded into the reactor tube and activated in-situ. Tests for the zeolites' activity toward the conversion of methanol to olefins were carried out on the same materials listed above. The methanol-to-olefin conversion studies were carried out at 400 °C with a methanol ( H P L C grade, Fisher Scientific) vapour flow rate of 100 cc/min. The reaction temperature of 400 °C and higher flow rate was used to maintain the conversion of the feed below 40%. The proviso for maintaining the conversion level above 5% and below 40% is outlined in the Mob i l test.[42] Values below the lower limit are within the error of the measurement while conversions above 40% wi l l suffer from overcracking, and mass and heat transport effects. 2.5 Simulated Powder Diffraction Patterns Simulated powder diffraction patterns were generated using a program called PowderCell. The software is available, free of charge, for individual or academic use from: www. lmcp.jussieu.fr/sincris/logiciel/prg-powdercell. html The software generates powder diffraction patterns from single-crystal X-ray structure data. Single-crystal X-ray structure data for the zeolite A and its derivatives were obtained from the Atlas of Zeolite Structure Types[43] which can be accessed on line from: www. iza-sc. ethz. ch/cgi-IZA-SC/collection 49 o o CD 0 r + C N r CD =3 o o CD : c o 2 0 o co CD (/> CD CD 2 3 ) 5 1_ o 0 LL o w w c CD o ^ O 0 I 5 ._ o 0 03 .fc CO o § O 0 X O CD CD Q. 0 >, or DQ .2 E < CU o cu N o WD C cu CU t s #cu CU CU JS e .2f " c / J CU ns s-o cu CU IT) c4 CU J-S o References E. Roland, in H . G. Karge, J. Weitkamp (Eds.): Zeolites as Catalysts, Sorbents, and Detergent Builders, Elsevier Science Publishers, B . V . , Amsterdam 1989, p. 645. W. Meise, F. E . Schwochow, in W. M . Meier, J. B . Uytterhoeven (Eds.): Adv. Chem. Ser. Vol 121, A C S , Washington 1973, p. 169. D . W. Breck, Zeolite Molecular Sieves, John Wiley and Sons, Inc, New York 1984, p. 333. J. Shi, M . W. Anderson, S. W . Carr, Chem. Mater. 8 (1996) 369. J. Ciric, J. Colloid and Interface Science 28 (1968) 315. L . Gora, R. W. Thompson, Zeolites 18 (1997) 132. S. P. Zhdanov, in E . M . Flanigan, L . B . Sand (Eds.): Adv. Chem. Ser., Vol. 101, A C S , Washington 1971, p. 20. G . T. Kerr, J. Phys. Chem. 70 (1966) 1047. S. Bosnar, B . Subotic, Micropor. andMesopor. Mater. 28 (1999) 483. T. Antonic, B . Subotic, N . Stubicar, Zeolites 18 (1997) 291. E . N . Coker, A . G . Dixon, R. W. Thompson, A . Sacco, Micropor. Mater. 3 (1995) 637. J. F. Charnell, J. Crystal Growth 8 (1971) 291. G . Scott, A . G . Dixon, A . Sacco, R. W. Thompson, in P. A . Jacobs, R. A . van Stanten (Eds.): Zeolites: Facts, Figures, Future, Elsevier Science Publishers, Amsterdam 1989, p. 363. E . I. Basaldella, J. C. Tara, Materials Lett. 34 (1998) 119. D. W. Breck, W. G . Eversole, R. M . Mil ton, T. B . Reed, T. L . Thomas, J. Am. Chem. Soc 75(1956) 5963. H . S. Sherry, H . F. Walton, J. Phys., Chem 71 (1967) 1457. R. M . Barrer, W. M . Meier, Trans. Faraday Soc. 54 (1958) 1074. P. E . Eberly, J. Phys. Chem. 66 (1692) 812. 51 [19] D . Drummond, A . De Jonge, L . V . C. Rees, J. Phys. Chem. 87 (1983) 1967. [20] R. P. Townsend, Pure andAppl. Chem. 58 (1986) 1359. [21] K . R. Franklin, R. P. Townsend, J. Chem. Soc. Faraday Trans. 1 81 (1985) 1071. [22] K . R. Franklin, R. P. Townsend, J. Chem. Soc. Faraday Trans. 1 81 (1985) 3127. [23] R. Harjula, A . Dyer, S. D . Pearson, R. P. Townsend, J. Chem. Soc. Faraday Trans. 88 (1992)1591. [24] R. Harjula, A . Dyer, R. P. Townsend, J. Chem. Soc. Faraday Trans. 89 (1993) 977. [25] G . H . Kuhl , J. Catal. 29 (1973) 270. [26] G . H . Kuhl , A . E . Schweizer, J. Catal. 38 (1975) 469. [27] W . Lutz, B . Fahlke, U . Lohse, M . Bulow, J. Richter-Mendau, Crystal Res. & Technol. 75(1983) 513. [28] P. K . Dutta, B . Del Barco, J. Phys. Chem. 89 (1985) 1861. [29] L . B . McCusker, K . Seff, J. Am. Chem. Soc. 103 (1981) 3441. [30] P. Fletcher, R. Townsend, J. Chem. Soc. Faraday Trans. 1 78 (1982) 1741. [31] A . Jentys, A . Lugstein, H . Vinek, J. Chem. Soc. Faraday Trans. 93 (1997) 4091. [32] H . G . Karge, G . Borbely, H . K . Beyer, G . Onyestyak, in M . J. Phillips, M . Ternan (Eds.): Catalysis: Theory to Practice, Interscience Publishers, New York, 1988, p. 396. [33] J. A . Rabo, M . L . Poutsma, G . W . Skeels, (Miami Beach, F L A ) , Proceedings of the 5th Int. Congr. on Catalysis, pp. 1353 (1972). [34] J. A . Rabo, P. H . Kasai, in L . O. McCaldin , G . Somorjai (Eds.): Prog, in Solid State Chem., Pergamon Press, Oxford 1975, p. 1. [35] H . G. Karge, in H . Chon, S.-K. Ihm, Y . S. U h (Eds.): Progress in Zeolite and Microporous Materials. Stud. Surf. Sci. Catal, Vol. 105, Elsevier Science, B . V . , Amsterdam 1997. [36] C. A . Fyfe, G . T. Kokotailo, J. D . Graham, C. Browning, G . C. Gobbi, M . Hyland, G . J. Kennedy, C. T. DeShutter, J. Am. Chem. Soc. 108 (1986) 522. [37] G . Borbely, H . K . Beyer, L . Radics, P. Sandor, H . G . Karge, Zeolites 9 (1989) 428. 52 [38] Q. Chen, J, J. Fraissard, Chem. Phys. Lett. 169 (1990) 595. [39] Micromeritics, Internal Publication: Application Note Number 105 (1998). [40] G . Englehardt, D . Michel , High-Resolution Solid-State NMR of Silicates and Zeolites, John Wiley & Sons, New York 1987. [41 ] Cambridge Isotope Labs Solvent Reference Chart. [42] J. N . Miale, N . Y . Chem, P. B . Weisz, J. Catal. 6 (1966) 278. [43] W. M . Meier, D . H . Olson, C. Baerlocher, Atlas of Zeolite Structure Types, Elsevier, London 1996. 53 Chapter 3 Characterization of the Zeolite Starting Materials 3.1 Introduction The preparation of N a A used in this work yields a sufficient amount of material that a single batch can act as the parent material for all of the subsequent ion-exchanged products. Working from a single batch imparts a degree of homogeneity that could be jeopardized i f several smaller syntheses were combined. Careful characterization of the state of the system, with regards to crystallinity and purity, prior to modification allows any changes in the framework from the treatment of the parent material to be tracked. Powder X-ray diffraction, M A S N M R , and S E M are ideal methods to study the precursor materials because they are non-destructive and do not require any sample treatment before analysis. Heating, evacuating, or grinding the sample could alter the material so that the data collected may not be representative of the pristine material. Using highly crystalline materials which are free from impurities makes it easier to identify any contributions from damaged material resulting from post-synthesis treatments (such as ion exchange) because the data are not complicated by detritus from a poor quality starting material. While zeolite A is available commercially as N a A , CaA, and K A the variance in the properties of the material from batch to batch and from supplier to supplier coupled with the lack of information about the history of the preparation and ion exchange of the zeolite makes the commercially available materials unattractive for use in the present study. The zeolite materials were characterized through each sequential ion exchange to ascertain whether the processes involved in generating the various exchanged materials were damaging the crystallites. 54 3.1.1 SEM S E M is indispensable for identifying the condition, in terms of crystallite quality and sample purity, of the zeolite material crystallized from the synthesis gel. Wel l resolved, individual crystallites with similar morphology and size are evidence of a good quality material. A n S E M image can show the presence of amorphous material in the sample that remains from unreacted synthesis gel. While S E M is a useful technique for gaining information about the morphology and homogeneity of the sample, it gives no direct information about the internal ordering of the crystallites. 3.1.2 Powder XRD Powder X-ray diffraction is sensitive to the long-range ordering of the zeolite lattice and is a key technique for assessing the crystallinity of the sample. The ease of use of powder diffraction with its rapid data collection and analysis combined with the fact that practical zeolite applications use zeolites as microcrystalline powders makes it a useful method for the study of zeolite systems. N o two zeolite structures share the same unit cell size and symmetry which makes the powder pattern for each material unique. This "fingerprint" allows the sample to be identified based on the characteristic angular positions of the reflections in the powder pattern. When the crystal structure of the zeolite is known, as is the case with zeolite A , single crystal data can be used to generate a simulated powder diffraction pattern; a collection of which has been published for known zeolite structure types.t 1] The simulated powder diffraction pattern can be compared to the experimental pattern; a highly crystalline material wi l l show sharp reflections at the expected positions while any reflections unaccounted for in the simulated pattern can be attributed to a crystalline impurity. 55 When materials share the same framework but contain different hydrated charge-balancing cations, as is the case with N a A , CaA, and N H 4 A , the powder patterns wi l l show the same collection of reflections but their positions may change i f the exchanging ion alters the size of the unit cell. The intensities of the reflections wi l l also be altered upon ion exchange due to the different scattering strengths of the various ions and their positions inside the framework. Amorphous materials have no long-range crystalline structure and, as such, contribute only to the background of X-ray powder diffraction patterns. The powder pattern of a completely amorphous sample wi l l show a broad hump in the baseline between about 15 and 35° 26. The powder pattern of a sample consisting of both crystalline and amorphous material wi l l feature the sharp reflections of the crystalline material superimposed on the broad background contribution from the amorphous component. A n important aspect of X-ray powder diffraction is the bias it shows towards the crystalline material in the sample. Using a single powder pattern as a measure of the amount of amorphous material in the sample wi l l cause the actual amount of amorphous material present to be underestimated. Figure 3.1 illustrates this discrimination. The reflections in the two powder patterns lie at the same positions and are equally sharp. The baseline of the lower plot, corresponding to the sample with half of the component mixture being amorphous, is less flat in the 15-35° 26 region than that of the pure zeolite material but by comparing the two patterns it is not evident that fully one half of sample (b) is amorphous. This effect wi l l be greatest when the crystalline material is most highly ordered as the reflections wi l l be narrower and more intense. 56 With aluminosilicate zeolites, disorder can be introduced into the lattice by partial dealumination in which some framework aluminum atoms are converted to extra-framework I I 20 25 30 "35" 40 45 20 (degrees) Figure 3.1 Powder patterns of a sample of (a) 100%CaA and (b) an equal mixture of CaA and amorphous material (decomposed N H 4 A ) . aluminum species. The extra-framework aluminum exists as an amorphous component in the material and contributes only to the background of the powder pattern. Silicon and aluminum atoms scatter X-rays with virtually equal intensity so the framework S i / A l ratio wi l l not affect the width of the reflections in the X R D pattern of a perfectly crystalline material. Line broadening of the reflections in the powder patterns only occur due to a disruption in the long-range order of the lattice. — r ~ 40 I 10 I 15 — r ~ -r~ 30 57 3.1.3 2 9Si MAS NMR While X R D is sensitive to the long-range order of the zeolite framework, solid-state M A S N M R is sensitive to the local ordering of the atoms in the lattice. For purely siliceous 9 0 zeolites the number of signals observed in the Si N M R spectrum reflects the number of crystallographically inequivalent nuclei in the unit cell and their relative abundance. When aluminum is present in the lattice, however, the signals of the individual silicon nuclei cannot be resolved as a result of signal broadening due to the disordered distribution of aluminum Al Al Al Al Si < > < > < > ( > < > AI-O-Si-O-AI Al-OSi-O-Si Al-O-Si-O-Si Si-O-Si-O-Si S i - O S i - O S i $ 6 <> CO <> Al Al Si Si Si Si (4AI) Si (3AI) Si (2AI) Si (1AI) Si (OAI) Figure 3.2 Possible Si connectivities to Al in an aluminosilicate zeolite framework and silicon in next-nearest neighbour and further coordination spheres. In aluminosilicate zeolites with a S i / A l >1, five signals arise in the 2 9 S i spectrum due to five possible combinations of next-nearest neighbour coordinations which are shown in Figure 3.2. The number of next-nearest neighbour aluminum atoms influences the 2 9 S i chemical shift as shown in Figure 3.3. 9 0 The intensities of the five peaks in the Si spectrum (with respect to each other) change with the S i / A l ratio of the material. A s the S i / A l ratio increases, less sites with a high aluminum coordination are present in the framework and the signals corresponding to silicon sites with lower aluminum connectivities are enhanced. The reverse is true with decreasing S i / A l ratios down to a S i / A l ratio of unity where there no longer exists a distribution of silicon sites and there is perfect alternations of Si and A l in the framework. The 2 9 S i spectrum 58 of zeolite A (S i /Al = 1) is simplified since each Si is surrounded by four A l nuclei so the Si spectrum shows a single peak in the Si (4A1) region. 2 9 S i M A S N M R is a powerful technique for the detection of amorphous aluminosilicate and to assess any damage to the material from treatment. The peak width of 90 the signal in the Si spectrum of zeolite A is primarily influenced by the S i - O - A l bond 90 lengths and angles. A n amorphous component wi l l appear as a broad peak in the Si spectrum resulting from the inherent range of bond lengths and angles present in such a material. Unlike an X R D pattern, the 2 9 S i spectrum can be quantitative in terms of the numbers of silicons in different environments so the contribution of an amorphous 90 component can be measured directly from a single spectrum. Si spectroscopy is also particularly diagnostic for framework damage because small changes in the ordering of the lattice introduced by dealumination wi l l cause a distortion of the S i - O - A l bond angles which wi l l manifest itself as a broadening of the Si(4Al) peak in the N M R spectrum. 3.1.4 Nomenclature Since this work involves the preparation, modification (through ion-exchange), and activation (via thermal treatment) of zeolite A materials, it is necessary to maintain a system to differentiate between the different classes of materials involved. The starting material, N a A , is referred to as the as-synthesized material because the zeolite crystallizes out of the gel in the sodium form. The C a A and N H 4 A materials, generated through aqueous ion-exchange, are referred to as the as-exchanged materials. The three mixed C a 2 + / N H 4 + materials, prepared through contact-induced ion exchange, have CaiNFL, ratios of 50:50, 60:40, and 70:30 and wi l l be referred to as the as-mixed materials owing to the nature of their 59 preparation. These latter materials may also be referred to as the 50%Ca, 60%Ca, and 70%Ca materials respectively. The present chapter deals with the characterization of these unactivated materials; thermally activated materials w i l l be dealt with in subsequent chapters. It should be mentioned that when referring to the mixed materials and their compositions, the ratios refer to the ion equivalents in the lattice and not their absolute numbers. In the 50:50 material, half of the charge-balance is by C a 2 + while the other half is by the ammonium ions. Within the framework there are actually twice as many ammonium ions as calcium ions on account of the double-charge on calcium. The results from electron probe microanalysis are listed in Table 3.1. The weight percentages reported by the instrument have been converted to mole percentages. Electron microprobe analysis is insensitive to light nuclei and no signal was observed for nitrogen (even in the case of N H 4 A ) and, as a result, the ammonium content of the materials could not be calculated. The microanalysis results indicate that the aqueous ion exchange procedure for C a A and N H 4 A was incomplete since N a + ions remain in the materials. Table 3.1 Electron microprobe results for zeolite starting materials Silicon Normalized Molar Compositions Sample Na Ca Al Si N H 4 A 0.10 ± 0 . 0 1 0.00 ± 0.00 0.94 ± 0.02 1.00 ± 0 . 0 2 50%Ca 0.08 ± 0 . 0 1 0.27 ± 0 . 0 1 0.97 ± 0.00 1.00 ± 0 . 0 1 60%Ca 0.07 ± 0 . 0 1 0.31 ± 0 . 0 3 0.96 ± 0.02 1.00 ± 0.02 70%Ca 0.06 ± 0.02 0.33 ± 0.06 0.97 ± 0.03 1.00 ± 0 . 0 2 C a A 0.04 ± 0.00 0.48 ± 0.02 0.97 ± 0.02 1.00 ± 0 . 0 1 61 N H 4 A contains a greater amount of residual N a ions than CaA and the mixed materials contain about an average of the two parent materials. Values for the weight percentage of oxygen were obtained but, because the samples were not dehydrated, were considered unreliable because of the large amount of water present in the samples. The S i / A l ratios calculated from the data agree well with the expected S i / A l ratio of 1 and deviations are likely due to the accuracy of the method rather than to a breakdown of the framework. Several measurements were taken at different magnifications and the results were averaged. Because the ammonium contents of the materials could not be measured, the best method to evaluate the cation content of the materials is to compare the calculated C a / A l ratio with their expected values because both the Ca and A l signals are reliable and the ratio of the two wi l l not be influenced by water adsorbed in the framework. The ideal and 2_j_ calculated Ca contents for the as-exchanged and as-mixed materials based on the unit cell formula C a 2 + x N H 4 + y Ali2Sii2048 are listed in Table 3.2. The C a 2 + contents agree with the expected values and establish that the procedure of mixing the two materials by weight is an accurate means of exactly controlling the ionic contents of the mixed materials. The reasons Table 3.2 Ideal and calculated Ca 2 + contents for the as-exchanged and as-mixed materials Material mole% Ca/Al (expected) mole% Ca/Al (observed) Expected Ca2+ Content/u.c Observed Ca2+ Content/u.c N H 4 A 0.00 0.00 ± 0.00 0 0.00 ±0.00 50%Ca 0.25 0.28 ±0.01 3 3.31 ±0.15 60%Ca 0.30 0.32 ±0.04 3.6 3.80 ±0.52 70%Ca 0.35 0.34 ±0.09 4.2 4.14+1.09 CaA 0.50 0.49 ± 0.04 6 5.90 ±0.46 62 for preparing mixed materials from C a A and N H 4 A rather than from N a A and N H 4 A w i l l be discussed in Chapter 4. 3.2 As-synthesized NaA. The S E M micrograph of a sample of as-synthesized N a A prepared in the present work is shown in Figure 3.4. The image indicates that the as-synthesized material consists of well-resolved crystallites with a uniform size distribution having the cubic morphology expected of zeolite A . While the general form of the crystallites is cubic they lack the sharp edges often seen in micrographs of zeolite A . [4-9] The lack of sharp edges is not an indication of poor crystal quality but is due to the crystal morphology of very small (<1 pm) crystallites where the contribution from the (111) crystal faces is more evident than with larger crystals. HO] The micrograph shows no evidence for any appreciable amount of non-crystalline material in the sample. The S E M image of N a A establishes that the material is of good quality and free of unreacted aluminosilicate gel. These properties combine to produce Figure 3.4 S E M image of (a) as-synthesized N a A and (b) commercial N a A (Aldrich) a material which is ideal for modification and activation studies since the changes in a clean sample wi l l be more easily observed than in a less homogeneous sample. For the purpose of 6 3 comparison, S E M images were acquired of a commercially available zeolite A powder (Aldrich) and a representative micrograph is shown in Figure 3.4 (b). The contrast between a carefully controlled synthesis demonstrated in this work and a commercial preparation is evident in the image. The powder X R D pattern of N a A is shown in Figure 3.5 The pattern is characteristic of a highly crystalline material, showing sharp reflections that remain well-resolved up to higher angles (50° 20). The region between 15 and 40° 28 shows little evidence of any contribution from an amorphous fraction in the sample. The slight roll in the baseline may be due to a small amorphous component in the parent material or it could be incoherent -JUUU 25 10 15 I 20 30 35 40 45 20 (degrees) Figure 3.5 X R D pattern of as-synthesized N a A . 64 scattering from the instrument. The baseline contribution does not change from preparation to preparation and is common to all of the as-synthesized, as-exchanged, and as-mixed materials. Figure 3.6 shows the powder patterns of the as-synthesized material and the simulated powder diffraction pattern of hydrated N a A . The powder pattern of the as-synthesized N a A shows all of the expected reflections and no unaccounted for reflections that would be evidence for crystalline impurities. i i 1 1 1 1 1 1 1 5 10 15 20 25 30 35 40 45 20 (degrees) Figure 3.6 (a) powder pattern of as-synthesized NaA and (b) simulated powder pattern of hydrated NaA. 65 3.3 As-Exchanged CaA and NH4A The ion-exchange reactions were carried out at room temperature to minimize the chance of damaging the material. Figure 3.7 shows the S E M images for C a A and NH4A. The micrographs are comparable with the images collected from N a A . The exchange process did not seem to be destructive to the material as there is no evidence of crystallite fragmentation. Figure 3.7 S E M images of (a) as-exchanged C a A and (b) as-exchanged N H 4 A The powder X R D patterns for the N a A , CaA, and N H 4 A are shown in Figure 3.8. The powder patterns of the as-exchanged materials are comparable to that of the as-synthesized N a A material in that all three patterns show sharp, well-resolved reflections. The flat baseline through the amorphous fingerprint region between 15 and 30° 29 indicates that the ion-exchange process was not destructive to the framework of the parent material. 3.4 Characterization of the Mixed Materials Establishing that two homo-ionic zeolites have undergone contact-ion exchange is not trivial. Bulk analysis of a sample containing equal quantities of C a A and N H 4 A would yield the same result as testing a sample where each crystallite contained 50%Ca 2 + and 5 0 % N H 4 + 66 ions. The contact-induced ion exchange process has been followed using X-ray diffract ionj 1 1 ] I R [ 1 2 1 and Ramant 1 3 ] spectroscopy, and l 2 9 X e t 1 4 ] and solid-state N M R l 1 1 ] Studying the process of contact-induced ion exchange is only possible i f the two materials being studied have properties dissimilar enough to be resolved by the instrumental technique used because, unlike a typical solid-state ion exchange, there are no volatile products or inorganic salts produced as part of the ion-exchange process. In the zeolite A system, the exchange of sodium ions in the as-synthesized material for calcium or ammonium ions changes the unit cell dimension slightly. The result is that the powder patterns for the C a A and N H 4 A materials are analogous but the positions and If K ^ A J J 22 I 26 I 30 24 28 " T -32 34 I 36 20 (degrees) Figure 3.9 Overlaid powder patterns of CaA(solid line) and NH4A(dashed line). 68 intensities of the reflections are different. Figure 3.9 shows an overlay of a section of the powder patterns of the two as-exchanged materials. The small separation of the reflections allows the contributions of each material to the powder pattern of a mixture to be assigned. A n intimate mixture of C a A and N H 4 A should produce a powder pattern that is, in principle, an algebraic addition of the two individual powder patterns. Figure 3.10 shows a portion of the powder pattern of a mixture containing equal amounts (by weight) of C a A and N H 4 A immediately after mixing. The contributions of the two materials to the powder pattern can be clearly established. The C a A pattern lies at slightly higher angles which is due to its slightly smaller unit cell. 1 1 1 1 1 1 1 1 22 24 26 28 30 32 34 36 29 (degrees) Figure 3.10 X R D pattern of a 50:50 mixture of CaA and N H 4 A after mixing. Figure 3.11 shows a stacked plot of a region of the X R D patterns taken of the 50:50 material at different times as it approached exchange equilibrium. The split peaks visible in 69 s-a © X cu "5 s 2 • JS 9t r-S •8 CU <D i-bo CD 3 CD CN -8 ^ -a CU m M g § >-cu « cu ^ JS cs U <« -a "o CU w a I o O T3 cu C fc- O S u •** ±~ .a fl fi -2 © - 0 ir, S © * O « a ^ o >r, S3 CU w '-5 3 § CU £ 1 •SP fi the early stages of exchange can clearly be assigned to a heterogeneous mixture of the two powders. A s the ion exchange progresses, the split peaks begin to coalesce into broad peaks. The broadened reflections in the powder pattern df the mixed material are a result of framework disorder brought about by an inhomogeneous distribution of the two charge-balancing cations in the crystallites and not by framework destruction. A t equilibrium the inhomogeneity is removed, the reflections sharpen as the powder pattern becomes an average of the two contributing powder patterns, and each crystallite contains a balance of Ca and NFL; + ions. The contact-induced ion exchange was followed using a single sample packed in a powder tray. As the Figure 3.11 suggests, the mutual exchange of the two ions proceeds slowly at room temperature. It was not determined how long the system requires to reach equilibrium at ambient conditions but based on the trend, the time required to ensure complete exchange would be unnecessarily long. The use of very mild hydrothermal conditions greatly facilitates the exchange process. Freshly mixed samples were exposed to an atmosphere saturated with water vapour at 45 °C, showed sharp, single reflections in their powder patterns after two days and were considered to have reached equilibrium after three days under these conditions. The powder patterns for the three mixed materials are shown in Figure 3.12. The powder patterns for the 50%Ca and 60%Ca materials are virtually indistinguishable except for very slight differences in the intensities of some of the reflections which could be due to preferred orientation effects. The 70%Ca pattern is offset by 0.1° 20 to higher angles, compared to those of the 50%Ca and 60%Ca materials, possible reflecting the larger C a 2 + contribution to the material. 71 The Si N M R results for N a A , CaA, and N H 4 A are shown in Figure 3.13. The C a A spectrum shows a peak which is much broader than the peak in the N a A spectrum. This may be the result of substituting a divalent ion for a monovalent ion in the zeolite A lattice since the N H 4 A spectrum shows a narrower peak. The chemical shifts of the three materials are similar with the N a A resonance at -88.9 ppm,[15] C a A at -89.4 ppm, and N H 4 A at -90.4 ppm. The N a A spectrum is clean with no broad component that would indicate the presence of amorphous material. The C a A and N H 4 A spectra are also free of a broad component which signifies the ion exchange process was non-destructive to the materials. The Si spectra of the mixed materials are also shown in Figure 3.13. The three spectra show resonances which are broader that the N a A resonance but are the same width as the peak in the C a A spectrum. The three mixed materials have the same chemical shift of -89.9 ppm which, unfortunately, prevents the chemical shift from being used to determine the contents of the as-mixed materials. The three as-mixed materials also show no evidence of an amorphous fraction. If the material were a heterogeneous mixture of C a A and N H 4 A the 29 * Si spectra would be expected to show two peaks with an intensity ratio equal to the mix composition. Since only one resonance is observed in all three materials, the data support the materials being a homogeneous mixtures of C a 2 + and N H 4 + . The S E M , X R D , and N M R data collected on the various materials are self-consistent and establish that the N a A parent material is of good quality, aqueous ion exchange need only be carried out at room temperature to achieve high degrees of exchange, and that contact induced ion-exchange occurs between C a 2 + and N H 4 + in the zeolite A system. The data further establish that neither the aqueous nor contact-induced ion exchange process is harmful to the zeolite framework. 73 NH 4A I | I I M | " " I " ' I I H " I " > I | I I I I | I I I I | U " I " " I " " I " " I " " I " " I " " I ' " I | I I I I | I I I I | I II I | I I I I | I I I I | I I I I | II I I | I I I I | I I I I | I I I I | I I I I | I I I I | I II I | I I I I | II I -60 -70 -80 -90 -100 -110 -120 -130 -60 -70 -80 -90 -100 -110 -120 -130 8 ( P P m ) . 6 (PPm) Figure 3.13 Si NMR spectra of the as-synthesized, as-exchanged, and as-mixed materials 74 3.5 References [I] M . M . J. Treacy, J. B . Higgins, v. B . J.B, Collection of Simulated XRD Powder Diffraction Patterns for Zeolites, Elsevier 1996. [2] E . Lipmaa, M . Magi , A . Samoson, A . R. Grimmer, G . Englehardt, J. Am. Chem. Soc. 702(1980) 4889. [3] C. Fyfe, J. M . Thomas, J. Kl inowski , G . C. Gobbi, Angew. Chem. Int. Ed. Engl. 22 (1983)259. [4] G . Scott, A . G . Dixon, A . Sacco, R. W. Thompson, in P. A . Jacobs, R. A . van Stanten (Eds.): Zeolites: Facts, Figures, Future, Elsevier Science Publishers, Amsterdam 1989, p. 363. [5] E . I. Basaldella, J. C. Tara, Materials Lett. 34 (1998) 119. [6] D . W . Breck, Zeolite Molecular Sieves, John Wiley and Sons, Inc, New York 1984. [7] J. Cir ic , J. Colloid and Interface Science 28 (1968) 315. [8] L . Gora, R. W . Thompson, Zeolites 18 (1997) 132. [9] J. F. Charnell, J. Crystal Growth 5 (1971) 291. [10] J. R. Agger, N . Pervaiz, A . K . Cheetham, M . W . Anderson, J. Am. Chem. Soc. 120 (1998)10754. [II] C. A . Fyfe, G . T. Kokotailo, J. D. Graham, C. Browning, G . C. Gobbi, M . Hyland, G . J. Kennedy, C. T. DeShutter, J. Am. Chem. Soc. 108 (1986) 522. [12] G . Borbely, H . K . Beyer, L . Radics, P. Sandor, H . G . Karge, Zeolites 9 (1989) 428. [13] Y . Huang, R. M . Paroli, A . H . Delgado, T. A . Richardson, Spectochim. Acta Part A 54(1998) 1347. [14] Q. Chen, J, J. Fraissard, Chem. Phys. Lett. 169 (1990) 595. [15] G . Englehardt, D . Michel , High-Resolution Solid-State NMR of Silicates and Zeolites, John Wiley & Sons, New York 1987. 75 Chapter 4 Activation and Thermal Stability of the Zeolite A Materials 4.1 Introduction Zeolites in their hydrated forms are not useful for many adsorption processes because much of the internal volume of the crystal is filled with water. Dehydrating the materials by exposing them to high temperatures makes possible the adsorption of other guest species into the void space of the framework. In a zeolite which contains only alkali cations, the thermal activation process involves the removal of adsorbed water while zeolites containing ammonium ions also undergo a simultaneous de-ammoniation (or decationization) process whereby the ammonium ions are converted into free ammonia and protons. The gaseous ammonia is released and the protons are thought to remain in the framework attached to the bridging oxygen of an S i - O - A l group. Temperatures around 500 °C are necessary to achieve complete activation of the zeolite A materials. It is important to establish the thermal behaviour of the materials rather than assume they are stable to the activation process. To be a viable sorbent or catalyst, the zeolite framework must remain intact through the activation process. Careful characterization of the materials through the activation process also helps to establish that the resulting sorption and catalytic properties come from a homogeneous, tailored material and are not complicated by the presence of a large amorphous fraction. The thermal stabilities of the homo-ionic forms of zeolite A have been established^'^] and the results show that the resident cation influences the thermal stability of the zeolite. The sodium form of zeolite A is stable to 700 °C after which the material begins to recrystallize to Carnegeite and Nephelinel^] Exchanging the sodium ions for C a 2 + increases the thermal stability of the framework to ca. 800 °C after which the material begins to 76 become amorphous. The ammonium form of zeolite A is unstable at relatively low temperatures and decomposes completely below 200 °C. 4.2 Thermal Stability of NH 4A Attempts at producing a stable, 100% protonic form of zeolite A have been unsuccessful. Attempts to generate the acid form of zeolite A have been made by slowly heating a sample of N H 4 A to remove adsorbed water prior to activation and by the activation of N H 4 A in a stream of ammonia. 17] Attempts at generating stable acid forms by activating a N a A sample in an ammonia/HCl vapour mixture!^] and by direct exchange of a N a A sample using p-toluenesulfonic acid[9] were similarly unsuccessful. It is seems possible, however, to remove the water from the N H 4 A framework under vacuum at room temperature and leave the ammonium ions, and hence the structure, intact.DO] The thermal behaviour of N H 4 A is illustrated by the X R D patterns in Figure 4.1. Samples of the material were placed in a muffle furnace, heated at 10°/min from ambient to 350 °C. The temperature program included an isothermal dwell of 15 min after each target temperature was reached at the end of which the sample was removed and cooled to room temperature prior to X R D analysis. A t temperatures greater than 150 °C, the framework begins to show a great deal of degradation. N H 4 A is not only unstable toward elevated temperatures but also to conditions that promote the removal of ammonia. The X R D pattern (shown in Figure 4.2) of a sample of N H 4 A maintained at 100 °C and evacuated overnight demonstrates that the material decomposes almost entirely into amorphous material. 77 . o • l O o co CU no •o ?ri W) e 93 JS u I cs X z o s-s o >• es JS E s-e s-oo The A l - O - S i bonds are subject to attack by the acidic framework protons and, in a lattice that is charge-balanced only by protons, one bond of each aluminum-based tetrahedron is subject to cleavage by the hydrolysis of a framework oxygen. Widespread hydrolysis of the framework would be sufficient to cause complete degradation. A mechanism for framework hydrolysis is shown in Equation 4.1 P1] which shows how a bridging, acidic hydroxyl group can be converted to a terminal silanol group. ? . N H ; <j> O — A l — O — S i — O A A O H + O I- I I O—Al O — S i — O I O O H O • O—Al O — S i — O (4.1) I • o A A A While it has been established that N H 4 A is unstable to activation, it nonetheless provides a useful reference with which to compare the behaviours of the mixed materials. 30 10 15 20 " T -25 35 40 45 29 (degrees) Figure 4.2 Room temperature powder XRD pattern of N H 4 A evacuated overnight at 100 ° C . 79 4.3 Thermal Stability of CaA A study of the thermal behaviour of C a A was conducted in the present work to establish, in part, the durability of the C a A framework to severe thermal conditions and to establish another reference for the thermal behaviours of the mixed materials. 4.3.1 High Temperature Transformation of CaA at 825 °C At temperatures above 800 °C, the crystallites of C a A begin a crystalline-to-amorphous phase transformation. The process begins at the outside of the crystal and continues inward until the entire crystallite has become amorphous. Such behaviour has been established with the N a A systemD2] by studying the water sorption characteristics of partially transformed samples of N a A . The partially transformed samples were ground to fragment the amorphous-coated crystallites which allowed them to regain a portion of their microporous capacity by exposing the crystalline interior. A t the early stages of conversion the material is composed of crystallites that consist of a crystalline interior surrounded by a thin amorphous shell which seals off the internal volume of the crystal. We have prepared such a material from C a A treated under similar conditions to those used in the present work and confirmed the structure by Reitveld refinement of the synchrotron X-ray powder pattern. [13] At room temperature and exposed to the atmosphere, the material remained dehydrated and the cations were in the exact positions found previously for C a A fully dehydrated and kept under vacuum. [14] These experiments show clearly that the C a A structure is completely stable at temperatures below 800 °C. The conversion of the crystallites can be monitored using physisorption and X R D . Figure 4.3 illustrates the loss of surface area with time for samples of C a A heated at 825 °C for various times. The physisorption measurements were carried out at 77 K using 80 nitrogen gas as the adsorbate. The surface area of a completely crystalline sample of the C a A starting material measured by N 2 physisorption is 630 m 2 /g. Figure 4.3 shows that after 4 hours there has been appreciable conversion of the crystallites, as evidenced by the low surface area of the material. The encapsulation of the crystallites is virtually complete at 10 hours after which the surface area remains constant (ca. 8 m 2/g). The surface area at this point corresponds to the external surface of the crystallites as the interior is inaccessible to the gas. Exposure for 30 hours at 825 °C causes the bulk of the material to become amorphous and little crystallinity remains. 700 0 2 4 6 8 10 12 14 16 18 20 Time (hrs) Figure 4.3 Surface area of CaA as a function of time at 825 °C The room temperature powder X R D patterns of samples of C a A heated at 825 °C for various times are shown in Figure 4.4. Samples heated at 825 °C were exposed to the atmosphere for one day at room temperature to allow them to rehydrate before the powder patterns were collected. The X R D patterns after heating show a pronounced increase in the 81 intensity of the first reflection (hkl = 200) with respect to the rest of the reflections. This happens when a zeolite is dehydrated because of the loss of water molecules in the framework which occupy crystallographic positions.P] When the water is removed by heating, the resultant loss of scattered intensity causes an increase in the relative intensity of the first reflections with respect to the rest. The fact that when the sample was heated for 18 hours did not return to its hydrated form after exposure to the atmosphere confirms that the interior of the crystallites is inaccessible to water; in agreement with the previous diffraction study. P 3] The dehydration of a zeolite can also cause a relocation of the cations to different positions in the lattice. The locations of the N a + ions, for example, in N a A lie at different positions in the hydrated and dehydrated forms. P^] The redistribution of the cations can also affect the relative intensities of the reflections in the powder pattern of a thermally treated material compared to the untreated material. The redistribution of divalent cations in zeolite A after thermal treatment has been studied by powder X R D for Ca 2 + -Na + P6 , 17] a n ( j Na + -Mg 2 + [18] systems, and by ion-exchange measurements for a C a 2 + - N a + system.[19] The studies report that C a 2 + ions relocate to new positions within the framework when the materials are dehydrated and that the relocation of the cations after thermal treatment is irreversible as they are not restored to their original positions when the materials were rehydrated. The X R D studies relied on powder X R D results and refinement techniques and the authors claim that the identification of the cations and their locations within the frameworks are not unambiguous. Overall, the studies suggest that the materials undergo some form of cationic rearrangement upon dehydration but the precise nature of the rearrangement is unknown. 83 4.3.2 Thermal Stability at 500 °C The effects of dehydration on the C a A material were studied to establish the materials' stability to conventional activation conditions. Samples of C a A were heated at 107min from room temperature to 500 °C and held at 500 °C for 24 hrs to complete the activation process. Figure 4.5 shows the powder pattern of the as-exchanged material above a sample which has been dehydrated at 500 °C and allowed to rehydrate in air for one day. i i i i i i i i i 5 . 10 15 20 25 30 35 40 45 29 (degrees) Figure 4.5 Room temperature powder XRD patterns of (a) as-exchanged CaA, (b) CaA activated at 500 °C and rehydrated for 1 day in air. The powder pattern of the rehydrated material shows a flat baseline and the reflections remain sharp but show slightly different relative intensities with respect to the as-exchanged material. The intensity differences could perhaps be due to the relocation of some of the C a 2 + ions as a result of the thermal activation. The structure of a single crystal of 84 dehydrated C a A has been solvedD^] and approximately one-fifth of the cations were found to be located inside the sodalite cages. The Ca ions in the as-exchanged material must be contained within the large cavity linked by the 8-membered rings because when Ca is hydrated it is too large to fit through the 6-membered rings of sodalite cages. The relocation of the cations must, therefore, occur as a result of the activation process. Through the Reitveld refinement of the synchrotron X-ray powder pattern of a sample of C a A treated at 825 °C, it was found in this laboratory that the electron density in the interior of the sodalite cages may be the result of an artifact in the refinement process and not from C a 2 + ions located at that position. [20] The crystal structure of hydrated C a A is unknown, so it is not possible to compare the positions of the ions before and after activation. The activation and rehydration procedure was carried out a second time on the hydrated, activated sample. The powder patterns of the material before and after the second activation/rehydration step were identical. This result indicates that any initial relocation of the cations in the structure is permanent and irreversible and that the C a A system is completely stable to activation at 500 °C. 4.4 Thermal Stability of the Mixed Materials The goal in using mixed C a 2 + / N H 4 + zeolite A materials was to impart the thermal stability of C a A to the unstable N H 4 A framework while retaining the Bronsted -acid character generated by activating an ammonium zeolite. Claims of producing a mixed material can be held in contention i f it is not clearly demonstrated that the material is a uniform hybrid of the two parent materials. The difficulty in establishing this lies in confirming whether the material is, indeed, a homogeneous mixture of the two ions and whether the protons produced upon activation reside within the crystalline framework or as 85 part of a separate amorphous fraction. Without establishing that the structure of the material remains intact through the activation process, the measured sorption and catalytic properties of the sample cannot be unequivocally attributed to the characteristics of a hybrid material Previous attempts have been made to produce thermally stable, mixed-ionic forms of zeolite A that, when activated, could have acidic character. The materials were generated by partial exchange of the alkali cation in a dilute aqueous solution containing ammonium ions except for one study that carried out the direct exchange of the alkali cation ion for protons. [9] Proposed mixed N a + / N H 4 + forms of zeolite A have been prepared for a thermal analysis study of zeol i tespl] and infrared (IR) spectroscopic studies of the hydroxyl groups in zeolite A . [22-24] A study of the water and «-hexane sorption characteristics has also been carried out on a series of N a + / N H 4 + and C a 2 + / N H 4 + zeolite A materials.t 7l The previous studies lack detailed structural data on the materials. Characterization of the mixed materials was limited to, in most cases, elemental analysis or sorption measurements. Some of the studies carried out powder X R D measurements but no data were presented. The principal s tudies^ 9, 22] W ere carried out in the 1970s, before conventional solid-state N M R spectroscopy and, as a result, no 2 9 S i or 2 7 A 1 M A S N M R results have been reported in the literature for mixed-ion systems of zeolite A that contain ammonium ions. Zeolite A materials charge-balanced by C a 2 + ions were preferred because of the high thermal stability of CaA, the additional void space created within the framework as a result of having only one half of the original ions present in the lattice, and the effective increase in pore diameter when C a 2 + is substituted for N a + . This last property is reflected in the fact that N a A does not adsorb linear hydrocarbons while C a A does. 86 In order for N a A to adsorb linear hydrocarbons, some of the N a ions need to be exchanged by a divalent cation or by ammonium (which converts to a proton upon activation). One sodium ion needs to be removed from at least two of the pores in the unit cell of N a A to allow linear hydrocarbons access W i t h l 2 N a + ions in the unit cell, a minimum exchange of roughly 17% is required to sufficiently open the pore structure. A t such low exchanges, the diffusion of hexane into the framework is restricted. Since the pore structure of C a A is capable of sorbing linear hydrocarbons at any NH44" loading, a mixed material prepared with C a A and N H 4 A can, i f required, be prepared with very low amounts of ammonium ions. 4.4.1 Structural Characterization The mixed C a 2 + / N H 4 + materials used in the present work were activated in air by heating at a rate of 10°/min from room temperature to 500 °C and were held at 500 °C for 24 hours to ensure that they were free from water and ammonia. The materials were allowed to rehydrate in air for one day before the X R D patterns and N M R spectra were collected. Materials activated for the physisorption study were transferred hot from the furnace, as outlined in Chapter 2, to ensure that the materials did not absorb water prior to analysis. The X R D patterns of the rehydrated activated mixed materials are shown in Figure 4.6. The powder patterns of the three activated mixed materials are indistinguishable. They show a baseline that is slightly less flat than that of the 100%Ca material which could suggest the presence of some amount of amorphous material in the sample. The X R D reflections are broader in patterns for the mixed materials than in the C a A reference indicating a small 87 in . o co o co t/2 ii ii N a T3 CD CN .o CM L in U O 93 U 0 s O ir, s-cu «3 S-TS >> J3 CU u i s CU S3 a cu cu 03 cu C3 0 0 oo degree of framework disorder. The removal of the ammonium cations and the distribution of the remaining cations in the lattice affect only the intensities of the reflections so the broadening may be due to disorder introduced by a small degree of framework dealumination (explained later). That the widths of the reflections do not noticeably sharpen with increasing C a 2 + content is unexpected since a higher C a 2 + content should cause the material to behave more like the pure C a A material. The fact that the activated 70%Ca and 50%Ca materials have similar powder diffraction patterns, though the ammonium content of the 50%Ca material is 40% greater than the 70%Ca material, suggests that the framework is stabilized by Ca rather than destabilized by ammonium ions. However, we found in the present study that replacing 50% of the alkali cations for ammonium ions is near the threshold exchange level beyond which the materials are unstable to activation. Figure 4.7 shows an overlay of a section of the powder patterns corresponding to as-mixed and activated/rehydrated 50%Ca samples. The region at higher angles was selected as the effect is clearer. The reduced intensities of the reflections seen in the powder pattern of the activated material compared to the as-mixed material are the result of the decomposition of the ammonium ions. The intensity of the diffracted X-ray beam from an atom or ion is proportional to the atomic number of the species. When the ammonium ions, with an effective atomic number of 7, are replaced by protons the scattered intensity from the sample decreases. A s well as being reduced in intensity, the reflections do not return to their original positions after activation and rehydration. The shift of the reflections in the activated/ rehydrated materials to higher angles indicates that the framework has undergone a small permanent contraction. 89 cu cs w "O CU 93 •a CU i . T3 93 T3 o 93 cu eS •O c 93 CU C o CU cu S-i DX) <u w C D CN CU = £ i S ce 93 93 05 CU — O. "a S S -2 <» 5 o c S3 fc» o cu CU C CU 93 S» CL 4* o rt J- £ IS « i * a 93 cu PH -C ^ H o ON CU CU & s The cause of the shift of the reflections in the powder pattern of the activated rehydrated material is not clear as no study has been made on the effect of partial cation substitution with respect to the unit cell parameter for zeolite A . The size of the charge-balancing cation influences the size of the unit cell in the homo-ionic forms of zeolite A p 5 ] so the contraction could be due to the framework changing in response to the decrease in the size of one of the charge-balancing cations (NH/—>H + ) . It is also possible, though, that the contraction could be due to partial dealumination of the framework. The removal of aluminum atoms from the structure of zeolite Y is known to cause a decrease in the length of the unit cell that is proportional to the degree of framework dealumination[26-29] D U ^ because the S i / A l ratio of H Y is considerably higher (>2.5) than that of zeolite A , a comparison between previous work on H Y and the present study is not possible. The contraction of the unit cells of the activated/rehydrated mixed materials estimated from the (600) reflection in the powder patterns at ca. 21.7° 29 is roughly equivalent for the three materials and corresponds to roughly 0.7% of the unit cell length. Framework dealumination of zeolites can occur when an ammonium-containing zeolite is activated in the presence of water. Although this has not been previously reported for zeolite A , it is well established for zeolite H Y and other systems of higher S i / A l ratios. A mechanism to explain the dealumination of zeolites upon activation has been proposed by Kerr based on studies of the acid form of zeolite Y.P9] Starting from the framework hydrolysis mechanism shown in Equation 4.1, the presence of water at elevated temperatures is thought to act on the remaining A l - O - S i bonds to completely remove the aluminum atom from the framework. Kerr has proposed that the reaction in Equation 4.3 must occur at nearly the same rate as that in Equation 4.2 to control the degree of dealumination otherwise 91 complete dealumination of the framework and a subsequent loss of crystal structure would result. However, the mechanism is simply a proposal; the actual process and mechanism of dealumination is poorly understood. Si f ? ? ? /• S h - O — A l O — S i — O + 3 H 2 0 • S i — O — H H—O—Si + A I (OH) , (4.2) I I I i_i O H O y Si O O O O I I I . I O—Al O — S i — O + A I (OH) , • O — A l — O — S h - 0 + AI(OH) 2 + + H 2 0 (4.3) I I I I I O H O 0 O 29 • 27 Si and A l M A S N M R spectra were recorded to examine the effect of activation on the frameworks of the activated zeolite materials. 2 9 S i N M R spectroscopy can be sensitive not only to the presence of amorphous material but also to the effects of framework dealumination (as outlined in Section 3.1.3). Whether the activation process partially dealuminates the framework might also be directly observed through their 2 7 A 1 spectra since the extra-framework aluminums could have different chemical shifts and/or coordinations. 27 The A l spectra of the activated materials are shown in Figure 4.8. A l l four spectra show a small signal at 0 ppm which is characteristic of aluminum with octahedral coordination. Since all of the aluminum atoms in a normal zeolite are bound tetrahedrally in the framework, the presence of aluminum in higher coordination states indicates that some of the framework aluminum atoms have been altered. Whether the octahedral aluminum atoms are still coordinated in some way to the framework or are part of a separate fraction cannot be determined from the 2 7 A 1 spectrum alone. The proportion of octahedral to tetrahedral aluminum in the spectra is very similar between the three mixed materials. The contribution 92 of octahedral aluminum is relatively small (<10%) compared to the large, tetrahedral peak which illustrates that even i f the frameworks have undergone some dealumination, the bulk of the frameworks remain intact. 29 The Si spectra of the activated materials are also shown in Figure 4.8. The large peak at ~ -90 ppm in the spectrum of each of the activated mixed materials broadens after activation indicating that the frameworks of the mixed materials are less highly ordered. One cause of the disorder is apparent in the small shoulders that are evident in all of the spectra and lie on the high field side of the main peak. The relative intensities of the shoulders to the Si(4Al) peak in the spectra of the activated materials is greatest in the 50%Ca material and decreases with decreasing N H 4 + content. The shoulders could be the result of partial dealumination of the framework and may correspond to Si(3Al) and Si(2Al) sites within the lattice. The C a A and 70%Ca materials show only a small contribution from Si(3Al) sites while the 50 and 60%Ca materials appear to show a larger contribution from the Si(3Al) sites as well as a small shoulder in the Si(2Al) region. The spectra of the 50 and 60%Ca materials also show a broad feature at the far-right edge of the spectra, centered approximately at -105 ppm, which may be due to a small amount of amorphous material. 29 The Si spectra were deconvoluted using composite line shapes where the relative proportions of Gaussian and Lorentzian character were varied to achieve the best fit to the experimental data. The widths of the various Si(nAl) peaks were set equal to the experimental Si(4Al) peak width. The peak areas were used as an estimate of the contributions of the various Si(nAl) sites to the spectra as well as the contribution of any amorphous material. Because of the limited resolution of the spectra, there are substantial errors in the deconvolutions. 93 29 Si 27 Al -60 -70 -80 -90 -100 -110 -120 -130 160 120 80 40 0 -40 -80 8 (PPm) 8 (PPm) Figure 4.8 2 9 Si and 2 7 A l solid-state MAS NMR spectra of the activated materials. Spinning sidebands are marked with an asterisk and an impurity is marked with an "i". 94 The deconvoluted Si spectrum of the activated/rehydrated 50%Ca material is shown in Figure 4.9. It was necessary to fit the experimental Si(4Al) peak with two independent curves to achieve a good fit to the data. This does not infer that there are two distinct Si(4Al) sites, however, but the sum of the two curves produces a good fit to the experimental anisotropic peak shape. The anisotropy of the experimental peak may be caused by a distribution of Si(4Al) environments throughout the framework but it could also be the result of inaccurate phasing. Without well-defined peaks, phasing the spectrum can be problematic which, as a consequence, may cause a skewing of the Si(4Al) peak and lead to the calculated contribution of the Si(4Al) sites to be underestimated. The spinning sidebands were assigned to the Si(4Al) peak because the sidebands were separated from the experimental Si(4Al) peak by a frequency equal to the spinning rate. Accordingly, the intensities of the spinning sideband peaks were added to the contribution of the Si(4Al) sites to the spectrum. The various estimations that were employed during the peak fittings and the errors that can occur from imprecise phasing combine to increase the weighting of the Si(3Al), Si(2Al), and amorphous contributions. The quantities of these species may be less than calculated. The S i / A l ratio can be calculated from the peak areas of the Si(nAl) peaks in the Si spectra using the following equation where n refers to the number of next-nearest neighbour aluminum nuclei and I is the peak area: [30] S i / A l = I I „ /Z0 .25-« - I„ (4.4) The results of the deconvolutions and the calculated S i / A l ratios are given in Table 4.1. 95 calculated 4^0 5o 1$0 ^ T u O 1^20 1^40 5 (ppm) Figure 4.9 Deconvolution results of the activated/rehydrated 50%Ca material. The experimental pattern is shown below with the individual peaks fits (dashed lines). The spectrum calculated from the sum of the individual peak fits is shown above, 'ssb' marks the spinning sidebands 96 Table 4.1 Contributions of the various Si(nAl) sites and amorphous material to the Si spectra of the activated mixed materials 50%Ca 60%Ca 70%Ca %Si(4Al) 68.5 ± 1.5 76 ± 4 91.5 ± 1.5 %Si(3Al) 14 ± 2 1 2 ± 2 8 ± 2 %Si(2Al) 5 ± .1 2 ± 1 -%Amorphous 12.5 ± 5 10 ± 3 -S i / A l 1.23 ± 0 . 4 5 1.16 ± 0.84 1.03 ± 0 . 2 7 The amount of aluminum that has been removed from the framework of the activated materials can be calculated from the S i / A l ratio. The reciprocal of the S i / A l ratio represents the moles of aluminum present in the framework for every mole of silicon. Since the starting materials had an A l / S i = 1, the difference between the A l / S i ratios of the starting materials and the activated materials represents the amount of extra-framework aluminum present in 29 • the sample. The amounts of extra-framework aluminum, as estimated from the Si spectra are given in Table 4.2. The error estimate for the 60%Ca material reflects an additional complication encountered with the deconvolution of this spectrum; it was found that the experimental spectrum could be fit equally well with both Si(2Al) and amorphous contributions or, solely, by a larger amorphous contribution. While a visible shoulder to high field of the Si(3Al) peak in the 2 9 S i spectrum suggests that the 60%Ca material contains Si(2Al) sites, the errors are a conservative estimate of the contributions of the various sites. 27 The A l spectra were also deconvoluted to determine the amounts of octahedral aluminum in the spectra of the activated mixed materials. The amount of octahedral aluminum would correlate with the E F aluminum i f all of the E F aluminum that has been 97 removed from tetrahedral framework sites is transformed to octahedral species. The results of the deconvolutions of the 2 7 A 1 spectra of the activated mixed materials and the amounts of 90 E F aluminum calculated from the Si spectra are given in Table 4.2. The percentages of E F 90 aluminum, calculated from the Si spectra, are higher (with the exception of the 70%Ca material) than the percentages of octahedral aluminum present in the 2 7 A 1 spectra. This result suggests that not all of the E F aluminum exists as octahedral aluminum. If the amount of E F aluminum calculated from the S i / A l ratio is adjusted by the amount of amorphous material estimated to be in the samples, the adjusted values for the percentage of E F aluminum lie closer to the amount of octahedral aluminum present in the 97 A l spectra. This result may indicate that the some of the E F aluminum in the activated Table 4.2 Octahedral and tetrahedral contributions to the 27A1 spectra of the activated mixed materials 50%Ca 60%Ca 70%Ca %Tetrahedral A l 92 ± 0.7 93 ± 2 95 ± 0 . 1 , %Octahedral A l 5.7 ± 1 . 4 3.9 ± 1.2 3.9 ± 1.0 % E F A l (from S i / A l ratio) 18.7 ± 6 . 8 13.8 ± 9 . 9 2.9 ± 0 . 8 % E F A l (adjusted) 6.2 ± 4 . 7 3.8 ± 3 . 9 2.9 ± 0 . 8 materials exists as part of a small amorphous aluminosilicate component while the remainder of the aluminum is present as a small, separate, octahedral fraction. If the aluminum present in the amorphous fraction exists in a tetrahedral environment, the 2 7 A 1 signal from the amorphous material would add to the tetrahedral framework peak and cause the amount of tetrahedral framework aluminum to be somewhat overestimated. 98 2 7 A l spectra collected at 208 M H z (800 M H z for *H) of the same activated samples of N H 4 A and 50%Ca previously run at 104 M H z are shown in Figure 4.10. The higher magnetic field produces a narrowing of the lines compared to the spectra run at 104 M H z and gives much better resolution of the different aluminum environments. The broad peaks in the spectrum of the activated N H 4 A material are a result of the disordered nature of the amorphous material. The sharpness of the tetrahedral resonance clearly indicates that the 50%Ca material is much more highly ordered and crystalline than the N H 4 A sample. At the higher field, a peak at ca. 30 ppm is apparent which is characteristic of aluminum in a 5-coordinate environment. The spectrum of activated N H 4 A shows almost equal contributions from tetrahedral and 5-coordinate aluminum but only a small contribution from octahedral aluminum. This is interesting because it is unusual to find large amounts of 5-coordinate A l in activated zeolites but it was not the subject of the present study to characterize this species. The 208 M H z 2 7 A 1 spectra were deconvoluted and the contributions of the various environments to the spectra are listed in Table 4.3. The contribution o f the 5-coordinate peak to the spectrum may be overestimated as a result of the manner in which the spectrum was deconvoluted. The 2 7 A 1 spectra were deconvoluted using line shapes having varying amounts of Gaussian and Lorentzian character to achieve the best fit with the experimental data. A l l the peaks in the spectra were fit with symmetrical peak shapes (as the resolution is not sufficient to do otherwise) but in fact the peaks in a 2 7 A 1 spectrum wi l l all show an anisotropic line shape to some degree that tails to the high field side of the peak. The 97 anisotropy of the peaks is greater at lower field and is somewhat more evident in the A l spectra in Figure 4.8. Without complete resolution of the 4- and 5-coordinate peaks it is not possible to estimate the anisotropy and, therefore, the exact amount of overlap. 99 it, VI L> CU -S CU on a « fl S w < CO O o • o CN O • O ho a a to o - O * 2 TJ fl cu S •a I 73 S S 2 cu s-C3 > c CU fl C3 S • • N •o bD 2 fl S3 X i cu "a 5 * a ~ # 2 o •3 s e3 -g I 2 fl cu - °" 93 cn U cu s ? x i c> O cu " X i •o fl cu 45 cu C3 X ! cu SB £ O O o - O CN o -o co s 93 cu 55 cu . a a £ — "O r l S-N 2 S3 S 0 a 1 S cu S CUD 12 If the 5-coordinate peak in the 50%Ca spectrum is due to an amorphous fraction then it is possible that there might be a corresponding 4-coordinate contribution that lies underneath the framework tetrahedral peak. If the amorphous component in the 50%Ca sample is similar in composition to the amorphous N H 4 A sample and the ratios between the environments are the same, then the contribution of amorphous material, based on the 5-coordinate peak, to the 2 7 A 1 spectrum of the 50%Ca sample could measure 24% ± 5%. Considering the possible errors in calculating the peak areas, particularly the asymmetry of the very large tetrahedral peak, this figure is considered comparable to the (less ambiguous) 90 estimate of - 1 3 % of amorphous material calculated from the Si spectrum. 27 Table 4.3 Contributions of the various aluminum environments to the 208 M H z Al spectra of samples of activated N H 4 A and 50%Ca NH4A 50%Ca 4-coordinate 54 ± 5 87 ± 2 5-coordinate 3 8 ± 4 9 ± 1 6-coordinate 8.7 ± 1.6 4.5 ± 0 . 3 The 2 9 S i spectra show that the proportions of Si(3Al) and Si(2Al) sites in the spectra increase with increasing ammonium content while the 2 7 A 1 spectra show a similar trend with respect to octahedral aluminum. The 2 9 S i and 2 7 A l M A S N M R results are therefore self-consistent as a greater amount of dealumination should produce more Si(3Al) and Si(2Al) framework sites. In general, all the spectra demonstrate that even materials with large ammonium contents are thermally stable and do not show significant decomposition upon activation. 101 The activated/rehydrated mixed materials are not stable to reactivation and i f the mixed materials are reactivated at 500 °C the X R D patterns show a marked destruction of the framework which increases with increasing ammonium content. This behaviour has been noted previously in a series of N a + - N H 4 + materials characterized through sorption measurements.U> 9] Activated samples of Na +/NH4 + zeolite A which showed high sorption capacities exhibited very low sorption capacities after rehydration and reactivation. The loss of sorption capacity increased with increasing ammonium exchange of the materials. The frameworks of mixed zeolite A materials containing ammonium seem to be sensitive to exposure to water at elevated temperatures after the initial activation. 4.4.2 Sorption Characteristics The sorption characteristics of the materials contribute to a wider understanding of their framework properties in terms of the internal volumes of the crystallites and the dimensions of the micropores. Measurements of the adsorption properties and micropore dimensions of zeolites have been carried out using a variety of probe molecules. [5] Previous studies establish the effective pore diameters and adsorption capacities for the homo-ionic forms of zeolite A . The adsorption of rc-hexane is used as a benchmark test for measuring the adsorption capacity of zeolite materials. Reported values for rc-hexane adsorption in C a A vary between 12.6 g hexane/100 g zeol i te^] and 14.1 g hexane'100 g zeoi i tet 9 ] at 25 °C. Hexane adsorption data for the samples used in the present work are presented in Table 4.4. The value for C a A is somewhat higher than the commonly accepted value of 12.6 g hexane/100 g z e o i i t e which is likely the result of the experimental conditions in the present study which maintained rigorously anhydrous conditions. Such conditions prevent the materials from adsorbing 102 water which, in turn, interferes with the adsorption of hexane. These results further establish that the mixed materials have not undergone extensive framework collapse since the activated materials all adsorb similar amounts of rc-hexane. The roughly 9% difference in sorption capacity between the 50%Ca material and C a A correlates well with the percentage 29 * of amorphous material in the activated 50%Ca material calculated from the Si spectrum. Table 4.4 «-Hexane adsorption on zeolite A materials at 1 atm and 30°C % C a n-Hexane uptake (g rt-hexane/100 g zeolite) 100 13.44 ± 0 . 0 1 70 13.05 ± 0 . 0 1 60 13.05 ± 0 . 0 1 50 12.51 ± 0 . 0 1 A typical plot for the adsorption of rc-hexane on an activated sample is shown in Figure 4.11. The material quickly adsorbs rc-hexane to its maximum value. The value of the plot at the end of the plateau was considered the value for maximum adsorption. The slight upward drift of the plot is likely the result of the furnace slowly cooling during the measurement; allowing further adsorption of n-hexane. After approximately 18 minutes the hexane vapour stream was turned off and a dry N2 stream admitted. A similar study on a zeolite X systemPl] reported a sharp transition with a pronounced weight loss upon purging which the authors attributed to the adsorbate loosely bound to the external surface of an amorphous component. The gradual weight loss observed with the zeolite A samples when the purge gas is introduced can be regarded as normal evaporative loss of hexane from the framework and not loss from hexane adsorbed on 103 a separate amorphous fraction. 114 10 15 20 Time (min) Figure 4.11 Adsorption of n-hexane at 30 °C as a function of time on the 50%Ca material The surface areas and pore sizes of the activated materials were measured through gas physisorption using argon gas at 87.3 K . The surface areas obtained for the various activated materials are presented in Table 4.5. The relatively low value for the surface area of the 50%Ca material was unexpected considering the results from the hexane adsorption experiments. The change in the specific surface area from 100%Ca to 50%Ca represents a decrease of almost 30% while less than a 10% reduction in sorption capacity was observed with the n-hexane adsorption experiments. The same trend in the surface areas with respect 2+ to C a content was observed when the experiments were carried out using N 2 as the adsorbate at 77 K . Zeolite frameworks undergo a slight contraction (on the order of 1 %) at 77 K . The additional decrease in the internal volume of the crystallites may be due to the 104 blockage of some of the channels by amorphous material and/or octahedral aluminum species which could prevent the gas from fully penetrating the framework. The channel blockage by the extra-framework species may have a greater effect at low temperatures than at 30 °C which could account for the discrepancy between the n-hexane and argon sorption measurements. Table 4.5 Surface areas of the activated materials from argon physisorption %Ca 2 Surface Area (m /g) 100 647 ± 1 70 590 ± 2 60 534 ± 2 50 460 ± 2 The pore size distributions from the argon physisorption experiments are shown in Figure 4.12. The pore size distributions were calculated using the Horvath-Kawazoe (H-K) method and the treatment of the data had a profound effect on the calculated pore dimensions. The original slit-model (H-K) method, used without accounting for the aluminosilicate framework, yielded pore dimensions that were closest to the values expected from the previous sorption work. The calculations were repeated varying both the adsorbent parameters and the underlying method, i.e., slit pore or cylinder pore t 3 2 ] model and a large variation was found in the position of the maximum of the differential plot of the pore diameter with respect to pore volume. The results indicate that the calculation of the pore dimension of zeolites is very dependent on the parameters used in the calculation which is in accordance with previous studies. [32-34] p o r m e purposes of the present study, the absolute 105 (yB/oo) Q P / A P (yBpo) Q P / A P C3 a « %-» u S3 .o 13 © a o s « S3 o 9H o (y6/oo) Q P / A P (y6/oo) QP/AP pore size, as calculated from gas adsorption, can be disregarded because, regardless of the parameters used, any trend of the pore diameters across the series of zeolite A materials remains the same. The derivatives of the pore volume with respect to size are shown as they give a clearer representation of the mean diameter and distribution of pores in the material. It is clear from the data that the mean pore sizes and distributions of the pores in the mixed materials are virtually indistinguishable from those of the C a A material. The mean pore diameter of the activated mixed materials is 0.2 A larger than the C a A material which could be the result of fewer C a 2 + ions being present around the pore openings in the structures of the mixed materials. 4.5 Summary The X R D , N M R , and sorption experiments are self-consistent and establish that mixed C a 2 + / N H 4 + materials with ammonium contents up to 50% are stable to thermal activation at 500 °C. The 2 9 S i and 2 7 A l spectra demonstrate that while the materials undergo a small degree of dealumination upon activation and minor amorphous components are formed in the 50 and 60%Ca materials, the materials' frameworks remain essentially intact. The sorption studies establish that the activated mixed materials have similar adsorption capacities for hexane and that the pore diameters are nearly identical to the C a A parent material. The sorption results establish that the C a 2 + / N H 4 + system is preferable to a N a + / N H 4 + system because mixed Ca / N H 4 materials can be prepared with high or low ammonium contents without affecting the sorption characteristics or pore sizes of the materials. These properties are desirable for catalytic applications because a wide range of catalyst compositions is possible. 107 4.6 References [I] D . W . Breck, W. G . Eversole, R. M . Mil ton, T. B . Reed, T. L . Thomas,./ Am. Chem. Soc 75(1956) 5963. [2] H . Fichtner-Schmittler, W. Lutz, S. Amin , A . Dyer, N . Wark, Zeolites 12 (1992) 750. [3] A . Colantuono, S. Dal Vecchio, G. Mascolo, M . Pansini, Thermochim. Acta 296 (1997) 59. [4] B . L . Y u , A . Dyer, H . Enamy, Thermochim. Acta 200 (1992) 299. [5] D . W. Breck, Zeolite Molecular Sieves, John Wiley and Sons, Inc, New York 1984, p. 333. [6] Y . Huang, unpublished results . [7] G . H . Kuhl , J. Catal. 29 (1973) 270. [8] R. M . Barrer, A . G . Kanellopoulos, J. Chem. Soc (A) (1970) 765. [9] D . P. Roelofsen, E . R. J. Wils , H . van Bekkum, J. Inorg. Nucl. Chem. 23 (1972) 1437. [10] L . B . McCusker, K . Seff, J. Am. Chem. Soc. 103 (1981) 3441. [II] A . P. Bolton, M . A . Lanewalla, J. Catal. 18 (1970) 154. [12] J. L . Thomas, M . Mange, C. Eyraud, Molecular Sieve Zeolites-I. Adv. Chem. Ser. 101, American Chemical Society, Washington, D C 1971. [13] H . Meyer zu Altenschildesche, Y . Huang, G. T. Kokotailo, C. A . Fyfe, D. Cox, NSLS Brookhaven Annual Report (1996) . [14] J. J. Pluth, J. V . Smith, J. Am. Chem. Soc 105 (1983) 1192. [15] M . M . J. Treacy, J. B . Higgins,-v. B . J .B, Collection of Simulated XRD Powder Diffraction Patterns for Zeolites, Elsevier 1996. [16] H . Siegel, R. Schollner, V . J J , W . J. Mortier, Zeolites 7 (1987) 148. [17] H . Siegel, R. Schollner, B . Staudte, J. J. Van Dun, W. J. Mortier, Zeolites 7 (1987) 372. [18] H . Siegel, W. Schmitz, R. Schollner, A . Dyer, H . Enamy, Thermochim. Acta 61 (1983) 329. 108 [19] A . Dyer, H . Enamy, Zeolites 5 (1985) 66. [20] H . Meyer zu Altenschildesche,. [21] A . Dyer, Thermochim. Acta 110 (1987) 521. [22] J. F. Tempere, D. Delafosse, J. Catal. 39 (1975) 1. [23] L . M . Kustov, Y . V . Borokov, V . B . Kazanskii, Kinet. Catal. 25 (1984) 120. [24] V . Dondur, V . Rakic, Thermochem. Acta 93 (1985) 753. [25] R. M . Barrer, W . M . Meier, Trans. Faraday Soc. 54 (1958) 1074. [26] G . T. Kerr,./. Phys. Chem. 73 (1968) 2780. [27] J. W. Roelofsen, H . Mathies, R. L . deGroot, in P. A . Jacobs, R. A . vanSanten (Eds.): Zeolites: Facts, Figures, Future, Eslevier Science Publishers, Amsterdam 1989, p. 643. [28] A . Macedo, A . Auroux, F. Raatz, E . Jacquinot, R. Boulet, in W. H . Flank, T. E. Whyte (Eds.): ACS Symp. Ser. 368, American Chemical Society, Washington 1988, p. 98. [29] G . T. Kerr, J. Catal. 15 (1969) 200. [30] G . Englehardt, D . Michel , High-Resolution Solid-Slate NMR of Silicates and Zeolites, John Wiley & Sons, New York 1987. [31] G . H . Kuhl , A . E . Schweizer,./. Catal. 38 (1975) 469. [32] A . Saito, H . C. Foley, J. AIChE 37 (1991) 429. [33] A . F. Venero, J. N . Chiou, Mat. Res. Soc. Symp. Proc. Ul (1988) 235. [34] A . Saito, H . C. Foley, Micropor. Mater. 3 (1995) 531. 109 Chapter 5 Characterization of the Acidic Sites in the Activated Mixed Materials 5.1 Introduction While Chapter 4 established that the C a 2 + / N H 4 + materials are stable to activation, it is equally important to determine that the activation process has produced framework hydroxyl groups, that the hydroxyl protons exhibit Bronsted acidity, and that the acid sites are within the open framework structure. Confirming that the material possesses Bronsted acidity is accomplished by exposing the material to a basic probe molecule and examining the system for an acid-base reaction. Ammonia was used as the probe molecule in the present work as it is small enough to diffuse though the 5 A pores of the frameworks. The additional benefit of using ammonia as the probe molecule is that the reaction of ammonia with the acidic framework sites should form ammonium ions. A s ammonium ions were originally present in the as-mixed materials, the "re-ammoniated" materials should return to their as-mixed forms. Ammonia is not selective to Bronsted sites, however, and is also strongly physisorbed to Lewis acid sites within the material. The physisorbed ammonia is in vast excess and renders the signal from the framework ammonium ions and hydroxyl groups imperceptible. This disadvantage can be overcome by rehydrating the material which displaces the Lewis bound ammonia. Early studies of hydroxyl groups in zeolite frameworks were carried out using IR spectroscopyt 1 ' 2] and thermal analysis^] but these methods have largely been replaced by the use of ' r l M A S N M R spectroscopy. The study of Bronsted acidity in zeolites by multinuclear solid-state N M R has recently been reviewed.[4] The characteristic chemical shifts of the various types of hydrogen nuclei found in zeolites are shown in Figure 5.1. [5] The hydroxyl groups produced by the decomposition of the ammonium ions are generally 110 referred to as framework hydroxyl groups because of their location at a framework A l - O - S i bridge. S i - O H groups are also present in zeolites and are located at the peripheries of the crystals and at defects within the framework. Other hydroxyl groups have been attributed to O H groups attached to extra-framework aluminum. The bars representing the chemical shift ranges of the various species in Figure 5.1 are labelled: S i O H represents terminal silanol groups, A l O H represents hydroxyl groups attached to extra-framework aluminum, while the S i - O H - A l bar represents the framework hydroxyl groups. The clear separation between the chemical shifts of the various species allows the contributions of the ammonium ions, the framework hydroxyl groups, and the non-acidic hydroxyl groups to be assigned in the spectra of the mixed materials. NH, ' | Si-OH-Al A OH SiOH | i i i i i i i i i | i i i i i i i r i j i I r i I I I I i "| " i r i i i i i i i | i i i i i i i i i j i I I I i i i i i | i i i i t i i i i ^ T1 i i i i i i i i j 8 7 6 5 4 3 2 1 0 8 (ppm) Figure 5.1 *H chemical shift ranges for framework protons in zeolites.l^] 5.2 Characterization of Acid Sites in Ca 2 +/NH 4 + Zeolite A The *H M A S N M R spectra of the as-mixed and activated/rehydrated materials are shown in Figure 5.2. The spectra for the three as-mixed materials are similar and show a broad peak centered at ca. 6 ppm from the ammonium ions and a small shoulder around 2 ppm from the terminal silanol groups. I l l 50%Ca 60%Ca 70%Ca ' ' ' ' ' T 1 1 1 T"'l T"T 1" T"'T' T"T I I I I I T 1 | f l I T I I I I I | I I I I I I 1 1 T ] T 1 I I I I I I I | I 1 I I 1 I I I I | 1 I I r ' T T ' T ' T ' T J I 1 I 1 I 1 I I I | 1 I I I I I I I1 I 50 40 30 20 10 0 -10 -20 -30 5 ( P P " i ) Figure 5.2 Room temperature 'H MAS NMR spectra of the as-mixed mixed materials as indicated. The asterisks mark spinning side bands (spinning rate ~5 kHz). The inset figures are expansions of the central portions of the spectra. 112 The ' r l spectra of the activated mixed materials are shown in Figure 5.3. The three spectra all show a prominent peak at roughly 4.5 which is associated with framework hydroxyl groups. The spectra also show two resonances in the non-acidic region at 2 ppm and 2.5 ppm; the first corresponds to terminal silanol groups while the latter is likely due to hydroxyl groups attached to extra-framework aluminum. The assignment of the latter peak is supported by the characteristic chemical shift of this species and the observation that some extra-framework aluminum is present in the samples. Unexpectedly, the 50%Ca and 60%Ca materials show a small shoulder at 6 ppm, perhaps indicating that a small fraction of the ammonium ions are not removed by activation. Thermal analysis of the mixed materials has been carried out and, when maintained at 500 °C, the materials reach a constant weight after only 2-3 hours. The presence of a peak at 6 ppm after activation for 24 hours at 500 °C (as is the case for the activated materials) suggests that a small amount of ammonium ions remain in the sample. The *H spectrum, at room temperature, of an activated, anhydrous N H 4 A sample is shown in Figure 5.4. While the large peak at 4 ppm demonstrates that the material contains a large proportion of framework hydroxyl groups, the sharp peak at approximately 7 ppm confirms that a measurable amount of ammonium ions remain in the material after activation. The sharpness of the ammonium peak suggests that the ammonium ions are not engaged in hydrogen bonding with water adsorbed in the framework or that the ions are very mobile and are rapidly exchanging between different sites within the framework. 113 I I I I I I I I I I I I I I I I I I I I I I 1 I I I I I I I 1 I I I I I I I I I 1 I I I I I I I I I I I I I I I I I I 1 I I I I I" I" I 1 I I I I I I I I I I I I I I I I I 1 I I I I p I t I I 50 40 30 20 10 0 -10 -20 -30 -40 5 (PPm) Figure 5.3 Room temperature ' H M A S N M R spectra of the activated/rehydrated mixed materials as indicated. The asterisks mark spinning side bands (spinning rate ~4.8 kHz). The inset figures are expansions of the central portions of the spectra. 114 11 I I ! ! II II I I | I I I I I I I I I [ I I I I I 1 I I I | ! I I II I I I I [ ! ! I I I I I I I [ I I I I I I I I I | I I' I I I ' I I I I | 1 I I I I I I I I j I I I I I I I I I | I I 50 40 30 20 10 0 -10 -20 -30 -40 5 (PPm) Figure 5.4 *H MAS N M R spectrum of N H 4 A activated at 500 °C. The asterisks mark spinning side bands. Samples of the activated mixed materials were exposed to gaseous ammonia to establish whether the peak at ca. 4.5 ppm is due to protons which exhibit Bronsted acidity. If the peak at ca. 4.5 ppm is due to Bronsted acid sites, it should disappear and the peak at 6 ppm return as the ammonia molecules are converted to ammonium ions. The activation, re-ammoniation, and rehydration were carried out simultaneously on the three mixed materials. Samples were calcined overnight at 500 °C, transferred hot from the furnace to a Schlenk tube, and cooled to room temperature under vacuum (10"3 Torr). The vessel was exposed to 1 atm of flowing NH3 for 10-15 minutes after which the vessel was purged with nitrogen gas for 15 minutes. The samples were hydrated by placing the samples, unstoppered, in a closed desiccator containing water and equilibrating the samples at 60 °C for one day. 115 50%Ca 60%Ca 70%Ca 1 I I I I 1 I I I 1 1 I | 1 I I I T T'THTI'V I I I II I I | I I I 1 I 1 I 1 I | I I I 1 I 1 I I I J 1 "I I 1 I T 1 1 T J J 1 I 1 1 I 1 1 1 J 1 I 1 I I 1 I 1 I J I 1 1 I 1 I 1 ! I | I I 50 40 30 20 10 0 -10 -20 -30 -40 5 (PPm) Figure 5.5 Room temperature 'H MAS NMR spectra of the activated/re-ammoniated/ rehydrated materials as indicated. The asterisks mark spinning sidebands (spinning rate ~4.4kHz) while the '#' marks a spectral artifact. The inset figures are the expansions of the central portions of the spectra. 116 The H spectra of the re-ammoniated/rehydrated activated mixed materials are shown Figure 5.5. The spectra show a single large peak at 6 ppm (see Figure 5.1) indicating the reappearance of the ammonium ions. The S i O H and A l O H groups still are present after re-ammoniation which confirms their non-acidic nature. There is no visible contribution remaining from framework hydroxyl protons at 4.5 ppm but a small contribution from these species could be hidden under the broad ammonium peak. The appearance of ammonium ions and the disappearance of the framework hydroxyl protons in the activated/re -ammoniated/rehydrated materials establish that the protons associated with the framework hydroxyl groups exhibit Bronsted acidity and that they are located at positions within the framework that are accessible to ammonia molecules. While the ' H M A S N M R results provide strong evidence that the acidic hydroxyl groups lie within the crystalline framework, this cannot be confirmed from the ' H spectra alone because N M R spectroscopy is sensitive to the local ordering in the materials. The effect of re-ammoniation on the long-range order of the activated mixed materials' frameworks can be determined directly through the powder X R D patterns of the samples. A s demonstrated in Chapter 4, the activation process causes a decrease in the intensity of the reflections in the powder patterns of the activated materials due to the loss of the ammonium ions. If the acidic framework hydroxyl groups reside within the open frameworks of the materials, the restoration of the ammonium ions would be expected to reverse this effect. Figures 5.6-5.8 show the as-mixed, activated, and activated/re-ammoniated/ rehydrated powder patterns for the 50, 60, and 70%Ca materials respectively. For brevity the activated/rehydrated materials are be labelled '500 -H2O', and the activated/ re-ammoniated/ rehydrated materials are labelled '5OO-NH3-H2O' . It should be mentioned that the as-mixed, 117 5OO-H2O, and 5OO-NH3-H2O patterns are from different samples but, for consistency, all three samples were taken from the same batch. It is apparent from the three figures that when the activated materials are exposed to ammonia and rehydrated, much of the change that occurred upon activation is reversed. Though the baselines of the 5OO-NH3-H2O patterns are still slightly less flat than their respective as-mixed patterns, the reflections have appreciably sharpened compared to those of the 5OO-H2O patterns. Figure 5.9 shows an overlay of the central portion of the same three powder patterns shown in Figure 5.6. It is evident from the Figure that the material has undergone a lattice expansion, evidenced by a shift of the reflections to higher angles, but the reflections do not exactly return to their original positions. The lattice expansion observed after re-ammoniation and rehydration was unexpected and indicates that the lattice contraction observed in the 500-H2O patterns is not due entirely to the partial dealumination of the frameworks. The re-ammoniation results indicate that the small degree of dealumination that occurs during activation does not have a large effect on the unit cell of the material. It is also evident that the reflections in the powder pattern of the 5OO-NH3-H2O sample sharpen appreciably and increase in intensity compared to the powder pattern of the 5OO-H2O sample. The widths of the reflections in the 5QO-H2O samples were, in Chapter 4, attributed to disorder in the framework brought about by partial dealumination of the framework. The increase in the intensity of the reflections in the 500-NH3-H2O pattern in Figure 5.9 can be explained by the restoration of the ammonium ions but the reflections do not, however, regain their full intensities compared to those in the as-mixed pattern. This is not unexpected since the dealumination that occurs during activation removes some of the associated acid sites from the framework. This factor, coupled with the formation of a small 121 amount of amorphous material during activation prevents the sample from making a complete recovery to its original form. The powder patterns for the 60% Ca and 70%Ca as-mixed and 5OO-NH3-H2O samples show similarities to the 50%Ca patterns in Figure 5.9 but the changes become less pronounced with decreasing NH/ content. Since the 70%Ca contained the smallest number of ammonium ions, experienced the smallest degree of dealumination upon activation, and contains no significant amount of amorphous material, the re-ammoniated/rehydrated 70%Ca sample makes the most complete return to the parent material. 5.3 Summary The 'rl MAS NMR data indicate that the removal of the ammonium ions by activation produces framework hydroxyl groups. The NMR data also demonstrate that the framework hydroxyl protons display Bronsted acidity since ammonium ions are formed when the activated materials are exposed to ammonia. That the acidic hydroxyl groups lie within the frameworks is established through the powder XRD patterns of the materials which demonstrate a clear recovery toward their as-mixed forms when the samples are re-ammoniated. Confirming that the acidic character of the materials is restricted to the framework supports that any resulting catalytic activities are the product of a crystalline, acidic zeolite. 123 5.4 References [1] J. B . Uytterhoeven, P. Jacobs, K . Makay, R. Schoonheydt, J. Phys. Chem. 72 (1968) 1768. [2] J. W . Ward, in E . M . Flanigan, L . B . Sand (Eds.): Adv. Chem. Ser. 101, A C S , Washington, D C 1971, p. 380. [3] J. L. Carter, P. J. Lucchese, D . J. C. Yates, J. Phys. Chem. 68 (1964) 1385. [4] M . Hunger, Catal. Rev. 39 (1997) 345. [5] G . Englehardt, D . Michel , High-Resolution Solid-State NMR of Silicates and Zeolites, John Wiley & Sons, New York 1987. 124 Chapter 6 Exploratory Catalytic Testing of the Mixed Materials 6.1 Introduction The results of Chapter 5 establish that when the zeolite A mixed materials are activated, the resulting framework protons exhibit Bronsted acidity; but the strength of the acid sites in the activated mixed materials remained to be established. It is known that zeolites can exhibit a wide range of acid strengths depending on the structure type, the nature and number of any metal cations, and the S i / A l ratio. D] Evaluating the acid strengths of the mixed materials is an important aspect of the characterization process which can help assess the potential of the materials as solid-acid catalysts. Originally, the study of acidic hydroxyl groups in zeolites was carried out on dehydrated zeolites using IR spectroscopy [2-5] but with the development of high-resolution ' H M A S N M R , the use of IR spectroscopy has been largely replaced. The ' H chemical shift of the framework hydroxyl protons is used as a measure of the acid strength of the protons and the most acidic protons lie above 5 ppm; at the upper part of the range for framework hydroxyl groups (see Figure 5.1). The framework protons in all three of the activated zeolite A mixed materials resonate at about 4.4 ppm, which suggests that the three materials should have the same acid strength. The chemical shift of the protons in the zeolite A mixed materials also indicates that they should be less acidic than those in zeolite H Y (5 = 4.8 ppm)^ 1 ] for example. The estimation of the proton's acidity through spectroscopic methods such as ' H M A S N M R can be used to infer the potential catalytic activity of the material. Catalytic tests, however, yield a direct measure of the performance of the catalyst and, as such, exploratory 125 catalytic activity tests were carried out in the present study to establish whether the Ca 2 + /NH.4 + zeolite A materials possess activity in two standard catalytic reactions. 6.2 Catalytic Activity Study for the Cracking of n-Hexane The cracking of rc-hexane has been a benchmark catalytic test for evaluating the acid strength of zeolites [6-10] x n e t e s t j named the a-test by M o b i l , D 1 ' 12] compares the activity of the sample to that of an amorphous aluminosilicate standard. The extent to which the n-hexane feed is converted to products under standard conditions is a measure of the materials' activities. The products of «-hexane cracking include methane, ethane, ethylene, propane, propene, and a number of aliphatic and olefinic C4 and C5 hydrocarbons. The gas chromatograph used in the present study was calibrated using gas mixtures containing known percentages of all of the isomers of the C l , C2 , C3 , and C4 hydrocarbons. The integrated peak values for the various calibration gases were used to generate scaling factors that were used to convert the products' signal values into percentages and the feed conversion was calculated by summing these percentages. Certain species co-eluted on the column, e.g., propane and propene had the same retention times and several of the C4 hydrocarbons co-eluted as well . In these instances, an average scaling factor for the group was calculated from the species' individual scaling factors. While this method does confer a certain amount of uncertainty to the measurements, it is quite adequate to establish whether the materials demonstrate catalytic activity and whether the activities for the mixed materials are the same. The remaining experimental conditions for carrying out the a-test were outlined in Section 2.4.7. The various product species from the cracking reaction and their retention times are listed in Table 6.1. 126 Table 6.1 Retention times of the various product species from the cracking of «-hexane Species R (min) Methane 0.7 Ethylene 1.7 Ethane 2.2 Propane/ene 4.9 C4 species 8 . 2 - 9 C5 species 11.2- 11.6 «-Hexane 13 Not all of the peaks in the chromatograms from the cracking reactions were accounted for by the calibration gases. Calibration gases for the C5 hydrocarbons were unavailable and the assignment to the peaks at 11.2 to 11.6 min was based on the species eluting between the C4 and C6 hydrocarbons. The contributions of the C5 peaks were negligible compared to the other product species and were not included in the calculation of the conversion of the feed. Other peaks were detected at 14 and 16 minutes but the identities of these species are not known. The integrated peak values for these species are orders of magnitude smaller than the main product signals and, as a result, were disregarded. The a-value of a material represents the ratio of the activity of the zeolite catalyst versus an amorphous aluminosilicate standard chosen to have an a-value of 1. The a-values for the zeolite A mixed materials were calculated using the following equationP^] where the first term is a constant to account for the activity of the amorphous aluminosilicate standard, the packing density, p, is set to 0.55 g/cc, F is the flow rate of the feed mixture, wt is the 127 mass of the unactivated sample, and X represents the fraction of n-hexane that has been converted to products. a J - ^ ) ( p { ^ } n ( l - X ) (6.1) ^ 60 J \wtJ The results for the conversion of hexane over the various zeolite A materials and their oc-values are given in Table 6.2. Table 6.2 Catalytic activity results and a-values for the cracking of w-hexane over the various zeolite A materials Sample wt(g) F (cc/min) Conversion(%) a C a A 0.75 18 3.69 ± 0.26 0.29 ± 0.02 70%Ca 0.75 18 3.56 ± 0 . 2 5 0.28 ± 0.02 60%Ca 0.75 18 3.06 ± 0 . 2 1 0.24 ± 0.02 50%Ca 0.75 18 2.92 ± 0.20 0.23 ± 0.02 N H 4 A 0.75 18 1.15 ± 0 . 0 8 0.09 ± 0 . 0 1 Very low feed flow rates and very large sample weights were required to obtain conversions that were measurably greater than the conversion from thermal cracking alone (which causes a conversion of 1-2% of the feed). The very low feed conversions and the low a-values for the mixed materials were unexpected in light of the evidence for Bronsted acidity presented in the previous chapters. None of the C a 2 + / N H 4 + materials exhibits substantial activity towards the catalytic cracking of n-hexane and the N H 4 A sample does not exhibit any measurable activity beyond what can be accounted for by thermal cracking of the feed. 128 Samples of C a A and the three Ca / N H 4 materials were sent to M o b i l O i l for analysis on a reactor system which is dedicated to «-hexane cracking measurements. The results obtained from these tests are given in Table 6.3. The a-values obtained from this independent testing of the zeolite materials used in the present study are in agreement with the values collected in this laboratory and further establish that the mixed materials have little activity toward the catalytic cracking of M-hexane. High-temperature «-hexane adsorption measurements were carried out (as outlined in Section 2.4.4) to investigate the cause of the mixed materials' poor activities. A single Table 6.3 Catalytic activity results and a-values for the cracking of w-hexane over the various zeolite A materials, courtesy of Mobil Oil Corporation. Sample wt(g) F (cc/min) Conversion(%) a C a A 0.75 7.27 6.8 0.39 70%Ca 0.75 7.27 7.9 0.46 60%Ca 0.75 7.27 4.9 0.28 50%Ca 0.75 7.27 7.2 0.42 sample of the 70%Ca material was used for all the measurements. The sample was activated by heating at 107min to 500 °C under nitrogen gas. After 15 minutes at 500 °C, the nitrogen gas was switched off and the sample was exposed to a continuous stream of «-hexane vapour for the remainder of the experiment. The cumulative adsorption of ^-hexane as a function of temperature was measured by sequentially cooling the sample in roughly 100 °C steps from 500 °C to 200 °C and measuring the sample weight at each temperature. The results of the high-temperature rc-hexane adsorption study are shown in Figure 6.1. 129 At 500 °C, the 70%Ca material does not adsorb any measurable amount of n-hexane which indicates that the zeolite A mixed materials are unable to adsorb n-hexane at the standard temperature for the oc-test (540 °C) and 1 atm. The materials begin to absorb n-hexane below 500 °C but the combination of lower temperature and moderate uptake precludes catalytic cracking taking place with the zeolite A materials. The results establish that the lack of catalytic activity seen in the rc-hexane catalytic activity tests is most likely the result of the feed being unable to diffuse into the frameworks to react with the acid sites. 200 250 300 350 400 450 500 550 Temperature (°C) Figure 6.1 /i-Hexane uptake at various temperatures for the 70%Ca material The calculated diameter of the 8-membered ring in zeolite A from X-ray data, is 4.3 A. The kinetic diameter of the pores in zeolite A can, however, be larger or smaller than this value depending on the number and type of metal ions present in the framework. The kinetic diameter of the 8-ring pores in zeolite A is estimated by determining the largest molecule that can diffuse though the pore openings. Linear hydrocarbons such as propane, n-130 butane, and rc-hexane, have a kinetic diameter (a) of 4.3 A D 4, 15] a r e t n e largest molecules that can diffuse into C a A and, correspondingly, the mixed materials. Branched hydrocarbons such as isobutane (a = 5 A) are excluded because their kinetic diameters exceed the pore size of C a A . It is not clear whether the kinetic diameter of n-hexane at high temperature increases due to increased contributions from less favourable conformations and or whether some change in the framework causes the molecules to be excluded. 6.3 Preliminary Catalytic Study of the Conversion of MeOH to Hydrocarbons by the Zeolite A Materials In order to probe the catalytic activity of the materials, methanol (a = 3.85 A),D 5] was chosen because it has a kinetic diameter significantly smaller than n-hexane and might not be excluded from the zeolite A framework at high temperature. Methanol is a very appropriate molecule for catalytic testing because it is used as the feedstock for a reaction of major commercial interest namely, the catalytic conversion of methanol to olefins over acidic zeolites. The chemistryD^] and process technologyD 7] of the methanol-to-olefin (MTO) process have been recently reviewed and recent studies have focussed on controlling the product distribution to produce a small range of low molecular weight hydrocarbons suitable as a petrochemical feedstock. The product distribution can be shifted to contain predominately C2-C4 olefins at low conversions by using zeolite catalysts with small (8-ring) poresD^, 18] s j n c e larger products, i f formed, are unable to diffuse out of the materials. The larger products may crack to smaller products or remain as carbonaceous deposits within the sample. The most promising small-pore material, to date, is the SAPO-34 molecular sieve (a Chabazite structure-type) which is made up of 8-membered rings that interconnect to form a 3-dimensional network of channels. A commercial M T O process has been developed jointly 131 between U O P and Norsk Hydro (Norway) to convert natural gas to ethylene and propylene using the SAPO-34 molecular sieveP**] as the M T O catalyst. High-temperature adsorption experiments were carried out (as outlined in Section 2.4.4) to determine whether the activated mixed materials were able to absorb methanol at elevated temperatures. The M T O process is typically run at 400 °C and, consequently, the high temperature adsorption measurements for the mixed materials were only carried out at this temperature. The plot for the adsorption of methanol with respect to time at 410 °C for the 70%Ca material is shown in Figure 6.2 and shows a sharp increase in the sample weight at ca. 4 min when the sample is exposed to the methanol vapour stream. The gain in weight continues while the sample is exposed to the vapour stream and the plots for all three of the mixed materials are virtually indistinguishable. The continuous increase in the weight of the sample is the result of the accumulation of coke deposits on/in the samples. In the process of coking, the white catalyst powder is blackened; a behaviour which was noted with all of the mixed materials when exposed to methanol vapour but was not observed when they were exposed to n-hexane vapour. Coking may be considered a sign of the materials' catalytic activity since coke is a byproduct of catalytic conversion processes. The high-temperature methanol adsorption studies establish that all of the activated mixed materials are capable of absorbing methanol at 400 °C. Coke is a general term given to the carbonaceous products of the reaction that are too large to diffuse out of the zeolite framework. As coke progressively fdls up the catalyst, the material is deactivated as the channel structure becomes clogged and reactants and products 132 0 10 20 30 40 50 60 Time (min) Figure 6.2 Methanol uptake at ca. 410 °C for the 7 0 % C a material cannot diffuse through the framework. Since coke is a byproduct of all zeolite acid-catalyzed reactions, controlling the rate at which the catalyst loses activity due to coke accumulation is very important because it influences the length of time that the catalyst can be used before having to be removed and replaced with fresh catalyst. The rate at which a zeolite 'cokes out' is dependent on the acid strength of the material and very active acid catalysts often coke out quickly. In fact, zeolites that have strong acid sites, and therefore high activities, are modified to reduce their activity specifically to reduce the rate at which the catalyst bed is deactivated. [13] The M T O catalytic activity tests of the zeolite A materials were carried out in the same reactor as the previous n-hexane cracking experiments and the experimental details are described in Section 2.4.7. The conversion of the feed was determined by difference. The methanol reference (0% conversion) was established by sampling the methanol vapour 133 stream through the reactor bypass (see Figure 2.5) and assigning the integrated peak value to represent 100% M e O H in the vapour stream. The percentage of the feed that was converted to products during the catalytic activity tests was determined by the difference between the values for the methanol bypass sample and the methanol in the product stream. The catalytic activity test for each material was compared to a separate bypass value which was measured shortly before the catalytic test was run. The reactant streams were sampled after 5 minutes on stream. Figure 6.3 shows a typical chromatograms for the M T O reaction over the zeolite A mixed materials below the chromatogram for the corresponding methanol bypass sample. The results of the M T O activity tests are given in Table 6.4. Two additional samples were run to compare the activities of the zeolite A mixed materials to other standard materials. The H Y sample is an acid form of zeolite Y prepared in this laboratory and the sample denoted S i A l is an amorphous aluminosilicate powder (Aldrich). There was no measurable thermal conversion of the feed when the methanol vapour was passed through the empty reactor tube at 400 °C. The observation that the three mixed materials differ in their activities toward the conversion of methanol to hydrocarbons was expected but that the 50%Ca material shows the least activity is surprising as it should contain the largest amount of framework protons. The 50%Ca and 60%Ca materials may have the same acid strength as the 70%Ca material but the presence of the amorphous component in these samples could reduce the activity of the samples by blocking some of the channels. The preliminary results indicate that greater activities may be achieved by increasing the Ca content of the mixed materials. A s expected, the zeolite A mixed materials have lower activities than H Y but enhanced activities 134 MeOH Impurity MeOH C3 C1 C2 C4 C • - C D C5 Impurity (a) Time (min) —> Figure 6.3 Chromatograms for (a) the MTO reaction at 400 °C after 5 minutes on stream for the 70%Ca material and (b) the methanol bypass sample. Peak markings are in minutes. 135 Table 6.4 Percent conversion of MeOH to hydrocarbons over the mixed zeolite A materials and reference materials Sample Weight (g) Flow Rate (cc/min) Conversion (%) 100%Ca 0.1 100 6.0 ± 0 . 1 70%Ca 0.1 100 28 .010 .6 60%Ca 0.1 100 23.0 + 0.5 50%Ca 0.1 100 15.0 + 0.3 H Y 0.1 100 83.0 ± 1.7 S i A l 0.1 100 4 . 0 ± 0 . 1 compared to the amorphous aluminosilicate sample. The Ca /NH4 zeolite A materials may be potentially well suited as M T O catalysts because they show moderate activity which should decrease the rate at which they coke out. The materials used in the present study seem to be most active at short times on stream (< 30 min) and progressively deactivate over the course of 2 hours on stream under the test conditions. The various products from the M T O reactions are given in Table 6.5. The same column and heating program were used as with the hexane cracking study so the product species could be identified based on their retention times. Only the C1-C4 products were included in the calculations of the product distribution because calibration mixtures were not available for the C5 hydrocarbons and, as a result, their minor contributions could not be calculated. C6 hydrocarbons were not observed in any of the tests using the zeolite A materials. The H Y sample did show a small peak around 13 min which is characteristic of n-hexane but the water peak at 13.8 min overlaps this position and, as a result, the identity of 136 this peak is unknown. At higher conversions, the presence of C6 hydrocarbons may be more prominent. Table 6.5 Product distributions from the conversion of methanol to hydrocarbons over the zeolite A materials and reference materials Species R (min) Percentage of Total Products (wt%) CaA 70%Ca 60%Ca 50%Ca HY SiAl Methane 0.7 4.4 8.2 9.1 10.9 5.9 29.3 Ethylene 1.7 4.9 15.3 18.3 14.1 41.6 16.5 Ethane 2.2 0.5 0.7 0.9 0.7 1.0 0.0 Propane/ene 4.9 44.6 40.2 39.8 35.7 3.4.8 38.6 C4 species 8 .2 -9 45.6 35.6 31.9 38.6 16.7 15.6 C5 species 11.2-11.6 — — — — — — The C a A material, itself, is slightly active in the methanol-to-hydrocarbons reaction and produces little ethylene but large relative amounts of C3 and C4 hydrocarbons. The product distributions from the Ca 2 + /NH.4 + materials are significantly different from the C a A material. The mixed materials show a moderate amount of ethylene production and roughly equal contributions of the C3 and C4 hydrocarbons. A l l three give similar percentages of the different products indicating a similar conversion mechanism. The product stream from the H Y sample, however, shows predominately ethylene and propane/ene and much less C4 hydrocarbons than the zeolite A materials. The S i A l sample, while yielding a much lower degree of conversion and producing less ethylene, gives similar C3/C4 product distribution as the H Y sample. The similarities of the product distributions between the H Y and the S i A l samples may be the result of the larger pore dimensions of these materials (~10 A for H Y 137 and ~ 45 A for S iAl ) . The product distributions from the activated mixed materials could be the result of the reduced pore and channel dimensions of the zeolite A framework compared to the zeolite Y and amorphous aluminosilicate samples. 6.4 Conclusions The results from the catalytic activity tests of the zeolite A mixed materials demonstrate the catalytic activities of the mixed materials and illustrate the need to carry out more than one catalytic activity test. The results obtained from the n-hexane cracking study illustrate that the uptake of «-hexane at room temperature does not necessarily mean that it wi l l be adsorbed at high temperature. n-Hexane is not an appropriate probe molecule for testing the activities of all microporous materials and, as a result, the very low literature De-values for zeolite A materials^ 1> 19] <j0 n 0 f reflect the activities of the materials towards other reactants. The C a 2 + / N H 4 + zeolite A materials could be viable M T O catalysts for several reasons. They are easily synthesized from inexpensive starting materials which is not the case with the SAPO-34 material which is synthesized using, in part, 85% orthophosphoric acid, aluminum isopropoxide, and tetraethylammonium hydroxide. The number of acid sites in the framework of SAPO-34 is controlled by adjusting the reagents in the synthesis gel or though post-synthesis dealumination. The acid content of the zeolite A materials, however, is easily controlled by adjusting the ammonium content of the materials which allows for considerable flexibility in the composition of the catalyst. Furthermore, using contact-induced ion exchange, cations other than C a 2 + or N F L / could easily be included in the mixed materials whose presence could affect the product distribution. Lastly, the zeolite A materials, like SAPO-34, have a 3-dimensional array of intersecting channels which is a considerable 138 advantage as it allows the reactants and products to diffuse more freely through the structure even i f parts of the framework are blocked by coke formation. 139 References M . Hunger, Solid-State Nucl. Mag. Res. 6 (1996) 1. J. B . Uytterhoeven, P. Jacobs, K . Makay, R. Schoonheydt, J. Phys. Chem. 72 (1968) 1768. J. L . Carter, P. J. Lucchese, D . J. C. Yates, J. Phys. Chem. 68 (1964) 1385. L . M . Kustov, Y . V . Borokov, V . B . Kazanskii, Kinet. Catal. 25 (1984) 120. H . G . Karge, Micropor. Mater. 22 (1998) 547. J. A . Rabo, M . L . Poutsma, in E . M . Flanigan, L . B . Sand (Eds.): Molecular Sieve Zeolites-II. Adv. Chem. Ser. 102, A C S , Washington, D C 1971, p. 284. S. M . Babitz, B . A . Will iams, J. T. Mil ler , R. Q. Snurr, W . O. Haag, H . H . Kung, Appl. Cat. A: General 179 (1999) 71. B . A . Williams, S. M . Babitz, J. T. Mil ler , R. Q. Snurr, H . H . Kung, Applied Catalysis A: General 177 (1999) 161. D . H . Olson, W. O. Haag, R. M . Lago, J. Catal. 61 (1980) 390. W . O. Haag, R. M . Lago, P. B . Weisz, Nature 309 (1984) 589. J. N . Miale, N . Y . Chem, P. B . Weisz, J. Catal. 6 (1966) 278. P. B . Weisz, J. N . Maile, J. Catal. 4 (1965) 527. B . Borghard, Personal Communication (1999). J. H . C. van Hooff, J. W . Roelofsen, in H . van Bekkum, E . M . Flanigan, J. C. Jansen (Eds.): Studies in Surface Science and Catalysis, Vol. 58, Elsevier, Amsterdam 1991, p. 241. A . Araya, A . J. Blake, I. D . Harrison, H . F. Leach, B . M . Lowe, D. A . Whan, Zeolites 72(1992)24. M . Stacker, Micropor. Mesopor. Mater. 29 (1999) 3. F. J. K e i l , Micropor. and Mesopor. Mater. 29 (1999) 49. S. Wilson, P. Barger, Micropor. and Mesopor. Mater. 29 (1999) 117. G . H . Kuhl , J. Catal. 29 (1973) 270. 140 Chapter 7 Conclusions and Suggestions for Further Work The work presented in this thesis establishes that thermally stable, acidic forms of zeolite A can be prepared by starting with materials containing C a 2 + and N H / ions in the same framework. The zeolite A materials presented in this work represent the first series of acidic forms of zeolite A that have been fully structurally characterized. The 2 9 S i and 2 7 A 1 M A S N M R spectra of the activated materials demonstrate that the frameworks of the zeolite A mixed materials undergo a small amount of dealumination upon activation. The dealumination is not accompanied by widespread damage to the structure; a result previously considered not possible for a zeolite A framework with a S i / A l = 1. P ] The ' H M A S N M R studies of the activated materials verify that framework bridging hydroxyl groups are generated by the decomposition of the ammonium ions and that these framework hydroxyl protons exhibit Bronsted acidity. Catalytic studies establish that the mixed materials are not active toward the catalytic cracking of n-hexane at 540 °C because the materials are unable to absorb n-hexane at temperatures above 500 °C. It would be of interest to establish the largest hydrocarbon that can diffuse into the frameworks at elevated temperatures by conducting high temperature adsorption experiments using a variety of aliphatic and olefinic hydrocarbons. Such experiments would be valuable as they would provide an indication of what hydrocarbons would be suitable as feedstocks for reactions involving the zeolite A mixed materials. Catalytic tests for methanol conversion over the zeolite A materials indicate that the mixed materials show promise as M T O catalysts since they displayed reasonable activities in the M T O reaction; producing, predominately, ethylene and C3 and C4 hydrocarbons. The product distribution from the methanol conversion tests are similar to that from SAPO-34, a 141 commercially used M T O catalyst, though the zeolite A materials seem to produce less ethylene and more C4 hydrocarbons than SAPO-34 at 400 °C. The powder X R D results from the contact-induced ion exchange study of the 50%Ca material establish that materials containing a mixture of C a 2 + and N r l / ions can be produced by mixing the powders of the two parent zeolites. The locations of these ions within the frameworks, however, is not known. The locations of the cations in dehydrated C a A have been reported!?] and a structure of a single crystal of N H 4 A which had been dehydrated at room temperature has also been reported.[3] There have been no reported structures for zeolite A containing both C a 2 + and N H / ions but by carrying out the same exchange procedures on large crystals that were used on the microcrystalline powders it should be possible to produce mixed-ionic crystals of zeolite A suitable for single-crystal X-ray structure analysis. The thermal behaviour of the mixed materials has not yet been fully explored. Preliminary studies suggest that the materials are stable to temperatures up to 700 °C but that they decompose at 800 °C without first passing through the amorphous-coated crystallite phase. Knowing the full thermal behaviour for the mixed materials wi l l help establish the maximum working conditions for the materials. It has been shown in the previous chapters that the mixed materials undergo a small amount of dealumination upon activation. That the materials are stable to partial dealumination suggests that they may be successively dealuminated; providing a material that should be more thermally stable and may be more acidic than the original material. Such highly siliceous forms of zeolite could be useful as supports for transition metals and such a material might be produced by successively re-ammoniating and re-activating the samples. If the zeolite A mixed materials are to be used 142 as M T O catalysts their stabilities to hydrothermal conditions at high temperature also needs to be further investigated since large amounts of water are produced during the M T O reaction. Preliminary powder X R D results seem to indicate that the 5 0 % C A material is stable to a saturated water atmosphere at 500 °C for 1-2 days. A more complete study of the hydrothermal effects on the frameworks of the materials using 2 7 A 1 and 2 9 S i M A S N M R should give valuable information about how the catalysts might be affected in the course o f the M T O process. The thermal behaviour of the N H 4 A material was used as a reference in the present study to demonstrate the difference between a thermally stable zeolite, i.e., C a A and one that is not stable to elevated temperatures. It is well established that the structure o f N H 4 A decomposes at elevated temperatures but it was found in the present work that the decomposed material has unusual structural features and several observations should be explored in more detail. After activation at 500 °C, the N H 4 A material completely lacks porosity; why this occurs and the manner in which the sample decomposes is not known. This was not a key topic of the present work but is of interest in its own right. A series of X R D , N M R , and gas adsorption experiments on N H 4 A samples exposed to temperatures ranging from 100-500 °C would establish the thermal behaviour of the zeolite and could establish how the material retains a small fraction of ammonium ions even after activation at 500 °C for 24 hours. Further, the 2 7 A 1 M A S N M R spectrum collected at very high field for the N H 4 A material activated at 500 °C shows a significant contribution from 5-coordinate aluminum which is nearly equal in intensity to that of the 4-coordinate peak. The presence of 5-coordinate aluminum in zeolites is relatively unusual and further studies would be interesting and should help determine the nature of the 5-coordinate environment. 143 The majority of the future work with respect to the zeolite A mixed materials is in the field of catalysis and their activities toward the acid-catalyzed conversion of small molecules. The preliminary studies demonstrate that the mixed materials are catalytically active and are capable of converting methanol to hydrocarbons. From the results of the M T O tests, the activated mixed materials seem to exhibit product selectivity since their product distributions differ from the large pore H Y and amorphous aluminosilicate samples in that they seem to produce more C4 hydrocarbons. In the present study, all of the samples were run at 400 °C with a methanol vapour flow rate of 100 cc/min to compare the activities of the zeolite A mixed materials toward the H Y and amorphous aluminosilicate samples under fixed conditions. For a better comparison of the product distributions it is necessary to compare them at equivalent feed conversions since the product distribution could change with the degree of conversion of the feed. The performance of the zeolite A mixed materials in the M T O reaction has yet to be fully explored and several aspects of the process can be varied. Firstly, the product distribution has yet to be examined as a function of temperature. The product distribution from SAPO-34 can be shifted to include more C2 and C3 olefins at the expense of C4 olefins when the temperature is lowered from 450 °C to 375 °C.[4] Lowering the reaction temperature should also decrease the amount of coking and, thereby, increase the life time of the catalysts. In the present study, activity tests were only carried out on three of the myriad possible C a 2 + / N H 4 + materials. The preliminary results indicate that the activities of the 2~l- "b mixed Ca /NH4 materials increase with decreasing ammonium content and 80 or 90%Ca materials may show greater activities than the 70%Ca sample. The conversion of M e O H to 144 hydrocarbons with respect to the Ca content of the mixed materials should be explored along with the effect that introducing a third type of cation into the Ca 2 + /NFL; + materials might have on the product distribution of the M T O reaction. Another relevant aspect of catalytic processes which has yet to be explored is catalyst recycling. When the catalyst 'cokes out', the material is regenerated by heating the sample in air to burn off the occluded organics. The regeneration process can produce temperatures in excess of the working temperature of the reaction which may be destructive to the zeolite frameworks and could cause partial dealumination which, in turn, would change the materials' activities. The stability of the mixed materials toward successive catalytic/regeneration cycles would give a broader perspective on how robust the materials might be over the long term in M T O and other catalytic processes. 145 References [1] G . Englehardt, D . Michel , High-Resolution Solid-State NMR of Silicates and Zeolites, John Wiley & Sons, New York 1987. [2] J. J. Pluth, J. V . Smith, J. Am. Chem. Soc 105 (1983) 1192. [3] L . B . McCusker, K . Seff, J. Am. Chem. Soc. 103 (1981) 3441. [4] S. Wilson, P. Barger, Micropor. andMesopor. Mater. 29 (1999) 117. 146 Appendix A : Physisorption Data Table A. l Argon Physisorption data at 87 K for the zeolite A materials 50%Ca 60%Ca 70%Ca 100%Ca Relative Volume Relative Volume Relative Volume Relative Volume Pressure Ads. Pressure Ads. Pressure Ads. Pressure Ads. (cm3/g) (cmVg) (cm3/g) (cmJ/g) 3.96E-06 3.0462 2.8E-06 3.0469 3.19E-06 3.0462 2.89E-06 3.0471 8.29E-06 6.0913 4.96E-06 6.0932 6.13E-06 6.0923 5.04E-06 6.0937 1.25E-05 9.1369 7.32E-06 9.1398 8.9E-06 9.1375 7.04E-06 9.1403 1.66E-05 12.1822 9.84E-06 12.1873 1.15E-05 12.1832 8.92E-06 12.1868 2.07E-05 15.228 1.24E-05 15.2329 1.41 E-05 15.2296 1.08E-05 15.2339 2.48E-05 18.2745 1.51 E-05 18.2788 1.67E-05 18.2756 1.26E-05 18.2805 2.89E-05 21.3209 1.79E-05 21.325 1.92E-05 21.3213 1.43 E-05 21.3277 3.31E-05 24.3671 2.1 E-05 24.3718 2.16E-05 24.3672 1.6E-05 24.375 3.73E-05 27.4139 2.45E-05 27.4181 2.41 E-05 27.4132 1.77E-05 27.4224 4.14E-05 30.4606 2.8E-05 30.4649 2.67E-05 30.4599 1.92E-05 30.4694 4.56E-05 33.5073 3.16E-05 33.5111 2.95E-05 33.5061 2.07E-05 • 33.5167 5.02E-05 36.5531 3.52E-05 36.5578 3.24E-05 36.5527 2.22E-05 36.5641 5.5E-05 39.599 3.89E-05 39.6049 3.53E-05 39.5991 2.36E-05 39.6115 6.01E-05 42.6452 4.27E-05 42.6516 3.83E-05 42.6453 2.5E-05 42.6586 6.55E-05 45.6916 4.67E-05 45.6984 4.14E-05 45.6916 2.65E-05 45.7056 7.14E-05 48.7374 5.09E-05 48.7455 4.46E-05 48.7385 2.8E-05 48.753 7.76E-05 51.784 5.52E-05 51.7919 4.81 E-05 51.7851 2.96E-05 51.8004 8.45 E-05 54.8295 5.98E-05 54.8386 5.17E-05 54.8314 3.12E-05 54.8484 9.2E-05 57.8747 6.46E-05 57.8854 5.53E-05 57.8777 3.27E-05 57.8966 0.000101 60.92 6.98E-05 60.9321 5.91 E-05 60.9247 3.44E-05 60.9444 0.00011 63.9652 7.54E-05 63.9786 6.31 E-05 63.9714 3.63 E-05 63.9917 0.000122 67.0102 8.15E-05 67.0251 6.72E-05 67.0177 3.83E-05 67.0393 0.000136 70.0543 8.81 E-05 70.0708 7.15E-05 70.0643 4.03E-05 70.0868 0.000154 73.0971 9.55E-05 73.1169 7.61 E-05 73.111 4.24E-05 73.1345 0.000178 76.1374 0.000104 76.1632 8.11 E-05 76.1574 4.46E-05 76.1821 0.000213 79.1746 0.000114 79.2091 8.66E-05 79.204 4.69E-05 79.2301 0.000264 82.2052 0.000126 82.2543 9.23E-05 82.25 4.95E-05 82.2782 0.000343 85.2258 0.00014 85.298 9.86E-05 85.2961 5.23E-05 85.3256 0.000464 88.2193 0.000159 88.3413 0.000105 88.3422 5.53E-05 88.3731 0.000654 91.215 0.000184 91.3811 0.000113 91.3887 5.85E-05 91.4204 0.000953 94.1737 0.000219 94.4182 0.000122 94.4344 6.21 E-05 94.468 0.001422 97.071 0.000268 97.4493 0.000132 97.4795 6.61 E-05 97.5151 0.002147 99.8694 0.000341 100.4714 0.000143 100.5241 7.06E-05 100.5625 0.003235 102.5198 0.000449 103.4797 0.000157 103.568 7.57E-05 103.6097 0.004824 105.014 0.000611 106.4659 0.000175 106.6107 8.15E-05 106.6566 0.006971 107.2833 0.000854 109.4068 0.000199 109.6519 8.82E-05 109.7035 0.009617 109.3722 0.00123 112.3399 0.00023 112.6907 9.61 E-05 112.7506 0.012278 111.1843 0.001801 115.1971 0.000273 115.7248 0.000106 115.7965 0.015818 112.7015 0.002652 117.9392 0.000334 118.7529 0.000117 118.8413 0.020039 114.0814 0.003872 120.5148 0.00042 121.7707 0.000131 121.8862 0.024601 115.2668 0.005569 122.9272 0.000543 124.7694 0.000149 124.9302 0.029435 116.3417 0.007669 125.1385 0.000724 127.7707 0.000173 127.9714 0.034678 117.2918 0.009898 127.1414 0.000992 130.7447 0.000204 131.011 0.040126 118.1587 0.013141 129.0397 0.001394 133.6714 0.000248 134.0467 0.045777 118.938 0.016914 130.5864 0.001993 136.5256 0.00031 137.0669 0.05143 119.6433 0.021059 131.9955 0.002877 139.2661 0.000404 140.0537 0.057219 120.3095 0.025176 133.1128 0.004103 141.8826 0.000551 143.0692 0.063076 120.9038 0.02994 134.2086 0.005484 144.2877 0.000743 146.0546 147 0.069137 121.4676 0.034896 135.2184 0.07524 121.9998 0.040196 136.0993 0.081397 122.4938 0.045152 136.8595 0.087528 122.9777 0.050452 137.5683 0.093726 123.4435 0.055614 138.2299 0.099993 123.8911 0.061251 138.8399 0.106328 124.2885 0.066626 139.4199 0.120284 125.1499 0.072407 139.9909 0.140521 126.3023 0.078188 140.5244 0.160676 127.3589 0.084107 141.02 0.180778 128.3878 0.090094 141.5058 0.200533 129.4181 0.096081 141.9601 0.220562 130.4413 0.101998 142.4278 0.240451 131.4925 0.119342 143.5849 0.260202 132.6116 0.140813 144.9308 0.280226 133.8351 0.161667 146.0264 0.30025 135.1791 0.181758 147.1329 0.320411 136.6012 0.201244 148.1902 0.340707 138.0045 0.221204 149.2605 0.360459 139.1619 0.241027 150.3713 0.38021 140.1181 0.261056 151.5607 0.420599 141.3687 0.281016 152.8339 0.461798 141.8754 0.301045 154.2512 0.501989 142.2172 0.321074 155.89 0.542057 142.5033 0.341365 157.6933 0.582016 142.7974 0.360444 159.3433 0.621928 143.1024 0.380336 160.9272 0.66186 143.4429 0.400434 162.2465 0.701752 143.8219 0.419905 163.2298 0.741595 144.2809 0.440354 163.896 0.781439 144.8099 0.460039 164.2783 0.821146 145.4952 0.48 164.5053 0.860718 146.4399 0.500029 164.6776 0.899819 147.8855 0.520196 164.8177 0.918876 148.9905 0.540156 164.9544 0.920108 149.3649 0.560254 165.0771 0.939942 150.8518 0.580214 165.2148 0.97181 156.8486 0.600243 165.3321 0.98174 161.075 0.620135 165.4619 0.640185 165.617 0.660269 165.7635 0.680215 165.9007 0.700182 166.0805 0.720349 166.24 0.740172 166.4244 0.76027 166.6169 0.780161 166.8601 0.800328 167.1007 0.820151 167.3673 0.84018 167.7135 0.860278 168.1019 0.880307 168.4694 0.899717 169.1032 0.920324 169.9631 0.939782 171.2278 0.959735 173.3136 0.00769 146.5073 0.001149 148.9739 0.010568 148.53 0.001934 151.7438 0.013995 150.319 0.003447 154.2532 0.01797 151.8658 0.006064 156.3466 0.022082 153.1727 0.01002 158.0012 0.026879 154.3448 0.014828 159.2532 0.031608 155.3388 0.020117 160.1974 0.036816 156.2978 0.02616 160.9732 0.042298 157.1694 0.032478 161.598 0.047781 157.9235 0.038995 162.1486 0.053469 158.6627 0.045319 162.6382 0.059294 159.3057 0.052186 163.0469 0.065325 159.9298 0.05912 163.4284 0.071424 160.5244 0.066055 163.7989 0.07766 161.1066 0.073126 164.1196 0.083773 161.6213 0.080197 164.4152 0.090064 162.1402 0.087337 164.6986 0.096163 162.63 ' 0.094476 164.9679 0.102604 163.0869 0.101615 165.2364 0.120134 164.2569 0.120014 165.8048 0.140364 165.5146 0.140474 166.427 0.160786 166.6715 0.160726 166.9677 0.180728 167.7907 0.180702 167.4947 0.200464 168.8695 0.200624 167.9159 0.220461 169.9474 0.220518 168.3764 0.240375 171.0117 0.240251 168.8836 0.260215 172.1306 0.259979 169.3874 0.280232 173.3446 0.280302 169.9264 0.300243 174.6759 0.300344 170.4728 0.320254 176.1636 0.320385 171.0115 0.34047 177.7764 0.340355 171.5586 0.360412 179.3408 0.360052 172.081 0.380286 180.8482 0.380024 172.5518 0.400029 182.1529 0.399925 172.9832 0.420369 183.2197 0.419964 173.3268 0.461082 184.3609 0.439796 173.6193 0.502063 184.9385 0.459833 173.8876 0.542153 185.3904 0.479867 174.1186 0.582106 185.859 0.51939 174.5091 0.622059 186.3072 0.539233 174.7684 0.662148 186.7082 0.539836 174.7885 0.701882 187.2195 0.559843 174.9953 0.74178 187.7816 0.599186 175.4036 0.781438 188.533 0.619082 175.6569 0.821412 189.3107 0.619695 175.6535 0.86096 190.3209 0.659503 176.1186 0.900015 191.8707 0.660121 176.1849 0.93759 194.8128 0.680162 176.4447 0.941798 195.5351 0.719284 176.9717 0.959177 198.5925 0.720011 177.0428 0.976529 204.3756 0.740217 177.3689 0.778996 178.0422 0.780087 178.1251 0.800479 178.5192 0.838503 179.44 0.840218 179.5854 148 0.978835 177.0877 0.860809 180.1923 0.897528 181.5805 0.90026 181.8875 0.921276 182.9538 0.954221 186.0981 0.961496 187.3687 0.980213 191.6649 149 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0059649/manifest

Comment

Related Items