UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

An alternative preparation of organofunctionalized silica gels and their characterization by solid-state… Zhang, Yugao 1991

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

Item Metadata

Download

Media
831-UBC_1991_A6_7 Z3.pdf [ 5.17MB ]
Metadata
JSON: 831-1.0059757.json
JSON-LD: 831-1.0059757-ld.json
RDF/XML (Pretty): 831-1.0059757-rdf.xml
RDF/JSON: 831-1.0059757-rdf.json
Turtle: 831-1.0059757-turtle.txt
N-Triples: 831-1.0059757-rdf-ntriples.txt
Original Record: 831-1.0059757-source.json
Full Text
831-1.0059757-fulltext.txt
Citation
831-1.0059757.ris

Full Text

AN ALTERNATIVE PREPARATION OF ORGANOFUNCTIONALIZED SILICA GELS AND THEIR CHARACTERIZATION BY SOLID-STATE NMR SPECTROSCOPY by YUGAO ZHANG M.Sc, Hubei Research Institute of Chemistry, P. R. China, 1984 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Department of Chemistry) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA September 1991 © Yugao Zhang, 1991 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. Department of C h e m i s t r y The University of British Columbia Vancouver, Canada D a t e Oct. 8, 1991 DE-6 (2/88) A B S T R A C T AN ALTERNATIVE PREPARATION OF ORGANOFUNCTIONALIZED SILICA GELS AND THEIR CHARACTERIZATION BY SOLID-STATE NMR SPECTROSCOPY The theoretical consideration and experimental techniques required to obtain high -resolution solid-state NMR spectra of silica gels and their organofunctionalized derivatives are presented. A series of organofunctionalized silica gels has been prepared by the copolymerization of the two components, tetraethoxysilane (TEOS), (EtO)4Si, and mefhyltriethoxysilane (MTES) (EtO)3SiCH3, which is an alternative to the conventional anchoring method. CP/MAS solid-state NMR techniques have been used to elucidate the structures of the organofunctionalized silica gels. The copolymerization method is superior to the anchoring method in that more functional groups can be incorporated into a silica gel matrix. CP/MAS solid-state NMR has been successfully applied in the investigation of the thermal stabilities of the functionalized silica gels as a complementary technique to the more traditional DSC technique and it has been concluded that in an inert gas (N2) environment, the methyl groups on the methylsilica gel made by copolymerization method are more thermally stable to decomposition than that made by the anchoring method and both are much more stable than in a normal atmospheric environment. Two-dimensional heteronuclear 1H- 2 9Si correlation NMR spectroscopy has been used to characterize three different types of modified silica gels; methylsilica gel prepared by the anchoring method; methylsilica gel prepared by the copolymerization method; and a mechanical mixture of silica gel and polymethylsiloxane. This technique successfully ii demonstrated the structural differences among the three materials which cannot be easily distinguished by other analytical techniques. It is shown that methylsilica gel prepared by the anchoring method consists of two distinct phases when the functional group loading is high (i.e., polymethylsiloxane built up "horizontally" on the surface of silica gel), and it is very like the mechanical mixture of silica gel and polymethylsiloxane. On the other hand, the methylsilica gel prepared by the copolymerization method consists of a distribution of the two components, i.e., there is no domain structure between silica gel and polymethylsiloxane. The methyl protons and hydroxyl protons in the functionalized silica gels were found to be sources of cross-polarization transfer to the Si(0-)4 framework silicons, but the physically absorbed water does not contribute to this transfer. iii T A B L E O F C O N T E N T S ABSTRACT .. ii T A B L E OF CONTENTS iv LIST OF TABLES viii LIST OF FIGURES ix ACKNOWLEDGEMENTS xii Chapter I. GENERAL INTRODUCTION 1 1.1. Functionalized Silica Gels 1 1.2. Solid-State NMR Spectroscopy 3 1.2.1. Nuclear Spin Interactions in the Solid State 4 a) . The Zeeman Interaction 5 b) . Dipole-Dipole Interaction 7 c) . The Chemical Shift Interaction 9 d) . Spin-Spin Interaction 14 e) . Summary 14 1.2.2. Techniques Used to Obtain High-Resolution NMR Spectra in Solids 15 a) . Introduction 15 b) . High Power Decoupling 16 c) . Magic Angle Spinning (MAS). 17 d) . Cross-Polarization(CP) 20 e) . High-Resolution Solid-State NMR Using Cross-Polarization and Magic-Angle Spinning Techniques (CP/MAS) 24 iv f). Application of CP/MAS Solid State NMR to the Study of Silica Gel and its Derivatives 25 1.3. Outline of the Research Project 28 1.4. References for Chapter I 29 Chapter II. PREPARATION OF FUNCTIONALIZED SILICA GELS AND A PRELIMINARY INVESTIGATION OF THEIR STRUCTURES BY SOLID STATE NMR 33 2.1. Experimental 33 2.1.1. Reagents 33 2.1.2. Preparation of High-Purity Silica Gel 33 2.1.3. Preparation of Functionalized Silica Gels 34 2.1.4. Solid State NMR 35 2.1.5. Elemental Analysis 35 2.2. Results and Discussion 36 2.2.1. Characterization of High Purity Silica Gel 36 2.2.2. Characterization of Functionalized Silica Gels 39 a) . Methyl-Functionalized Silica Gel 39 b) . Other Examples of Functionalized Silica Gels 41 2.2.3. Quantitative NMR Investigations of Functionalized Silica Gels 48 2.3. References for Chapter II 51 v Chapter III. THE CHARACTERIZATION OF FUNCTIONALIZED SILICA GELS BY TWO-DIMENSIONAL HIGH-RESOLUTION SOLID-STATE HETERONUCLEAR CORRELATION NMR SPECTROSCOPY 52 3.1. Introduction 52 3.1.1. Two-Dimensional NMR 52 3.1.2. Two-Dimensional Heteronuclear Solid-State Correlation NMR Spectroscopy 54 3.1.3. Pulse Sequence for Two-Dimensional Correlation Spectroscopy 55 3.1.4. Quadrature Detection incoi 58 3.2. Experimental Section 59 3.2.1. Sample Preparations 59 a) . Deuterium Exchange of Protons in Hydroxyl Groups 59 b) . Preparation of Methylsilica Gel by the Anchoring Method 59 c) . Preparation of Methylsilica Gel in a Deuterium Environment 60 3.2.2. Instrumental Details 60 3.3. Results and Discussion 61 3.3.1. Proton NMR Spectra of the Modified Silica Gels 61 3.3.2. Characterization of Different Types of Functionalized Silica Gels by 2D Correlation Experiments 66 a) . General Considerations 66 b) . Two-Dimensional ^ - ^ S i Correlation Spectra of Silica Gel 69 c) . 2D 1 H- 2 9 Si Correlation Spectra of Gel Mixtures 72 d) . 2D Correlation Spectra of Methylsilica Gels 79 dl). Methylsilica gels prepared by the copolymerization method 79 d2). Methylsilica gels prepared by the anchoring method 85 3.4. Conclusions 92 3.5. References for Chapter HI 95 Chapter IV. A PRELIMINARY INVESTIGATION OF THE THERMAL STABILITIES OF THE METHYLSILICA GELS BY SOLID-STATE NMR AND DIFFERENTIAL SCANNING CALORIMETRY , 97 4.1. Introduction 97 4.2. Experimental 97 4.2.1. Thermal Treatment 97 4.2.2. Solid-State NMR 98 4.2.3. Differential Scanning Calorimetry 98 4.3. Results and Discussion 100 4.3.1. Thermal Behavior of Methylsilica Gel Made by Copolymerization 100 4.3.2. Thermal Behavior of Methylsilica Gel Made by the Anchoring Method 107 4.4. Conclusions 109 4.5. References for Chapter IV I l l vii LIST OF TABLES Table 1.1. Approximate ranges of the different solid-state nuclear spin interactions 5 Table 1.2. NMR parameters of some common nuclei 8 Table 2.1. 2 9 Si chemical shifts of silica gel model systems 38 Table 2.2. 1 3 C NMR data for functionalized silica gels and precursors 42 Table 3.1. Phase-cycle for suppression of mirror image signals 58 viii LIST OF FIGURES Figure 1.1. The energy diagram of the Zeeman interaction for I > ^  : 6 Figure 1.2. Schematic representation of the 1 3 C chemical shielding of a carbonyl group for different situations 11 Figure 1.3. Schematic representation of the theoretical powder line shapes for the chemical shift anisotropy 13 Figure 1.4. A schematic representation of the geometric arrangement for mechanical sample spinning 19 Figure 1.5 . 2 0 7 Pb spectra of polycrystalline Pb(NC«3)2 obtained at 41.7 MHz showing the effect of the spinning rate on MAS spectra 21 Figure 1.6A. Pulse sequence used for cross-polarization(CP) of a dilute nucleus ( 1 3C) from the abundant spin system (^ H) 22 Figure 1.6B. Schematic representation of the behavior of the lH and 1 3 C spin magnetizations during the CP pulse sequence in A 22 Figure 1.7. 1 3 C CP/MAS spectrum of adamantane 26 Figure 2.1. 2 9 Si CP/MAS NMR spectrum of high-purity silica gel 37 Figure 2.2. 1 3 C and 2 9 Si CP/MAS NMR spectra of methylsilica gel (Gel 3) 40 Figure 2.3. 2 9 Si and 1 3 C CP/MAS NMR spectra of Gel 6 43 Figure 2.4. 2 9 S i and 1 3 C CP/MAS NMR spectra of Gel 11 45 Figure 2.5 A. 1 3 C CP/MAS NMR spectrum of Gel 7 obtained at 100.6 MHz 47 Figure 2.5B. 1 3 C CP/MAS NMR spectrum of Gel 4 obtained at 100.6 MHz 47 Figure 2.6A. Quantitatively reliable one-pulse 2 9 Si MAS NMR spectrum of Gel 18 obtained at 79.5 MHz 50 Figure 2.6B. Deconvolution of spectrum A in terms of Gaussian peaks 50 ix Figure 3.1. Pulse sequence used for 2D correlation experiments in the solid state 57 Figure 3.2. 400 MHz lH MAS NMR spectra of Gel 18 obtained with a single-pulse-acquisition pulse sequence 63 Figure 3.3. 2 9 Si CP/MAS NMR spectrum of Gel 18 65 Figure 3.4. 2 9 Si CP/MAS NMR spectra of the different types of modified silica gels 67 Figure 3.5. Contour plot of the 2D iH-^Si correlation experiment on unfunctionalized silica gel (Gel 1) 70 Figure 3.6. Comparison of ID spectra and the corresponding 2D projections as indicated from Figure 3.5 71 Figure 3.7. Spectra of the heated silica gel (Gel 1) 73 Figure 3.8. Contour plot of the 2D !H- 2 9Si correlation experiment on silica gel after deuterium exchange with D 2 0 74 Figure 3.9 Contour plot of the 2D !H- 2 9Si correlation experiment on the mechanical mixture of silica gel and polymethylsiloxane 75 Figure 3.10 Contour plot of the 2D *H-2 9Si correlation experiment on the mechanical mixture of silica gel and polymethylsiloxane after D2O exchange 77 Figure 3.11. 2 9 Si ID CP/MAS spectra of a mixture of silica gel and polymethylsiloxane 78 Figure 3.12. Schematic representation of the microstructure of the mixture of silica gel and polymethylsiloxane 80 Figure 3.13. Contour plot of the 2D correlation experiment on Gel 18 which was synthesized by the copolymerization of TEOS and MTES 81 Figure 3.14. Contour plot of the 2D *H-2 9Si correlation experiment on the D2O exchanged Gel 18 83 Figure 3.15. Contour plot of the 2D !H- 2 9Si correlation experiment on D2O exchanged Gel 3 made by the copolymerization of TEOS and MTES 84 x Figure 3.16. Schematic representation of the microstructure of the gels prepared by the copolymerization method 86 Figure 3.17. Contour plot of the 2D lH- 2 9Si correlation experiment on Gel 33 synthesized by the copolymerization of TEOS and MTES in a deuterium environment 87 Figure 3.18. Contour plot of the 2D lH-29Si correlation experiment on Gel 26 synthesized by the anchoring method 88 Figure 3.19. Contour plot of the 2D !H- 2 9Si correlation experiment on D2O exchanged Gel 26 90 Figure 3.20. Contour plot of the 2D !H- 2 9Si correlation experiment on Gel 34 91 Figure 3.21. Schematic representation of the microstructure on Gel 34 93 Figure 4.1. Schematic representation of the apparatus used for heating the functionalized silica gels under N2 protection 99 Figure 4.2. Stacked 2 9 Si CP/MAS NMR spectra of Gel 3 heated at the different temperatures indicated for four hours and then cooled to ambient temperature 101 Figure 4.3. Stacked NMR spectra of Gel 3 after heating at different temperatures under N2 for four hours 103 Figure 4.4. DSC thermogram of the high-purity silica gel 105 Figure 4.5. The DSC thermogram of Gel 3 106 Figure 4.6. Stacked NMR spectra of Gel 26 heated at the different temperatures indicated for four hours 108 Figure 4.7. Stacked NMR spectra of Gel 26 heated at different temperatures for four hours under N2 protection 110 xi A C K N O W L E D G E M E N T S I would like to sincerely thank my supervisor, Dr. C A . Fyfe for his guidance, support and encouragement throughout the course of my graduate studies and during the preparation of this thesis. I am grateful to Dr. Nick Burlinson for his enthusiastic advice and continuous encouragement throughout this research. I also would like to thank Dr. R. Thompson for allowing me the use of the DSC equipment. I am indebted to numerous friends and colleagues for their collaboration, advice, friendship and discussions on several aspects of this work. In particular, I wish to thank Mrs. Stephanie Isbell for her proof-reading of this thesis and suggestions. Lastly, I especially wish to thank my wife Jenny and my son Tony, for without their continued love, patience, and understanding over the past three years this work would not have been possible. xii Chapter T GENERAL INTRODUCTION 1.1. Functionalized Silica Gels Silica gel is a coherent, rigid,three dimensional network of connected colloidal silica particles'^, having a macroscopic structure that can be expressed by the chemical formula Si02-xH20. Surface silanol groups (=SiOH) are present and molecular water is normally adsorbed on the surface of the solid. Organofunctionalized silica gel consists of silica gel with organic functional groups covalently bonded to it. Because of the variabilities of the organic group attached as well as the characteristically large surface areas and the unique physical properties of silica gel itself, organofunctionalized silica gels have been used in many areas of applied science and technology. Examples are as follows^ 2"4!: 1) . Reinforced polymer composites. 2) . Stationary phases for chromatography. 3) . Ion exchange resins. 4) . Immobilization of peptides and proteins on solid carriers. 5) . Immobilization of transition metal complex catalysts. The usual preparation of an organofunctionalized silica gel is by treating silica gel with an organofunctional silane coupling agent X3SiRY (where X is a hydrolizable group such as EtO- or C1-, R is a alkyl group acting as a link between the silicon atom and the functional group and Y is a functional group such as -Nt^ -Pifo or -H). The reaction proceeds via condensation of the surface hydroxyl group of the silica gel with the X groups of the coupling agent and can be described as: l OH OH S i 0 2 ^ — OH + X3S1RY OH SiO, — O' <C— OH + HX (1.1) OH 1— O —Si—RY This kind of silanized silica gel has been such an important area of study in recent years that a huge quantity of literature on this subject has been published, including some very important reviews^ 3"6]. Despite the extensive studies carried out in this area, there are still some difficulties in the characterization and use of these kinds of materials. 1) . Anchoring metal complexes on to the surface of a functionalized silica gel usually results in heterogeneous catalysts which are easily separated from reaction media, while keeping the high activities and selectivities of the homogeneous metal complex precursors used in the liquid phase. However, because of the relatively low concentration of hydroxyl groups on the surface of silica gel, there are limited concentrations of active functional groups^] and this can affect the catalytic properties of the supported metal complex catalysts. The low concentration of functional groups may also restrict the application of the functionalized silica gels as stationary phases for chromatography 2) . Another effect of the low concentration of hydroxyl groups on the surface of silica gel is that few bonds will be formed between the organofunctional silanes and the support. On average, only 1-1.5 of the three possible bonds are formed^ which may cause the functional groups to be kinetically mobile^10!. 2 3). Normally used coupling silanes usually have multiple functional groups. These functional groups can condense with each other and thus reduce the accessibility of the silane for the anchoring reaction. Because of the disadvantages discussed above, we have investigated an alternative method for the preparation of organofunctionalized silica gels. This takes advantage of the trifunctional nature of the silane coupling agent X3SiRY and uses this to incorporate the substituted silane directly into the gel matrix during the synthesis of the gel, i.e. the final product is made in a one step process as an inorganic copolymer as in equation (1.2): H + Si(OR')4+(R"0)3SiRY • Copolymer (1.2) H 2 0 1.2. Solid-State NMR Spectroscopy While high resolution solution-state NMR has long been established as a powerful tool for the elucidation of the structures and dynamics of chemical systems, studies of magnetic resonance phenomena in the solid state have, until quite recently, been primarily restricted to wide line applications, usually on abundant nuclei such as lH and 1 9 F . The special problems associated with solid-state NMR are well known. The NMR spectra of solid systems show typically very broad, mainly featureless adsorptions. The "high-resolution" characteristics of spin-spin couplings, chemical shifts and intensities (which relate directly to the chemical and molecular structure) of the solution state are obscured in the solid state spectrum. The critical difference between solution and solid-state NMR is that in the former case, the fast and essentially isotropic motion of the molecules gives an averaging of the interactions of the nuclei both with other nuclei and with electric field gradients if present. This averaging, for some interactions (e.g. chemical shift, O", and scalar spin-spin couplings, J) gives a discrete value, while for other interactions (e.g. dipolar interactions 3 and quadrupolar interactions) a zero value results which means that the interactions are not observable in the solution spectrum. In solids, however, the above isotropic averaging motions do not usually exist although in some cases limited motions may occur. The features of the NMR spectra of solids are such that the various interactions are dependent on the orientation of the nuclear spin vector to the magnetic field vector. The approach in 'high-resolution' NMR of solids is to give a man-made motion to the solid system so as to remove or average the characteristic solid-state interactions and simplify the spectrum to obtain the isotropic shifts for structural investigations. A number of texts are available which provide detailed treatments of the theory and practice of solid-state NMRt 1 1" 1 7]. The following discussion of nuclear spin interactions in the solid-state and the practical techniques used in high-resolution solid-state NMR spectroscopy is intended only as a brief introduction to the subject. 1.2.1 Nuclear Spin Interactions in the Solid State The main interactions involving a nucleus with a magnetic moment which may occur in the solid state are: a) . The Zeeman interaction with the magnetic field; b) . Direct dipole-dipole interactions with other nuclei; c) . Magnetic shielding by the surrounding electrons giving chemical shifts; d) . Spin-spin coupling to other nuclei(Scalar coupling); e) . Quadrupolar interactions. Quadrupolar interactions will be present for nuclei with spin I > ^ only and will not be discussed further here since this interaction is not present in the systems currently being studied. Considering all the possible interactions, a general Hamiltonian for a nucleus of spin I in the solid state may thus be written as in equation (1.3) 4 HTotal = H Z + H D + H c s + H s c + H Q (1.3) In a particular solid state system, one or two of the terms will usually dominate the Hamiltonian and hence determine the characteristics of the spectrum. Typical ranges for the values of the various solid state nuclear spin interactions are presented in Table 1.1. We will discuss some important interactions present in our systems in more detail in the following sections. Table 1.1. Approximate ranges of the different solid-state nuclear spin interactions (adapted from ref. 13) Interaction Hamiltonian Magnitude (Hz) Zeeman H z lOMO 9 Dipolar H D 0-105 Chemical Shift H c s 0-104 Scalar coupling Hsc 0-104 Quadrupolar H Q 0-109 a). The Zeeman Interaction The Zeeman interaction is the basis of the nuclear magnetic resonance phenomenon and occurs for all nuclei with odd atomic mass or odd atomic number. It is the interaction between the magnetic moment of the nucleus (UN) and the appUed field H Q and yields 21+1 energy levels (where I is the nuclear spin quantum number) of separation ficoo = Y N ^ H Q (COQ is the precession frequency and Y N the magnetogyric ratio of nucleus N respectively) (see Figure 1.1). The Hamiltonian for the Zeeman interaction (Hz) may be written as: 5 Figure 1.1. The energy diagram for the Zeeman interaction for I = 1/2. 6 OA H z = - n N - H 0 = - ^ h H 0 - I (1.4) The applied magnetic field is conventionally chosen to lie along the Z direction, and the Hamiltonian describing the Zeeman interaction then becomes: where g^ is the nuclear g factor for the particular nucleus N and (3^  is the Bohr Magneton. The Zeeman interaction determines the observation frequency of a particular nucleus at a given magnetic field strength (e.g. the observation frequency for *H at 9.4T is 400.00 MHz), and via the Boltzman distribution between the levels, the fundamental sensitivity of NMR to that nucleus. The Zeeman interaction is linear with the applied magnetic field and thus the seperation between the energy levels is larger at higher fields, which causes an increase in the population difference between them and results in an increase in the signal-to-noise ratio (S/N) of the spectrum. Different magnetically active nuclei have different and characteristic magnetogyric ratios and Larmor frequencies. Values of the Zeeman interaction (in MHz) at an applied field of 9.4T for some common nuclei are given in Table 1.2. b). Dipole-Dipole Interaction The dipolar interaction (Hp) arises from the direct dipole-dipole interactions between nuclei. For a solid containing a single type of spin I with magnetogyric ratio Yi> the homonuclear dipolar interaction for two isolated spins may be written as: H Z = - Y N ^ H Q ^ = " E N P N ^ Z (1.5) 1 Yi 2ri 2 (l-3cos26 i j) (Ii-Ij - 3IizIjz) (1.6) 7 Table 1.2. NMR parameters of some common nuclei (adapted from Ref. 13) Natural Relative Absolute Nucleus y/2jt Frequency Spin abundance sensitivity3 sensitivitvb (MHz/Tesla) (MHz at field of9.4T) (Units of n) (%) IH 42.57 400.00 1/2 99.98 1.0 1.00 2H 6.54 61.40 1 1.5x10-2 9.65xl0-3 1.50X10-6 llB 13.66 128.34 3/2 80.42 0.17 0.13 »3C 10.71 100.58 1/2 1.11 1.59x10-2 1.76X10"4 15N 4.31 40.53 1/2 0.37 1.04 xlO"3 3.85X10"6 19p 40.10 376.31 1/2 100.0 0.83 0.83 27A1 1.09 104.23 5/2 100.0 0.21 0.21 29Si -8.46 79.46 1/2 4.7 7.84x10-3 3.69x10^ 31p 17.24 161.92 1/2 100.0 6.63x10-2 6.63x10-2 a at constant field for equal number of nuclei. b product of relative sensitivity and natural abundance. where ry is the distance between nuclei i and j, and 6jj is the angle between the internuclear vector ij and the magnetic field. For unlike spins I and S with magnetogyric ratios of Yi and Ys respectively, the heteronuclear dipolar interaction of two isolated spins may be written as: HD(I-S) = (1 - 3cos29)IZSZ (1.7) The total dipolar interaction will be the summation of all possible homonuclear and heteronuclear pairwise interactions. Important features of the dipolar interaction are : 1) it depends on the products of magnetogyric ratios and hence will be most important for spin ^ nuclei with large magnetic moments (e.g. lU and 1 9F); 2) it falls off very rapidly with the internuclear distance ( r _ 3 dependence); 3) it is independent of the appUed magnetic field HQ c). The Chemical Shift Interaction The chemical shift interaction is due to the shielding of the nucleus by the fields induced by the surrounding electrons and can be described as in equation (1.8) where a is the chemical shift tensor. The interaction is Unear with the applied field and wiU be proportionally larger and more important at higher magnetic field strengths. Since it involves the surrounding electrons, it is the interaction which is most sensitive to the geometry, bonding and the identity of other atoms surrounding a particular nucleus and wiU usually be the most 'chemically diagnostic ' measurable feature in NMR studies. Also, since the chemical shielding of the nucleus by the surrounding electrons is a three-dimensional H c s = Y l f t I - £ - H o (1.8) 9 quantity, the chemical shift will be anisotropic (i.e. it will depend on molecular orientation with respect to the magnetic field). In solution, an average (isotropic ) value of the chemical shift is observed, resulting from the isotropic motion of the molecules. Empirical chemical shift-structure relationships which have been developed have made solution NMR the most widely used technique for the elucidation of unknown molecular structures. In the solid state, however, the molecular motion is restricted, and the chemical shift will be dependent on the orientation of the molecule in the magnetic field and thus will potentially contain more information regarding structure and bonding in the solid state than the isotropic (solution) value, but will be more difficult to interpret. In equation (1.8), since I and HQ are both vector quantities, a must be a 3 x 3 matrix or second rank tensor. In a suitable axis system, called the principal axis system, a is converted to a diagonal form with three principal elements O J J , O 2 2 and O 3 3 which serve to characterize the three-dimensional nature of the shielding. The chemical shielding anisotropy Ao\ a measure of the orientation dependence of the chemical shift, can be defined from these elements: Ac= G33-^(on+a22) (1.9) If we have a single crystal with one magnetically equivalent carbon per unit cell, we will observe a single line. If we change the orientation of the crystal with respect to the magnetic field, the position of this line changes. In polycrystalline samples, signals will result from each of the random orientations and a broad line will be observed, the shape of which will depend on the principal elements of the shielding tensor. This orientation dependence of the chemical shift interaction is illustrated schematically in Figure 1.2 for the 1 3 C resonance of a C=0 group. Figure 1.2A illustrates that in a single crystal in a fixed 10 Figure 1.2. Schematic representation of the 1 3 C chemical shielding of a carbonyl group for different situations. A. Oriented single crystal as indicated. B. Powder. C. Liquid. 11 A c=o JL B orientation to the magnetic field, a single sharp line will be observed for each magnetically unique orientation of a particular nucleus with respect to the applied magnetic field. The frequencies of these lines will change as the orientation of the single crystal is changed. In the case of a polycrystalline sample, where all possible orientations are present simultaneously, a broad line will be observed due to the summation over all of the possible random orientations of the individual chemical shielding tensors (Figure 1.2B). In solution, (Figure 1.2C), the random and rapid motion of the molecules produces an isotropic average value for the chemical shift. Theoretical powder patterns for the general case (i.e. O j ^ 0"22 * 0*33 ) and for an axially symmetric shielding tensor (i.e. O r j = 022 * 0*33 ) are presented in Figure 1.3. The first of these (Figure 1.3A) results from a completely asymmetric shielding of the nucleus, and in this case, each of the three principal elements of the shielding tensor possesses a unique value which can be identified from the singularities as indicated. A shielding tensor powder pattern is described as being "axially symmetric" (Figure 1.3B) when two of the elements of the shielding tensor are identical. The shielding element of the unique axis is often defined as C\\ and the other two as elements O\L as shown in Figure 1.3. The total width of the chemical shift anisotropy can vary from several parts per million for a carbon of a methyl group to several hundreds parts per million (ppm) for a carbonyl carbon. Although the chemical shift anisotropy contains valuable information, several overlapping chemical shielding tensor powder patterns from chemically distinct nuclei often complicate the observed spectrum. 'Magic angle spinning' is the technique that is used to reduce the chemical shift anisotropy to its isotropic average. We shall deal with magic angle spinning in detail in Section 1.2.2c. In solution, the isotropic motion of the molecules average the shielding anisotropy with 0"AV> the trace of the tensor as defined in equation (1.10) o~AV = 5 +CT22 + CT33) 12 Figure 1.3. Schematic representation of theoretical powder line shapes for the chemical shift anisotropy. A) Asymmetric shift anisotropy. B) . Axially symmetric shift anisotropy. The principal tensor components described in the text are indicated by arrows. 13 d) . Spin-Spin Interaction(Scalar Interaction) Spin-spin interactions arise from the indirect interaction of two spins via their surrounding electrons. It occurs because a bonding electron's spin tends to correlate with the spins of its neighboring nuclei. The electron has been influenced, and will thus influence other neighboring nuclear spins. The spin-spin interaction is observed in a solution NMR spectrum as the J coupling which splits lines into "doublets", "triplets" or "multiplets". The spin-spin interaction between a pair of spins I and S may be represented as in equation (l.H) H S C = I-J-S (1.11) A where J is the spin-spin interaction tensor. The interaction is field independent and its magnitude is similar in solution and in the solid state. It is not averaged to zero by the rapid tumbling that occurs in solution. However, it is not often observed in solid state NMR spectroscopy since its magnitude is usually much smaller than the other interactions between nuclei. e) . Summary There are a number of nuclear spin interactions that contribute to the NMR spectra of solids whose relative importance will depend on the nucleus being observed and the sample being studied. They are anisotropic (i.e. orientation dependent), in a solid due to the relatively fixed orientations of the molecules, whereas in solution averaged values are observed due to the fast and isotropic motion of the molecules. 14 1.2.2. The Techniques Used to Obtain High-Resolution NMR Spectra in Solids a). Introduction The basic idea in obtaining high resolution solid-state NMR spectra is to use some special techniques to remove the dipolar interactions and to produce the isotropic average values for the other interactions as occurs in solution. However, which techniques are appropriate depends on the nature and relative magnitudes of each of the spin interactions in the solid state. The nature and magnitude of these interactions will be determined by the characteristics and natural abundance of the particular nucleus being observed and also by the basic physical properties of the system itself. In the description of the development of the experimental techniques used to obtain high resolution NMR spectra of solids, it is convenient to divide the nuclear spin systems in solids into two types: "abundant" and "dilute" spin systems. In abundant spin systems, such as *H , there is a high concentration of magnetically active nuclei of high natural isotopic abundance. In such systems (eg *H and 1 9F), the spin Hamiltonian is completely dominated by the dipolar term (HQ) resulting from the direct dipole-dipole interaction between the nuclei, since it is much larger than the chemical shift and spin-spin coupling terms. The general Hamiltonian of this kind of spin system can be written as: H Total = H z + H D(H-H) + Hothers (1-12) The proton NMR spectra of solid systems generally show only a single broad featureless absorption. Although various techniques have been developed for the removal of the homonuclear H-H dipolar coupling^4-17!, the finite efficiencies of these experimental techniques, coupled with the small chemical shift range (10 ppm) of the lH nucleus, has precluded their application as a completely general analytical technique. 75 A much simpler situation exists for a "dilute" spin system which can result either from the absolute low concentration of a particular nucleus in the sample or from the low natural abundance of the particular isotope under study. The total Hamiltonian of a dilute spin solid system may be represented, for a nucleus X, as in equation (1.13) HTotal = H z + H D ( X - X ) + H D ( H - X ) + H C S (1-13) There are some very important simplifications for a dilute spin system compared to the abundant spin system and these allow high resolution solid state NMR spectra of dilute nuclei to be obtained under the appropriate experimental conditions. The heteronuclear dipolar interaction HD(H-X ) * s usually large, but since it is between the X nuclei and protons, it can be removed by the application of a powerful decoupling field at the proton resonance frequency while observing at the X-nucleus frequency. The critical simplifying feature comes from the "dilute" nature of the spin system: the homonuclear dipolar interaction H D ( X - X ) i s usually negligible due to the r _ 3 dependence of the interaction. After removal of the heteronuclear and the homonuclear interactions, the major remaining line broadening interaction for non-quadrupolar nuclei is the chemical shift anisotropy of the nucleus induced by the surrounding electrons. In most cases, this interaction can be averaged by using the magic angle spinning technique. b). High Power Decoupling The heteronuclear dipolar interaction Hrj(H-x) c a n be removed by the application of a strong radiofrequency field at the Larmor frequency of the protons. This applied field causes the protons to undergo rapid transitions, or spin flips, which cause the dipolar field generated by the protons at the X-nucleus to become zero 'in the time average' and thus it "decouples" the protons from the nucleus of interest18]. This approach is analogous to 16 scalar decoupling experiments used to remove indirect J-couplings in the high resolution NMR spectra of liquids. An important experimental difference between the decoupling in liquids and that in solids is the strength of the decoupling field required to remove the heteronuclear interactions in solids. A crude estimate of the decoupling field intensity required to effectively decouple protons can be made by noting that the "flip"rate of the protons must be comparable to the heteronuclear dipolar linewidths of the X-nucleus. Since protons "flip" at the rate of 4.3 kHz per Gauss and a typical solid-state dipolar linewidth is a few kHz, it requires a decoupling field on the order of a Gauss to be effective t19L This is a fairly high field, usually requiring kilowatt amplifiers (compared to the five watts or less required in the typical solution NMR experiment) especially considering that the decoupling field is required to be on for a long time compared to the transmitter pulse. Homonuclear dipolar interactions H(x-x) cannot be eliminated by high power decoupling, but as mentioned previously, they are usually negligibly small in magnetically dilute spin systems because of the dependence of the magnitude of the interaction on the inverse cube of the internuclear distance. c). Magic Angle Spinning (MAS) As we have mentioned earlier, the dipolar interaction is reduced by the presence of molecular motion, and completely removed by isotropic motion as in solution. Experiments to produce the same effect in the solid state by macroscopic rotation of the whole sample were proposed and implemented independently by Andrew, Bradbury and EadesC20'21! and Lowe^22]. The basic part of this experiment consists of fast rotation of the sample about an axis R inclined at an angle a to the magnetic field Ho by an angle p* to the vector r which is the internuclear vector ry in the case of dipolar interaction, as shown in Figure 1.4. The average of (3cos26 -1) about the conical path indicated for the vector r is given in Equation 1.14: 17 (3cos20 - 1) = \ (3cos2p - l)(3cos2a -1) (1.14) where the extremes of the angle 0 are oc+f3 and a-p\ Fortunately the angle a is under the control of the experimentalist. If a = 54°44' is chosen in the experiment, then cosa = (^) and (3cos2a -1)=0, so that (3cos26 - 1)=0 for all orientations (i.e. all values of 6 ).Since the dipolar interaction is proportional to the term (1 - 3cos26), at a = 54°44', the dipolar interaction vanishes, hence the label 'Magic Angle'. The scalar coupling does not vanish with MAS. However, these couplings are generally not observed in solid state NMR spectroscopy because the spectrum of the X-nucleus is obtained with high power decoupling. It was also shown independently by Andrew and Bradbury and Eades t 2 1! and Lowet 2 2! that magic angle spinning can also be used to average the chemical shift anisotropy. Under the condition of sample rotation about an axis inclined at an angle a to the magnetic field, the time average of the chemical shift tensor can be described as: 3 1 o*t = j ( sin2a) o"jso + ^ (3cos2a -1) a p (1.15) where o p is the projection of the chemical shift tensor onto the spinning axis. If a =54°44', then sin2a = ^ and (3cos2a - 1) = 0, so that the time average of the chemical shift tensor will be equal to the isotropic average value,OjS0. Spinning the sample at the magic angle and at a rate comparable to the frequency spread of the chemical shift anisotropy will yield a single isotropic average peak, while spinning at lower rate will yield a spectrum where the isotropic peak is accompanied by a series of sharp absorptions spaced at multiples 18 Figure 1.4. Schematic representation of the geometric arrangement for mechanical sample spinning. The solid sample is rotated with an angular velocity of C0r about R which is inclined to the magnetic field by the angle a. A typical vector r is inclined at the angle (5 to the rotation axis. Its inclination to HQ varies periodically with time (adapted from reference 23). 19 14 A of the sample rotation frequency and centered on the isotropic peak with intensities approximately reflecting the profile of the chemical shift anisotropy powder pattern. These peaks are called spinning 'sidebands' (SSB) (See figure 1.5). Spinning at angles other than 54°44' will result in the chemical shift tensor being reduced in width by the factor ^ (3cos2cc -1). In summary, the net effect of magic angle spinning is similar to that of the rapid and random molecular tumbling present in solution and can be used to reduce the dominant line broadening interactions present in spin 1/2 nuclei to their isotropic values, provided that the rate of sample spinning is comparable to the interaction energies in frequency units. d). Cross-Polarization(CP) The spin dilution discussed earlier is essential for obtaining high resolution NMR spectra of solid materials. However, it introduces the major disadvantage of reduced sensitivity. The reduction of sensitivity due to magnetic dilution makes the use of pulse-FT techniques mandatory. There are, however, many nuclei of interest that are of low absolute sensitivity and some have very long spin lattice relaxation times, making it even more difficult to obtain high resolution NMR spectra with good signal-to-noise ratio (S/N). These problems can be overcome by the technique known as "cross-polarization" (CP), first introduced by Pines, Gibby, and Waugh [25-27]. CP substantially increases the S/N in the spectra of these magnetically dilute spin systems. In the CP technique, the magnetization of rare spins is enhanced by the transfer of magnetization from the abundant spins. This is accomplished by the pulse sequence illustrated in Figure 1.6, using the example of *H as the abundant spin and 1 3 C as the rare spin. In the first step of the sequence, the proton magnetization is rotated by 90° by a strong on-resonance radiofrequency pulse applied along the x' axis to align with the y' axis in the 20 Figure 1.5 . 2 0 7 Pb spectra of polycrystalline Pb(N03)2 obtained at 41.7 MHz showing the effect of spinning rate on MAS spectra, (adapted from reference 24). A. The spectrum of a static sample with a chemical shift anisotropy pattern having a spectral width of 200 ppm; B. The spectrum at the magic angle spinning rate of 3.2 kHz; C. The spectrum with spinning at 500 Hz showing spinning sidebands 21 B J \ JU 3-IA Figure 1.6. A. Pulse sequence used for cross-polarization of a dilute nucleus ( 1 3C) from the abundant spin system ^ H). B. Schematic representation of the behavior of the *H and 1 3 C spin magnetizations during the CP pulse sequence in A, (steps 1-5 in A and B correspond). 22 .9 P° 90° phase pulse / 1H shift C P Decoupling 1 2 3 i t c -Wait Period V F I D / / / / Repeat rotating frame (the rotating frame is a coordinate system which rotates about the z-axis colinear with the applied magnetic field Ho at the same frequency as the nuclear moment precession) by a strong on-resonance radiofrequency pulse applied along the x* axis. The magnetization is then "spin-locked" along y' by an on-resonance "spin-locking" pulse H i H applied along y'. While the *H spins are kept along y' for a time period tc, a strong on-resonance pulse ^13c i s simultaneously applied to the 1 3 C spins. The magnitudes of the two spin-locking fields, and HI3Q, are chosen such that the 'Hartmann-Hahn' condition t28J is established (equation 1.16). Y i H H l H = Y i 3 c H i 3 c (1.16) The net effect of fulfilling the 'Hartmann-Hahn' or matching condition above is that during the contact time (tc), the energy required for spin flips is identical for both the *H and 1 3 C spins. Since the 1 3 C spins are very dilute in the system, they will adopt the more favorable spin distribution of the proton spin system, while the total proton magnetization will be little affected. Thus a maximum enhancement equal to the ratio of the magnetogyric ratios of the abundant and rare spins (Yi/Ys) may be obtained. For example, the maximum enhancement is equal to 4 for iH/^C systems and -5 for 1 H/ 2 9 Si systems. At the end of the "matching" or "contact" time, tc, the ^ 13^ RF field is switched off and the free induction decay (FID) of the X-nucleus is recorded during the acquisition time. The proton field is kept on during this period for heteronuclear decoupling of the 1 H -1 3 C dipolar interactions. After the acquisition of the 1 3 C FID, the proton field H i H is also turned off, followed by a delay time before the whole CP sequence is repeated again. Besides the signal enhancement by direct magnetization transfer from the abundant spin reservoir to the rare spins, there is another contribution to the enhancement of the rare spins. Since the growth of the magnetization of the X-nuclear spins depends only on the proton magnetization during the contact time, the rate at which the experiment can be repeated is 23 controlled only by the proton spin-lattice relaxation time, which is generally much shorter than the X-nucleus spin-lattice relaxation time. Thus the recycle time for efficient spectral accumulation is generally much reduced compared to that of a conventional pulse-FT experiment, resulting in much better S/N in a given time period. Cross-polarization is a signal enhancement technique which does not affect the widths of the signals and therefore does not affect the resolution of the experiment. The signal enhancement resulting from the cross-polarization technique may be quite different for different nuclei, producing spectra which will not be usually quantitatively reliable, although in general, they tend to be more quantitative than the comparable solution pulse Fourier transform experiments. This increased spectral reliability in the solid state is due to the signal enhancement resulting from intermolecular interactions ( The two nuclei I and S are not necessarily bonded). Thus, nuclei without directly attached protons (e.g. carbonyl) will still show some enhancement in the solid state compared to the solution NMR spectrum. e). High-Resolution Solid-State NMR Using Cross-Polarization and Magic Angle Spinning (CP/MAS) Techniques The complete CP/MAS experiment involves the use of: i). Cross-polarization to increase the S/N and to overcome the long Tj relaxation processes of the dilute nuclei being observed, ii). High-power proton decoupling during acquisition to remove the dipolar interactions, iii). Magic angle sample spinning to average the chemical shift anisotropy. The first experiment combining all these techniques which showed the general applicability of the method was reported by Schaefer and Stejskatf29] in 1976. The additive effects of the different components of the complete CP/MAS experiment are illustrated in Figure 1.7 for the 1 3 C spectrum of adamantane (adapted from reference 30). The top spectrum (Figure 1.7a) is of a stationary sample recorded under normal conditions. The spectrum shows a single broad peak with considerable 'noise'. The second 24 spectrum (Figure 1.7b) is again of a stationary sample, but high-power heteronuclear proton decoupling was employed during the acquisition. The resolution is greatly improved since the heteronuclear dipolar interactions are removed. Magic angle spinning can reduce both the dipole-dipole interactions and the chemical shift anisotropy, so the spectrum obtained with MAS (Figure 1.7c) shows a even better resolution than the one with high-power proton decoupling. The fourth spectrum (Figure 1.7d) was acquired using both MAS and high-power decoupling, the resolution is much better because of the complete removal of both the dipolar interactions and the chemical shift anisotropy. The last spectrum in the figure (Figure 1.7e) was acquired using the total combined experiment of cross-polarization, high-power decoupling and magic angle spinning. The signal/noise ratio of the spectrum has been greatly improved compared to the previous ones with no loss of resolution. f). AppUcation of CP/MAS Solid State NMR to the Study of SiUca Gel and Its Derivatives Silica gel is an amorphous powder (dried gel). The lack of the regularity of a crystal lattice makes it impossible to determine the structure of silica gel by X-ray diffraction techniques. The study of the surface properties of silica gel, surface-immobilized species and their reactions has been an area of intense research in recent years. However, these studies have suffered from the lack of a good, high-resolution spectroscopic technique capable of characterizing these chemically active surfaces and their derivatives. Solid state NMR has been shown to be a very efficient technique in characterizing the structures and reactions of silica gel and its derivatives. Sindorf and Macielt31"36! reported a series of studies of the surface compositions and reactions of silica gel using CP/MAS 2 9 S i and 1 3 C solid state NMR. Fyfe and coworkers^] have shown that a quantitatively reliable analysis of silica gel and its surface derivatives can be obtained by 2 9 S i solid state NMR using the magic angle spinning technique alone at high magnetic field strengths. 25 Figure 1.7. 1 3 C CP/MAS spectrum of adamantane. (adapted from reference 30) a) . Spectrum of static sample. b) . Proton-decoupled spectrum. c) . MAS spectrum. d) . Spectrum with MAS and proton decoupling. e) . CP/MAS spectrum with proton decoupling. The respective linewidths are indicated 26 4 0 0 0 H z d 2 H z The framework of silica gel consists of siloxane bonding units in the bulk of the gel and silanol groups on its surface, with some physically absorbed water. 2 9 S i solid state NMR is ideally suited to studies of the structure of silica gel for several reasons, is a spin 1/2 nucleus and satisfies the requirement of being a "dilute" nucleus due to its low natural abundance (4.7%) which results in very small dipolar interactions between the nuclei. When silica gel is well dried, there are only a small number of silanol groups as the proton source. The heteronuclear dipolar interactions between the *H and 2 9 S i nuclei are small. Thus a conventional one pulse experiment can be used to obtain a quantitatively reliable 2 9 S i spectrum using the MAS technique alone at high magnetic field strengths^7] using long recycle times. Proton decoupling during acquisition is not necessary. Proton NMR should, in principle, be an extremely useful tool in the characterization of the surface groups of silica gel and its derivatives. However, ^ - i H magnetic dipolar interactions usually give featureless broad line spectra. To obtain high resolution *H NMR spectra of solids, one can lower the *H concentration by isotopic exchange with 2 H resulting in reduced homonuclear dipolar interactions where relatively narrow *H NMR peaks can be obtained with sample spinning alone. Otherwise it is possible to use extremely fast MAS (above 5 kHz) [38,39] o r multiple-pulse line-narrowing techniques^40'41! or the combination of MAS with multiple-pulse line-narrowing (CRAMPS: Combined Rotation and Multi-Pulse Spectroscopy) f 4 2l techniques. 1 3 C is a spin 1/2 nucleus with natural abundance of 1.1%. It fulfils the requirement of being "dilute", so 1 3 C solid state NMR techniques can be easily used in the study of systems containing 1 3 C nuclei. Carbon atoms exist usually along with protons, which makes it possible to use cross polarization techniques to enhance the S/N of the 1 3 C spectra. Thus one has to use high power decoupling to remove the *H- 1 3 C dipolar interactions and MAS to reduce the chemical shift anisotropy to obtain high resolution 1 3 C spectra of solid samples. Several papers'^ 43"47] have been published on the study of alkyl group modified silica gels using 1 3 C CP/MAS solid state NMR. 27 If the functional group used for modifying silica gel contains nuclei which fulfil the requirements of the "dilute" condition, these nuclei can also provide some very useful information regarding the surface characteristics of the modified silica gels. One such example is 3 1 P , which is a spin 1/2 nucleus. Even though its natural abundance is 100%, it is usually present at a low absolute concentration and can yield high resolution 3 1 P NMR spectra using CP/MAS and high power decoupling techniques. Phosphines have long been recognized as excellent ligands for transition metal complexes and are commonly used in the immobilization of transition metal complexes on insoluble supports'48^ where the 3 1 P nuclei in the ligands should, in principle, act as sensitive probes of both the electronic and the geometric structure of the whole immobilized complex. Fyfe and coworkers t49"52! have investigated the use of 3 1 P CP/MAS NMR for the detection and characterization of transition metal complex catalysts supported by organic polymer (polystyrene) as well as inorganic polymer, (silica), matrices. Their results have shown that 3 1 P CP/MAS is a sensitive and very diagnostic technique for probing the structure of polymer supported catalysts. 1.3. Outline of the Research Project Our aims in this project were to synthesize a variety of organofunctionalized silica gels using the copolymerization method as well as the conventional anchoring method and to use cross polarization, magic angle spinning solid state NMR spectra of appropriate nuclei (mainly 2 9 Si , 1 3 C) to investigate the structures of the modified silica gels. We also thought to use two-dimensional solid-state heteronuclear correlation NMR techniques to study the structural differences between the modified silica gels synthesized by the two different methods. The thermal stabilities of the chemically modified silica gels were also to be investigated by the use of other experimental techniques, for example, Differential Scanning Calorimetry (DSC). 28 1.4. REFERENCES 1.1) . R K. Her, "The Chemistry of Silica: Polymerization, Colloid and Surface Properties and Biochemistry," Wiley-Intersciences, John Wiley & Sons, New York, (1979). 1.2) . E.P. Plueddemann in "Kirk-Othmer Encyclopedia of Chemical Technology", Vol.20, Wiley, New York, 3rd. Ed.(1982), p967. 1.3) . Vaclav Chvalovska in "Carbon-Functional Organosilicon Compounds", Editor: Valav Chvalovska and J.M. Bellama, Plenum Press, New York and London, (1984). 1.4) . V. Deschler, P. Kleinschmidt and P. Panster, Angew. Chem. Int. Ed. Eng. 25, (1986) 236. 1.5) . D.E. Leyden and W. T. Collins (Editors)," Silylated Surfaces", Gordon and Breach Science Publishers, New York, London and Paris, (1978). 1.6) . E.P. Plueddemann, "Silane Coupling Agents", Plenum Press, New York, London (1982). 1.7) . D.W. Sindorf and G.E. Maciel, J. Phys. Chem., 88, (1982) 5208. 1.8) . H.A. Claessens, LJ .M. Van de Ven, J.W. de Haan, C A . Cramers and N. Vonk, J. High Resolution. Chromat. & Chromat. Comm., (1983) 433. 1.9) . D.E. Sindorf and G.E. Maciel, J. Am. Chem. Soc. 105, (1983) 3767. 1.10) . M.E. Gangona and R.K. Gilpin, J. Mag. Res. 53, (1983) 140. 1.11) . A. Abragam, "Principles of Nuclear Magnetism", Claredon Press, Oxford, (1961). 1.12) . CP. Slichter, "Principles of Magnetic Resonance", Second Ed., Springer-Verlag, Berlin, (1978). 1.13) . C.A. Fyfe, "Solid-State NMR for Chemists", C.F.C. Press, Guelph, (1984). 29 1.14) . M. Mehring, "Principle of High Resolution NMR in Solids", Springer-Verlag, Berlin, Heidelberg, New York (1983). 1.15) . C.R. Dybowski and B.C. Gerstein, "Transient Techniques in NMR, An Introduction to the Theory and Practice", Academic Press, New York, (1982). 1.16) . V. Haeberlen, "Advances in Magnetic Resonance", Suppl. 1, Academic Press. New York (1976). 1.17) . E R. Andrew, in "Progress in NMR Spectroscopy," Vol 8, Eds., J.M. Emsley, J. Feeney and L H. Sutcliffe, Pergamon Press, Oxford, (1971). 1.18) . F. Bloch, Phys. Rev., I l l , (1958) 841. 1.19) . E. Fukushima and S B. W. Roeder, "Experimental Pulse NMR. A Nuts and Bolts Approach", Addison-Wesley Publishing Company Inc, Toronto, (1981). 1.20) . E.R. Andrew, A.Bradbury and R.G. Eades, Nature (London), 182, (1958) 1659. 1.21) . E.R. Andrew, A. Bradbury and R.G. Eades, Nature (London), 183, (1959) 1802 1.22) . I.J. Lowe, Phys. Rev. Lett., 2, (1959) 285. 1.23) . R.K. Harris, "Nuclear Magnetic Resonance Spectroscopy, A Physical View", Pitman, London, (1983). 1.24) . J.K. M. Saunders and B.K. Hunter, "Modern NMR Spectroscopy," 2nd Ed., Wiley Interscience, New York, (1980). 1.25) . A. Pines, M.G. Gibby and J.S. Waugh, J. Phys. Chem., 56, (1972) 1776. 1.26) . A. Pines, M.G. Gibby and J.S. Waugh, Chem. Phys. Lett., 15, (1972) 373. 1.27) . A. Pines, M.G. Gibby and J.S. Waugh, J. Phys. Chem., 59, (1973) 569. 1.28) . S.R. Hartmann and E.L. Hahn, Phys. Rev. 128, (1962) 2042. 1.29) . J. Schaefer and E.O. Stejskal, J. Am. Chem. Soc.,_98, (1976) 1031. 1.30) . CXP Application Note: "High Resolution NMR in Solids by Magic Angle Spinning," Fig. 3, Page 5, Spectrospin., A.G., Zurich. 1.31) . D.W. Sindorf and G.E. Maciel, J. Am. Chem. Soc.,J02, (1980) 7606. 30 1.32) . D.W. Sindorf and G.E. Maciel, J. Am. Chem. Soc. 103, (1981) 4263. 1.33) . D.W. Sindorf and G.E. Maciel, J. Phys. Chem., 86 (1982) 5208. 1.34) D.W. Sindorf and G.E. Maciel, J. Phys. Chem. 87, (1983) 5516. 1.35) . D.W. Sindorf and G.E. Maciel, J. Am. Chem. Soc, 105, (1983) 1487. 1.36) . D.W. Sindorf and G.E. Maciel, J. Am. Chem. Soc, 105, (1983) 1848. 1.37) . C A . Fyfe, G.C Gobbi and G. J. Kennedy, J. Phys. Chem., 89, (1985) 278. 1.38) . S.F. Dec, R.A. Wind and G.E. Maciel, Macromolecules, 20, (1987) 2754. 1.49) . S.F. Dec, R.A. Wind and G.E. Maciel, J. Mag. Reson., 70, (1986) 335. 1.40) . J.S. Waugh, L.M. Huber and U. Haeberlin, Phys. Rev. Lett., 20, (1968) 180. 1.41) . D.P. Durum and W.K. Rhim, J. Chem. Phys., 71, (1979) 944. 1.42) . C E . Bronnimann, R.C. Zeibler and G.E. Maciel, J. Am. Chem. Soc, 110, (1988) 2023. 1.43) . G.R. Hays, A.D.H. Glague, R. Huis and G. Van Der Velden, Appl. Surf. Sci., 10, (1982) 247. 1.44) . G.E. Maciel, D.W. Sindorf and V. J. Bartuska, J. Chromat., 205, (1981) 438. 1.45) . E. Bayer, K. Albert, J. Reiners, M. Nieder and D. Muller, J. Chromat., 264. (1983) 197. 1.46) . W.E. Rudznski, T.L. Montgomery, J.S. Frey, B.L. Hawkins and G.E. Maciel, J. Catal., 28, (1986) 444. 1.47) . C H . Chiang, N.I. Liu and J.L. Koenig, J. Colloid, and Surf. Sci., 86, (1982) 27. 1.48) . R.H. Grubbs, CHEMTECH, (1973) 560. 1.59). C A . Fyfe, H.C Clark, J.A. Davies, P.J. Hayes and R.E. Wasylishen, J. Am. Chem. Soc, 105, (1983) 6577. 1.50) . H.C. Clark, J.A. Davies, C A . Fyfe, P.J. Hayes and R.E. Wasylishen, Organometal. 2, (1983) 177. 31 1.51) . L. Bemi, H.C. Clark, J.A. Davies, D. Drexler, C A . Fyfe and R.E. Wasylishen, J. Organomet. Chem., 224, (1982) c5. 1.52) . L. Bemi, H.C. Clark, J.A. Davies, D. Drexler, C A . Fyfe and R.E. Wasylishen, J. Am. Chem. Soc, 104, (1982) 438. 32 Chapter II PREPARATION OF FUNCTIONALIZED SILICA GELS AND A  PRELIMINARY INVESTIGATION OF THEIR STRUCTURES BY SOLID STATE NMR 2.1. Experimental 2.1.1. Reagents All of the functionalized silanes were purchased from Pertrarch System Inc. and were used in the preparations without further treatment. Silica gel (Fisher S-157) used as substrate for the anchoring reaction was purchased from the Fisher Chemical Company. All other chemicals used as solvents were used directly as received, except toluene, which was distilled from sodium under N2 before use. 2.1.2. Preparation of High-Purity Silica Gel The high-purity silica gel samples were synthesized by the acid hydrolysis of tetraethoxysilane (TEOS), using the method described by Peri and Hensley A 60-ml volume of concentrated HCl was added dropwise, with constant stirring, to 100 ml of a solution of TEOS (40 vol.%) in methanol. The solution gelled in about 1 hour. After standing overnight, the gel was broken into small pieces and placed in 200 ml of 50 vol.% methanol in distilled water. After standing for 12 hours, the gel was washed with distilled water several times until the pH of the eluate was 7, then dried at room temperature for a week. It was then ground to a fine powder and characterized by solid-state NMR. 33 2.1.3. Preparation of Functionalized Silica Gels i) . Copolymerization Method All functionalized silica gels made by this method were synthesized by the cocondensation of a functionalized silane ( e.g. methyltriethoxysilane (MTES), CH3Si(OEt)3 ), with tetraethoxysilane (TEOS), Si(OEt)4, in methanol as solvent using concentrated HC1 as catalyst. A typical example (Gel 3) of the procedure is as follows: 0.2 mole of TEOS was mixed with 0.05 mole of the functionalized silane in 60 ml methanol and 40 ml of concentrated HC1 was added into the mixture with constant stirring. The solution normally gelled in 20 minutes to one hour depending on the type of functionalized silane used. The gel was dried in air for a period of one week after thorough washing with methanol and water. Further treatments of the gels were carried out according to the nature of the experiment. ii) . Anchoring Method Silylation of the silica gel (Fisher S-157, surface area 750 m2/g) was accomplished by refluxing ~6 g of the substrate, which had been previously dried in vacuo at ~150°C overnight, with an excess of trifunctional silane, (CH3Si(OR)3), coupling agent in dry toluene for 12 hours. After filtration, the solid was Soxhlet-extracted in toluene for a further 12 hours and washed with acetone and water successively several times. The solid was then refluxed in water for 12 hours and then dried at ~100°C in vacuo for 24 hours. 34 2.1.4. Solid State NMR A Bruker MSL-400 high-power NMR spectrometer with a 9.4T wide-bore magnet was used to obtain the 2 9 Si , 1 3 C , *H and 3 1 P NMR spectra at 79.5 MHz, 100.6 MHz, 400.0 MHz and 160.9 MHz respectively. A Bruker CP/MAS broadbanded (variable x-nucleus tuning frequency) probe was used in these experiments. Ceramic spinners with an internal volume of approximately 450 uL were used in all cases and were spun with compressed air at frequencies of 3-5 kHz. Optimization of the magic angle (54° M) was achieved using the technique described by Frye and Maciel^2] where the intensities of the spinning sidebands of KBr are maximized while observing the 7 9 B r resonance. Secondary external reference standards were used in all the spectral measurements. For 2 9 S i spectra, solid QsMs (Cubic octamer silicic acid trimethylsilyl ester) was used as secondary standard with the 2 9 S i chemical shift of the high field peak assigned as -109.7 ppm with respect to tetramethylsilane (TMS)t3]. For 1 3 C spectra, solid adamantane was used as secondary standard with the 1 3 C chemical shift of the low field peak assigned as 38.56 ppm with respect to T M S ^ . For *H spectra, liquid H 2 O was used as a secondary standard and the *H chemical shift was assigned as 5.20 ppm with respect to TMS. Thus chemical shifts of all the spectra in this thesis were reported in ppm with respect to TMS. 2.1.5.Elemental Analysis Microscale elemental analyses for C and H were performed by Mr. P. Borda of the Microanalytical Laboratory, Chemistry Department, University of British Columbia. All samples were heated for at least 12 hours at ~110°C in vacuo before being sent for microanalysis. 35 2.2. Results and Discussion 2.2.1. Characterization of High Purity Silica Gel Figure 2.1 shows the 2 9 S i solid-state CP/MAS spectrum of the high purity silica gel synthesized by the method described in Section 2.1.2. The basic feature of the spectrum is that there are three well separated peaks which are at -110.6 ppm, -100.6 ppm and -90.8 ppm with respect to TMS. The three resonances labelled A, B and C are separated by intervals of approximately 10 ppm, strongly suggesting that the local chemical environments indicated in the figure are in agreement with the results of previous studies^ 5]. The assignments of the chemical shifts for the solid-state NMR resonances of the silica gel are in good agreement with values previously reported in 2 9 Si studies of related siloxane materials. A representative subset of such work involving both solid-state and conventional 2 9 S i solution NMR is summarized in Table 2.1. 2 9 S i studies of siloxane polymers in solution suggest that a 9 to 10 ppm chemical shift difference should separate resonances associated with groups differing by the substitution of a silicon hydroxyl (=SiOH) for a siloxane bond (^SjiOSi)^6"10]. The data in this table indicate that the absolute chemical shifts of silicon environments assigned in this study fall comfortably within the range expected for each structural type. It also indicates that the resolution obtained by solid-state 2 9 Si CP/MAS NMR is good enough to distinguish the different local chemical environments of the silicon atoms in silica gel in agreement with the results of previous studies. 36 Figure 2.1. 2 9 Si CP/MAS NMR spectrum of high-purity silica gel obtained at 79.5 MHz. 300 scans, contact time 20.0 ms, repetition time 5.0 s, and a sample spinning rate of 3.2 kHz. 37 Table 2.1.2 9Si Chemical Shifts of Silica Gels and Related Model Systems3 Species Si(OSi=)4 HOSi(OSi=)3 (HO)2Si(OSi=)2 Environment Shift* Reference Aqueous silicic acid polymers -107 6 Quartz CP/MAS -107.4 8 Low cristobalite CP/MAS -109.9 8 Silica gel CP/MAS -109.3 5 Silica gel CP/MAS -110.6 This study Aqueous silicic acid polymers -95 to -100 6 Solid silicates CP/MAS -98 to -99 8 Silica Gel CP/MAS -99.8 5 Silica gel CP/MAS -100.6 This study Aqueous silicic acid polymers -86 to -91 6 Solid silicates CP/MAS -86 to -88 8 Silica gel CP/MAS -90.6 5 Silica gel CP/MAS -90.8 This study a). Reported in ppm with respect to TMS 38 2.2.2. Characterization of Functionalized Silica Gels a). Methyl-Functionalized Silica Gel Methyl-functionalized silica gel (Gel 3) was prepared by the copolymerization method described in section 2.1.3. The chemical reaction can be described qualitatively as in Equation 2.1. I H3CSi(OEt)3 + Si(OEt)4 (2.1) i where represents the siloxane (silica) polymer matrix. Another methyl-functionalized silica gel (Gel 18) was also prepared by the copolymerization method, but a mole ratio of TEOS and MTES of 1:1 was used instead of 4:1 as in the preparation of Gel 3. Gel 3 was characterized by both 2 9 S i and 1 3 C solid-state CP/MAS NMR. Figure 2.2A shows the 2 9 S i CP/MAS spectrum of Gel 3. The resonances in the high field region of the spectrum are similar to those of the pure silica gel and correspond to silicon atoms with four siloxane bonds, Si(OSi=)4 at -110.5 ppm, to silicon atoms with three siloxane bonds and one hydroxyl group, (HO)Si(OSi=)3, at -101.3 ppm, and to silicon atoms with two siloxane bonds and two hydroxyl groups, (HO)2Si(OSi=)2, at -91.8 ppm. The low field resonances at -61.3 ppm and -54.2 ppm can be assigned to methyl substituted surface silicons with three siloxane bonds, H3CSi(OSi=)3 and with one hydroxyl group and two siloxane bonds, H3C(HO)Si(OSi=)2 respectively according to the literature t6'11]. From the spectrum, it can be seen that the two components TEOS and MTES have copolymerized together to some 39 Figure 2.2. 1 3 C and 2 9 Si CP/MAS NMR spectra of methylsilica gel (Gel 3). A. 2 9 Si CP/MAS NMR spectrum obtained at 79.5 MHz, 320 scans, 22.0 contact time and 5.0 s repetition time. The sample was spun at 3.2 kHz. B. 1 3 C CP/MAS NMR spectrum obtained at 100.6 MHz, 800 scans, 1.0 ms contact time, 2.0 s repetition time and a 3.2 kHz sample spinning rate. 40 B. 1 3 C —I—'—i—'—I—'—I—'—I—1—I—' I 1 I 80 60 40 20 0 -2 0 - 4 0 - 6 0 F F H AOA degree although we can not tell the microstructure of the material at this stage from the spectral data. Figure 2.2B is the 1 3 C CP/MAS spectrum of Gel 3. The fact that only one methyl resonance appears in the spectrum suggests that the hydrolysis reactions have been quite complete, since there are no residual ethoxy groups remaining in the gel. b). Other Examples of Functionalized Silica Gels Silica-attached long chain organics are widely used as stationary phases in reverse-phase chromatography. As an alternative to the traditional synthetic method of preparing this kind of material by bonding of the reactant functionality, R, to the silica surface, materials of these organosilanes were prepared by the copolymerization method. The preparation can be described as follows: where R = -C6H5 (Gel 4), -CH 2(CH 2)6CH 3 (Gel 6), -CH2(CH2)i6CH3 (Gel 7) and - C H 2 C H 2 C H 2 N H 2 (Gel 11). Figure 2.3 shows the CP/MAS NMR spectra of Gel 6. From the 2 9 S i CP/MAS NMR spectrum of the gel (Figure 2.2A), we can see that the resonances of all the silicon species are well resolved. The main features of the 2 9 Si spectrum are the same as those of Gel 3 in terms of the chemical shifts, as expected. Figure 2.3B is the 1 3 C CP/MAS NMR spectrum obtained at 100.6 MHz of Gel 6. The assignments of the solid-state NMR peaks are based on the 1 3 C NMR spectrum of the precursor silane in solution at 75.5 MHz (Table 2.2). From the solid-state 1 3 C NMR data, (2.2) 41 Table 2.2. 1 3 C NMR data for functionalized silica gels and precursors X = -OH or -0-Si=, = silica-siloxane polymer substrate. Compound 1 3 C NMR data (ppm) Ref. 2_3 ( C H 3 - C H 2 0 ) 3 S i - ^ ~ ) 4 Q C 2 C 3 C 4 18.2 58.6 ^ — ' 130.2 134.7 127.8 127.8 Gel 4 (CH3-CH20)Si-CH2-CH2-CH2-CH2-CH2-CH2-CH2-CH3. 18.4 58.2 10.8 23.2 33.5 29.7 29.7 32.3 23.0 14.1 Gel 6 ^ 0 - S i - C H 2 - C H 2 - C H 2 - C H 2 - C H 2 - C H 2 - C H 2 - C H 3 . ^ I 13.0 23.8 33.1 30.5 30.5 33.1 23.8 14.1 X (CH3-CH20)3Si-CH2-CH2-CH2-(CH2-)i2CH2-CH2-CH3 18.4 58.1 10.8 23.2 33.5 29.7/30.2 32.4 23.0 14.2 X Gel 7 J ) - 0-Si-CH 2 -CH 2 -CH 2 -(CH2) 1 2 -CH 2 -CH 2 -CH 3 I 15.2 23.9 33.1 31.2 33.1 23.9 13.1 (CH3-CH20)Si-CH2-CH2-CH2-NH2 18.4 58.2 7.8 27.7 45.3 (17.6 57.7 7.0 26.6 44.5) 13 T Gel 11 P)-0-Si-CH 2 -CH2-CH 2 -NH 3 ^ I 10.6 21.9T 43.6 42 Figure 2.3. 2 9 S i and 1 3 C CP/MAS NMR spectra of Gel 6. A. 2 9 Si CP/MAS NMR spectrum obtained at 79.5 MHz, 320 scans, 22.0 ms contact time and 5.0 s repetition time. The sample was spun at 3.2 kHz. B. 1 3 C CP/MAS NMR spectrum obtained at 100.6 MHz, 1714 scans, 1.0 ms contact time, 2.0 s repetition time and a 3.2 kHz sample spinning rate. 43 I V 1 2 3 4 5 6 7 8 p)-0—Si-CH2CH2CH2CH2CH2CH2CH2CH3 B.13C O I (2+7) 183 S0 PPM -SB (HO)5i(OSis)3 A. 2 9 Si I 43A we can see that the chemical shifts resemble those of the precursor silane in solution. The resolution obtained in the solid-state is good enough to distinguish some of the different carbon sites in the alkyl chain. The NMR data indicate that the chemical shift of C-l (=Si-C_H2-) is moved downfield about 2.0 ppm after the silane becomes a gel, the deshielding being apparently due to the change of the environment of these carbon nuclei after condensation. In the 1 3 C CP/MAS spectrum of Gel 6, there is a small additional peak which appears at ~52 ppm. Because silane-attached methoxy groups absorb at about 50 ppm, we assign the resonance at 52 ppm to =Si-OCH3 groups, which could be derived from the starting material during the preparation of the functionalized silica gel. A possible explanation is an equilibrium reaction of free silanol groups with methanol: =Si-OH + C H 3 O H « =Si-OCH.3 +H20 (2.3) Relatively speaking, the 1 3 C CP/MAS spectrum of the functionalized silica gel made by copolymerization of TEOS and octyltriethoxysilane is the same as those of the similar gels made by anchoring methods which were reported by Bayer and coworkers^12! and Hays and coworkers^13]. Figure 2.4 shows the results of the solid-state NMR characterization of Gel 11 which was made by copolymerization of TEOS and 3-aminopropyltriethoxysilane (APS). The 2 9 Si CP/MAS NMR spectrum of the gel (Figure 2.4A) indicates that it contains five silicon species, which are identified in the figure inserts. This spectrum is similar to that of Gel 3, although the functional groups of the two gels are quite different since the point of attachment (Si-O-C) is the same in all cases. Figure 2.4B is the 1 3 C CP/MAS NMR spectrum of Gel 11, from which we can see that the three different carbons of the organic functionalities are well resolved in the solid state spectrum. The spectrum indicates that the 44 Figure 2.4. 2 9 S i and 1 3 C CP/MAS NMR spectra of Gel 11. A. 2 9 S i CP/MAS NMR spectrum obtained at 79.5 MHz, 320 scans, 22.0 ms contact time and 5.0 s repetition time. The sample was spun at 3.2 kHz. B. 1 3 C CP/MAS NMR spectrum obtained at 100.6 MHz, 280 scans, 1.0 ms contact time, 2.0 s repetition time and a 3.2 kHz sample spinning rate. 45 ~ ? 1 2 3 P)-0-Si-CH2CH 2CH 2NH 2 O B. 1 3 C J l _ -1 I I L. 100 50 P P M A. 2 9 Si (HO)Si(OSUi)3 I Si(OSis)4 ft / I I I I I 1 L . -50 -100 _! I I I I 1 1 -150 PPM 4£X hydrolysis of the ethoxy groups is complete since no alkoxy resonances remain after the copolymerization. Comparing the NMR data of Gel 11 and the solution spectrum of APS, it is seen that after condensation, Ci shifts ~3 ppm to a lower field and the C 2 and C3 resonances shift to higher fields as shown in Table 2. These results are consistent with the work of Chiang and his coworkers^14!, who studied the structures of the poly-APS by CP/MAS NMR and suggested that the chemical shift change after the polymerization of APS is related to differences in the APS structure. To provide further examples of functionalized silica gels by the copolymerization method, two further samples were prepared. Figure 2.5A shows the 1 3 C CP/MAS NMR spectrum of Gel 7 from the cocondensation of TEOS and octyldecyltriethoxysilane. The corresponding surface modified silica gel made by the surface reaction of silica gel and octyldecyltriethoxysilane is well known as the C-18 phase in reverse-phase chromatography and has been well characterized by solid-state NMR [12,15]. The assignment of the 1 3 C spectrum is confirmed by the solution NMR data of the precursor, octydecyltriethoxysilane (Table 2). As in the case of Gel 6, the C-l resonance of Gel 7 also shifts about 3 ppm downfield after condensation. Figure 2.5B is the 1 3 C CP/MAS NMR spectrum of Gel 4 which was prepared by copolymerization of TEOS and phenyltriethoxysilane. From the 1 3 C chemical shift data of the phenyltriethoxysilane in solution (Table 2.2), we assigned the resonance peaks of the aryl carbons in Gel 4 as Ci (the carbon atoms connected with silicon atoms) at 130.2 ppm, C 2 (ortho carbons ) at 134.7 ppm and C3 and C4 (para and meta carbons) at 127.8 ppm respectively. With the aid of the MAS technique, all the three carbon resonance peaks observed in the solution spectrum of phenyltriethoxysilane are resolved in the solid state NMR spectrum. In contrast to the alkyl carbons in the previous functionalized silica gels, the aryl carbons in phenyltriethoxysilane are little affected by the copolymerization in terms of their chemical shifts. 46 Figure 2.5. A. 1 3 C CP/MAS NMR spectrum of Gel 7 obtained at 100.6 MHz, 9680 scans, 1.0 ms contact time and 2.0 repetition time. The sample was spun at 3.3 kHz. B. 1 3 C CP/MAS NMR spectrum of Gel 4 obtained at 100.6 MHz, 1500 scans, 1.0 ms contact time and 2.0 repetition time. The sample was spun at 3.3 kHz. 47 B Q 2 3 i ' i ' 1 1 1 1 1 1 1 1 i 1 r 150 100 PPM I i 2 3 4 5-14 15 16 17 18 ^0-Si-CH2CH2CH2CH2(CH2)ioCH2CH2CH2CH3 1 00 50 PPM 4?A 2.2.3. Quantitative NMR Investigations of Functionalized Silica Gels As has been shown in the previous sections, the combined techniques of cross-polarization, high power dipolar decoupling and magic angle spinning have made it possible to obtain NMR spectra of functionalized silica gels of moderate resolution. Cross-polarization itself helps greatly to enhance the signal-to-noise ratios of the 1 3 C and 2 9 S i spectra of the systems studied. However, the spectra obtained using the cross-polarization technique can not be quantitatively interpreted because the proton polarization sources and the dynamics of the enhancement may be different for different individual nuclei in the system. In studies of similar systems by 2 9 S i CP/MAS, Maciel et al t5'15] concluded that the cross-polarization sequence detects silicon nuclei that are at or near the surface where the OH groups are located since the 2 9 Si-!H cross-polarization efficiency decreases as the Si-H distance increases and the internal Si atoms of the silica framework should not contribute to the CP/MAS spectra. The peak intensities in these CP/MAS spectra cannot be assumed to be quantitatively reliable because they depend on 2^Si-lH cross-polarization dynamics, and as pointed out by the authors t 1 5 l "should not be used quantitatively until detailed studies of these factors have been carried out". Without such studies, it is not possible from CP/MAS spectra alone to determine how "surface sensitive" the CP/MAS technique is. Based on the fact that in some systems like silica gels and their derivatives where the dipolar interactions are not large it is possible to remove intermolecular dipolar interactions by MAS, Fyfe and coworkers t 1 6l demonstrated that using the MAS technique alone can provide completely quantitatively reliable structural information of silica gels and their surface-modified derivatives as long as the delay time is ~5 times the Ti of the slowest relaxing Si nuclei. The loss in S/N by not using the CP technique may be alleviated by working at very high HQ values. 48 In this section, we report the results obtained for the functionalized silica gels using Fyfe et al's method^16]. To get a quantitatively reliable MAS NMR spectrum, one has to first determine the spin-lattice relaxation times (Ti) of the silicon nuclei and then record the spectrum using a single-pulse technique with a repetition time 5 times the longest Ti value to ensure complete remagnetization of all of the 2 9 Si nuclei between pulses. Figure 2.6A is the quantitative spectrum of the functionalized silica gel Gel 18. The resolution obtained is sufficient for accurate deconvolution of the spectrum, which enables one to determine the relative proportions of the different silicon nuclei present in the system being investigated. The relative peak areas from a deconvolution of the spectrum into overlapping Gaussian peaks are indicated in Figure 2.6B. The results indicate that in Gel 18, there are 28.7% of the silicon atoms with four siloxane bonds £i(OSi=)4,13.7% of the silicon atoms with three siloxane bonds and one hydroxyl group (HO)£i(OSi=)3, and 1.5% silicon atoms with two siloxane bonds and two hydroxyl groups(HO)?Si(QSi=)?, totalling 43.9% of the silicon atoms from the condensation of the TEOS component, while there are 44.7% of surface silanes with three siloxane bonds CH3Si(OSi=)3 and 11.4% of surface silanes with two siloxane bonds and one hydroxyl group CH3(HO)Si*(OSi=)2 totalling 56.1% of the silicon atoms from the condensation of the MTES component. To check the reliability of the observed relative intensities of the 2 9 S i peaks, the functionalized silica gel was analyzed for carbon content. If we assume that the sample is made up entirely of Si02 and CH3Si02/3 units, the elemental analysis (11.2 wt % C) corresponds to a SiO2-t0-CH3SiO2/3 ratio of 0.4. The ratio of the intensity in the silica region to the intensity of the surface silane peaks in the 2 9 Si MAS NMR spectrum of Gel 18 (Figure 2.6B) is 0.44, suggesting that the observed peak intensities are both quantitative and reliable. 49 Figure 2.6. A. Quantitatively reliable one-pulse 2 9 Si MAS NMR spectrum of Gel 18 obtained at 79.5 MHz, with a repetition time of 107 s and 563 scans. The sample was spun at 4.0 kHz. B. Deconvolution of spectrum A in terms of Gaussian peaks. 50 RSi(OSia)j 2.3. REFERENCES 2.1) . J.B. Peri and A.L. Hensley, Jr, J. Phys. Chem., 72,(1968) 2926. 2.2) . J.S. Frye and G.E. Maciel, J. Mag. Res.,48, (1982) 125. 2.3) . G. Engelhardt and D. Michel, "High Resolution Solid-State NMR of Silicates and Zeolites", John Wiley & Sons, Chichester, (1987) pi 13. 2.4) . W.L. Earl and D.L Van der Hart, J. Mag. Reson., 48, (1982) 35. 2.5) . G.E Maciel, D.W. Sindorf, J. Am. Chem. Soc, 102, (1980) 7606. 2.6) . H.C. Marsmann, Z. Naturforsch, 296,(1974) 495. 2.7) . R.K. Harris and R.H. Newman, J. Chem. Soc, Faraday Trans. II, 73, (1977) 1204. 2.8) . E. Lippmaa, M. Magi, A. Samoson, G. Engelhardt and A.R. Grimmer, J. Amer. Chem. Soc, 102, (1980) 4889. 2.9) . Von.D. Hoebbel, G. Garzo, G. Engelhardt and A. Till, Z. Anorg. Allg. Chem., 450. (1979) 5. 2.10) R.K Harris and B.E.M. Bann, "NMR and the Periodic Table", Academic Press, New York, 1978. 2.11) . G. Engelhardt, H. Jancke, E. Lippmaa and A. Samoson, J. Organomet. Chem. 210. (1981) 295. 2.12) . E. Bager, K. Albert, J. Reiners, M. Nieder and D. Miiller, J. Chromat., 264. (1983) 197. 2.13) . G.R. Hays, A.D.H. Clague, R. Huis and G.van der Valden, Appl. Surf. Sci., KL (1982) 247. 2.14) . C-H. Chiang, N-I. Liu and J.L. Koenig, J. Colloid & Interf. Sci., 86, (1982) 27. 2.15) . G.E. Maciel, D.W. Sindorf and V.J. Bartuska, J. Chromat., 205, (1981) 438. 2.16) . C A . Fyfe, G.C Gobbi and G.J. Kennedy, J. Phys. Chem., 89, (1985) 281. 51 Chanter TIT1 THE CHARACTERIZATION OF FUNCTIONALIZED SILICA GELS  BY TWO-DIMENSIONAL HIGH-RESOLUTION SOLID-STATE  HETERONUCLEAR CORRELATION NMR SPECTROSCOPY 3.1. Introduction 3.1.1. Two-Dimensional NMR A two-dimensional (2D) spectrum S(fi,f2) is defined as a spectrum which is a function of two independent frequency variables fi and f2 and is obtained after a two-dimensional Fourier transformation of a time function S(ti,t2). The two frequency axes of such a spectrum will be labelled Fi and F 2 . Jeener HI was the first (1971) to propose the idea of two-dimensional Fourier transformation of an NMR signal obtained as a function of two time variables, yielding a spectrum that is a function of two frequency variables. It was much later (1976) before a detailed and rather complex theoretical description of 2D Fourier transformation NMR was presented by Ernst and coworkers 121. Since then, 2D NMR has been highly investigated and has proven to be one of the most powerful NMR techniques for the elucidation of chemical structures, including organic, inorganic, organometallic and biological molecules. Its usefulness is clearly manifested in the rapid growth in the number of publications related to the development and application of the technique. For example, seventeen papers were published in 1977, whereas 197 were published in 1985. By the end of 1988, the total was 1308! . A paper based on this chapter has been sent to J. Am. Chem. Soc. for publication. 52 This section will attempt to provide a very brief introduction of the principle of 2D NMR. More detailed treatments of the theory and practice of the technique can be found in a number of texts and excellent reviews listed in the references^3-12]. In principle, in all two-dimensional experiments, four different time-intervals can be distinguished, (a) a preparation period, (b) an evolution period, (c) a mixing period and (d) a detection period, which are represented as in the following scheme: Preparation Evolution Mixing Detection ti h During the preparation period a nonequilibrium state of the system is created by applying suitable pulses. The preparation can be with or without decoupling in the heteronuclear case. It could be a pair of closely spaced 7t/2 pulses, etc. During the evolution period ti, the system is allowed to evolve under a suitably tailored Hamiltonian. This variable is important because whatever modulation takes place in this period is going to be "carried over" to further periods of evolution. During this period the spin system may freely evolve or may be subject to interrupted evolution under suitable pulses. In the third period of mixing, one can employ suitable pulses for effecting coherence transfer, or monitoring cross-relaxation or chemical exchange, etc. The fourth period, t2, which corresponds to acquisition of the final response of the observed spins, contains all of the experimental information. This period is usually free from any pulses of the observed spins, although decoupling of the hetero-spins can be applied during acquisition. 53 In a 2D NMR experiment, one acquires data as a function of the (fixed) detection period t2 for a progressively incremented period ti. By repeating the experiment for a large number of different ti times, keeping all other variables constant, a two-dimensional time-domain signal S(ti,t2) is obtained. One can get a 2D frequency-domain signal by the process of double Fourier transformation which can be described as in equation 3.1. FT over t2 FT over ti S(ti,t2) > S(ti,f2) > S(fi,f2) (3.1) In the corresponding two-dimensional frequency spectrum, the behavior of the spin system during both evolution and detection periods is shown simultaneously. Often the correlation between these two is a useful source of extra information. 3.1.2. Two-Dimensional Heteronuclear Solid-State Correlation NMR Spectroscopy The observation of high-resolution solid-state NMR spectra of abundant nuclei such as protons still represents a formidable challenge because of the strong dipolar interactions between them. Although sophisticated pulse sequences can largely remove homonuclear dipolar interactions, the residual linewidths remain significantly large in comparison to the dispersion of the chemical shift, which is relatively small for protons. The residual linewidths obtained with multiple-pulse sequences coupled with MAS are typically ~2 ppm t13"5]. When compared to the normal 10 ppm proton chemical shift range, it is apparent that only four or five nonequivalent signals at best are distinguishable in a given experiment. This fact severely constrains the types of samples amenable to analysis by these techniques. 54 There are a number of approaches involving two-dimensional (2D) techniques^ which can be employed to increase spectral resolution. For example, in proton solution NMR spectroscopy, the Jeener experiment and extensions of it have been used for this purpose t1"2*16]. Recently, 2D 1 3 C - ! H chemical shift correlation spectroscopy has been demonstrated to be very effective in increasing the resolution of proton spectra^17 .^ Although these experiments have been used primarily to study small molecules, they have been extended to the examination of a number of proteins^]. More recently, 2D heteronuclear correlation spectroscopy coupled with MAS techniques has been used in solid-state systems^19]. Ernst and coworkers t19^ first implemented the 2D heteronuclear * H - 1 3 C correlation technique in a solid state system using a single crystal of ferrocene as the sample. Zumbulyadist20] and Vegat21] have subsequently reported MAS !H- 2 9 Si heteronuclear solid state correlation NMR spectra of oc-Si:H and Zeolite Rho respectively. 3.1.3. Pulse Sequence for Two-Dimensional Correlation Spectroscopy The general Hamiltonian relevant to chemical shift correlation spectroscopy in solids is given in equation 3.2. Htotal = Hcs(I) + Hcs(S) + H J + Hd(I-I) + Hd(I-S) ( 3 - 2 ) Where H c s(i), H c s(s) , Hj , H^.j) and H^T-S) represent the proton chemical shift (I spins), the silicon chemical shift (S spins), the heteronuclear proton-silicon scalar (J) coupling, the homonuclear proton-proton dipolar interactions, and the heteronuclear proton silicon dipolar interactions respectively. Successful completion of a two-dimension NMR experiment requires manipulation of this Hamiltonian to eliminate one or several of the terms during each time period. This procedure is illustrated below with the general pulse sequence used in solid-state correlation spectroscopy. 55 In accordance with the general scheme of the two-dimensional spectroscopy, the experiment consists of four distinct time period (Figure 3.1). (P). The experiment begins with a preparation period.which in this case involves a 90° pulse for proton spins along the x' axis in the rotating frame with a 4 scan phase cycling scheme. (E). During the evolution period, which is represented as t^  in the sequence, the proton spins evolve without any decoupling under the influence of their chemical shifts. (M). In the mixing period, the I spin (1H) coherence is transfered to the S spins (in our case 2 9 S i ) by cross-polarization. The most substantial difference between the liquid and solid-state heteronuclear correlation experiments is the way in which the coherence transfer or mixing is performed during the mixing period. In the liquid-state heteronuclear correlation experiment, the coherence transfer is achieved through the scalar coupling (spin-spin interaction). That is, the magnetization is transfered between the neighboring (bonded) pairs of nuclei I and S. Hence the 2D correlation spectra provide information about scalar interactions. In the solid-state correlation experiment, the coherence transfer is carried out through cross-polarization, the mechanism of which is the dipole-dipole interaction. Therefore the corresponding 2D correlation spectra provide information about heteronuclear dipolar interactions. Since the dipolar interactions are strongly dependent on the distance between the two nuclei I and S (see equation 1.7), the spectra thus give information about the proximity of the "correlated" nuclei I and S. (D). In the final detection period, the S magnetization ( 2 9 Si in this case) is monitored under conditions of heteronuclear I spin decoupling to remove the 1 H - 2 9 S i dipolar interactions (if necessary). 56 Figure 3.1. Pulse sequence used for the 2D heteronuclear correlation experiments in the solid state. 57 P E H n — t i — -M D CP Decoupling Pi 2 9 S i P 2 CP -—h P3 \^AQ—^ p 4 3.1.4. Quadrature Detection in GJ± If positive and negative frequencies are not distinguished in the evolution period, all signals appear mirrored in the 2D spectrum about 0)j = 0. Since large offsets Ao)j must be avoided in spin-locking experiments, real and rnirror image signals will tend to overlap. In order to suppress the rnirror image in the coj dimension in the 2D experiment, the FID should be accumulated with quadrature detection in o&i which can be accomplished by proper phase cycling in analogy to solution-state NMR experiments [3,22] in all 2D experiments in this study, the phases of the I and S pulses have been cycled in four successive scans according to the scheme in Table 3.1. Table 3.1. Phase-cycle for suppression of mirror image signals Phase Scan 1 2 3 4 Pulse Pi X Y -X -Y P2 Y Y Y Y P3 X X X X P4 X -Y -X Y In summary, 2D lH-^Si heteronuclear solid-state correlation experiments can yield information about the distance between the "correlated" pair of nuclei, proton and silicon. Also because the superior dispersion and resolution of the silicon chemical shifts effectively separates overlapping proton resonances in the second dimension, the overall proton resolution can be improved over that of the conventional one-dimensional experiment. 58 3.2. Experimental Section 3.2.1. Sample Preparations a) . Deuterium Exchange of Protons in Hydroxyl Groups Deuterium exchange of the protons in hydroxyl groups in silica gel and methylsilica gels was performed by refluxing the samples with D2O of approximately 15 times the volume of the gel samples. Each sample was refluxed twice to make the exchange as complete as possible. All exchanges were done under dry N 2 . Deuterium samples were dried in vacuo at temperatures of 100-150°C, depending on the sample. The samples were packed into NMR spinners and sealed with Teflon tape in a glovebag purged with N 2 . b) . Preparation of Methylsilica Gel by the Anchoring Method 1). Gel 26. A Methylsilica gel (Gel 26) was synthesized by anchoring the methyl group onto the surface of silica gel using the procedure described by Fyfe and coworkers f23]. The Silica gel (Fisher S-157; ~ 5g ; predried in vacuo at ~150°C overnight) was refluxed with an excess of MTES in dry toluene for 12 hours.The mole ratio of Si0 2 to MTES was about 1 to 3. After filtration, the solid was Soxhlet-extracted in dry toluene for 12 hours and washed with acetone and water several times successively. The solid was then refluxed in H 2 0 for 12 hours and dried in vacuo at ~100°C for 12 hours. The reaction can be described approximately as follows: (3.3) 59 2). Gel 34. This anchored methylsilica gel was prepared by the same procedure as that for Gel 26. The only difference was that the silica gel used was not predried for the preparation of Gel 34. c). Preparation of Methylsilica Gel in a Deuterium Environment. A methylsilica gel (Gel 33) was synthesized by the same procedure used for the preparation of Gel 18 except that the preparation scale was reduced and concentrated HCl and methanol were replaced with the corresponding deuterated ones. The reaction can be described as: All steps of the preparation were performed under N 2 . Extensive precautions were taken to exclude moisture from the sample. 3.2.2. Instrumental Details One-dimensional *H and 2 9 S i spectra were obtained at 400 MHz and 79.5 MHz respectively on a Bruker MSL-400 Spectrometer under the same conditions as described earlier. Two-dimensional !H- 2 9 Si heteronuclear correlation experiments were also carried out on the Bruker MSL-400 Spectrometer using the Bruker MAS probe. Resonance frequencies (3.4) 60 of the two dimensions Fj and F 2 were set to 400 MHz and 79.5 MHz respectively. The pulse sequence described in Figure 3.1 was used. The optimum CP matching condition was found empirically for the different samples by adjustment of the 2 9 S i radiofrequency amplitude for maximum CP signal intensity. For each 2D experiment, a set of 64 FID's was obtained with increments of 16.7 \is in ti. Each FID was accumulated with 256 data points in t2. Both dimensions were zero-filled to 512 data points before Fourier transformation to increase resolution. The other conditions, such as the number of transients per FID and the contact time, varied from experiment to experiment as indicated in the figure captions. 3.3. Results and Discussion 3.3.1. Proton NMR Spectra of the Modified Silica Gels. The surface characteristics of silica gel and its derivatives play an important role in catalysis and chemical separations and the nature of the surface sites has been extensively studied, revealing structurally distinct surface groups including physically absorbed water and lone (isolated single) (1 ), geminal ( 2 ) and vicinal silanols ( 3 ) and siloxanes ( 4 ) (shown from left to right): O /0 N Si Si Si k\ S i Si (I) (2) (3) (4) Both lone and vicinal silanols are included in the category of single silanols i.e., one covalently bound hydroxyl group per silicon atom. If a silica gel is modified by a silane 61 coupling agent, for example, H3CSi(OEt)3, the surface characteristics of the product will be more diverse. It will contain other kinds of hydroxyl groups such as the lone silanol that shares a silicon atom with a methyl group (5) and terminal silanols ( 6 ) (i.e. the germinal hydroxyl group that is connected to a terminal silicon atom of the polymer network): I H Si Si / \ I (5) (6) As mentioned earlier, proton NMR experiments on solid systems, in general, are encumbered by the severe line-broadening effects of strong lH-lH magnetic dipolar interactions. However, if these effects are not too large, because of large ^H-lH internuclear distances and/or fast motional averaging, then moderately sharp *H NMR resonances can often be achieved by magic angle spinning, which averages both the homonuclear dipolar interactions among protons and the chemical shift anisotropy. In this case, the spinning rate for MAS need be only a few kilohertzt24]. The results for proton NMR experiments on the modified silica gels demonstrate that the homonuclear dipolar interactions are not very strong in this system. Figure 3.2 shows proton NMR spectra of Gel 18 (a methylsilica gel prepared by copolymerization). The spectrum of sample Gel 18 in equilibrium with the atmosphere (Figure 3.2A.) shows that there are mainly three peaks in the spectrum, a somewhat broad peak at 0.45 ppm with a large number of spinning sidebands, and two relatively narrow peaks at 3.2 ppm and 4.8 ppm (all with respect to TMS). To assign these peaks, a series of treatments were carried out on this sample. Figure 3.2B shows the spectrum of the sample after heating in vacuo at ~100°C for 24 hours. The peak at 4.8 ppm has been almost completely removed compared with Figure 3.2A. As there should be no physisorbed water present on the surface after this 62 Figure 3.2. 400 MHz lH MAS NMR spectra of Gel 18 obtained with a single-pulse-acquisition pulse sequence and a 4.0 kHz sample spinning rate, a 2.0 s repetition time and 40 scans. A. Original untreated sample spectrum. B. Heated sample. C. D 2 0 exchanged sample. D. Shortly after exposure to the normal atmosphere. 63 treatment in vacuole, the 4.8 ppm resonance is assigned to protons in physically absorbed water. Figure 3.2C is the spectrum of a sample that was refluxed in D 2 0 twice and then heated in vacuo at ~100°C under which conditions most protons of the hydroxyl groups would be expected to have been deuterium-exchanged. This spectrum shows only one broad peak at 0.45 ppm with a small shoulder. Thus this peak can be assigned unambiguously to the protons in the methyl groups. From the combination of Figures 3.2A, B and C, it is clear that the resonance at 3.2 ppm can be assigned to protons of silanol moieties. Figure 3.2D is the spectrum obtained shortly after exposing the deuterium exchanged and heated sample to the atmosphere. The physisorbed water builds up very fast on the surface further confirming the assignment of the resonance at 4.8 ppm. This assignment of the proton resonance peaks of the silica gel is consistent with that of the 2 9 S i CP/MAS spectrum of this sample (Figure 3.3), which can be assigned to silicons connected with a methyl group, a hydroxyl group and two siloxanes (-53 ppm), silicons connected with a methyl group and three siloxanes (-64 ppm), silicon with two hydroxyl groups and two siloxanes (-91 ppm), silicons connected with one hydroxyl group and three siloxanes (-100 ppm), and silicons connected with four siloxanes (-109.7 ppm) respectively. There are thus mainly three types of protons present in this system, protons in methyl groups, protons in hydroxyl groups and protons in the physically absorbed water. Spinning sidebands accompanying the central proton signals reveal that the spinning rate of MAS is not large enough to completely remove the dipolar interactions and the chemical shift anisotropics. The broad proton resonance of the methyl group is due mainly to the residual homonuclear dipolar interactions between protons within the methyl group. Although the resolution of the spectra in Figure 3.2 is not good enough to completely resolve the different proton signals present, they reveal the three potential proton sources for the cross-polarization which is the basis for the 2D heteronuclear !H- 2 9 Si solid-state correlation experiments. 64 Figure 3.3. 2 9 S i CP/MAS NMR spectrum of Gel 18 obtained at 79.5 MHz, 600 scans, a 4.3 kHz spinning rate, a 22.0 ms contact time and a 3.0 s repetition time. 65 3.3.2. Characterization of Different Types of Functionalized Silica Gels with 2D  Correlation Experiments a). General Considerations In the characterization of the functionalized silica gels, two questions must be answered: firstly, it must be established that the organic functionality is chemically intact and that it is present in the correct proportion in the mixture; secondly, the distribution of the functionality throughout the matrix should be described. That is, it must be established that there is not an isolated 'domain structure' of silica gel and polymethylsiloxane formed. The first problem is relatively straightforward and can be solved using one-dimensional high-resolution solid-state 1 3 C and 2 9 S i NMR experiments as we have shown in Chapter II. A much more difficult task is to answer the second question, that is, to describe the nature of the incorporation of the functionalized silicons into the gel matrix. This is illustrated in Figure 3.4, which shows the 2 9 S i CP/MAS NMR spectra of three materials combining the silica gels and polysiloxane: (A) a methylsilane-functionalized silica prepared by reaction of the methylsilane with silica gel by the anchoring method (here the methyl groups should be on the surface), (B) a methylsilica gel prepared by the direct incorporation of the methylsilane during gel synthesis by the copolymerization method (here there should be a more random distribution of methyl groups throughout the matrix), (C) a mechanical mixture of unfunctionalized silica gel and polymethylsiloxane, where there can be no interactions between the methyl groups and the silica gel. As can be seen from the figure, all three spectra are identical in terms of the chemical shift values of the resonances (the peak intensities vary because of different proportions of the components and functionalities). To some degree this is due to limited resolution, as the signals are relatively broad because of the amorphous nature of the materials, but it is also to be expected that the differences in 66 Figure 3.4. 2 9 S i CP/MAS NMR spectra of different types of modified silica gels. A) . Methylsilica gel (Gel 26) prepared by the anchoring method. B) . Methylsilica gel (Gel 3) prepared by the copolymerization method. C) . A mechanical mixture of silica gel and polymethylsiloxane. 67 / local silicon environments will be small. The critical difference is that between the local structures (7_) and (8 ). In these, the effect on the local environment of the silicon marked (*) is only whether or not the adjacent silicon is substituted by a methyl group or an oxygen functionality. ' I I 0 O CH, O 1 I I I O - S i - O - S i - O - - O - S i - O - S i - 0 JL i I I ? o o o 1 I I I (Z) (8) One possible approach to this problem is to take advantage of the very strong distance dependence of the cross-polarization process (1/r3). Thus, cross-polarization from the methyl group to the silicon nuclei would only be efficient if they were in close proximity, and the observation of an Si*(0-)4 resonance could be taken as evidence for the presence of structure ( 8 ). A complicating factor in these experiments is that the hydroxyl groups can also act as a source for polarization transfer. In a one-dimensional 2 9 S i CP/MAS spectrum of the methyl functionalized silica gel (e.g., Figure 3.4B), one cannot tell whether the observation of the Si*(0-)4 resonance comes from the magnetization transfer from protons in the methyl groups or from the protons of hydroxyl groups. Attempts were made to eliminate the protons of hydroxyl groups by repeated refluxing of the samples with D 2 0 as described in the experimental section. Although the correct trends were observed (i.e., greatly reduced polarization transfer to backbone silicons after D 2 0 washing), removing all of the hydroxyl protons from functionalized silica gel is very difficult because of the strong affinity of the gel for water and other unknown reasons. (We will discuss this subject in more detail later in this section). Thus, the use of 2D heteronuclear correlation NMR spectroscopy, a two-dimensional NMR experiment that can couple chemical shift resolution with the cross-polarization technique becomes mandatory. 68 At first sight, it might seem that this two-dimensional experiment would be limited by the same factors as the one-dimensional CP experiments. However, because the proton spins are relatively isolated, MAS alone gives enough resolution to clearly distinguish OH and CH3 signals, thus making it possible to identify the sources of the observed polarization transfers. In addition, because there are two related frequency scales in the experiment, the chemical-shift resolution is better than in the simple one-dimensional experiments. In this section, we will demonstrate how the systematic application of these techniques yields information on the distribution of the functionalized silicons throughout the gel matrices formed by using the different preparations. b). Two-dimensional ^ - ^ S i correlation spectra of silica gel At the beginning of the 2D correlation experiments, a simple system of known structure and composition was investigated to establish the reliability of the technique in the present context. Figure 3.5 shows the results of the !H- 2 9 Si correlation experiment carried out on an unfunctionalized silica gel stored in equilibrium with normal atmosphere using the experimental conditions detailed in the figure caption. The figure is the contour plot of the 2D correlation spectrum with projections of the data onto F^ (*H chemical shift) and F 2 ( 2 9 Si chemical shift) axes which can be used to establish the connectivities. Figure 3.6 displays comparisons of the proton ID MAS spectrum with the corresponding 2D projection (Figure 3.6A), and silicon ID CP/MAS spectrum with the corresponding 2D projection (Figure 3.6B). The results show that the 2 9 Si 2D projection is essentially identical with the ID spectrum. However, the *H 2D projection is different from the ID spectrum. There are two peaks in the ID proton spectrum which arise from water (large peak at 4.2 ppm) and hydroxyl groups (small peak at 2.1 ppm). However, in the 2D projection, there is only one resonance peak at 2.1 ppm, which means that only the structural hydroxyl protons contribute to the cross-polarization magnetization transfer to the silicon nuclei. The water 69 Figure 3.5. Contour plot of the 2D ^H-^Si correlation experiment on unfunctionalized silica gel (Gel 1) obtained with 22.0 ms contact time, 3.0 s repetition time, 4.0 kHz sample spin rate. The vertical axis represents the proton chemical-shift scale; the horizontal axis the 2 9 Si chemical-shift scale. The spectrum was obtained from 64 FID's and 80 scans were accumulated for each FID. 70 (HO)Si(OSi% (HOfeSXDSi-fe j 1 f ° S i s ) 4 ts • to i IS) i IS) • CNI ' Csl I J3 CO -a 1 6 ' CD ' OO -i 1 1 r -i 1 1 r -i 1 1 r 0 -50 ' -100 F 2 , 2 9 S i Chemical shift (ppm) Figure 3.6. Comparison of ID spectra and the corresponding 2D projections as indicated from Figure 3.5. A) . spectra. B) . 2 9 S i spectra. 71 molecules on the surface of silica gel are very mobilet25] and their residence times in specific framework positions are thus very shortf2!], greatly reducing the efficiency of dipolar couplings. The cross peak in Figure 3.5 shows that there are correlations between protons and three different silicon environments in the gel. Figure 3.7 shows the NMR spectra of a silica gel heated in vacuo at ~150° for 24 hours. Figure 3.7A displays proton ID single-pulse-acquisition MAS spectrum and the 2D projection. Comparing this to the proton ID MAS spectrum of the unheated gel sample (figure 3.6A), the proton resonance peak of the physically absorbed water has disappeared. This indicates that the physisorbed water has been successfully removed by the evacuation. The resonance at 2.0 ppm can be assigned to the protons in isolated silanols and the shoulder peak at ~3.0 ppm can be assigned to protons in germinal silanols, according to the data obtained by Bronnimann et al. using the CRAMPS technique f 2 5l. Figure 3.7B shows the contour plot of the 2D 1 H- 2 9 Si correlation spectrum of the gel. In addition to the cross peaks between the two dimensions appearing in the contour plot, which show strong coupling between protons in hydroxyl groups and all three chemically different silicon nuclei, there are strong spinning sidebands along the Fj axis in the contour plot. This is a clear indication that there are dipolar interactions between hydroxyl protons. Figure 3.8 is the contour plot of the 2D correlation spectra of the silica gel that was refluxed twice in D2O and dried in vacuo at ~150°C for 24 hours. As can be seen from the figure, there is a very marked decrease in the intensity of the heteronuclear connectivity peak, but there is still a residual interaction arising from the trace amounts of hydroxyl protons indicated in the projection onto Fi . It is this situation that leads to possible ambiguities in the one-dimensional experiments. c). 2D lU-29Si Correlation Spectra of Gel Mixtures Figure 3.9 shows experiments carried out on a physical mixture of unfunctionalized 72 Figure 3.7. Spectra of the heated silica gel (Gel 1) A) . *H ID MAS spectrum and 2D projection. B) . Contour plot of the 2D lH-29Si correlation spectrum obtained with a 22.0 ms contact time, 3.0 s repetition time, 4.0 kHz sample spinning rate. The spectrum was obtained from 64 FID's and 280 scans were accumulated for each FID. 73 HO-Figure 3.8. Contour plot of the 2D !H- 2 9Si correlation experiment on silica gel after deuterium exchange with D2O. A 22.0 ms contact time, 3.0 s repetition time and 4.0 kHz sample spinning rate were used. The spectrum was obtained from 64 FID's and 200 scans were accumulated for each FID. 74 Figure 3.9 Contour plot of the 2D !H- 2 9Si correlation experiment on the mechanical mixture of silica gel and polymethylsiloxane, obtained with a 22.0 ms contact time, 3.0 s repetition time and a 4.0 kHz sample spinning rate. The spectrum was obtained from 64 FID's.and 120 scans were accumulated for each FID. 75 silica gel and polymethylsiloxane. In the 2 9 Si spectrum, the low field signals are all due to silicons with attached methyls in the polymethylsiloxane and the high field signals to silicons with no attached groups in the silica gel. In this case, there should be no connectivities between the methyl protons and any of the silicon nuclei in the gel. From these experimental results, we can see that there are clear connectivities to both groups of silicons. In the case of the lower-field signal of the methyl substituted silicons in polymethylsiloxane, the connectivities are mainly to the methyl protons and the very substantial intensity from spinning sidebands is consistent with this. In the case of the three high-field signals due to the silica gel, the connectivities must be to hydroxyl protons, as is borne out by the very small sideband pattern observed. Most importantly, however, the sources of polarization can be unambiguously identified from the chemical shifts. Thus, as indicated in the figure, the two different sets of silicon nuclei are polarized from two distinct proton sources: methyl groups for the polymethylsiloxane silicons, and hydroxyl groups for the silica gel silicons. The projection on Fj identifies their proton chemical shifts as 5 = 0.45 ppm and 8 = 1.9 ppm respectively. Consistent with this interpretation, D20 exchange of the mixture completely removes the connectivities to the silica gel, while a very strong connectivity to the polymethylsiloxane with its associated sideband pattern remains (Figure 3.10 ). The 2D correlation experiment results are consistent with the corresponding 2 9 S i ID CP/MAS spectra. Figure 3.11 shows the ID CP/MAS 2 9 Si spectra obtained before and after D 2 0 exchange of the mixture. Figure 3.11 A is the spectrum of the mixture without D2O exchange. Resonances on the high-field region of the spectrum due to the gel silicons are observed. We cannot tell at this stage whether the polarization transfer to the Si*(0-)4 silicons is from methyl protons in the polymethylsiloxane or the hydroxyl protons in silica gel. However, if we wash out protons in the hydroxyl groups by D2O exchange, the 2 9 S i CP/MAS spectrum (Figure 3.1 IB) shows no signal in the high-field site, which indicates there is no polarization transfer to the framework silicon Si* ( 0 ) 4 from methyl protons in 76 Figure 3.10 Contour plot of the 2D !H- 2 9Si correlation experiment on the mechanical mixture of silica gel and polymethylsiloxane after D 2 0 exchange. A 22.0 ms contact time, 3.0 s repetition time and 4.0 kHz sample spin rate were used. The spectrum was obtained from 64 FID's.and 200 scans were accumulated for each FID. 77 Figure 3.11. 2 9 Si ID CP/MAS spectra of mixture of silica gel and polymethylsiloxane obtained at 79.5 MHz, with 320 scans, a 22.0 ms contact time and a 5.0 s repetition time. The sample was spun at 3.2 kHz. A) . Original sample. B) . D 2 0 exchanged sample. 78 the polymethylsiloxane. Thus it is clear that there is no correlation between the methyl protons and the silicon Si*(0-)4. In other words, methyl groups in the polymethylsiloxane are far away from the framework silicon Si*(0-)4 in the silica gel as expected. Its microstructure can thus be qualitatively described as in Figure 3.12. It should be noted, however, that this situation is a relatively simple one in that the hydrophobic/hydrophilic natures of the two materials are quite different and the clear distinction observed here cannot be assumed for systems where there is a more intimate mixture of the two components. From the self-consistency of results of the 2D experiments on this mixture, it was considered that the technique provided a reliable probe of the microstructure of these systems. d). 2D Correlation Spectra of Methylsilica Gels dl). Methylsilica gels prepared by the copolymerization method. If we copolymerize two components CH3Si(OEt)3 and Si(OEt)4 and if the reactivity ratios of the components are very close to 1, then the material made by the copolymerization would contain the two components in a random distribution although affected by the reactivity ratios. Figures 3.13 and 3.14 show the results of 2D correlation experiments on Gel 18 (a methylsilica gel prepared by copolymerization of Si(OEt)4 and CH3Si(OEt)3 in 1:1 mole ratio). Microanalysis indicates that it contains 11.3% C. Figure 3.13 is the contour plot of the 2D correlation spectra of Gel 18 without D2O exchange. From the figure, it can be seen that there is an intense correlation of the silicons Si*(0-)4 to the methyl protons, as well as to the hydroxyl protons as indicated in the figure. Although the gel contains much physically absorbed water on its surface, as shown by the ID spectrum, there is no such peak with lH chemical shift at ~4.0 ppm in the 2D projection, nor in the corresponding 79 Figure 3.12. Schematic representation of the microstructure of the mixture of silica gel and polymethylsiloxane. 80 Polymethylsiloxane i l l 1 <p p C H 3 O C H 3 O H ^ O c H , H 3 C - Si - O - S i - O - S i -O-Si - O - s V O - S i - O - S i - O ( C H 3 0 C H 3 OH C H 3 O C H 3 C H 3 9 - O - S i - O - S i - C H 3 H ^ - S i - O - S i - O-Si - O - S i - C H 3 t b OH C H 3 ? H O O O ? - O - S i - O - S i - O - S i - O - S i - O - S i - O - S i - O - S i - O - S i - O - S i - C H 3 C H 3 C H 3 C H 3 OH C H 3 C H 3 C H 3 C H 3 OH OH OH OH OH OH OH OH OH OH OH I I I \ / I I I I I - O - S i - O - S i - O - S i - O - S i O - S i - O - S i - O - S i - O - S i - O - S i -- O - S i - O - S i - O - S i - O - S i - O - S i ^ / S i - O - S i - O - S i - O - S i -i b i If b H O i i i 4 I I I I I I I I I - O - S i - O - S i - O - S i - O - S i - O - S i - O - S i - O - S i - O - S i - O - S i -I I I I I I A JL 1 O O O O O O O O O I I I I I • I I I I Silica gel Figure 3.13. Contour plot of the 2D correlation spectrum of Gel 18 which was synthesized by copolymerization of TEOS and MTES. A 22.0 ms contact time, 3.0 s repetition time and 4.0 kHz sample spin rate were used. The spectrum was obtained from 64 FID's and 200 scans were accumulated for each FTD. 81 - I — . — i — . — i — i — i — i — i — i — i — i — i — i — r ~ 80 60 40 20 0 -20 -40 -60 lH Chemical shift (ppm) cross peak in the 2D contour plot. We therefore may conclude that the physically absorbed water on the surface of the gel does not contribute to the cross-polarization of the 2 9 Si spins for the same reasons as in the case of silica gel. In order to enhance the heteronuclear ^H-^Si correlation from the methyl group, Gel 18 was refluxed twice in D 2 0 and dried in vacuo at ~100°C for 24 hours. Figure 3.14 is the result of the 2D correlation experiment on this sample which shows that in addition to the connectivity of methyl protons with the silicons to which they are directly bonded, there is a clear heteronuclear connectivity between the Si*(0-)4 silicons and the methyl protons, as well as a smaller connectivity to residual OH protons. As previously, the two types of protons may be distinguished by their *H chemical shift in the Fj axis. In addition, there are a large number of spinning sidebands in the connectivities consistent as previously with the polarization source being the methyl groups. We may thus conclude that the methyl groups are close to the framework silicons Si*(0-)4-In order to make the correlation of the methyl protons to Si*(0-)4 silicons quite definite, we have prepared a methylsilica gel (Gel 3) by copolymerization of Si(OEt)4 and CH3Si(OEt)3 in 3:1 mole ratio using the same procedure as for Gel 18. The microanalysis indicates that it contains 4.9% C, compared to 11.3% for Gel 18. Figure 3.15 shows the contour plot of the 2D correlation experiment of Gel 3, which was D 2 0 exchanged twice to get rid of the hydroxyl protons. From the figure, we can see that despite the small quantity of the methyl groups in the gel matrix, there is a clear heteronuclear connectivity between the methyl protons and the framework Si*(0-)4 silicons. As in the case of Gel 18, there is a series of strong sidebands connected with the Si*(0-)4 silicon resonance which indicates clearly that the source of polarization transfer is the methyl protons. Thus in the cases of both Gel 18 and Gel 3, there is clear evidence for the incorporation of the methyl substituted silicons throughout the matrix; i.e., the absence of clear domain structures and the presence 82 Figure 3.14. Contour plot of the 2D J H- 2 9 Si correlation spectrum of the D 2 0 exchanged Gel 18. A 22.0 ms contact time, 3.0 s repetition time and 4.0 kHz sample spin rate were used. The spectrum was obtained from 64 FID's and 200 scans were accumulated for each FID. a 83 Figure 3.15. Contour plot of the 2D 1 H- 2 9 Si correlation spectrum of the D 2 0 exchanged Gel 3 made by copolymerization of TEOS and MTES. A 22.0 ms contact time, 3.0 s repetition time and 4.0 kHz sample spin rate were used..The spectrum was obtained from 64 FLD's and 320 scans were accumulated for each FID. 84 i H Chemical shift (ppm) of substantial numbers of local structural units of type (8 ). These systems thus represent a unique type of functionalized silica gel. The micro structure of these systems can be qualitatively described as in Figure 3.16. It can be noted from both figures 3.14 and 3.15 that though the samples have been refluxed in D20 twice to exchange the hydroxyl protons, there is still a small number of hydroxyl protons left on the surface of the gels. It is possible that the hydroxyl protons remaining after D2O exchange in these systems may be due to those hydroxyl groups which are "capped" or hindered by the hydrophobic methyl groups and are thus not easily accessible to D2O. To test this we prepared a methylsilica gel (Gel 33) by copolymerization of TEOS and MTES in C D 3 O D as solvent and using DC1 /D 2 0 as catalyst. Figure 3.17 shows the results of the 2D correlation experiment on this sample. From the figure, it can be seen that the heteronuclear correlation arising from hydroxyl groups is completely removed. d2). Methylsilica gels prepared by the anchoring method. Gel 26 was prepared as described in Section 3.2. Lb. 1 by anchoring the methyl substituted silane onto the surface of silica gel through the reaction of 5g silica gel (predried in vacuo at ~150°C for 12 hours) with 6.1 ml of CH3Si(OEt)3 in dry toluene. Under these anhydrous conditions, the reactions of individual organosilanes with the surface are optimized and self reaction minimized. From the ID 2 9 Si CP/MAS spectrum of the modified silica gel (Figure 3.4A), it can be seen that there are some hydroxyl groups left on the surface of the silica gel after the condensation reaction between the trifunctional silane and the hydroxyl groups of silica gel, since there are resonance peaks at -101.0 ppm and -91.0 ppm which arise from silicons connected to single hydroxyl and geminal hydroxyl groups respectively. Figure 3.18 shows the results of a 2D correlation experiment on gel 26. 85 Figure 3.16. Schematic representation of the microstructure of the Gels prepared by the copolymerization method. 86 If If CH3 <j> CH3 9 9 C H 3 H 3 C- Si-O-Si-O-Si -O-Si -O-^i -O-Si -O-Si -O- ( ? ^ C * H O 0 H ? ? ? _ 0 - S i - 0 - S i - O H OH/^ 3 1 HjC-Si-O-Si-o-Si-O-Si-I I Si-O-Si I 1 I I ? ? <T ?\ ? ? ? ? H 3 C-Si -0 - Si-O-Si-O - Si -O-S i -O-S i - O - Si-O-Si-O- Si-0 O O ^ P H O O HC* _ O 6 1 1 I /CH 3 1 1 /CH 3 , 1 -O-Si -O-Si -O- Si-O-Si-O-Si -O-Si -O-Si - O - Si-O-Si-I / 1 I / 1 I 1 1 O H3C 0 9 HO 6 9 0 9 - O-Si-O-Si-O- Si-O-Si-O - Si-O -Si-O-Si -O-S i -O-Si -b 6 ( p o o i i p o I T 1 1 I I I 1 Figure 3.17. Contour plot of the 2D lH-29Si correlation experiment on Gel 33 synthesized by copolymerization of TEOS and MTES in a deuterium environment. A 22.0 ms contact time, 3.0 s repetition time and 4.0 kHz sample spin rate were used. The spectrum was obtained from 64 FID's and 200 scans were accumulated for each FID. 87 Figure 3.18. Contour plot of the 2D lH-29Si correlation experiment on Gel 26 synthesized by anchoring method. A 22.0 ms contact time, 3.0 s repetition time and 3.4 kHz sample spin rate were used..The spectrum was obtained from 64 FID's.and 240 scans were accumulated for each FID. 88 The contour plot in the figure indicates that there is a clear heteronuclear correlation between framework Si*(0-)4 silicons and hydroxyl protons, as well as a smaller connectivity to the methyl groups. As previously, the two types of protons can be distinguished by their chemical shifts in Fj. From this result, we can see that only a relatively small number of methyl groups has been anchored onto the surface of the silica gel. These methyl groups may be distributed in low concentrations on the surface of silica gel and by rotation can contact with the surrounding framework Si*(0-)4 silicons. They can thus correlate with these silicons. In this case, if we remove the hydroxyl protons by refluxing the sample in D 2 0 , we should see only the heteronuclear correlation between framework silicons and methyl protons if the deuterium exchange is complete. Figure 3.19 is the 2D correlation spectra for the D 2 0 washed Gel 26. The existence of only the methyl proton peak on the Fj (*H) projection indicates that the D 2 0 exchange is quite complete. Comparing the 2D correlation experiment for D 2 0 exchanged Gel 3, which was prepared by the copolymerization method, with that of the D 2 0 exchanged Gel 26, there are no great differences in terms of the correlations and spinning sideband patterns, but less hydroxyl protons appear in the latter case. This is consistent with our previous results, showing that the residual hydroxyl protons in Gel 3 after D 2 0 exchange are due to those hydroxyl groups capped or hindered by the hydrophobic methyl groups and thus are not accessible to D 2 0. However, if we anchor more methyl-functionalized silane onto the surface of silica gel, or if water is present, two distinct phases between the silica gel and the polymethylsiloxane will result from the self polymerization of the trifunctional silane on the surface of silica gel, giving a multilayer coating on the surface. With this in mind, we prepared a methylsilica gel (Gel 34) by reacting 1.2g silica gel (in equilibrium with the atmosphere and thus containing considerable physisorbed water) with 2.0 ml of CH3Si(OEt)3. Microanalysis indicates that this sample contains 9.86% C. Figure 3.20 is the contour plot of the 2D correlation experiment on this sample and shows only a very small heteronuclear connectivity between 89 Figure 3.19. Contour plot of the 2D !H- 2 9Si correlation experiment on D 2 0 exchanged Gel 26. A 22.0 ms contact time, 3.0 s repetition time and 4.0 kHz sample spin rate were used. The spectrum was obtained from 64 FID's and 280 scans were accumulated for each FID. 90 T — I 1 1 1 1 1 1 J 1 1 1 1 1 1 1 1 1 J-0 -50 -100 2*Si Chemical shift (ppm) Figure 3.20. Contour plot of the 2D ^ - ^ S i correlation experiment on Gel 34. A 22.0 ms contact time, 3.0 s repetition time and 4.0 kHz sample spin rate were used..The spectrum was obtained from 64 FID's.and 200 scans were accumulated for each FID. 91 the Si*(0-)4 groups and methyl protons and between Si*(0-)4 groups and hydroxyl protons. These results indicate that: 1) most of the hydroxyl groups on the surface of silica gel have been replaced by the trifunctional silane coupling agent; 2) the material prepared by the anchoring method contains two domain phases of the silica gel and polymethylsiloxane with the former covered by the latter. Thus the microstructure of the gel 34 can be qualitatively described as in Figure 3.21 according to the 2D correlation experiment results. 3.4. Conclusions Two-dimensional heteronuclear iH-^Si correlation NMR spectroscopy has been used to characterize three different types of modified silica gels: methylsilica gel prepared by the anchoring method, methylsilica gel prepared by the copolymerization method and a mechanical mixture of silica gel and polymethylsiloxane. This technique has successfully demonstrated the structural differences among three materials which cannot be easily detected by other analytical techniques. Our results show that methylsilica gel prepared by the anchoring method consists of two distinct phases when the functional group-loading is high, i.e,.when polymethylsiloxane is built up "horizontally" on the surface of silica gel and its characteristics are very similar to that of mechanical mixture of silica gel and polymethylsiloxane. The methylsilica gel prepared by the copolymerization method consists of two components with a more "random" distribution, i.e., there is no domain structure between silica gel and polymethylsiloxane. These experiments also demonstrated that methyl protons and hydroxyl protons in the functionalized silica gels are sources of cross-polarization transfer to framework Si(0-)4 silicons, but the physically absorbed water does not contribute to this process. 92 Figure 3.21. Schematic representation of the microstructure of Gel 34. 93 i i i A 0 9 CH 3 O CH 3 qiJ,CH39cH3 H 3 C - Si-O-Si-O-Si - O - S i - O - ^ i - O - S i - O - S i - O | >0-Si-0-Si-CH 3qi7H 3 I CHsHjC-Si-O-Si-O-Si-O-SiH . . . . . . . . . . . . . o . . . . . . . ^ O . . . . . V . . . - O - S i - O - Si-O-Si-O -S i - O - S i - O - S i - O - S i - O - S i - O - S i -I I I I I I I I 1 0 0 6 0 0 0 0 0 6 1 I I I I I I I I _0_Si-0-Si-0- S i - O - S i - O - S i - O - Si-O-Si - O - Si-O-Si-i 1 1 1 k k L 1 1 O O O O 6 6 O n O I 1 1 I I I I T I - O - S i - O - S i - O - Si-O-Si-O - S i - O -S i -O-S i - O - S i - O - S i -I I 1 1 1 1 A 1 1 6 0 0 0 0 0 6 0 0 1 1 l l 1 1 1 1 43A The functionalized silica gels prepared by copolymerization of two components, e.g., TEOS and MTES, represent a unique type of functionalized silica gel. Their structures may form the starting point for the generation of new materials, especially when high loadings of functionalized silanes are desirable. It should be noted that the general problem encountered in the present work of distinguishing between a domain structure and a true blend or mixture is a common one in many areas of materials research and the two-dimensional NMR techniques used in the present work may be of more general applicability. 94 3.5. REFERENCES 3.1) . J. Jeener, Ampere International Summer School, Basko Polje, Yugoslavia, Sept, 1971. 3.2) . W.P. Aue, E. Bartholdi and R.R. Ernst, J. Chem. Phys., 64, (1976) 2229. 3.3) . A. Bax, "Two-Dimensional Nuclear Magnetic Resonance in Liquids", Delft University Press, (1982). 3.4) . A.E. Derome, "Modem NMR Techniques for Chemistry Research", Pergamon Press, Toronto, (1987). 3.5) . R.R. Ernst, G. Bodenhansen and A. Wokaun, "Principles of Nuclear Magnetic Resonance in One and Two Dimensions", Clarendon Press, New York (1987). 3.6) . W.S. Brey (Ed), "Pulse Methods in ID and 2D Liquid NMR", Academic Press, San Diego, (1988). 3.7) . J. Schraml and J.M. Bellama, "Two-Dimensional NMR Spectroscopy", Wiley, New York, (1988). 3.8) . G.E. Martin and A.S. Zektzer, "Two-Dimensional NMR Methods for Establishing Molecular Connectivity: a Chemist's Guide to Experiment Selection, Performance and Interpretation", VCH, New York, (1988). 3.9) . N. Chandrakumar and S. Subramanian, "Modem Techniques in High-Resolution FT-NMR", Spring-Verlag, New York, (1987). 3.10) . H. Kessler, M. Gerhrke and C. Griesinger, Angew. Chem. Int. Ed. Engl., 27, (1988) 490. 3.11) . B. Blumich and H.W. Spiess, Angew. Chem. Int. Ed. Engl., 27, (1988) 1655. 3.12) . O.W. S<|>rensen, G.W. Eich, M.H. Leritt and R.R. Ernst, Progress in NMR Spectroscopy, 16, (1983) 163. 3.13) . R.E. Talor, R.G. Pembleton, L.M. Ryan and B.C. Gerstein, J. Chem. Phys., 71, (1979) 4541. 3.14) . L.M. Ryan, R.E. Talor, A.J. Paff and B.C. Gerstein, J. Chem. Phys., 72, (1980) 508. 3.15) . G. Scheler, U. Hanbenreisser and H. Rosenberger, J. Mag. Reson., 44, (1981) 134. 3.16) . K. Nagayama, A. Kumar, K. Wuthrich and R.R. Ernst, J. Mag. Reson., 40, (1980) 321. 3.17) . A.A. Mandsley, L. Muller and R.R. Ernst, J. Mag. Reson., 28, (1977) 463. 3.18) . T-M. Chan and J.L. Markley, J. Am. Chem. Soc., 104, (1982) 4010. 3.19) . P. Caravatti, G. Bodenhausen and R.R. Ernst, Chem. Phys. Lett., 89, (1982) 363. 3.20) . N. Zumbulyadis, Phys. Rev. B. 33, (1986) 6495. 3.21) . A. J. Vega, J. Am. Chem. Soc, HO, (1988) 1049. 3.22) . A. Bax and C A . Morris, J. Mag. Reson., 42, (1981) 501. 3.23) . C A . Fyfe, G C Gobbi and G E. Kennedy, J. Phys. Chem., 89, (1985) 278. 3.24) . R. Echman, J. Chem. Phys., 76, (1982) 2767. 3.25) . C E . Bronnimann, R.C. Zeigler and G. E. Maciel, J. Am. Chem. Soc, 110. (1988) 2030. 96 Chapter nv A PRELIMINARY INVESTIGATION OF THE THERMAL  STABILITIES OF THE METHYLSILICA GELS BY SOLID-STATE NMR  AND DIFFERENTIAL SCANNING CALORIMETRY 4.1. Introduction As we have demonstrated in the previous chapter, organofunctionalized silica gels from different preparative procedures can display different chemical microstructures. Apart from the chemical structure, the physical characteristics of such silica gels also play an important, and sometimes crucial role in selection of an appropriate material for a specific purpose. For example, when selecting a stationary phase for HPLC analysis, one should consider the chemical properties of a given functionalized silica gel as well as its pore structure and its resistance to oxidationt1!. In this chapter, we will investigate the thermal stabilities of the methylsilica gels made by the two preparative methods, using CP/MAS NMR spectroscopy and differential scanning calorimetry (DSC) techniques. 4.2. Experimental 4.2.1. Thermal Treatments Two types of furnaces were employed for thermal treatment of samples. A laboratory box furnace was used to heat the functionalized silica gels in a normal atmospheric environment. A quartz tube furnace, illustrated schematically in figure 4.1, was used to heat the gels under a N2 environment. The purity of the N2 used was 99.9%. Before heating, N2 97 gas was flowed through the quartz tube loaded with the functionalized silica gel for 10 minutes to expel air. 4.2.2. Solid-State NMR 2 9 S i and 1 3 C CP/MAS NMR spectra were obtained using a Bruker MSL-400 spectrometer at 79.5 MHz and 100.6 MHz respectively. All samples were packed into identical spinners, to ensure that the sample weights were approximately same. All experiments were at room temperature. Detailed experimental conditions are given in the figure captions. 4.2.3. Differential Scanning Calorimetry Differential scanning calorimetry was carried out on a Mettler TA 300 Thermal Analysis System which consists of a TC10 TA processor, a control and evaluation unit and a plotter. Heat flow to the sample is measured under thermally controlled conditions. Samples of approximately 20 mg of material were sealed in an aluminium pan. An empty aluminium pan was used as the reference sample. A heat flow rate of 5°C/min. was used, and two kinds of furnace atmospheres were employed, air and N 2 . In the case of N 2 gas, a flow rate of 0.5 ml/min. was used. 98 Figure 4.1. Schematic representation of the apparatus used for heating the functionalized silica gels under N 2 protection. 99 Heat element (Vertical tube v Furnace) Sample — Quartz Disc •o -o 0 —4-— Quartz tube 'Densestone' I o o oo oo o c O O O O O O O O O O O O O O C O O O O O O O |0 O O O O 0 0 0} O O O O O O O O O O O O O O O O O O O O O O p o o o o o o q O O O O O O O O O O O O O O C JXQ-Q-0 O O O O O O O O O O O O O O O O O Glass wool N 2gas 4.3. Results and Discussion 4.3.1. Thermal Behavior of Methylsilica Gel Made by Copolymerization Figure 2 shows the stacked 2 9 S i CP/MAS NMR spectra of Gel 3 which was heated at the different temperatures indicated for four hours. The gel was made by the copolymerization of TEOS and MTES in a mole ratio of 4:1 as described in Section 2.1.3. The spectra were taken after the gel had been heated at the temperatures indicated and then cooled to room temperature. The spectra are plotted in the absolute intensity mode with the spectrum of the unheated sample as a reference. As mentioned earlier, the resonance peaks in the low field region arise from silicon atoms connected to methyl groups, while those in the high field region arise from silicon atoms without attached methyl groups. More precisely, the resonance peak at -54.2 ppm is from silicon atoms with one methyl group, one hydroxyl group and two siloxane bonds (CH3(OH)Si(OSi=)2); and the peak at -61.3 ppm is from silicon atoms with one methyl group and three siloxane bonds (CH3Si(OSi=)3). The peaks at -91.8 ppm, -101.3 ppm and -101.3 ppm come from silicons with two hydroxyl groups and two siloxane bonds ((HO)2Si(OSi=)2), silicons with one hydroxyl group and three siloxane bonds (HOSi(OSi=)3) and silicons with four siloxane bonds (Si(OSi=)4) respectively. An important feature of the stacked spectra is that as the heating temperature is increased, the intensities of all the peaks gradually decrease. The intensity of peaks arising from methyl-connected silicons decrease faster than those arising from silicons without methyl groups connected. In the figure, a dramatic intensity decrease in peak intensity was found at about 400°C. As we have described in Chapter Ul, the proton sources for the cross-polarization to enhance the signal-to-noise ratio of the 2 9 Si spectrum of the silica gel are both methyl protons and hydroxyl protons. Thus we can predict from figure 4.2, that as the heating temperature increases, the hydroxyl groups in the methylsilica 100 Figure 4.2. Stacked 2 9 Si CP/MAS NMR spectra of Gel 3 heated at the different temperatures indicated for four hours and then cooled to ambient temperature. The spectra were obtained at 79.5 MHz , with 10.0 ms contact time, and 10.0 s repetition time. 320 scans were accumulated for each spectrum with a sample spinning rate of 3.2 kHz. The spectra are plotted in the absolute intensity mode with the spectrum of the untreated sample as the reference. 101 gel gradually decompose or condense to siloxane bonds. This contributes to the initial decrease of the resonance intensities of both silicon atoms with and without attached methyl groups. This result is in agreement with other studies on silica gelst2-3]. In addition, the resonances become broader, reflecting a higher degree of disorder and distortion which also contributes to the lower intensities. When the heating temperature reaches about 400°C, the methyl groups on the silica gel start decomposing, which causes the dramatic decrease in the resonance intensity of the low field silicon signals. After the sample was heated at 600°C, the methyl groups completely decomposed and only a very small number of hydroxyl groups were left, so only resonances from silicons without attached methyl groups are detected. Figure 4.3 shows 2 9 S i and 1 3 C CP/MAS NMR spectra of Gel 3 heated at different temperatures under N 2 protection using the apparatus illustrated in Figure 4.1. The most important feature of the stacked 2 9 S i spectra (Figure 4.3A) is that the intensities of both resonance regions of the silicons with and without attached methyl groups decrease somewhat as the heating temperature is increased, However, there is no dramatic intensity change compared with the 2 9 S i spectra in Figure 4.2. From Figure 4.3A, we can see that a visible intensity decrease occurs for the resonance peaks at -54.2 ppm and -91.3 ppm, which arises from silicons with hydroxyl groups. This is, obviously, due to the successive condensation of the hydroxyl groups in the silica gel with increasing heating temperature. It is this condensation that is responsible for all of the 2 9 Si resonance intensity decrease since there is no apparent loss of methyl groups during the heating under N 2 protection. This result is confirmed by the corresponding 1 3 C CP/MAS NMR spectra of Gel 3 heated up to 600° C (Figure 4.3B), which show no apparent intensity decrease of the methyl resonance peak at about -4.0 ppm. Comparing Figure 4.2 with Figure 4.3, we can conclude that the dramatic decrease of the resonance intensities of 2 9 S i CP/MAS in Figure 4.2 at heating temperatures above 400°C is probably due to an oxidative decomposition of methyl groups in the methylsilica gel. 102 Figure 4.3. Stacked NMR spectra of Gel 3 heated at the different temperatures indicated under N 2 for four hours. A. 2 9 S i CP/MAS spectra obtained at 79.5 MHz , with 10.0 ms contact time, and 10.0 s repetition time. Each spectrum was the result of 320 scans with a sample spinning rate of 3.2 kHz. B. 1 3 C CP/MAS spectra obtained at 100.6 MHz, with 1.0 ms contact time, 2.0 s repetition time, 800 scans and a sample spinning rate of 3.2 kHz. 103 In order to confirm the conclusions from the solid-state NMR experiments, the thermal behavior of high-purity gel and Gel 3 was investigated by DSC. Figure 4.4 is the DSC thermogram of the high-purity silica gel, made by condensation of TEOS, which has been described in Chapter II. Two peaks were observed in the experimental temperature range from room temperature to 500°C. The first endothermic peak, with its highest point at ~105°C is due to the evaporation of the physically absorbed water on the surface of the gelt3!. The second peak at 490°C, having an exothermic character, might be due to some kind of crystalline inversion (perhaps a-»(3 cristobalite inversion); Frodelf4! has suggested that amorphous silica powders can be formed by subcrystalline particles having the cristoblite structure. ' Figure 4.5 shows the DSC thermogram of Gel 3. Curve A was obtained with air as flow gas and curve B with N 2 as flow gas. Compared to the DSC curve of the high-purity silica gel, in addition to the the same endothermic peak at 105°C and the exothermic peak at 490°C as in Figure 4.4, curve A features an additional endothermic peak centered at about 465°C. The main difference between the high-purity silica gel and Gel 3 is that the later was condensed with two components, TEOS and MTES, instead of the simple monomer TEOS. We thus can deduce that the endothermic peak at 465°C in the DSC Curve A of Gel 3 is due to a structural feature related to the methyl groups in the gel. It seems reasonable that the 465°C endotherm is due to the decomposition of the methyl groups from the main frame of the functionalized silica gel as indicated by the NMR studies. When the DSC measurement of Gel 3 sample is conducted with N 2 as flow gas, the thermogram (Curve B) showed a greatly reduced endothermic peak centered at about 465°C. In the DSC thermogram (not shown) of Gel 3 preheated at 450°C for 4 hours, obtained under the same conditions as that in Figure 4.5, there was no such endothermic peak at 465°C. Thus we can conclude that the endothermic peak at 465°C in the thermogram 104 Figure 4.4. DSC thermogram of the high-purity silica gel. 105 i i i I i i i i I i i r [ i i i i i i i i i i i 100.0 200.0 300.0 400.0 500.0 Temperature C Q Figure 4.5. The DSC thermogram of Gel 3 A. Obtained with air as flow gas. B. Obtained with N 2 as flow gas. 106 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 i 100.0 200.0 300.0 400.0 500.0 Temperature (*C) of Gel 3 is due to an oxidative decomposition of the functional group =Si-CH-3. This result is in complete agreement with that obtained from the 2 9 Si and 1 3 C CP/MAS NMR on Gel 3 (Figures 4.2, 4.3). 4.3.2.Thermal Stabilities of Methylsilica Gel Made by the Anchoring Method Figure 4.6 illustrates the 2 9 S i and 1 3 C CP/MAS NMR spectra of Gel 26 heated at different temperatures as in the previous studies. Gel 26 was made by the anchoring method described in Chapter II. The assignment of each resonance peak of the CP/MAS NMR spectrum is also as given in Chapter II. The general feature of both the 2 9 S i and 1 3 C CP/MAS spectra is that, as the heating temperature increases, the resonance intensities of all peaks decreases. An abrupt intensity decrease occurs for the sample that was heated to 600°C. As we have seen in the preceding section, the peaks at -54 ppm and -101.5 ppm arise from silicon atoms with one hydroxyl group, one methyl group and two siloxane bonds and silicon atoms with one hydroxyl group and three siloxane bonds. These peaks decrease considerably with increasing temperature, due to the successive condensation of the hydroxyl groups on the silica getf2'4!, while the concentration of the methyl groups on the surface of the modified silica gel does not change until the heating temperature exceeds that where the surface group =Si-CH3 begins to decompose. It is this decomposition that contributes to the dramatic intensity decrease in the 2 9 S i spectrum for the Gel 26 sample heated at 600°C. This decomposition can be more clearly observed in the 1 3 C CP/MAS NMR spectra (Figure 4.6B), where the peak intensity of the methyl resonance for the sample heated at 600°C (top spectrum) decreases dramatically. The greater thermal stability of this gel may reflect the reduced accessibility of the Si-CH3 functionalities in surface attached polysiloxane and oligomers. 107 Figure 4.6. Stacked NMR spectra of Gel 26 heated at the different temperatures indicated for four hours. A. 2 9 S i CP/MAS spectra obtained at 79.5 MHz, with 10.0 ms contact time and 10.0 s repetition time. 320 scans were accumulated for each spectrum with a sample spinning rate of 3.2 kHz. B. 1 3 C CP/MAS spectra obtained at 100.6 MHz, with 1.0 ms contact time, 2.0 s repetition time, 800 scans and a sample spinning rate of 3.2 kHz. 108 A.2*Si B. 13C Figure 4.7 shows the stacked 2 9 S i and 1 3 C CP/MAS spectra of Gel 26 heated at different temperatures under N 2 protection. The general features of this figure are quite similar to those of figure 4.6 except that the intensity reductions in both the 2 9 S i and 1 3 C spectra for the sample heated at 600°C are not as great as those in figure 4.6. Thus we are not confident in predicting that the decomposition of the surface =Si-CFJ.3 groups by heating at 600°C is due to an oxidation reaction, although other researchers^4'5] have drawn such a conclusion. 4.4. Conclusions Comparing the thermal stabilities of the methylsilica gels made by copolymerization (Gel 3) and by the anchoring method (Gel 26), we can draw some conclusions as follows: (1) . CP/MAS solid-state NMR can be successfully applied in the investigation of the thermal stabilities of the functionalized silica gels as a technique complementary to the more traditional DSC technique. (2) . In an inert gas (N2) environment, the methyl groups on Gel 3 are more thermally stable to decomposition than Gel 26 and both are much more stable than in a normal atmospheric environment (3) . The two gels show a common behavior on heating: the physically absorbed water first evaporates, then the hydroxyl groups on or inside the gels gradually condense to form siloxane bonds. (4) . In a normal atmospheric environment, the methyl groups on Gel 3 decompose at a lower temperature than that on Gel 26, although the methyl group loadings on both gels are almost same. 109 Figure 4.7. Stacked NMR spectra of Gel 26 heated at different temperatures for four hours under N 2 protection. A. 2 9 S i CP/MAS spectra obtained at 79.5 MHz, 10.0 ms contact time, and 10.0 s repetition time. 320 scans were accumulated for each spectrum with a sample spinning rate of 3.2 kHz. B. 1 3 C CP/MAS spectra obtained at 100.6 MHz, 1.0 ms contact time, 2.0 s repetition time, 800 scans and a sample spinning rate of 3.2 kHz. 110 4.5. REFERENCES 4.1) . J. Nawrocki and B. Buszewski, J. Chromat. Sci., 449, (1988) 1. 4.2) . T. Zmijewski, M. Mioduska and B. Pacewska, J. Thermal Anal., 32, (1987) 1755. 4.3) . C. Frodel, "The System of Mineralogy of DANA, Silica Minerals", Wiley, New York, Vol. 3, (1962) P 154. 4.4) . H. J. Huhn, J. Thermal Anal., 33, (1988) 851. 4.5) . T. J. Fripiat, J. Uytterhoeven, U. Schobinger and H. Deuel, Helv Chim. Acta. 43, (1960) 176. ill 

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:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0059757/manifest

Comment

Related Items