Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Multiple field solid state MAS NMR structural investigations of ultrastable Zeolite-Y and Aluminophosphate… Bretherton, Jeremy L. 2002

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

Item Metadata

Download

Media
831-ubc_2003-792021.pdf [ 14.05MB ]
Metadata
JSON: 831-1.0061234.json
JSON-LD: 831-1.0061234-ld.json
RDF/XML (Pretty): 831-1.0061234-rdf.xml
RDF/JSON: 831-1.0061234-rdf.json
Turtle: 831-1.0061234-turtle.txt
N-Triples: 831-1.0061234-rdf-ntriples.txt
Original Record: 831-1.0061234-source.json
Full Text
831-1.0061234-fulltext.txt
Citation
831-1.0061234.ris

Full Text

Multiple Field Solid State MAS NMR Structural Investigations of Ultrastable Zeolite-Y and Aluminophosphate Materials. Jeremy L . Bremerton M.Chem., St. Catherine's College, Oxford, 1996.  A THESIS SUBMITTED IN P A R T I A L F U L F I L L M E N T OF THE REQUIREMENTS FOR THE D E G R E E OF DOCTOR OF PHILOSOPHY in THE F A C U L T Y OF G R A D U A T E STUDIES Department of Chemistry  We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH C O L U M B I A November 2002 © Jeremy Laurence Bretherton, 2002.  In  presenting  degree  this  thesis  in  at the University of  partial  fulfilment  British Columbia,  of  the  requirements  for  an  advanced  I agree that the Library shall make it  freely available for reference and study. I further agree that permission for extensive copying  of  department  this or  thesis by  for scholarly  his  or  publication of this thesis  her  The University of British Columbia^ Vancouver, Canada  DE-6 (2/88)  It  be is  granted  by the head of  understood  that  for financial gain shall not be allowed without  Department of  3 7 ^ / / 0 2 .  may  representatives.  permission.  Date  purposes  copying  my or  my written  Abstract Solid state M A S N M R investigations of ultrastable zeolite-Y (USY) materials were performed at magnetic fields of 18.8 T, 17.6 T, 14.4 T and 9.4 T. A l 3QMAS and quantitative 2 7  27  high field 'A1 M A S N M R spectra were acquired and can be interpreted in terms of four distinct Z  aluminium environments. Simulation of t h e A l M A S N M R spectra at multiple fields enabled 27  the accurate parameterisation of the aluminium environments, as demonstrated by the good agreement between calculated and experimental isotropic A l 3QMAS shifts at both high and 2 7  low field. Quantitative A l M A S N M R spectroscopic data is supported by S i M A S 2 7  NMR  2 9  spectroscopy, powder X-ray diffraction patterns and nitrogen physisorption isotherms in the study of several series of U S Y materials. The effects of acid washing and steam calcination on the structure of U S Y and the proportions and nature of the aluminium present were investigated. The effects of repeated ammonium exchange and calcination on the structure of U S Y are shown to be consistent with a new model describing the progressive structural changes taking place during hydrothermal treatments. This information is used to provide a preliminary interpretation of high temperature (540°C) »-hexane catalytic cracking data acquired for the U S Y materials. These multiple field A1 M A S N M R techniques were applied to other mixed crystalline27  amorphous systems. In the study of a series of complex aluminophosphate ceramic materials, the quantitative reliability of these techniques was further demonstrated by the agreement of the results of multiple field A1 M A S N M R and P M A S N M R experiments. A1 M A S 27  31  27  NMR  spectroscopy has enabled the quantification of all crystalline and amorphous phases present in these mixtures, which is not possible using any other techniques. Multiple field A1 and P M A S N M R spectra and powder X-ray diffraction data were 27  3 1  obtained for calcined, hydrated A I P O 4 - I 8 molecular sieve. Al-P bonding connectivities were derived from A 1 - P refocused INEPT M A S N M R spectra. These data have facilitated the 27  31  assignment of the crystallographic space group and the unit cell parameters and the number of possible assignments of the N M R resonances has been reduced from 5.2 x 10 to eight. 4  11  Table of Contents Abstract  i  Table of Contents  ii  List of Figures  ix  List of Tables  xiv  Notation and Abbreviations  xvi  Acknowledgements  xix  iii  C H A P T E R 1. INTRODUCTION 1.1  1  SOLID STATE NUCLEAR MAGNETIC RESONANCE  1  1.1.1  The Zeeman Interaction  1  1.1.2  The Effect of Radio Frequency Pulses and the Concept of Coherence  2  1.1.3  Fourier Transform N M R  4  1.1.4  Off-Resonance N M R  6  1.1.5  Quadrature Detection  7  1.1.6  Perturbations to the Nuclear Spin Hamiltonian  9  1.1.6(a) Chemical shielding  10  1.1.6(b) Dipolar coupling  12  1.1.6(c) Scalar coupling  13  1.1.6(d) Quadrupolar coupling  14  1.1.6(e) Conclusion  20  1.1.7  Relaxation  21  1.1.8  Solid State N M R of Static Powdered Samples  25  1.1.9  Magic Angle Spinning  29  1.1.9(a) The refocusing effect of M A S  30  1.1.9(b) Effect of M A S on dipolar coupling to an abundant spin system  31  1.1.9(c) The effect of M A S on second order quadrupolar coupling  33  1.1.10  N M R Experiments  37  1.1.10(a) Single pulse experiment  37  1.1.10(b) Spin-echo experiment  38  1.1.10(c) The inversion and saturation recovery experiments  40  1.1.10(d) The INEPT experiment  42  1.1.10(e) The M Q M A S experiment  47  1.2  X - R A Y DIFFRACTION  58  1.2.1  Powder X-Ray Diffraction  58  1.2.2  Basic Crystallography  62  1.2.2(a) Three dimensional arrays  62  1.2.2(b) The asymmetric unit, lattice and unit cell  62  1.2.2(c) Evaluating the unit cell dimensions  63  iv  1.2.2(d) Experimental apparatus  65  1.2.2(e) X R D and N M R equivalence  65  1.3  MATERIALS  66  1.3.1  Zeolite Molecular Sieves  66  1.3.2  Aluminophosphate Molecular Sieves  71  1.3.3  Aluminophosphate Ceramic Mixtures  72  1.4  REFERENCES FOR CHAPTER 1  74  C H A P T E R 2. E X P E R I M E N T A L 2.1  83  SOLID STATE N M R SPECTROSCOPY  2.1.1  83  Spectrometers  2.1.1 (a)  83  General spectrometer layout  84  2.1.1 (b) Experimental setup  86  2.1.1(c)  86  2.1.2  M A S equipment  Spectral Processing  87  2.1.2(a) Zero filling and apodization  87  2.1.2(b) 2D spectroscopy  90  2.2  POWDER X - R A Y DIFFRACTION  94  2.3  THERMOGRAVIMETRIC ANALYSIS  96  2.4  PHYSISORPTION MEASUREMENTS  96  2.5  « - H E X A N E CATALYTIC CRACKING ACTIVITY MEASUREMENTS  2.6  REFERENCES FOR CHAPTER 2  99 101  v  C H A P T E R 3. M U L T I P L E FIELD A l M A S A N D 3QMAS N M R C H A R A C T E R I Z A T I O N 2 1  A N D QUANTIFICATION OF THE A L U M I N I U M E N V I R O N M E N T S IN U L T R A S T A B L E ZEOLITE Y  103  3.1  INTRODUCTION  103  3.2  MATERIALS AND METHODS  105  3.2.1  Synthesis of Amorphous Alumina and U S Y  105  3.2.2  Experimental Apparatus  105  3.3  RESULTS AND DISCUSSION  106  3.3.1  Changes in the Framework  106  3.3.2  Low Field  A l N M R Investigation of U S Y  108  3.3.3  High Field  A l N M R Investigation of U S Y  113  3.3.4  Simulation, Deconvolution and Parameterisation of A l N M R Data  3.3.5  2 7  2 7  2 7  2 7  A l N M R Investigations of Amorphous Alumina  116 121  3.4  CONCLUSIONS  123  3.5  REFERENCES FOR CHAPTER 3  124  C H A P T E R 4. SOLID STATE N M R A N D N I T R O G E N PHYSISORPTION INVESTIGATIONS OF T H E EFFECT OF A C I D W A S H I N G OF U S Y M A T E R I A L S  126  4.1  INTRODUCTION  126  4.2  MATERIALS AND EXPERIMENTAL METHODS  128  4.2.1  Materials  128  4.2.2  Experimental Apparatus  128  4.3  RESULTS AND DISCUSSION  129  4.3.1  Changes in the Framework  129  4.3.2  High and Low Field A1 M A S N M R Data  132  4.3.3  Discussion  138  4.4  CONCLUSIONS  141  4.5  REFERENCES FOR CHAPTER 4  142  27  vi  C H A P T E R 5. M U L T I P L E FIELD A1 N M R S T U D Y TO P R O B E T H E EFFECTS OF Z/  C A L C I N A T I O N O N T H E STRUCTURE A N D w-HEXANE C R A C K I N G A C T I V I T Y OF USY  144  5.1  INTRODUCTION  144  5.2  MATERIALS AND METHODS  145  5.2.1  Materials  145  5.2.2  Experimental Apparatus  146  5.3  RESULTS  5.3.1 5.3.2  2 9  148  S i M A S , A l M A S and 3QMAS N M R Spectroscopy 2 7  Physisorption and Powder X R D Data  5.3.3 rc-Hexane Catalytic Cracking Data 5.4  DISCUSSION  5.4.1  148 159 163 164  Structural Changes Indicated by Physisorption and Powder X-Ray Diffraction Data 164  5.4.2  Relationship between S i and A l M A S N M R Spectra and Structural Changes 168  5.4.3  Relationship Between Structural Changes and «-Hexane Cracking Activity  2 9  2 7  173  5.5  CONCLUSIONS  177  5.6  REFERENCES FOR CHAPTER 5  179  vii  C H A P T E R 6. DETECTION, C H A R A C T E R I Z A T I O N A N D QUANTITATION OF THE M U L T I P L E COMPONENTS IN C O M P L E X A L U M I N O P H O S P H A T E MIXTURES  CERAMIC 183  6.1  INTRODUCTION  183  6.2  MATERIALS AND METHODS  184  6.2.1  Materials  184  6.2.2  Experimental Apparatus  186  6.3  6.3.1  RESULTS AND DISCUSSION  187  Powder X-Ray Diffraction  187  6.3.2  3 1  P M A S N M R Spectroscopy  188  6.3.3  2 7  A l M A S and 3QMAS N M R Spectroscopy  195  6.3.4 6.4  Metaphosphates-B and -C  200  CONCLUSIONS  205  REFERENCES FOR CHAPTER 6  207  C H A P T E R 7. A L A N D P M A S N M R S T R U C T U R A L INVESTIGATION OF C A L C I N E D , 2 7  3 1  HYDRATED ALPO4-I8  209  7.1  INTRODUCTION  209  7.2  MATERIALS AND METHODS  212  Synthesis of AIPO4-18  212  7.2.1 7.2.2 7.3  7.3.1 7.3.2 7.3.3  2 7  A l and P M A S N M R Spectroscopy 3 1  RESULTS AND DISCUSSION 2 7  214  A l M A S and 3QMAS N M R Spectroscopies  Simulation of Single Pulse A1 and P M A S N M R Spectra 27  2 7  3 1  A l - ' P refocused INEPT M A S N M R Spectroscopy 3  7.3.4  Assignment of A l and P M A S N M R Resonances  7.3.5  Powder X-Ray Diffraction  2 7  212  3 1  214 216 221 225 229  7.4  CONCLUSIONS  230  7.5  ACKNOWLEDGEMENT  231  7.6  REFERENCES FOR CHAPTER 7  232  viii  PROPOSALS FOR F U R T H E R W O R K  234  P1.  ULTRASTABLE ZEOLITES  234  P2.  ALUMINOPHOSPHATE M O L E C U L A R SIEVES  236  Appendices: Table of Contents  Al  APPENDIX I. Laboratory to Rotating Frame Transformation Matrix  A3  APPENDIX II. Normalised Spherical Harmonic Functions  A4  APPENDIX III. Pulse Sequences and Macros  A5  APPENDIX TV. Simulation Parameters for 17.6 T Single Pulse and Echo A l M A S N M R 2 7  Spectra of U S Y Materials  A59  APPENDIX V . Fractional Co-ordinates of the Framework Atoms of Calcined, Hydrated AIPO4-I8  A66  ix  List of Figures Figure 1.1 Schematic representation of the possible nuclear spin states of a nucleus with spin I = V2  1  Figure 1.2 The effect of a n/2 radiofrequency (rf.) "pulse" applied at the Larmor frequency  3  Figure 1.3 Rotating frame diagram of a single pulse N M R experiment  5  Figure 1.4 Fourier transformation of a radio frequency pulse  7  Figure 1.5 Graphical representation of a tensor property, P  10  Figure 1.6 Schematic representation of the sequential effect of first and second order quadrupolar coupling (H  0)  and H ) on the resonance frequencies of a spin ( 2) Q  / = 5^ nucleus  19  Figure 1.7 Position of bulk magnetisation in the rotating frame of reference (i) before (ii) during (iii) after relaxation  22  Figure 1.8 Schematic representation of the relationship between relaxation time and correlation time  24  Figure 1.9 Plot of g(\Av\) against Av.  26  Figure 1.10 Examples of simulated first order powder lineshapes plotted on an arbitrary chemical shift scale  28  Figure 1.11 Schematic representation of slow M A S N M R spectroscopy  32  Figure 1.12 Progressive changes in the N M R spectrum of a nucleus dipolar coupled to an abundant spin system as a result of increasing M A S rate  33  Figure 1.13 Plot of the 2nd and 4th Legendre polynomials between 0° and 90°  34  Figure 1.14 Simulated M A S powder spectra of the central transition of a non-integer spin quadrupolar nucleus  36  Figure 1.15 Rotating frame diagram of 90°-180° spin-echo experiment  40  Figure 1.16 Rotating frame diagram of (a) the inversion recovery experiment, (b) the saturation recovery experiment  41  Figure 1.17 Energy level diagram of a coupled heteronuclear spin / = A two spin system l  43  Figure 1.18 (a) Rotating frame diagram and (b) pulse sequence diagram of the INEPT experiment  44  x  Figure 1.19 Pulse sequence diagram of the refocused INEPT experiment  45  Figure 1.20 Pulse sequence diagram of the 2D refocused INEPT experiment  46  Figure 1.21 Pulse sequence diagram of (a) the "perfect" DAS experiment with acquisition at 02, (b) the "perfect" DAS experiment with acquisition at experiment with acquisition at 6^  OMAS  and (c) the DAS 49  AS  Figure 1.22 (a) Pulse sequence diagram and (b) coherence pathway diagram of the nutation 3QMAS experiment for a spin 1= 5/2 nucleus Figure 1.23 Spin locking of a spin I- Vi nucleus in the rotating frame  51 54  Figure 1.24 RIACTII 3QMAS experiment  56  Figure 1.25 Z-filtered 3QMAS pulse sequence  58  Figure 1.26 Schematic representation of coherent scattering of X-rays leading to (a) constructive and (b) destructive interference  60  Figure 1.27 Examples of some of the lattice planes that pass through the atoms in a 2D array  61  Figure 1.28 The crystallographic axis system  64  Figure 1.29 Example of the relationship between zeolite structures and secondary building units  70  Figure 2.1 Component parts of a solid state N M R spectrometer  84  Figure 2.2 Fourier transformation of a truncated FID  88  Figure 2.3 (a) Un-sheared and (b) Sheared z-filtered A1 3QMAS spectrum of calcined, 27  hydrated A1P0 -18  92  4  Figure 2.4 Plan view of the Bragg-Brentano geometry of the Rikagu Rotorflex diffractometer  95  Figure 2.5 Flow diagram of the H-hexane cracking unit  100  Figure 3.1 S i M A S N M R spectra of (a) NaY (b) the U S Y material recorded at 9.4 T  107  Figure 3.2 Powder X R D patterns of (a) NaY and (b) U S Y  109  Figure 3.3 A1 M A S N M R (a) single pulse (b) 90°-180° echo spectra of U S Y at 9.4 T  110  Figure 3.4 Nutation A l 3QMAS N M R spectrum of U S Y acquired at 9.4 T  112  2 9  27  2 7  Figure 3.5 (a) Single pulse and (b) echo A l M A S N M R spectra of U S Y acquired at 2 7  18.8 T Figure 3.6 Al 27  114  RIACTII 3QMAS spectrum of U S Y at 208.4 M H z  xi  115  Figure 3.7 Experimental, simulated and deconvolved single pulse A l M A S 2 /  NMR  spectrum of U S Y  117  Figure 3.8 Experimental, simulated and deconvolved single pulse A l M A S N M R spectra 2 7  of U S Y acquired at (a) 14.4 T and (b) 9.4 T  120  Figure 3.9 A l (a) 90°-180° echo M A S and (b) nutation 3QMAS N M R spectra of 2 7  amorphous alumina acquired at 17.6 T  122  Figure 4.1 S i M A S N M R spectra of (a) U l and U l w (b) U2 and U2w, recorded at 9.4 T.... 131 2 9  Figure 4.2 Single pulse A1 M A S N M R spectra of U S Y samples acquired at (a) 17.6 T 27  and(b) 9.4 T  133  Figure 4.3 Total integrated intensity of A l M A S spectra of U S Y samples  134  Figure 4.4 Al  135  2 7  27  3QMAS N M R spectra of U l recorded at (a) 17.6 T and (b) 9.4 T  Figure 4.5 A1 3QMAS N M R spectra of U2w recorded at (a) 17.6 T and (b) 9.4 T 27  136  Figure 4.6 Single pulse A l M A S N M R spectra, simulation and deconvolution of U l 2 7  acquired at (a) 17.6 T and (b) 9.4 T  139  Figure 4.7 (a) Percentage and (b) relative percentage of aluminium environments in U S Y materials Figure 5.1 Si 29  140  M A S N M R spectra of the series of U S Y materials calcined at (a) 550°C  and (b) 650°C, recorded at 9.4 T  150  Figure 5.2 Single pulse A1 M A S N M R spectra of the series of U S Y materials calcined at 27  550°C and 650°C, recorded at 17.6 T  151  Figure 5.3 A l 3QMAS N M R spectra of U S Y materials ammonium exchanged and 2 7  calcined at 550°C (a) once (b) six times, recorded at 17.6 T  153  Figure 5.4 A l 3QMAS N M R spectra of U S Y materials ammonium exchanged and 2 7  calcined at 650°C (a) once (b) six times, recorded at 17.6 T  154  Figure 5.5 A l single pulse M A S N M R spectra and spectral simulations of U S Y 2 7  ammonium exchanged and calcined four times at 550°C, recorded at (a) 9.4 T and(b) 17.6 T Figure 5.6 A\ 21  156  single pulse M A S N M R spectra and spectral simulations of U S Y  ammonium exchanged and calcined five times at 650°C, recorded at (a) 9.4 T and (b) 17.6 T  157  xii  Figure 5.7. Differential pore size distributions based on nitrogen adsorption isotherms of U S Y materials  161  Figure 5.8 Selected X R P D patters of U S Y materials calcined at (a) 550°C and (b) 650°C  162  Figure 5.9 Si:Al calculated for U S Y materials calcined at (a) 550°C and (b) 650°C  169  Figure 5.10 (a) Percentage abundance of aluminium species in U S Y materials calcined at 550°C. (b) Relative percentages of aluminium species in U S Y materials calcined at 550°C normalized to a surface area of 1000 m g"' 2  170  Figure 5.11 (a) Percentage abundance of aluminium species in U S Y materials calcined at 650°C. (b) Relative percentages of aluminium species in U S Y materials calcined at 550°C normalized to a surface area of 1000 m g"' and corrected for a 2  trend in the percentage of extra-crystallite amorphous material  171  Figure 5.12 Proposed reaction scheme for the hydrothermal dealumination of U S Y  172  Figure 5.13 High temperature w-hexane catalytic cracking activity of U S Y materials calcined at (a) 550°C and (b) 650°C  174  Figure 6.1 Powder X-ray diffractograms of alumina sol, a-alumina and composite sol-gel derived ceramic mixtures  188  Figure 6.2 Single pulse P M A S N M R spectra of alumina sol, a-alumina and composite 3 1  sol-gel derived ceramic mixtures at (a) 9.4 T and (b) 17.6 T  190  Figure 6.3 Single pulse P M A S N M R spectrum of alumina sol derived ceramic mixture 3 1  at 9.4 T  194  Figure 6.4 Single pulse A l M A S N M R spectra of alumina sol, a-alumina and composite 2 7  sol-gel derived ceramic mixtures at (a) 9.4 T and (b) 17.6 T Figure 6.5 Single pulse  195  A l M A S N M R spectra of alumina sol derived ceramic mixture at  9.4 T and 17.6 T  197  Figure 6.6 Nutation Al 3QMAS N M R spectrum of alumina sol-derived ceramic mixture 27  at 17.6 T  198  Figure 6.7 Powder X R D patterns of metaphosphate-B and metaphosphate-C and a simulated pattern of "metaphosphate"  201  Figure 6.8 (a) Deconvolution (of the isotropic region) of, (b) simulation of, and (c) experimental single pulse P M A S spectrum of metaphosphate-C, recorded at 3 1  17.6 T  202  xiii  Figure 6.9 A 1 M A S spectra of metaphbsphate-C at (a) 17.6 T and (b) 9.4 T  204  Figure 6.10 A l 3QMAS N M R spectra of metaphosphate-C at (a) 17.6 T and (b) 9.4 T  206  Figure 7.1 Topological structure of calcined, dehydrated  AIPO4-I8  210  Figure 7.2 Topological structure of calcined, dehydrated  AIPO4-I8  Z/  2 7  viewed along (a) the c-  axis channel system and (b) the a-axis channel system Figure 7.3 Single pulse A l M A S N M R spectra of calcined, hydrated 2 7  211 AIPO4-I8  recorded  at (a) 9.4 T and (b) 18.8 T  214  Figure 7.4 Z-filtered Al 3QMAS N M R spectrum of calcined, hydrated 27  AIPO4-I8  recorded at 9.4 T  215  Figure 7.5 Nutation A l 3QMAS N M R spectrum of calcined, hydrated 2 7  AIPO4-I8  recorded  at 17.6 T  216  Figure 7.6 Single pulse P M A S N M R spectra recorded at (a) 18.8 T and (b) 9.4 T 3 1  218  Figure 7.7 Simulations and deconvolutions of single pulse A l M A S N M R spectra of 2 7  AIPO4-I8  recorded at (a) 18.8 T and (b) 9.4 T  220  Figure 7.8 A 1 - P refocused INEPT M A S N M R spectra of calcined, hydrated A I P O 4 - I 8 , 27  31  recorded at (a) 9.4 T and (b) 18.8 T  222  Figure 7.9 A 1 - P refocused INEPT M A S N M R (a) experimental (b) schematic spectra of 27  31  calcined, hydrated  AIPO4-I8,  recorded at 9.4 T  227  Figure 7.10 A 1 - P refocused INEPT M A S N M R (a) experimental (b) schematic spectra 27  31  of calcined, hydrated A I P O 4 - I 8 , recorded at 9.4 T Figure 7.11 X R P D pattern of calcined, hydrated  AIPO4-I8  228 230  Figure PI. A 1 - P refocused INEPT M A S N M R spectrum of calcined, hydrated 27  31  AlP0 -36, recorded at 18.8 T  237  4  FigureP2. A 1 - P 3 QMAS-refocused INEPT M A S N M R spectrum of calcined, hydrated 27  31  A1P04-14 recorded at 18.8 T  238  xiv  List of Tables Table 1.1 Values of  and k for spin / = 5/2 and / = 3/2 nuclei  53  Table 1.2 The Bravais Lattices  63  Table 3.1 Integrated total intensities o f A l single pulse and echo M A S spectra  Ill  2 7  Table 3.2 Summary of the spectral simulation parameters for the single pulse and echo  2 7  Al  M A S N M R spectra of U S Y at 18.8 T  118  Table 3.3 Calculated and experimental 3QMAS isotropic (Fl) shifts of U S Y at 9.4 T and 18.8 T  119  Table 3.4 Summary of the spectral simulation parameters for the 14.4 T and 18.8 T single pulse A l M A S spectra of amorphous alumina gel  121  Table 4.1 Si:Al values calculated for the series of U S Y samples  130  2 7  Table 4.2 Physisorption data obtained from N2 adsorption  i s o t h e r m . 1 3 1  Table 4.3 Integrated total intensities of single pulse and echo A l M A S N M R spectra at 2 7  17.6 T and 9.4 T  132  Table 4.4 Quadrupolar lineshape parameters for the simulation of A l M A S N M R spectra of (a) U l , (b) U l w , (c) U2 and (d) U2w at 9.4 T and 17.6 T Table 4.5 Percentage abundance of aluminium environments in U S Y samples  138 140  Table 5.1 Si:Al values calculated for the series of U S Y materials calcined at (a) 550°C and (b) 650°C  149  Table 5.2 Percentage of total aluminium in U S Y materials calcined at (a) 550°C and (b) 650°C, observed using A1 M A S N M R  152  27  Table 5.3 Percentage abundance" of aluminium species in U S Y materials calcined at (a) 550°Cand (b) 650°C  158  Table 5.4 Physisorption data from N2 isotherm analysis of U S Y materials calcined at (a) 550°C and (b) 650°C  160  Table 5.5 n-Hexane catalytic cracking product conversion percentages per unit mass of dry U S Y materials calcined at (a) 550°C and (b) 650°C  163  Table 6.1 Summary of the spectral parameters for P M A S N M R spectra of ceramic 3 1  mixtures at 9.4 T and 17.6 T  192  xv  Table 6.2 Percentages of the different phosphorous containing species present in the ceramic mixtures, from integration of the P M A S N M R spectra  193  3 1  Table 6.3 Summary of the spectral simulation parameters for the A l M A S N M R spectra 2 7  of ceramic mixtures at 9.4 T and 17.6 T  196  Table 6.4 Percentages of the different aluminium containing species present in the ceramic mixtures, from A l M A S N M R  199  2 7  Table 6.5 Summary of the spectral simulation parameters for the P M A S N M R spectra of 3 1  metaphosphate-C at 9.4 T and 17.6 T  203  Table 6.6 Summary of the spectral simulation parameters for A1 M A S spectra of 27  metaphosphate-C sample at 9.4 T and 17.6 T  203  Table 7.1 Integrals of peaks used to simulate the P M A S N M R resonances, recorded at 3 1  9.4 T and 18.8 T, from the crystalline component of calcined, hydrated AIPO4-I8  214  Table 7.2 Summary of the spectral simulation parameters for A1 M A S spectra of 27  calcined, hydrated A I P O 4 - I 8 at 9.4 T and 18.8 T  216  Table 7.3 Nearest neighbour A1<-»P T-site connectivities for the A I P O 4 - I 8 framework in the in the (a) P2/c and (b) P i space group  220  Table 7.4 Observed nearest neighbour A1<-»P T-atom connectivities from A 1 - P 27  31  refocused INEPT M A S N M R spectra of calcined, hydrated A I P O 4 - I 8 , recorded at 9.4 T and 18.8 T  221  Table 7.5 Nearest neighbour A1<-»P T-atom connectivities, derived from A 1 - P refocused 27  31  INEPT M A S N M R spectra of calcined, hydrate A I P O 4 - I 8 , recorded at 9.4 T and 18.8 T  223  Table 7.6 Possible peak assignments of the resonances of A l and P M A S N M R spectra 2 7  of calcined, hydrated A I P O 4 - I 8  3 1  223  xvi  Notation and Abbreviations ~  Approximately  =  Approximately equal to  Al  B r T e t  Broad tetrahedral aluminium  ^jFwk.Tet  Framework tetrahedral aluminium  Al  0 c t  6 co-ordinate aluminium  Al  P e n t  5 co-ordinate aluminium  AIPO4  Aluminophosphate molecular sieve  Al  Tetrahedral aluminium  T e t  aq  aqueous solution  ca.  Circa, approximately  cf.  Confer, compare to  CSA  Chemical shift anisotropy, chemical shielding anisotropy  CSG  Composite sol-gel  CW  Continuous wave  DAS  Dynamic angle spinning  DOR  Double rotation  e  unit electrical charge of an electron  e.g.  Exepli gratia, for example  =  Equivalent to  EFG  Electric field gradient  Eqn.  Equation  et al.  Et allii, other people  FID  Free induction decay  FT  Fourier transform  GC  Gas chromatograph  h  Plank's constant, 6.626 x 10" Js  H  Hamiltonian operator: Hy/ = Ey/, yields energy, E, in Joules.  h  h/2n, 1.054 x 10" Js  34  34  xvii  HY  Acid zeolite Y  I  Total angular momentum operator: hly/ = ti{l{l +1)) y/  I  z-component of angular momentum operator: hl y/ = -hm,y/  i.e.  Id est, that is to say  in situ  In its original place  INEPT  Insensitive nuclei enhanced by polarisation transfer  IR  Infra-red  J  Scalar coupling (Hz)  MAS  Magic angle spinning  MQMAS  Multiple quantum magic angle spinning  //„  Vacuum permeability: 1.26 x 10" mkgC"  nM  n molar solution, concentration, moldm"  N  Avagadro's number: 6.023 x 10  NaA  Sodium zeolite A  NaY  Sodium zeolite Y  NMR  Nuclear magnetic resonance  PAS  Principal axis system  PFG  Pulsed field gradients  ppm  Parts per million, chemical shift  R  Universal gas constant: 8.3145 JK" moi"  rf.  Radiofrequency  RIACT  Rotationally induced adiabatic coherence transfer  SBU  Secondary building unit  Si:Al  Silicon to aluminium ratio  S:N  Signal to noise ratio  s.t.p.  Standard temperature and pressure (273 K , 1 atmosphere)  7/*  Effective transverse relaxation time (s)  Ti  Longitudinal / spin lattice relaxation time constant (s)  T  Transverse / spin-spin relaxation time constant (s)  3QMAS  Triple quantum magic angle spinning  1  z  2  A  6  23  moi"  2  xviii  3  1  1  2  2  1  TEDOR  Transferred echo double resonance  TGA  Thermogravimetric analysis  TPPI  Time proportional phase incrementation  T-site  Framework tetrahedral site  2D  Two dimensional  USY  Ultra-stable acid zeolite Y  co  Frequency in radians per second  |X|  Magnitude of value X Quantum mechanical operator. Operates on the wavefunction of the  X  system (yf) to yield an expectation value X .  X  Tensor property  X  Vector property  XRD  X-ray diffraction  XRPD  X-ray powder diffraction  z-filtered  Zero quantum filtered Frequency width at half height  2  X  Nuclear quadrupole coupling constant  v  Frequency in cycles per second (Hz)  V  M A S  Magic angle spinning rate (Hz) Nutation frequency, radio frequency power (Hz)  xix  Acknowledgements I would like to thank Professor Colin Fyfe for his assistance and advice over the course of my degree and for providing all the time, instrument time and equipment that was necessary for its completion. Special thanks are due to Tom Markus and Milan Coschizza for helping to figure out "why there is no signal". Without the probes designed and built by Oskar Greiner and tuned by Tom Markus, the great majority of this work would not have been possible. It has always been greatly appreciated that Fyfe Group projects have received such a high degree of thought, effort and attention to detail. I would like to thank Dr. James Sawada for his considerable help navigating the minefield of physisoption measurements. Thanks also to Jeffrey Alvaji for his fanatical devotion to the catalytic cracking unit. I would especially like to thank Dr. Andrew Lewis for his help with a plethora of odds and ends and for being the last surviving custodian of archival Fyfe Group and Bruker information that would otherwise have required archaeological assistance to learn. Part of this research was performed in the Environmental Molecular Sciences Laboratory (a national scientific user facility sponsored by the U.S. DoE Office of Biological and Environmental Research) located at the Pacific Northwest National Laboratory, operated by Battelle for the DoE. I would like to thank the staff of the E M S L for their kind assistance and for access to their facility.  xx  Chapter 1. Introduction  1.1 Solid State Nuclear Magnetic Resonance  1.1.1  T H E Z E E M A N INTERACTION 1-3  In order for a nucleus to be N M R active, it must posses nuclear angular momentum, or "spin". The angular momentum of a particle in a given state can only be defined in one dimension, which shall define the z-axis. With respect to z, a nucleus with spin / (where the underline denotes a vector property) may take 21+1 orientations described by the quantum number mj (Figure 1.1). The angular momentum in the ry-plane is undefined.  Figure 1.1 Schematic representation of the possible nuclear spin states of a nucleus with spin I = 'A. The nuclear spin / is described with respect to the z-axis by the quantum number mj. For any value of mj the orientation of / in the xy-plane cannot be specified and so the state /»/ describes a cone of angular momentum vectors.  The nuclear magnetic dipole moment y. (a vector property) defined by the charge and spin, interacts with an applied magnetic field, producing the Zeeman Splitting (AE) of the energy levels. Where X is a quantum mechanical operator that acts upon the wavefunction describing  1  the system in a state defined by the set of quantum numbers  W(ri), to  yield the physical  property of the system X, the Hamiltonian of the Zeeman Interaction (i.e. the energy operator) is given by:  H ee =filB man  Z  H  =}*Bj  Zeeman  E  (1.1)  0  (1.2)  z  =-^B m  Zeeman  o  (1.3)  i  Where Bo is the applied magnetic field, defining the z-axis, / is the nuclear spin operator (underlined to signify that it is a vector property) and yis the magnetogyric ratio, / i s an empirical property and by convention +m states lie at lower energy than -nil states. 7  AE —' m  m  _, vB = —- = v h In '  m  L  Larmor frequency (Hz)  (1.4)  Eqn 1.4 gives the Larmor frequency of a bare nucleus in a magnetic field. Superconducting N M R magnets typically have field strengths of B = lfj'T producing Larmor frequencies on the 0  order of 10 -10 Hz. Boltzmann equilibrium population differences of the nuclear spin states 9  H  are therefore at most on the order of 10" %, making N M R a low sensitivity technique. 3  1.1.2  T H E EFFECT OF RADIO FREQUENCY PULSES AND THE CONCEPT OF COHERENCE. " 2  5  The concepts of bulk magnetization, the "rotating frame of reference", the interaction of nuclei with electromagnetic radiation and the generation of "coherences" are required to understand the practical implementation of N M R spectroscopy. Consider a sample containing one type of nucleus with spin 1= Vi in a magnetic field. At equilibrium there is a small excess of nuclei with mi = +Vi and the individual magnetic moments of the nuclei sum to give a bulk magnetization M . The phases of individual nuclear magnetic z  2  moments in the xv-plane are random and sum to zero. M i s a macroscopic property and can be treated in terms of classical mechanics, but its behaviour can also be understood in terms of a microscopic quantum mechanical picture.  time, t = 0  z,  t=  2yB  x  7t  Bo  z,  B  0  M  y  co  L  Static (laboratory) frame.  Static frame  Rotating frame  Figure 1.2 The effect of a n/2 radiofrequency (rf.) "pulse" applied at the Larmor frequency ("on resonance") at angular velocity m (in units of radians/sec). In the static frame of reference the bulk magnetization lies along z (i.e. M = M ) and the magnetic field component of the radiation (fl/) rotates around z at a . In the rotating frame of reference fl is a static field (arbitrarily) along x. Under the influence of fl/, M precesses around x at a rate of = -yBj and after a time t = n/ 2yB,, Mlies along v. At this time, observedfromthe staticframeof reference, M is processing around z at the Larmor frequency. L  z  L  ;  Classically, the interaction between a magnetic moment y. and a magnetic field B results in a torque (i.e. an angular force) given by — = // x B, where I is angular momentum. The dt  classical picture of the sample is therefore that it comprises individual nuclear magnetic moments precessing around z at the Larmor frequency at random phase, such that M i s static along z. Coherent radiofrequency (rf.) radiation perpendicular to B at the Larmor frequency 0  subjects the sample to an additional magnetic field that rotates around z at v . Observed from an L  axis system that itself rotates precisely at v , the "rotating frame of reference", the Bj field of the L  rf radiation is static in the xy-plane (Figure 1.2) and B does not appear. Therefore, resonant rf. 0  radiation (i.e. h v = AE) forces M to precess around Bj, after which it precesses around Bo (Figure 1.2). The precession of Maround B is known as nutation. The rf. power (v^) is the x  "nutation frequency" in units of Hz.  3  The quantum mechanical interpretation of this interaction is the generation of a coherence. The rf. pulse causes the populations of the nuclear spin states to change. For example, a n/2 pulse produces a saturation of the energy levels and a % pulse produces a population inversion. The coherence of the radiation ensures that all transitions occur in phase with one another and therefore after a n/2 pulse the summation of all of the nuclear spin wavefunctions is the precessing vector M. The nutation of M b y a rf. pulse or pulses is the basis of the N M R experiment. A pulse can be generated by current passing through a coil around the sample (with internal coil diameter typically 2.5-14 mm). The precessing bulk magnetization can be experimentally detected using the same coil before the system returns to equilibrium.  1.1.3  FOURIER T R A N S F O R M N M R  2 , 3  The relaxation of the system to its equilibrium state is facilitated by perturbations to the Zeeman interaction (1.1.6). The exponential decay of the ^-component of Minduces a current, known as a free induction decay (FID), in the rf coil. Once equilibrium is re-established, the experiment or "scan" can be repeated in order to build up signal to noise (Figure 1.3). If n is the number of scans, the signal to random noise ratio (S:N) is proportional to the square root of the number of scans, i.e. S: N oc The FID contains time domain data,/ft). Conversion of these data to the frequency 6  domain f(co) is effected by a Fourier transformation: 6  (1.5)  Experimentally,/^ is recorded as a set of equally spaced, discrete data points and the Fourier transform is performed between finite limits:  /(®) = £ / ( ' ) e -t  4  ( t a , )  *  (1.6)  Equation 1.6 can be broken up into orthogonal real and imaginary components:  /(*>) = h (®) + if04 (®) = Z  +  (1.7)  3(0*  Recycle delay, m t =0  Y Z  M' ^  i  y  / x% V ,  Receiver (y)  FID (rotating frame)  / \  V  | ^ FID (static frame) r  Figure 1.3 Rotating frame diagram of a single pulse N M R experiment. A rf. pulse is applied along x and Mis  nutated aroundx towards the^-axis. Let t=0 be defined as the time at which Mis aligned alongy, at which time the axes of the rotatingframeand the staticframecoincide. The receiver is aligned along y and the FID appears as an exponentially decaying cosine signal in the staticframe.The rotatingframeFID results from the subtraction of co from the signal acquired in the staticframe.In this "on-resonance" case the rotatingframeFID is a simple exponential decay. The system returns to equilibrium during the "recycle delay" and the experiment can be repeated. L  This is the origin of the terms "real" and "imaginary" FID or spectra. When /(f) = e x p ( - r / r ) , s i n ^ O e x p ^ r / r ) or cos(fl>,f)exp(-f/r) the Fourier transform results in a Lorentzian function:  3  f (0)) R  =  0) +l  MyB{T \+  An T\co -o) y 2  x  5  Q  Absorption  (1.8)  l + 47t T (co -o) ) l  0  N M R spectra are displayed in their real forms (e.g. Eqn 1.8) as absorption signals. The Lorentzian lineshape results from the "perfect" N M R experiment on a sample containing a single type of nucleus. It is an excellent approximation for many experiments on liquids and forms a component of the frequency domain in almost all N M R data.  1.1.4  OFF-RESONANCE N M R  2  4  The rf. pulse must be able to simultaneously excite nuclei over a range of frequencies in order to excite nuclei in a range of environments. The Heisenberg Uncertainty Principle (Eqn 1.10) governs the relationship between pulse energy and time:  AE.At>h  (1.10)  Acv>— At  (1.11)  7  Therefore, to generate nuclear resonance frequencies spanning a typical experimental sweep width of lO'-lO kHz, for example, requires pulse widths of 10 -10 //s in order for 2  1  2  to be  approximately uniform across the entire spectrum. The frequency spread of the rf. pulse B (aj) is x  given by the Fourier transformation of that pulse, i.e. a sine function lasting for a time T (Figure 1.4). The maximum amplitude of B](a>) is proportional to TBI and therefore the condition that Bi(aJ) be approximately constant over an experimental sweep width becomes increasingly difficult to attain at smaller values of r. This condition may be unattainable in practice, particularly for solid state N M R experiments, which frequently require very large sweep widths.  6  Figure 1.4 Fourier transformation of a radio frequency pulse. The resulting distribution of B/ in the frequency  domain takes the form of a sine x function (sin xix) centred at (0/ with maximum amplitude proportional to TBJ. 8  In practice, a degree of self correction occurs. In the off-resonant rotating frame of reference, a proportion of Bo is introduced, SBo. The nutation of Mthen occurs around the vector sum of B] and SBo, #<# The magnitude of B^/is greater than the magnitude of Bi and therefore when CO] ± coo, this compensates for the decay in the amplitude ofBi(co) and spectral distortions that result from this kind of non-uniform excitation, known as "power roll-off, are rarely experimentally observed. 5 ^ does not lie along the same axis as Bj and this produces "phase errors" associated with the change in the orientation of the nutation axis as a function of co. This effect can be corrected for during data processing, but it is desirable that these errors be minimised. In order for both power roll-off and phase errors to be minimised, high rf. power is routinely employed for solid state N M R spectroscopy.  1.1.5  QUADRATURE DETECTION ' 4  9  The rf. coil detects a signal from the sample in the laboratory frame. N M R data in the rotating frame, the FID, are generated by the subtraction of the transmitter frequency from the signal. Since spectral sweep widths are usually less than 1% of the Larmor frequency, this greatly reduces the required rate of data sampling. However, using this method it is not possible  7  to distinguish between the frequencies (co + cd) and (co - co) and the two directions of x  x  precession of M with respect to co appear identical to a receiver in phase with co . To obviate the x  x  need to keep co to one side of the whole spectrum requires that data is sampled by orthogonal x  receivers. Such "quadrature detection" can be achieved by alternating the phase of the carrier frequency that is subtracted from the signal. Equivalent, orthogonal data sets are treated as real and imaginary components of the Fourier transform (Eqn 1.7) and real and imaginary spectra (Eqns 1.8 and 1.9) are separated by varying the linear combination of transformed orthogonal data sets, in a process known as phasing. Quadrature detection requires the signal to pass through different parts of the N M R spectrometer hardware to obtain the real and imaginary data. No spectrometer is capable of generating rf. pulses with perfect phase and amplitude control. Similarly, perfect control over the receiver phase and the subtraction of the transmitter frequency is not possible. To minimise the accumulation of these errors, N M R experiments are "phase cycled" with successive scans. Phase cycling simply involves switching the pathway of the rf pulse, the transmitter frequency and the signal through different permutations of equivalent spectrometer hardware components in order that each pathway is employed equally for all pulses. For example, consider a pulse sequence with successive pulses that must differ in phase by Till. The amplitudes of pulses with a phase of +x, -x, +y and —y will vary somewhat. The relative phases of +x and +y pulses will be different from the relative phases of +y and —x pulses, and so on. Starting successive scans with pulse phases of +x, -x, +y and - y will yield FIDs differing only because of the pulse imperfections, providing that the receiver phase also changes as +x, -x, +y and - y with successive scans. Cycling of the phases of the pulses and the receiver in this manner and adding the data from successive scans results in a FID that is influenced far less by any one combination of phase and amplitude errors than a FID composed of the same number of scans without phase cycling. This is the basis of quadrature phase cycling. The importance of phase cycling has diminished with the advance of digital pulse control, but it remains an important consideration for N M R spectroscopy, particularly for low sensitivity experiments. The specific phase cycling schemes employed in recording the spectra presented in this work are given in Appendix III.  8  1.1.6  PERTURBATIONS TO T H E N U C L E A R SPIN H A M I L T O N I A N  2 1 0 1 1  The local chemical environment of a nucleus influences the resonant frequency of that nucleus. In the strong magnetic fields used for N M R spectroscopy these effects are perturbations to the Zeeman interaction.  H =H  Zeeman  + H + H + H j + HQ a  D  (1.12)  Where; H  Chemical shielding  H  Dipolar coupling  Hj  Scalar coupling  H  Quadrupolar coupling  a  D  Q  Equation 1.12 represents all of the interactions present in the systems studied in this work. In general the perturbations are anisotropic. Whereas a vector property X can be described by three parameters, the anisotropic nuclear spin interactions require nine parameters for a full characterisation and are described by a 3x3 matrix, or tensor X. The nuclear spin system is an example of a special case in perturbation theory where all of the states of the system have the same Hamiltonian ( H  Zeeman  ) and to first order, all perturbations can be described by diagonally  symmetric tensors with just six unique elements.  Such tensors have two important properties;  the trace of the tensor is independent of the axis system and the tensor can be diagonalised, i.e. expressed with respect to a principal axis system (PAS) in which all off-diagonal matrix elements are zero. Following this transformation, the relationship between the lab axis system and the PAS is described by the three Euler angles (Figure 1.5). In this case, the transformation facilitates an additional simplification because the axial symmetry of the lab frame requires the use of just two Euler angles. In addition, the PAS usually has a significant chemical relevance, such as a shared axis with a bond. With the exception of H , these interactions only produce significant perturbations to v to Q  L  first order.  9  Figure 1.5 Graphical representation of a tensor property, E. The Euler angles a and prelate the lab axis system (x,y,z) to the PAS (a,b,c), in which the magnitude of E in any direction is the distancefromthe origin to the surface of an ellipsoid defined by the diagonal matrix elements P P and P . aw  1.1.6(a)  bb  cc  Chemical Shielding  Momentum of the electrons surrounding a nucleus is influenced by the applied field in the same way as the momentum of the nucleus, generating magnetic fields that oppose the applied field. The field experienced by the nucleus is reduced by a proportion equal to the chemical shielding, c o r chemical shift S, where 8= 1 - cr.  B = B (l-a) Q  10  (1.13)  N M R spectra are usually plotted in chemical shift units of parts per million (ppm) of the transmitter frequency v (the centre of the spectrum), because this scaling renders the frequency 0  scale independent of the magnetic field strength with respect to chemical shift. With respect to resonance frequency v, chemical shift in ppm is given by:  8 ppm  = 1 0  6  ^  (1.14)  The diagonalised chemical shielding tensor i i can be broken down into three components:  '1 *=(*)  o  0  0  N  1 0 +A 0  <o  0  K  0  ^2  ,0  +i 0  ^2  0 + 7jA 0  -I  ,0  0 1 2  0  (1.15)  0  Where; (*) = \Tr*  Isotropic chemical shielding  (1.16a)  A=<7  Anisotropy  (1.16b)  Asymmetry  (1.16c)  „  c c  -(CT)  bb-°'aa  tT =  The isotropic chemical shielding is a property that is independent of the orientation of the molecule with respect to the magnetic field and is the value of chemical shielding observed in liquid N M R spectroscopy when the rate of molecular motion is rapid on the N M R timescale. The "timescale" of any given interaction simply relates to the order of magnitude of the reciprocal of the range of frequencies (Hz) covered by that interaction. The utility of the anisotropy and asymmetry parameters (Eqns 1.16b and 1.16c) becomes clear when applied to the solid state N M R of powdered samples (Section 1.1.8).  11  When expressed in lab frame polar co-ordinates (Appendix I) chemical shielding 13  becomes:  c = (CT) + - A ( 3 c o s / ? - l + /7sin /?cosar) 2  (1.17)  2  Therefore the resonance frequency of a particular nucleus (Eqn 1.4) becomes:  v =v (l-(cr)) + - v A ( 3 c o s / 9 - l + 7sin /?cosQr) 2  L  1.1.6(b)  0  Dipolar Coupling  2  0  (1.18)  2  Dipolar coupling is the through space interaction between nuclear magnetic moments. The classical expression for the coupling energy between each pair of nuclei is:  D  4x\  r  (1.19)  r  5  3  9  ft  Where the vacuum permeability, /JQ = 1.26 x 10" mkgC" and r is the vector joining the two nuclei. Permeability is the constant of proportionality between magnetic induction (B, in units of Tesla) and the magnetic field strength (H, in units of Am" ). The quantum mechanical 1  expression for H  D  can be derived from Eqn 1.19 by substituting fi^ = -/„[„.  The non-  negligible terms of this substitution Eqn 1.20 share the same angular dependence. Additional terms result from the expansion of //./? in this special case where z is fixed, ' but they have a 14 15  small influence on the nuclear energy levels. k  ° f^{ =t  Iuilz  +  12  Va,  )}(l-3cos 0) 2  +  (1.20)  Where; =D  Dipolar coupling constant (Joules) ' 1 0 4nr  (1.21)  0^  0  1  0  0  0  (1.22)  The PAS of the dipolar coupling tensor is always coincidental with the lab axes and the dipolar coupling interaction is traceless and is therefore not observed directly in solution, where molecules are subject to rapid isotropic motion.. /, I _ +I _I j is zero for homonuclear coupling and non-zero for +  2  X  2+  heteronuclear coupling, such that for an isolated pair of dipolar coupled / = Vi nuclei:  L  ~o  y(  v  l)  + D i,2 (3 m  0-1)  c o s 2  Heteronuclear dipolar coupling (1.23)  L  V  ) =  v  o ' + 2 i,i ( cos 19 -1) }  Dm  3  2  Homonuclear dipolar coupling (1.24)  The dipolar coupling involving nuclei where I >Vi takes the same form but may have several values, corresponding to the possible values of mj.  1.1.6(c)  Scalar Coupling  Scalar coupling (or J-coupling) describes the interaction between nuclei via the local electronic environment. Whereas dipolar coupling is through-space, scalar coupling is through intervening chemical bonds. Coupling between electronic orbital angular momentum and electron spin with nuclear spin is transmitted throughout the electronic orbital to any other  13  nuclei within that orbital since the interaction with a nucleus modifies the local spin distribution within the orbital that must be compensated for elsewhere. J-coupling is usually much smaller than the dipolar coupling and its anisotropy is almost always ignored. Unlike dipolar coupling, scalar coupling does not average to zero under the influence of molecular motion and has a field-independent effect on the resonance frequencies of nuclei in liquid state N M R . This isotropic term is a result of Fermi contact interaction; the communication between nuclei as a result of non-zero electron density within the nuclei. Scalar coupling decays very rapidly with the number of bonds separating the nuclei and is rarely observed between nuclei separated by more than three bonds. In contrast to liquid state N M R , in solid state N M R spectroscopy, line widths are such that J-coupling is often not observed. In the present study, the effects of J-coupling are limited to its role in coherence transfer experiments (Section 1.1.10(d)).  1.1.6(d)  Quadrupolar Coupling ' 13  16 1 9  Nuclei with spin I > Vi are known as quadrupolar nuclei. Investigations of A l (1= 5/2) are 2 7  an important part of this thesis. Quadrupolar nuclei have an electric quadrupole moment, in addition to a magnetic dipole moment, that results from an asymmetric charge distribution in the nucleus. The classical interaction energy between an electric field gradient //and an electric quadrupole Q is given by:  U  0 Q  (1.25)  =-V.Q 6= =  17  The nuclear electric quadrupole can be described by the sum of two dipoles and, since the electric dipole moment of all nuclei is zero, the dipoles must be aligned antiparallel to one another, resulting in an axially symmetric quadrupole tensor:  2=—^—/ =  2/(27-1)-  14  Nuclear quadrupole tensor  (1.26)  Where Q is the quadrupole moment of the nucleus and eQ can be regarded as the charge distribution difference between the z-axis and the xy-plane of the quadrupole ellipsoid. By convention, a prolate quadrupole has a positive sign and an oblate quadrupole a negative sign. The PAS of the nuclear quadrupole tensor is aligned with the laboratory axes in a magnetic field. Note that Q and Q are not directly equivalent.  H  -  Q LV.I 2/(27-1) e  6  Quadrupolar Hamiltonian  (1.27)  Quadrupole tensor  (1.28)  Where; Q=  eQ 21(21-1)  This is something of a misnomer since it is really a scaled E F G tensor and should not be confused with the nuclear quadrupole tensor (Eqn 1.26). In a similar fashion to the chemical shielding tensor, the PAS of the quadrupole tensor often has a close relationship to the molecular or crystal geometry surrounding the nucleus. K itself is scaled with respect to the nuclear charge and by convention V = eq and V >Vbb>V a in its PAS (a.b.c), such that; cc  cc  a  '-i(l-i7 ) V = eq\  0^  0  0  l(l-/7 ) 0  0  e  0  0  (1.29)  h  Where;  ?Q  T  =  Vbb  V  aa  15  Asymmetry  (1.30)  Expanded in terms of its Cartesian components in the PAS:  iy-i=Kai;+v i;+Kj;  0.31)  bb  Due to the Uncertainty Principle (Eqn 1.10) only the total angular momentum, I and one component, £ can be specified. Therefore the results of the operators I and I must be expressed x  y  in terms of raising and lowering operators (Eqn 1.32) . 15  I =I +  + ii  X  y  i_=l -il x  y  Raising operator  (1.32a)  Lowering operator  (1.32b)  Using these relationships it can be shown that: 15  / . F . / = (3/ -I )v 2  (1.33)  2  The quadrupolar Hamiltonian can be expressed in the PAS Cartesian (Eqn 1.34) and laboratory frame spherical polar (Eqn 1.35) co-ordinates and to first order:  (3I -I )v 2  H  H = Q  g  /  ^  g  Q  =  2/(2/-1)  (1.34)  2  (3 cos B -1 + TJQ sin ft cos a%I -I ) 2  2  2  (1.35)  2  2  The conversion between Eqn 1.34 and Eqn 1.35 is achieved using the rotational matrix elements given in Appendix I . Therefore when considering quadrupolar coupling only: 13  =  V  °  +  8/(2/-^ ^ "  l  16  m  '^ ° C  S  2  P  ~  1 +  V q  S l n 2  P  C  °  S  (  1  3 6 )  This expression leads to two commonly used definitions, given here in units of Hz: 16  C =—~ h 0  v = 2^2/ Q  Nuclear Quadrupole Coupling Constant  (1-37)  Quadrupolar coupling frequency  (1.38)  i)  There are several important properties of Eqn 1.36:  For a quadrupolar nucleus with spin I, there are (21+1) values of the quadrupolar coupling, just as with dipolar coupling. The quadrupole tensor is traceless to first order (Eqns 1.28 and 1.29) and therefore there is no first order "quadrupolar shift", cf. chemical shift. When mi - A, there is no first order quadrupolar coupling at a l l because 1-2/M/ = 0. This l  20  provides an important distinction between quadrupolar nuclei with half-integer spins and those with integer spins, for which m / ^ A (Figure 1.6). l  The angular dependence terms of the first order quadrupolar coupling (Eqn 1.36) are identical to those of chemical shielding (Eqn 1.18).  The quadrupolar coupling interaction is often considerably larger than any of the other perturbations that have been discussed and it is common for the magnitude even of the second order perturbations to be similar to, or larger than, the dipolar coupling or chemical shielding. Therefore, the properties of second order quadrupolar coupling must also be considered (Figure 1.6). Application of second order perturbation theory to the interaction of tensor properties is equivalent to the examination of the interaction between symmetrical nuclear quadrupole and electric field gradient tensors each with 5 uniquely defined elements, i.e. a matrix with six nonzero matrix elements, each of which can be expressed in terms of the other five and where T p = Tp . In general such tensors can always be broken down as T = T +T l  a  a  17  l  +T  2  where;  T = \Tr(T), L is traceless and diagonal and Z is symmetrical but not diagonal (cf. J ) . 1  u  2  2 1  This  separation produces the same results as first and second order perturbation theory, respectively. The full second order quadrupolar Hamiltonian can be expressed in Cartesian co-ordinates by expansion of Eqn 1.34:  zQ  \K^L-I )+(V -V il -I ) 2  2  aa  4/(2/ -1) |  2V (lj  +  ab  bb  2  x  y  + I I )+ 2V (lj  y  y  x  ac  + IJ )+ 2V (lj  z  X  bc  + lj  2  )  y  (1.39)  A substitution for the raising and lowering operators and a change from the rotational PAS to laboratory frame yields the spherical polar form of Eqn 1.39:  2  2  -«#< > =^-/ (o7 - 4 / + l ) Q ( a , / ? , ; 7 ) + 2 ^ / ( 2 / _ / 2  A  2  z  2  2  z  e  z  2  2  +l)n'( r,/?,i ) fl  7Q  (1.40)  Where v and v are in Hz and the Hamiltonian yields energies in Joules. The functions Q  0  Cl{a,p,ri ) and Q  contain the spatial information. The second order terms'  Q'(a,fi,T] ) n  contributions to the transition frequencies are:  Vn  / _  .  /  .\  \  \  2V  v%l _ = -^-{2Am,(m, - 1 ) - 2 / ( 2 / +1)+ 9)Q + —^-(6m,(m, -1)-2/(2/ m  r  x  +1)+3p' (1.41)  The most important transition in the solid state N M R spectroscopy of quadrupolar nuclei of non-integer spin is the central m, = Vi++- /i transition as it is not influenced by first order x  quadrupolar coupling and is therefore considerably narrower than all other allowed transitions. As a result the central transition offers the highest resolution and the greatest sensitivity, to the extent that it is often the only resonance that can be observed for these nuclei.  18  (First order)  (Bo) H = H Zeeman  (Second order)  + HQ(i)  H - H  ZEEMAN  _5  H =H  „+H$  ZEEMA  HV  +  Q  v +2v£  2'  0  _3 '  2'' v  I= 5  , 2 '/  21+1 =5  +  V  ( D  +  (  V  2  )  • 2 1  \  d e g e n e r a t e states  .  V  °  +  V  X . - X  + —•  2  _  V  v -2v  m  v  n  0  V  0 )  zv  +  v  ( 2 )  e  m, = + —^  1 )  (J  »vi  ,  2 ) m  m^mj-l  Figure 1.6 Schematic representation of the sequential effect of first and second order quadrupolar coupling (  and  ) on the resonance frequencies of a spin / = 5^ nucleus. Without quadrupolar coupling, the  energy levels are equally spaced. First order coupling modifies the resonance frequency by an amount proportional to {mi -'A) (Eqn 1.36) and therefore the central transition remains unaffected. All transitionfrequenciesare affected by second order quadrupolar coupling (Eqns 1.41-1.43). The figure cannot convey the orientation dependence of all mi->mi- 1 transitions, but the magnitudes of the frequency perturbations represent the potential frequency spread of each transition as a function of the spatial orientation of the sample.  The resonance frequency for non-integer spin quadrupolar nuclei, considering only the quadrupolar coupling and the Zeeman interaction, is given by:  vie  • 1(1 + l))(A(a,n )cos ft + B(a, tj )cos p + c{a,rj )) 4  Q  19  2  Q  Q  (1.42)  Where:  . 81 27 „ 9 „ A = — +—Tj cos2or + — 7 7 c o s 2 « 16 8 16 2  n 2  (1.43a)  2  n  6  45 3 9 5 = - - - + - 7 + 3 ^ c o s 2 « + —77 cos 2 a  (1.43b)  ^ 9 1 3 C = — + -Tj +-rj cos2a  (1.43c)  2  o  2  e  4  e  2  Q  e  2  e  to  9 , + -Ti cos 2a 2  2  Q  The relevant characteristics of the second order quadrupolar coupling are:  The magnitude is inversely proportional to magnetic field strength. The anisotropic terms of second order quadrupolar coupling are different from anisotropic terms of the first order perturbations H , H a  and H . X)  D  Q  The second order anisotropy is not traceless and therefore provides a contribution to the isotropic chemical shift that is inversely field dependent:  V  1.1.6(e)  M  V  30  4  Jv  1+  (1.44)  0  Conclusion  In addition to their effect on resonance frequencies, by which the majority of the useful chemical information is conveyed, these interactions (1.1.6(a)-(d)) provide the mechanism for and determine the rate at which the system returns to equilibrium, the process of spin "relaxation", discussed below. N M R spectra of liquid samples are usually characterised by narrow, symmetrical resonances. If the random reorientation of species in solution is rapid on the N M R timescale, as 20  is commonly the case for small molecules in solution, all of the interactions are reduced to their isotropic (i.e. trace) values and the resonance frequencies contain information about the isotropic chemical shift, the scalar coupling and the isotropic second order quadrupolar coupling only. Dipolar coupling information and anisotropic information cannot be extracted from the isotropic peak positions, but where present, these terms do contribute to relaxation and will indirectly affect the appearance of the spectrum. N M R spectra of solid samples by contrast are generally broad, with far less resolution between signals from nuclei in different environments, but containing information from all isotropic and anisotropic terms. In addition, nuclei in the solid state may have significantly different relaxation properties from nuclei in the liquid state.  1.1.7  RELAXATION " 2  4  The relaxation properties of the nuclei in a sample have profound experimental implications. They affect the S:N that can be achieved in a given time, the choice of experimental conditions (e.g. extremely rapid transverse relaxation may preclude the use of a spin-echo experiment (1.1.10(b))) and differences between the relaxation properties may facilitate the separation of overlapping N M R signals. Relaxation is the result of nuclear spin state transitions stimulated by fluctuations of local magnetic fields at the resonance frequency. Molecular motions and crystal vibrations cause relaxation via the distance and angular sensitivity of dipolar and quadrupolar coupling and chemical shielding interactions. Even where these interactions are isotropic, asymmetric vibrations cause relaxation. Translational molecular motion and sample rotation produce relaxation as the nuclei experience different regions of an inhomogeneous magnetic field and chemical exchange processes stimulate relaxation in the nuclei directly affected and also in nuclei to which they are coupled.  21  Two relaxation time constants are defined to describe the approximately exponential recovery of M:  dM  "Longitudinal" or "spin-lattice" relaxation time  "Transverse" or "spin-spin" relaxation time  dM  x  _  d  M  dx  0  ^ _  (M -M ) x  0  by  ;  M  7  k  /  X  2  dz  iii)  ^  {M -M )  2  —  >~y X  (b)i)  iii)  ii)  /  M  ^ 7/  3 -  y  •y  y  7-  X  X  Figure 1.7 Position of bulk magnetisation in the rotating frame of reference (i) before (ii) during (iii) after relaxation, when (a) T = T, and (b) T,» T . 2  2  Longitudinal relaxation results from changes in the populations of the nuclear energy levels. It is also called "spin-lattice" relaxation because energy is dissipated from the nuclear spin system to the surroundings, or "lattice", in the form of thermal energy. This non-adiabatic process is a result of fluctuations in the x andy components of the magnetic field. Consider a system of identical spins subject to random fluctuations in the local magnetic field (i.e. fluctuations in the magnetic field average to zero over time) characterised by the following properties:  3  f  G(T) -  exp  ^  Autocorrelation function  22  (1.45)  J(a) =  2r  c  G(r)  Spectral density function  (1.46)  The correlation time (r ) is the average time of one fluctuation, with r.m.s. amplitude c  B =B ,B i  x  y  or B and frequency co. r can only be defined precisely with respect to a particular z  c  contribution to relaxation, e.g. the average time for a molecule in solution to rotate through one radian, the inverse of the rate constant for a molecular/atomic jump motion in a solid. G(r) (Eqn 1.45) is a dimensionless function that measures the "persistence" of a fluctuation f(t). G(T) = f*(t + r)/(f) and is the "overlap" between the fluctuation at time t and at some timer later, averaged over all values of t. When the property r is defined and G(r) is assumed to take c  the formG(r)= e x p ^ - ^ / j , then Eqns 1.46)-1.48 result. J(a>) is a physical property of the system and represents the power associated with motions in the system as a function of frequency.  G(r) =  Bfj(co )  2 r  0  (1.47)  Transverse relaxation contains a non-adiabatic contribution identical to transverse relaxation and an additional adiabatic contribution. The latter is a result of "flip-flop" transitions; a change in the state of one nucleus as a result of an opposite change in the state of another nucleus with zero change in total energy. Assuming that r « T?. c  (1.48)  T2 contains an adiabatic "spin-spin" and a non-adiabatic "spin-lattice" term, but whenever the values of T\ and T2 are significantly different, it is the spin-spin term that dominates T2. Eqns 1.47 and 1.48 are not quantitative for real systems, but they do provide a qualitative understanding. The important characteristics of relaxation are:  23  The greater the amplitude of the fluctuation, the greater the rate of relaxation. Quadrupolar coupling relaxation is often very rapid. Nuclei of heavier elements tend to have faster relaxation times and nuclei in asymmetric environments have more rapid relaxation than nuclei in cubic environments (all other factors being equal). Relaxation is temperature and field dependent. Figure 1.8 schematically illustrates the variation of relaxation time with correlation time, due to the temperature dependence of molecular motions assuming a dipolar coupling based relaxation process.  log7/„  Increasing temperature  \ s  \  \  \  \  \  \  s  ---•7/2  1  l0gT  COoTc^l  c  Figure 1.8 Schematic representation of the relationship between relaxation time and correlation time. r 2  c  decreases with increasing temperature. Solid state NMR at and around room temperature is most often in the region to the right of the T minimum, such that a decrease in temperature results in an increase in T and a decrease of T . Relaxation in liquid state NMR spectroscopy is usually dominated by dipolar coupling and frequently lies to the left hand side of the minimum in T where T\ « T and the already rapid fluctuations' increase in rate with increasing temperature cause T\ and T to increase. The position of the T minimum is a function of the magnetic field. {  {  h  2  2  2  {  The final important consequence of relaxation is lifetime broadening; another consequence of the Uncertainty Principle. A resonance has an "intrinsic" linewidth determined by its spin state lifetime, defined by T . When T is substituted into Eqn 1.8 the expression for the lifetime2  2  broadened line shape is obtained. In reality, the observed rate of transverse relaxation " T " is 2  24  more rapid than T due to contributions from magnetic field inhomogeneity and fluctuations in 2  the magnetic susceptibility of the sample and/or its surroundings. Where Acou is the frequency width of a Lorentzian peak at half height:  A &), -  1  (1.49)  nT  2  In the solid state this intrinsic linewidth is usually far smaller than broadenings due to distributions in environments in amorphous materials or crystalline imperfections. Lifetime broadenings are therefore usually only considered in liquid state N M R .  ,2,22  1.1.8  SOLID STATE N M R OF STATIC POWDERED SAMPLES  It is not usually possible to obtain single crystals of the materials studied in this thesis that are large enough to perform single crystal solid state N M R . In the more general case, N M R spectroscopy of single crystals is time consuming due to the large number of crystal orientations for which spectra must be acquired in order to gain an understanding of the system. Solid state N M R is therefore normally performed on powdered samples and this is the case for all of the materials and experiments described in this thesis. In the absence of the isotropic molecular tumbling seen in liquid N M R , static solid state N M R spectra are the sum of approximately Lorentzian signals from individual crystallites summed over all possible crystallite orientations to produce characteristic powder lineshapes. The simple case is that of a polycrystalline sample with a random distribution of crystallites consisting of uncoupled nuclei in one axially symmetric environment, i.e. cr^ =  < <r and r\ zz  = 0. If A vis the frequency distance from the isotropic value v , then the anisotropy takes the iso  form:  2  Where v A is the frequency distance from v 0  iso  25  and A is given by Eqn 1.16(b).  If  dp  dAv  = g(JA v|), changing variables from sin/7 to A v yields:  gM - T 7 7 vA  Av-6  (1.50)  0  The function g(JA v|) is the "angle density" per frequency unit at a given frequency distance A v from v and is therefore proportional to the number of crystallites per frequency unit and to iso  the signal intensity (hence the constraints used to keep the value of g(JA v|) positive). The maximum and minimum values of A v correspond to discontinuities in g; one an asymptote and the other imposed by the physical model (Figure 1.9). Note that A can take positive and negative values and the lineshape may appear reflected about zero. When g(JAv|) = 0, v = v  iso  and the  maximum and minimum values of v occur at v (1 - <r ) and v (l - tr ). Therefore the diagonal 0  aa  0  cc  elements of the C S A can be taken directly from discontinuities in the powder lineshape.  *(M)  vA  vA  0  0  F i g u r e 1.9 P l o t o f g(JA v|) a g a i n s t A v.  26  A similar treatment for the rj # 0 case yields a slightly more complex result because the function g(JA \) v  m  u  s  t  D e  the sum of two partial derivatives over each of the two angles.  However, in all cases the discontinuities in the spectra of powdered samples that are dominated by one of the anisotropic interactions are either equal to, or functions of, the principal matrix elements of that interaction. Examples of first order powder lineshapes are given in Figure 1.10. The coupling patterns arising from dipolar coupling are analogous to those of scalar coupling, but in practice it is only possible to resolve coupling parameters (and therefore internuclear distances) from the most simple spin systems. When the spin system includes an abundant N M R active nucleus, dipolar coupling can reduce the spectrum to a broad, featureless peak; a result of the superposition the signals resulting from many pairwise, anisotropic interactions. Heteronuclear dipolar coupling can be eliminated by broadband decoupling at the Larmor frequency of the coupling nuclei. This removes the effects of dipolar coupling by nutating the nuclei at a rate that is rapid on the dipolar coupling timescale. The technique of magic angle spinning ' (discussed in Section 1.1.9) was devised to remove /jomonuclear dipolar coupling. 24 25  For non-integer spin quadrupolar nuclei, when the magnitude becomes large (Figure 1.10(e)) the central transition becomes increasingly dominant and the discontinuities in its second order quadrupolar powder lineshape, ' centred at 26 27  parameters.  27  Vy°™y,  can be used to extract tensor  Figure 1.10 Examples of simulated first order powder lineshapes plotted on an arbitrary chemical shift scale.  (a)i) Isotropic, ii) axially symmetric and iii) asymmetric CSA lineshapes, characteristic of the values of the asymmetry A and the anisotropy 77 quoted beneath each spectrum, (b) A "Pake doublet" resulting from one of a pair of dipolar coupled spin/= 54 nuclei. X = D in the case of heteronuclear coupling and X = D/2 in the case of homonuclear coupling. The quadrupolar coupling anisotropy of a spin / = 1 nucleus produces the same powder lineshape and X = % Discontinuities occur at gi1 and Q , cf. axial CSA. (c) Powder lineshape resulting from a spin / = 1 nucleus with asymmetric quadrupolar coupling, (b)-(c) Quadrupolar coupling frequencies of VQ and - VQ have the same powder lineshapes. (d) Powder lineshape resulting from a nucleus dipolar coupled to two equivalent, non-interacting nuclei. X) and X are defined as for (b). The spectrum of a spin / = 54 nucleus dipolar coupled to a single spin / = 1 nucleus consists of superimposed Pake doublets separated by D and 2D. (e) Spin / = 5/2 axially symmetric powder lineshape. 23  }i  2  28  1.1.9  M A G I C A N G L E SPINNING ' ' 2  9  28  The widths of solid state N M R resonances are in general much greater than the differences in isotropic chemical shifts and the resolution of resonances from nuclei in different environments can often be difficult or impossible. The technique of magic angle spinning (MAS) is ubiquitously employed to partially or fully remove the effects of anisotropic interactions from solid state N M R spectra. This technique of rotating the whole sample at the "magic" angle of 54°44" to B at a rate 0  that is rapid on the N M R timescale, removes all first order anisotropics and produces the 24  25  isotropic lineshapes. Attainable spinning rates may less than or comparable to the anisotropy timescale and it is important to understand the effect that this has on solid state N M R spectra. The nuclear spin Hamiltonians (Section 1.1.6) are time independent. Although the nuclear spin system evolves with time (e.g. precession around z, relaxation, response to rf. pulses) the energies are time independent (in the absence of through-space motion of the nuclei). In this general case the time dependence of the Hamiltonian is separable  from the time-independent  Hamiltonians described above. The nuclear energy levels then take the form f[x,y,z)g(t) and g = 1 in the absence of external perturbations (e.g. molecular motion, rf radiation, sample rotation). The anisotropics (Eqns 1.18, 1.23, 1.24 and 1.36) are linear combinations of the second order spherical harmonic functions (Appendix II). /th order spherical harmonics Y 14  / m  can be  expressed as a linear combination of the /th order spherical harmonics in any other axis system:  YjM=£a ,Y,>,/?)  (1.51)  m >  Where a  m>M  can also be expressed in terms of the basis set (a result specific to spherical  harmonics) The expression for C S A (Eqn 1.18) can be re-cast in the lab frame as: 28  1  f  * r  V ^2,0  29  Y  + ^2,+2 ^2,-l) ^2,±2 +  2 0  (1.52)  Where Tv^.m are normalisation constants. Eqn 1.52 can be now be expressed in terms of Y  2 m  in the sample spinning axis system at an angle 0 to z, where 0 = J3+0 and q> = a+(0-Qt). A l l R  R  Y {0,(p) can themselves be expressed in terms of Y {<f>-Qt,0 ) and these two changes of 2m  2m  R  axis system give the result:  i  r.  /  ^fc  0  j  m  Y ^-fi^j}  2,0 I m  ^  +  2  J  iV  ^  -' 2,±2 I m'  J  V  (1.53) Eqn 1.53 contains five types of angular terms, from m = 0, ± 1 , ±2.  Y , 2  =^ |^sin" ^exp{±/m(^-QO} ,  ± m  7  v 1  2,0  =  m^O  ^-(3cos 0 -l)  (1.54)  (1.55)  2  s  The exponential time dependent terms (Eqn 1.54) oscillate in phase and have time averages of zero. Therefore i f Qt» v A all Y 0  Y  2 0  2± m  {0 ,<f>- Cit) vanish with any value of 0 . The remaining R  R  terms vanish when 0 = 54.74°, the magic angle. R  1.1.9(a)  The Refocusing Effect of M A S ^ Z4  Now consider a sample rotating at a rate that is slow on the C S A timescale. Following a n/2 pulse, the magnetisation precesses around z and decays approximately according to exp(- t/T ') 2  providing that Q" » T *. The rate of rotation is then so slow as to have a negligible effect on 1  2  the resonance frequency of each nucleus during relaxation and the signal Fourier transforms to give a normal powder lineshape. However i f Q" « T the magnetisation oscillates 1  2  approximately according to exp(-/Q/) and refocuses when t = n—  30  and the FID becomes a  train of "rotational echos" (Figure 1.11). The rise and fall of intensity of each echo is Z4  approximately independent of the contribution to T* from C S A and the envelope of echo intensities decays according to a value of T * that has been modified by the sample rotation. The 2  FID now contains the same information as the FID from the static sample, but with an additional periodicity.  The Fourier transformation of the whole echo train yields a spectrum consisting of  isotropic peaks with widths determined only by isotropic chemical shielding induced relaxation, plus distributions of chemical shifts, separated by Q/2n Hz with intensities under the envelope of the static powder lineshape (Figure 1.11). The envelope of intensities of the sideband manifold ceases to describe a static powder lineshape in the intermediate regime, when Q" « T *, because there is insufficient time for the 1  2  full decay of the rotational echoes. Most experimental M A S conditions lie within this intermediate regime. The spectrum still breaks into narrow sidebands (becoming progressively dominated by the isotropic peak) and still contains C S A information. The total spectral intensities under static and M A S conditions are equal. Therefore M A S greatly improves resolution and S:N. These arguments are applicable to all of the first order interactions. However, there are two important cases requiring further discussion; dipolar coupling to an abundant spin system and second order quadrupolar coupling.  1.1.9(b)  Effect of MAS on Dipolar Coupling to an Abundant Spin System.  Nuclei that are both chemically and isotopically abundant comprise an abundant spin system, i.e. a large number of N M R active nuclei of a given isotope per unit volume. The most common examples of this are homonuclear and heteronuclear dipolar coupling to ' H . Spectra of nuclei dipolar coupled to an abundant spin system do break into sidebands but do so gradually.  30  Different dipolar couplings will refocus at different sample orientations, although the respective echo maxima will still be Q" s apart. A difference in time in the FID Fourier transforms to a 1  frequency difference in the spectrum. Therefore the powder M A S spectrum at intermediate spinning rates is a summation of spinning sideband manifolds centred at v (Figure 1.12). 0  31  ~~1  4Q  I  I  I  2Q  1  v  lm  1  1  -2Q  1  1—  -4Q  Figure 1.11 Schematic representation of slow MAS NMR spectroscopy, (a) The FID is comprised of a train of rotational echoes separated by Q" seconds. Echo maxima are contained within an exponential decay envelope defined by the MAS relaxation time T * • (b) The MAS spectrum consists of a "spinning sideband manifold" of 1  2  Lorentzian "«th order sidebands" at frequencies of «Q Hz from the isotropic peak, always at the "zeroth order sideband").  32  v  iso  (sometimes called  Figure 1.12 Progressive changes in the NMR spectrum of a nucleus dipolar coupled to an abundant spin system as a result of increasing MAS rate, (a) Static spectrum, (b) Slow spinning; Cl' »  r *. (c) Intermediate  spinning rate; Q~ «  T* . The static  1  (d) Intermediate spinning rate; Q  l  _ 1  < T * (e) Fast spinning; Q~' « 2  2  t  spectrum is broad and featureless. As the rate of MAS increases, spinning sideband regions begin to resolve and the signal coalesces at v when v s is greater than the largest dipolar coupling. M  0  1.1.9(c)  The Effect of MAS on Second Order Quadrupolar Coupling.  The majority of the work presented in this thesis involves 2 7  131618  i y  A l M A S N M R spectroscopy.  A l is a spin / = 5/2 nucleus and therefore the effects of M A S on the second order quadrupolar  coupling of the ntj =  + /2—->-54 1  transition is necessary for an understanding of the present work.  Indeed, the great majority of all N M R active nuclei are non-integer spin quadrupolar nuclei. Consider the central transition of a non-integer spin quadrupolar nucleus rotating at a rate that is rapid on the quadrupolar coupling timescale (i.e. v  ms  > v ). This resonance is broadened Q  by the second order quadrupolar coupling only, because the first order quadrupolar coupling is zero for the central transition (Eqn 1.36). Eqns 1.42 and 1.43 can also be expressed in terms of spherical harmonics and if the first of the two changes of axis system described in Section 1.1.9  33  are performed and the time dependent terms discarded, the second order quadrupolar coupling takes the general form:  X+X  V  v  o  ^  ;  Jv [+{y(^,^)+z(/7 ,^)cos4c?}p (cosc9j 0  Where ( X - / ( / - l ) ) ^ i ? ( 7 e ) = y°™y  e  7  v  2  (  E c m  4  J  1-44) and P„(cos6?«) are «th Legendre  polynomials: P (cos^) = i ( 3 c o s ^ - l )  (1.57)  2  2  P [cos9 ) = -(35cos 9 -30cos 4  4  R  2  R  30.56°  54.74°  0) R  (1.58)  70.12°  Figure 1.13 Plot of the 2nd and 4th Legendre polynomials between 0° and 90°  There is no value of OR where both P2 and P4 are zero (Figure 1.13). At the magic angle, the P2 terms vanish and when Eqn 1.56 is evaluated, the second order quadrupolar coupling takes  34  the same form as Eqn 1.42 but with functions A', B' and C" (Eqn 1.60) in the place of A, B and C (Eqn 1.42):  / ( / + l)L'(0,7fi)cosV  + 5 (^7  ) c o s ^ + C (0,7 )) (1.59) 2  ,  G  ,  G  Where:  A  B<  n cos 20 + — n i cos 20 8 48  (1.60a)  2  n  16  ~8  e  +  12 ^  77,,  16  e  8  +  7  e  C 0 S W  COS20  2  H  " 24 ^  77  48  2  COS  C 0 S 2  2  2 6 ?  20  (1.60b)  (1.60c)  e  Thus, second order M A S anisotropy takes the same general form as the static anisotropy and is referred to as a "reduced" second order quadrupolar anisotropy. The isotropic second order quadrupolar shift is at a higher frequency than the whole of the reduced anisotropy (Figure 1.14). The best resolution between the N M R resonances of non-integer spin quadrupolar nuclei is not achieved at the magic angle, but it is normally used because of its averaging effect on first order anisotropics, most importantly the CSA. These expressions for the M A S second order quadrupolar anisotropy and the powder lineshapes that they produce (Figure 1.14) assume an infinitely fast spinning rate. Second order powder lineshapes and the isotropic quadrupolar shifts are dependent on m,, but the differences in isotropic shifts are often smaller than the width of the anisotropies and the isotropic peaks and the spinning sidebands of the M A S N M R spectra of quadrupolar nuclei overlap. Due to the different widths of the central and each of the satellite transitions, quadrupolar lineshapes depend on the spinning rate. However, for non-integer spin quadrupolar nuclei, lineshapes quickly become close to the infinite spinning speed approximation due to the dominance of the central transition. 35  2  2  Figure 1.14 Simulated M A S powder spectra of the central transition of a non-integer spin quadrupolar nucleus. Spectra were simulated using the same values of VQ and the value of T]Q indicated in the figure.  Although the relationships to quadrupolar coupling parameters (v and rf ) are more Q  Q  complex than is the case for first order interactions, all of the discontinuities in the reduced second order quadrupolar lineshapes reflect the coupling parameters.  36  1.1.10  N M R EXPERIMENTS  1.1.10(a)  Single Pulse Experiment.  This is the simplest N M R experiment and has already been described in Section 1.1.2. There are some additional considerations particular to quadrupolar nuclei that must be taken into account in single pulse experiments. In principle, the complete M A S spectrum of a non-integer 18  spin quadrupolar nucleus is a complex summation of the central transition, 21 satellite transitions and all of the sidebands associated with each transition. In practice the intensity of the satellite transitions is divided between many spinning sidebands, because of the magnitudes of the first order quadrupolar couplings, such that satellite transitions are quite difficult to observe. The behaviour of quadrupolar nuclei in the solid state is dependent on the comparative values of v  Q  and Vrf, where v  rf  = yB j2n. There are three regimes: x  v » v : A l l of the transitions have the same nutation frequency. In the solid state this rf  Q  only applies to very small quadrupolar couplings. v « v : The rf. pulse is selective for the central transition only, due to the effect of rf  Q  power roll-off (Figure 1.4) and the quadrupolar nucleus exhibits "pseudo spin / = Vi" behaviour. In the solid state, because all of the absorbed radiation results in w/ = Vz<-^Vi transitions, the effective amplitude of the radiation, \B |, is greater than its value in X  solution by a factor equal to the number of transitions, 2/4-1, leading to the concept of a solid pulse length r& where 6\s the pulse angle:  (liquid)  T« =TT77 0!id)  a  n  d  vy>=(2/ + l>/J>  (1.61)  v &v ; The most common regime encountered experimentally. The satellite x  Q  transitions have longer solid state pulse lengths and are discriminated against in experiments optimised for the central transition. Eqn 1.61 is not accurate and therefore  37  the pulse length is a function of Vrf. It is for this reason that the pulse lengths and powers of quadrupolar nuclei are quoted in terms of the liquid state value. In order to obtain quantitative spectra, small tip angles (~ 12° liquid pulse) are employed in order to minimise the differences between pulse angles on resonances with different quadrupolar couplings.  The N M R signal is detected by the same radio frequency electronics used to generate the rf. pulse. Current generation in the magnetic field causes vibration within the electronics that themselves induce current. This "acoustic ringing" is extremely short in duration (ca. 5-30//s) and if present in the acquired data will Fourier transform as a broad signal which appears as a distortion of the spectral baseline. Therefore a minimum "ringdown delay" must be inserted between the final pulse and the acquisition. T values in solid state N M R can be so small that a 2  significant amount of signal from the sample is lost during this time, which is manifested as loss of intensity, distortions of the lineshape and a type of baseline distortion resulting from the assumption that the start of the acquisition is in fact t - 0. The single pulse experiments can often be a trade off between these different types of distortion. The spin-echo provides a solution to these problems and is useful for systems with long enough T 's. 2  1.1.10(b)  Spin-Echo Experiment  The spin-echo experiment provides a means of moving the signal clear of the ringdown period. Two rf. pulses, separated by a delay r (which is rotor synchronised under MAS), are followed by the formation of an echo after an additional time r (Figure 1.15). Fourier transformation from the echo maximum onwards yields an undistorted spectrum. Consider a system consisting of a single C S A broadened resonance. Observed from a rotating frame at v , following a x/2rf. pulse iso  will separate out into diverging component  vectors according to the CSA. Vectors with frequencies greater than v  iso  will precess  anticlockwise from y and vectors will frequencies lower than v will precess clockwise from y. Uo  After a TC pulse of the same phase at time r later, the magnetisations will still precess  38  anticlockwise and clockwise respectively, which now causes the vectors to converge after a further time r at -y. Therefore, the effects of C S A have been delayed by 2r. Transverse relaxation takes place during both r delays. After the first delay the magnetisation vectors have tilted out of the xy-plane and the total magnetisation can be represented by vectors in the xy-plane originating from the initial 7d2 pulse and a component M  z  due to longitudinal relaxation during r. The n pulse orients M along - z and it decays to zero z  during the second r delay. In practice, pulse imperfections contaminate the FID with components of M , but these can be eliminated by alternating the phase of the n pulse to +x and z  -x with successive scans. The effects of magnetic field inhomogeneity behave in the same way as CSA, provided that the inhomogeneities felt by each nucleus are identical during both r delays. This condition can be imposed by making ran integer number of M A S rotor cycles such that the magnitude of the echo decays according to the "true" T2 rather than T *. 2  Quadrupolar coupling is also refocused in the 90°-180° echo experiment and it is 31  particularly useful when quadrupolar relaxation prevents the acquisition of undistorted single pulse experiments. The spin-echo experiment was originally observed and named by Hahn to measure T by 32  2  varying r and is also known as the Hahn echo. When T » r , the ^pulse may be applied several 2  times to create an echo train. This modification, the Carr-Purcell experiment, enables the 33  evaluation of T2 within a single measurement and represents the first appearance of the 90°-180° spin-echo. The same result is obtained if the (n) pulse is changed to a (n) pulse (the Carrx  y  Purcell-Meiboom-Gill ' experiment) with the advantage that the experiment is then far less 33 34  sensitive to the accuracy of the pulse lengths.  39  Figure 1.15 Rotating frame diagram of 9 0 ° - 1 8 0 ° spin-echo experiment. Refer to text for discussion.  1.1.10(c)  The Inversion and Saturation Recovery Experiments.  In order to obtain quantitative N M R spectra, the value of T must be known. The inversionx  recovery experiment is used to measure T\ for rapidly relaxing systems (Figure 1.16(a)) and the saturation recovery experiment is more appropriate for the measurement of longer T\ values (Figure 1.16(b)). A train of pulses saturates the system prior to the r delay and therefore full relaxation of the system is not required between scans. For both experiments, signal intensity as a function of rcan be fit to an exponential function (given in figure) in order to evaluate T\.  40  fe v^y±*  1 /l(r)  M  (b)  =  7^.  M  0  FT  l-2exp  -y  Receiver (±y)  [(2)1x^^1,  -y  Receiver (±v)  I(T) =  M  0  l-2exp  Figure 1.16 Rotating frame diagram of (a) the inversion recovery experiment and (b) the saturation recovery  experiment, (a) M is inverted along -z and the signal sampled after a variable delay r. Intensities as a function of r can be fit to the exponential relationship given in the figure. The only phase cycling constraint is that the receiver be orthogonal to the final pulse, (b) Till pulses are applied in rapid succession to saturate nuclear spin states. The signal is allowed to recover for a time rand then sampled. Signal intensity can be fit to the same relationship. This method obviates the need to wait for &5T\ between increments of r.  41  1.1.10(d)  The INEPT Experiment.  The insensitive Nuclei .Enhanced by Polarisation Transfer experiment ' (INEPT) was devised as a means of increasing liquid N M R signals by the transfer of magnetisation from a sensitive nucleus (e.g. H) to a less sensitive nucleus (e.g.  C) via the scalar coupling. More  recently, the experiment has found use in solid state N M R as a means of determining bonding connectivities between nuclei.  39  Consider an isolated heteronuclear pair of spin 1= 14 scalar coupled nuclei. The total nuclear spin system consists of 4 energy levels, labelled in Figure 1.17 according to the spin states of the uncoupled nuclei. The equilibrium population difference corresponding to a 35  transition of frequency v is A„. Since v <*>v' , it is assumed that A « A' . n  n  n  n  n  If a nil pulse is applied at v , the population of all energy levels is changed; the (14,14) and 2  (-14,14) states' populations have changed by -A /2 and the (14,-14) and (-14,-14) states' populations 2  have changed by +A /2. If the system is manipulated to invert the population of just one of the 2  transitions, v then the (-14,14) population will have increased by Ai and the (14,14) will have u  decreased by A i . The result of this inversion is to change the population differences corresponding to v and v to (A +Ai) and (A -Ai) respectively. By this mechanism, the bulk 2  2  2  2  magnetization of nucleus 1 has been transferred to nucleus 2. If Ai»A2 the intensity of spectrum 1 is increased by this polarisation transfer (or "coherence transfer"). The INEPT experiment achieves the selective inversion of one transition of a doublet using a heteronuclear spin-echo sequence that mimics the homonuclear case and allows the magnetisation vectors to diverge due to scalar coupling before transferring them back to the z axis (Figure 1.18). Following a nil pulse, two components of M i diverge at a rate of J/2 Hz, where J is the scalar coupling. After a time r a n pulse of the same phase places the components of M i in the opposite half of the rotating frame. If a n pulse is applied simultaneously to nucleus 2, the components of M i also swap places and continue to diverge. If r = 1/4./, the components are aligned along ±x in opposite directions at the time that chemical shifts refocus. Simultaneous nil pulses transfer magnetisation from nucleus 1 to nucleus 2 and into the xy-plane, where it can be detected.  42  Figure 1.17 Energy level diagram of a coupled heteronuclear spin / = 'A two spin system in a magnetic field.  Transitions v  n  and v' correspond to the peaks in scalar coupled doublet in the spectrum of nucleus n. v n  « v'  n  n  and to a first approximation produce equal population differences between the energy levels An, and a spectrum of nucleus n consisting of a doublet (two peaks of equal intensity).  The result of the most basic INEPT experiment is a spectrum of nucleus 2 with ./-coupled doublets out of phase and of perturbed intensities (Figure 1.18(c)). However, if the phase of the final pulse on nucleus 2 is alternated by 180° with respect to the coherence transfer pulse and receiver in alternate scans, the equilibrium magnetisation originating from nucleus 2 is eliminated. The peaks in the resulting spectrum are then of equal and opposite intensity, where that intensity is related to the intensity of the single pulse experiment by a factor of \y /y 1. In x  2  reality, even when there is only one value of scalar coupling to consider, the optimal value of T may not be 1/4J because the magnitude of Mi decays according to Ti during the time IT. Further phase cycling is required to eliminate unwanted magnetisation from pulse imperfections and relaxation during the r delays. The phase cycling scheme employed in the present work is given explicitly in Appendix III. In this form, for a sample showing multiple signals in the single pulse spectra of both nuclei, the INEPT experiment only gives signals in the spectrum of 2 from nuclei with a through-bond connections nuclei 1, thereby conveying connectivity information. The choice of r is necessarily a compromise in experiments of this type. Quadrupolar nuclei behave as pseudo spin / = Vi nuclei with respect to the INEPT experiment. '  39 80  43  o  <1> >  T3 IN  ID  3  _  +  60 ' <L> >  3  a.  3 C  s  in 3 O  o  u  <fci  c  1o S S .2 fl 5 w  U  T3  A  u "C J3 4> *B. £  8  A  j— ft.  Z (U  "H °  I  2  (D  3  M ca  ° •2 Si .3 ca u  ea £•  3 "§  o J3 O is  o  a) ^  E u  CO  •s  * g  as 0  o <u 13 5  •fl Vi  S" s fl "3  sr  m  *» Q. ^  3  5, R  3  •*-» -*-»  § 2i s  El S8  • ' 3 s"  o  o tu Q.  sj . a 1- &o M  «  O 3 o 5 <D " o J3 3  u ^ •o a a S -,|rs  c/5  a a  u J3 13  S S S 2  «w  <u « N •fl -fl  M K •5 =^ •S <N +-»  -3  Pti  60  60  ?  :§  O  cd  o,  2 8 • u • •J l_ 3  .SP  O  s  b -5  44  =2  Coherence transfer by dipolar coupling can also occur from this experiment. In the solid state, dipolar coupling is usually eliminated by M A S . Under M A S conditions, the INEPT experiment must be rotor synchronised in order that dipolar coupling is not re-introduced (cf. the TEDOR experiment).  1.1.10(d)!)  The refocused I N E P T experiment.  Solid state N M R linewidths are frequently greater than the scalar coupling and therefore although INEPT spectra can still appear as inverted doublets, the magnitude of the scalar coupling cannot be accurately determined. Additionally, the intensity of the spectrum is compromised by the overlap between the lines in the doublets. This modification, devised to simplify spectra, removes the effects of scalar coupling from the final spectrum. '  36 40  Addition of a further delay r (r * ti) and simultaneous ;r pulses causes the components of Mj to refocus after 2 r . The resulting spectrum of nucleus 2 shows no scalar coupling but has intensity originating only from nucleus 1. F i g u r e 1.19 P u l s e s e q u e n c e d i a g r a m o f t h e r e f o c u s e d I N E P T e x p e r i m e n t . 2  45  2  2  The refocusing of the scalar coupling is achieved by the addition of a second echo experiment (Figure 1.19). The ideal value of r during this section of the pulse sequence may be 2  significantly different from r because r is governed by T2 of nucleus 2 instead of nucleus 1. x  2  Additional phase cycling is required to ensure that no magnetisation originating form the final npulse contaminates the signal. A refocused INEPT spectrum is identical to a spin-echo spectrum except that scalar coupling has been removed and intensity is only present from nuclei directly coupled to source nuclei.  2D I N E P T and 2D refocused I N E P T  l.l.lO(d)ii)  A great deal of additional information can be obtained from an N M R experiment if the spectrum can be correlated against the spectrum of another nucleus, techniques known generically as two dimensional (2D) heteronuclear correlation spectroscopy. 2 D N M R spectroscopy was first proposed " as a concept of correlating homonuclear coupling data to 41  43  chemical shift data and later expanded to include heteronuclear coupling and other dimensions  4 3 - 4 5  In the present context, the INEPT and refocused INEPT spectrum (or  "dimension") can be correlated against the spectrum of the source nucleus by the introduction of a preparation pulse and delay prior to the INEPT part of the sequence (Figure 1.20).  (fl  (fl  M,  (fl  M,  2  Figure 1.20 Pulse sequence diagram of the 2D refocused INEPT experiment, h is the directly detected time domain, t, is the incremented delay that produces the indirectly detected time domain. Refer to text for discussion.  46  M\ is nutated into the xy plane, allowed to evolve for a time t and the x component nutated x  back along z by a (nf2) pulse. The (nl2) pulse samples Mi(t ) so that the magnitude and phase y  y  x  of source magnetisation for the INEPT experiment is a function of t . The two dimensional x  experiment is constructed from a series of experiments at incremented values of t . These data x  can be Fourier transformed in t , the directly detected dimension, to produce a mixed time 2  domain/frequency domain "interferogram". The intensity of each set of v datapoints in the 2  interferogram, /( vz) is modulated by magnetisation from the source nucleus at intervals of t . x  Each sets of v (t ) datapoints can be Fourier transformed to produce a two dimensional spectrum 2  x  of the directly detected dimension (F2) of nucleus 2, against the indirectly detected dimension (Fl) of nucleus 1. Note that even in the 2D refocused INEPT experiment, the scalar coupling information can be present in F l . In order to obtain quadrature detection in F l , phase cycling of the second, selection pulse and the receiver must generate the conditions described in Section 1.1.5. Methods of phase sensitive 2D spectroscopy used in this study are discussed in Section 2.1.2(b). The INEPT experiment does not distinguish between heteronuclear dipolar and scalar couplings. It was originally applied to liquid state N M R where motional averaging removes dipolar coupling. The same is true in solid state M A S N M R providing that the delay periods of the experiment are rotor synchronized.  1.1.10(e)  The MQMAS  ' Experiment.  For non-integral spin quadrupolar nuclei, several techniques have been devised to remove the remaining second order coupling (Section 1.1.9(c)) and obtain high resolution spectra with isotropic resonances. A l l techniques approach the problem of zeroing the 2  nd  and 4 Legendre th  Polynomials (Eqns 1.57 and 1.58) within the same experiment. The technique of DOuble .Rotation (DOR) rotates the sample simultaneously about two 47  axes; the magic angle and either 30.56° or 70.12° (Figure 1.13). The technique is successful and has the advantage of being quantitative and applicable to all N M R pulse sequences. However it is beset by technical difficulties; the apparatus required to perform the experiment consists of a sample rotor within a sample rotor which severely restricts the maximum sample size. 47  Additionally, the maximum velocity of the outer sample chamber (rotating at the magic angle) is ~1.5 kHz and DOR spectra are typically made up of many overlapping spinning sidebands from rotation at both angles. DOR is therefore only useful for samples that give intense signals that have small quadrupolar couplings; conditions which may give well resolved spectra under M A S alone, particularly at high magnetic fields. Additionally, DOR cannot resolve resonances from amorphous materials whose DOR linewidths can easily exceed the spinning rate, due to distributions of chemical shifts and quadrupolar couplings. The technique of Dynamic ^ngle Spinning ( D A S )  19,48  '  49  is a 2D spin-echo experiment in  which the sample is rotated around each of two "complimentary angles" in two evolution periods. Complimentary angles satisfy the condition: r,P (cos^ ) = - r P (cos 0 ) 2  2  (1.62)  2  4  Or: P (cos 6>, ) = -A:P (cos <9) 2  2  2  and  P (cos0,) = - £ P ( c o s 0 ) 4  4  2  (1.63)  Where k = r / r , and r„ is the echo delay. Eqn 1.62 cannot be satisfied to include the magic 2  angle. The "perfect" D A S experiment in which the angle hop occurs instantaneously is shown in Figure 1.21(a). Refocusing occurs as a result of the angle hop, since the evolution of magnetisation due to both P2(cos# ) and P^costf,) is reversed following the hop, according to 1  Eqn 1.63. Notice that this condition also refocuses all first order interactions. The magnitude and phase of the echo at t = 0 are modulated by the isotropic chemical and second order 2  quadrupolar shift frequencies. Most commonly, 6 - 37.38° and Q = 79.19° are chosen such that k = 1. In this case a 2D X  Fourier transform yields a spectrum with an F2 dimension in which resonances have lineshapes characteristic of 9 , i.e. containing reduced CSA, dipolar and first and second order quadrupolar 2  anisotropics, against an isotropic indirectly detected dimension with resonances at DAS  V  - o^iso V  +  y°^y - ° ^  v  5  2  m  s  F2 dimension can be inconvenient due to the reduced first order  quadrupolar anisotropy, giving broad lines and low S:N, therefore D A S experiments often include an additional angle hop to the magic angle (Figure 1.21(b)) prior to detection.  48  In a real D A S experiment (Figure 1.21(c)) the angle hop takes ca. 30 ms and the magnetisation must be stored along +z to prevent the signal decaying according to T * during 2  the hop. Each pulse that tips magnetisation out of the xy-plane selects only a component of M, such that by comparison to the fictitious "perfect" experiment, the S:N is reduced by -Jl for each angle hop. A n additional problem with DAS is that the hop duration is so long that the longitudinal relaxation that occurs during the storage period can introduce magnetisation of greater magnitude than that remaining from the preparation pulse after t (which places high x  demands on the phase cycling) and in the extreme case, the system can equilibrate during the (a)  Figure 1.21 Pulse sequence diagram of (a) the "perfect" DAS experiment with acquisition at 61, (b) the "perfect" DAS experiment with acquisition at 6Jv and (c) the DAS experiment with acquisition at &MAS- r, is the incremented delay that provides the F l time domain. Magnetisation that builds up during the angle hop periods must be eliminated by phase cycling of the storage pulses with respect to the preparation pulse on alternate scans. Note that the indices x and y are not strictly accurate due to the change inframeof reference implied by the angle hops. 1AS  49  angle hop and destroy the stored magnetisation. DAS is also a 2D experiment and requires special spinning apparatus. Additionally the sample spinning can be stopped by the vibrations from the angle hops, making long experiments problematic. The Multiple Quantum Magic Angle Spinning ( M Q M A S )  46  experiment is the most recent  technique to be developed for second order quadrupolar line-narrowing. It is closely related to DAS and circumvents many of its problems. This 2D echo experiment selects only magnetisation that passes through a particular coherence pathway. Aside from the coherences from the central transition and the satellite transitions, for which Am/ = ±l,arf.  pulse applied to  a non-integral spin quadrupolar nucleus generates coherences from transitions where Ami  ±2,  =  ±3, etc.. These multiple quantum coherences are "forbidden", inefficiently excited transitions. The first order quadrupolar coupling contribution to m = 1 and m = -1 is equal (Eqn 1.35) and 7  7  therefore symmetrical transitions (i.e. +m ^>-m ), like the central transition, are free of first I  I  order quadrupolar anisotropy. It can be shown that the more general form of Eqn 1.56 for second order anisotropy under spinning conditions of any (+TO,-»-m/) transition is given b y : ' 51  2  C_  52  2  =2m v ^C {m )R(T ) ^C™(mMv ,^ ° m  m;  I  V  °  0+  I  jQ  +  Q  (1.64)  V  + ^ -C (/n ){Ffo ,^)+z(/7 ,^)cos4^}p (cosc?J e  (4)  /  e  fi  4  Where C (m,) are the coefficients associated with the nth order terms of the linear (n)  combination of weak coupling angular momenta, the Clebsch-Gordan coefficients: '  15 53  C  ( 0 )  (m,) = 2m, {/(/ +1) - Im)}  (1.65a)  C ( / w ) = 2 w { 8 / ( / + l)-12m -3}  (1.65b)  C ( / n ) = 2 m { l 8 / ( / + l ) - 3 4 w -5}  (1.65c)  (2)  2  /  /  /  (4)  2  /  7  Again the P (cos6?) terms vanish under M A S and the multiple quantum analogy of the 2  complimentary angle condition (Eqn 1.63) is given by:  50  54  C (m )r =-C (m )r w  (1.66)  w  x  x  2  2  C (m ) = -kC (m ) {4)  (1.67)  w  x  2  (a)  x  anti-echo  Figure 1.22 (a) Pulse sequence diagram and (b) coherence pathway diagram of the nutation 3QMAS  experiment for a spin / = 5/2 nucleus. Pulses of duration cA and $ generate coherences of level p and -p. Refer to  text for discussion.  There are several variations of the M Q M A S experiment ' " relating to the method of 46 54  58  coherence generation and the precise coherence pathway that is selected. Figure 1.22 illustrates the simplest; a triple quantum (3Q) M A S , 2-pulse nutation experiment for a spin / = 5/2 nucleus. Higher quanta experiments can be performed on nuclei with spin I> 5/2, but the higher the coherence level to be generated, the less efficient the experiment and therefore 3QMAS is by far the most common M Q M A S experiment in practice.  51  The coherence pathway (Figure 1.22(b)) can be understood as follows; fa stimulates the rrij = \ -> - \ transition; a coherence corresponding to a change in mi of +3. The state of the system can be said to have gone from the equilibrium coherence level ofp = 0 to p = +3. fa stimulates the m, = - \ - » - \ transition which has a coherence level ofp = -1. Thereafter, the system relaxes back to equilibrium. In the same way as the D A S experiment, an echo forms after a time \k\t where k is given u  by Eqns 1.66 and 1.67. The common values of k are given in Table 1.1. The rf. pulses fa and fa stimulate all possible coherences and magnetisation that passes along the desired coherence pathway must be selected by phase cycling of the pulses and receiver. This phase cycling is complex and is different for each variation of the M Q M A S experiment, but based on the same principle. To a first approximation the rate of precession of n-quantum coherences is proportional to n v (Eqn 1.64). Therefore, if the phase of fa is incremented by 6°, the phase of 0  the magnetisation at the time offahas changed by 36?° compared to the single quantum coherences and 30°'/5 compared to the 5-quantum coherences. The phases of all but the desired coherence add to zero i ffais incremented through 360° in 60° steps that correspond to 180° steps in the phases of the receiver and fa. This method does not distinguish between the +p and -p coherences (corresponding to the echo and anti-echo pathways) since they precess at the same rate. A n additional phase cycling of the receiver must be employed to make the distinction, increasing the minimum phase cycle of the 3QMAS experiment to 12. The choice between echo and anti-echo pathways (Figure 1.22(b)) is arbitrary since they differ only in the sign of their isotropic shifts. The duration of fa and fa can be optimised for the desired pathway. Generation of the "forbidden"/? = ±3 coherence requires a long pulse and the efficiency increases with rf pulse power. The p = -1 coherence is far off resonance and therefore fa is longer than the solid 90° pulse. Both pulse lengths are a function of v and although there are Q  theoretical treatments that purport to show the relationship between v and the rf. power, the Q  best values of fa and fa for a given system and N M R isotope must be obtained empirically. Separation of coherence pathways is further aided by differences in the value of k for each coherence pathway. Echoes formed from unwanted coherence pathways occur at different times. However, the fact that k ^ 1 complicates data processing. This is discussed in Section 2.1.2(b)ii).  52  Analogous to the DAS experiment, the magnitude and phase of the 3Q echo is modulated by isotropic frequencies. In M Q M A S experiments, this is given by the time average of the isotropic chemical shifts, the C  ^  ( 0 )  terms of Eqn 1.64.  =^ t e -  +  2  m  '  ° )  V  m, +3 /  =  ^ ( < £ - i + "o)  (1-68)  m,  k  +1  5  /=!  +  -3  -1  19  2  228  144  2  -42  54  _19  12  -228  -144  12  42  -54  9  -1  9  1  Table 1.1 Values of C* and k for spin / = 5/2 and / = 3/2 nuclei. The sign of the value of k determines the sign of the isotropic 3QMAS shift, i.e. the definitions of the echo and anti-echo pathways. The refocusing time is dependent only on the magnitude of k. 4)  Km^-m  c  a  n  be also be expressed in terms of v'™ \ (Eqn 1.44), with the result that the (2  +  2  _ >  2  isotropic 3QMAS shift, ' ' the sign of which has been arbitrarily chosen to be positive in this 50 56 58  work, for spin I = 5/2 nuclei is given in Hz by: 17  32  *  _8_^ + •93 v  2 (  1+1  3A  (1.69)  n  The advantages of M Q M A S over DAS lie in the facts that the total duration of the pulse sequence prior to t = 0 is orders of magnitude smaller (which more than compensates for the 2  inefficiency of multiple quantum coherence generation) and that the experiment can be performed at the magic angle at high spinning rates with conventional apparatus. The main disadvantage of the M Q M A S experiment is that the efficiency of coherence selection is dependent on the magnitude of quadrupolar coupling, the transmitter frequency,  53  and the  spinning rate and therefore it is not possible to obtain M Q M A S N M R spectra that quantitatively reflect the relative intensities of resonances with different quadrupolar couplings. Nutation 3QMAS experiments were performed on the Bruker DMX400 spectrometer using the programme z3q2pseq.jlb (Appendix III A(viii)) and processed using mqxfb. Nutation 3QMAS experiments were performed on Varian Inova systems using the programme 3qmasl.c  60  (Appendix III.B(ix)) and processed using daslp. Both programmes are component  parts of the V N M R software.  l.l.lO(e)i)  (a)  The R I A C T I I 3 Q M A S experiment.  (b)  z Ml  •y  ~2L  (d)  00  Figure 1.23 Spin locking of a spin I = Vi nucleus in the rotating frame, (a) Magnetisation M is nutated into the  xy-plane by a (^/2) rf. P l appears as a static field 5, along x. (b) Following the rf. pulse M lies along y. (c) Where there are many nuclear environments present, M is the sum of many components that have been nutated approximately into the xy-plane close toy. (d) The application of a rf. pulse (^) causes the components of Mto u  s e t n a t  x  nutate about y and they are said to be spin locked.  This experiment utilizes the phenomenon of "spin-locking" of the p = ±1 coherence (central transition) of non-integer spin quadrupolar nuclei to facilitate adiabatic transfer of that magnetisation to other spin states. Spin locking is achieved by the application of two successive, orthogonal rf. pulses (Figure 1.23). The first pulse nutates the magnetisation along y and since B is the only field that appears x  54  in the rotating frame of reference, the second pulse "holds" that magnetisation along y where it decays according to the relaxation time T . lp  Spin locking is complicated in the case of quadrupolar nuclei under M A S conditions by the existence of multiple coherences. During sample rotation, the energy of \m,\ > Vz states are 61  caused to cross the energy of the |m | = Vz states. Where this occurs, the possibility arises that the 7  magnetisation will be transferred to the higher state by a relaxation mechanism driven by the non-equilibrium populations of the |/M/| = Vz states, i.e. a coherence transfer from the p = 1 to some p > 1 coherence. Where conditions allow this coherence transfer, the spin locking of the 63  central transition decays more rapidly.  61,64  It is generally accepted that coherence transfer occurs  with an efficiency determined by Vega's "adiabaticity parameter", or. '  61 62  a=  v  2  (2.6)  v  (2)  2  If the condition a » 1 is met, then the level crossings can be considered adiabatic and coherence transfer will be efficient. Further work has shown that the transfer will oscillate between coherences throughout the rotor cycle. ' Less efficient, but non-oscillatory spin 61 64  locking will result i f a « 1. Finally, spin locking decays rapidly to higher order coherences when a*l. This principle has lead to the development of the iJotationally /nduced Adiabatic Coherence Transfer (RIACTII) version of the M Q M A S experiment (3QMAS version illustrated in Figure 51  1.24). The authors recognised that the condition for adiabatic transfer would be met by some of the nuclei in a powder sample for almost any set of conditions. The coherence pathway of the RIACTII experiment (so called because both transfer pulses are adiabatic transfers) is slightly more complex than for the nutation experiment as &p = 0—> ±1 transition precedes the generation of the p = ±3 coherence. It is otherwise analogous to the nutation experiment. Interest in the RIACTII experiment has been in the potential to acquire M Q M A S spectra ' ' with intensities that are less sensitive to quadrupolar coupling and experimental 51 65 66  conditions. The majority of such work has been performed on "ideal" test samples, investigating the merits of soft and hard pulses and pulse duration, for example, but it is not possible to obtain quantitatively reliable 3QMAS spectra of "real" samples containing any distribution of  55  chemical/quadrupolar shifts and quadrupolar couplings. However, RIACTII is often found to be less sensitive to such distributions and can be useful in the observation of all signals from samples containing nuclei with a particularly wide range of couplings, particularly when some of those couplings are unusually large.  (a)  x  anti-echo  Figure 1.24 R I A C T I I 3 Q M A S experiment. Triple quantum magnetisation is generated by a (n/2) pulse followed by the application of a spin locking pulse ^, for a time l/vi, . Conversion of the triple quantum coherence to the observable single quantum coherence is achieved with a similar spin locking pulse foand the phase of this pulse 6 and the phase of the receiver are cycled with respect to each other such that the desired coherence pathway is selected. This phase cycling is identical to that of the simple nutation 3QMAS experiment. x  AS  Studies utilizing RIACTII have almost exclusively focused on spin / = 3/2 nuclei because theoretical calculations predict that there is no energy level crossing between the m, = \C. <-> VL  56  states of spin / = 5/2 nuclei such as A 1 and O. Work using RIACTII 3QMAS experiments on Z/  u  spin / = 5/2 nuclei have been largely regarded as showing intensity from nutation by the ^ pulses. However, it has been independently verified that a significant percentage of the 57  intensity from such experiments (ca. 40%) originates from true adiabatic transfer, confirmed 67  by more recent work that suggests that the coherence transfer is facilitated by a mixing of nuclear spin states at high rates of sample spinning. For RIACTII 3QMAS experiments performed in this work, spin locking pulse duration was set to l/4v  MAS  . This value was optimised experimentally, but has some theoretical  justification. ' '  51 64 66  RIACTII 3QMAS experiments were performed on the Bruker DMX400 using the programme e3q3pseq.jlb (Appendix III.A(ix)) and on Varian Inova spectrometers using the programme 3qRIACT2p.c (Appendix III.B(x)) written by Dr. Jorgen Skibsted.  l.l.lO(e)ii)  Zero quantumfiltered3 Q M A S Experiments.  69  The zero quantum filtered (z-filtered) 3QMAS experiment more efficiently selects the desired coherence pathway by the storage of the echo magnetisation along the z-axis using a soft pulse selective for the central transition for a period during which unwanted magnetisation is allowed to decay (Figure 1.25(a)). The coherence pathway (Figure 1.25(b)) differs from the other 3QMAS experiments used in this work in that the echo and anti-echo pathways are symmetrical. In principle this should confer the advantages of improved S:N (because no single pulse length can be optimized for both parts of an asymmetric coherence pathway) and fewer spectral distortions, compared to the other 3QMAS experiments discussed here. The improvement in S:N observed for some systems was found in the present work to be 69  outweighed by the loss of signal during the additional pulse (cf. D A S storage pulses). The zfiltered 3QMAS experiment is therefore employed in this work at 9.4 T on materials giving 3QMAS spectra with high S:N, in order to obtain spectra with minimal distortions. The programme mqzqf.cf written by Professor Christian Fernandez was used to acquire z-filtered 3QMAS spectra on the Bruker DMX400 spectrometer. At high field, quadrupolar coupling 57  effects were found to be sufficiently reduced that relatively undistorted spectra resulted from the nutation 3QMAS experiment.  (a)  echo  (b)  anti echo F i g u r e 1.25 Z - f i l t e r e d 3 Q M A S p u l s e s e q u e n c e . M a g n e t i s a t i o n that h a s p a s s e d a l o n g t h e p = 0 — • ± 3 p a t h w a y is s t o r e d a l o n g z b y <jh T h e f i n a l c o n v e r s i o n t o p = -1 b y a s o f t nil p u l s e g e n e r a t e s a n e c h o a f t e r |jfc|f.  1.2 X-Rav Diffraction  1.2.1  POWDER X - R A Y DIFFRACTION  70  X-ray diffraction (XRD) is a powerful solid state analytical technique that is complimentary to N M R . Whereas N M R is sensitive to the ordering of the local environment of a nucleus, X R D  58  is sensitive to long range repetitive order. The two ordering regimes are often concomitant, but the comparison and contrast between the information obtained from the two techniques can be extremely instructive. There is a strong interaction between the electric field component of X-rays and electron density. Crystals act as diffraction gratings to incoming X-rays as a result of coherent scattering 71 77  (also known as Thompson scattering) between planar arrays of atoms '  separated by distances  of the same order of magnitude as the X-ray wavelengths (ca. 10°-10' A). There are several possible outcomes from the interaction between X-ray light and matter, but unless the frequency of the radiation is close to an electronic absorption band of the sample, coherent scattering predominates. Coherent scattering is the consequence of an elastic collision between an electron and a photon that form a transient excited state, sometimes described as a "complex" between the photon and the matter, and the subsequent ejection of a photon with the same energy and phase as the incident photon. Scattering from a planar array of atoms can behave in the same way as reflection from a plane. Consider coherent X-ray radiation incident on an infinite number of partially transparent parallel planes (Figure 1.26). If the X-rays reflected from successive planes travel an additional path that is an integer number of wavelengths, constructive interference will result (Figure 1.26(a)) and if the additional path is a half integer number of wavelengths, destructive interference will result (Figure 1.26(b)). The relationship between the angle of incidence 0, the wavelength X and the distance between the planes d, that leads to constructive interference is Braggs' L a w  73  (1.60). nX = 2dsin0  (1.60)  Where n is an integer. Coherent scattering from atoms occurs radially and therefore a diffraction pattern is built up as a function of 0. Crystals can be considered to be comprised of atoms located at the intersections of many sets of parallel planes of atoms, the "lattice planes", illustrated for a simple two dimensional array in Figure 1.27. Each set of lattice planes, which must pass through all of the lattice points that make up the structure, generates a diffraction pattern with an intensity that is directly  59  proportional to the number of electrons surrounding the atoms that comprise the planes. Therefore the total diffraction pattern of a sample is sensitive to the positions and chemical nature of all of the atoms in that sample.  Figure 1.26 Schematic representation of coherent scattering of X-rays leading to (a) constructive and (b) destructive interference.  60  In the present work, X R D is mainly employed to give qualitative "fingerprints" of powder samples. A simple X-ray powder diffraction (XRPD) pattern gives a guide to the overall crystallinity of the sample (the width and shape of peaks) and to the amount of amorphous material present in the sample (appearing as a broad signal in the baseline) and the information is useful for comparisons between related materials. X R P D is widely used to identify the crystalline phase or phases in unknown samples by comparison of experimental data to databases ' of peak positions. However, X R P D cannot in 75 76  general be used to quantify the relative proportion of the components of a mixture. Although the relative intensity of peaks is proportional to the amount of each phase present in sample mixtures, peak intensities are also very sensitive to experimental setup (collimation of the beam, sample geometry, diffractogram resolution) ' and X-ray diffraction must be carefully 70 77  calibrated in order to be quantitative in this manner. Therefore X R P D cannot be regarded as being quantitatively reliable in the same sense as N M R .  (a)  (11)  (b)  (2 3)  Figure 1.27 Examples of some of the lattice planes that pass through the atoms in a 2D array. Planes are  labelled according to their Miller indices. The magnitude of the Miller index is the reciprocal of the number of planes per unit cell (dimensions a and b), i.e. the number of planes that intersect the unit cell axis between adjacent lattice points. Refer to Section 1.2.2(c) for discussion. Figure adapted from reference 74.  61  1.2.2  1.2.2(a)  BASIC CRYSTALLOGRAPHY  Three Dimensional Arrays and Space Groups  Crystal structures are composed of repeating patterns of atoms, molecules or ions in three dimensions that could in principle continue indefinitely in all dimensions. The relationship between identical points in the structure are described by symmetry elements. There is a restricted number of symmetry elements that may be present in an infinitely repeatable three dimensional array and a finite number of ways that these symmetry elements can be combined. These permissible combinations are the 230 crystallographic space groups.  The structure of  any crystalline material is defined by knowledge of its space group and the co-ordinates of the atoms in the unit cell (Section 1.2.2(b)).  1.2.2(b)  The Asymmetric Unit, Lattice and Unit Cell.  The "asymmetric unit" or "basis" of a crystal structure is the smallest repeating group of atoms required to build up the whole structure. Each asymmetric unit is associated with a lattice point. The three dimensional array of lattice points, related to each other in a manner described by the space group, is known as the "space lattice" and the "unit cell" is the contents of the volume encompassed by adjacent lattice points. The unit cell is the preferred "minimum representation" of a structure since a space-filling arrangement of unit cells will build up the complete structure, whereas the asymmetric unit can only reproduce the entire structure with knowledge of the space group symmetry operators. The unit cell is thus a far more convenient basis for understanding a structure. There is no unique unit cell to describe any given lattice. Unit cells that contain the minimum number of lattice points can be categorised as one of the fourteen Bravais Lattices,  72  defined according to the characteristics of the unit cell dimensions (which by themselves define the crystal system) and the relative positions of lattice points within the unit cell (Table 1.2). A l l of the 230 possible space groups fall into one of these fourteen classifications. It is the goal of crystallography to identify the crystal system, Bravais lattice, space group, unit cell dimensions, unit cell angles and ultimately the positions of all atoms comprising the 62  unit cell (expressed in terms of fractional co-ordinates of the unit cell dimension in each axis), each successive step requiring additional experimental data. In Chapter 7 of this thesis, the unit cell parameters (dimensions and angles) are extracted from X R P D data with prior knowledge of the space group, using the principles discussed in Section 1.2.2(c).  1.2.2(c)  Evaluating the Unit Cell Dimensions.  The lattice planes in crystal structures (illustrated for a 2D array of lattice points in Figure 1.27) , including those planes that comprise crystal surfaces, are defined according to the unit cell parameters and the Miller Indices. Miller indices are the reciprocal of the steps in fractional co-ordinates of the intersection between successive lattice planes with the unit cell axes (Figure 1.28) , (h,k,l):  h=— Aa  k= — Ab  and  /=— Ac  Crystal System  Unit cell parameters  Lattice Type  Cubic  a = b = c a = p=y= 90°  P, I,F  Tetragonal  a = b±c  Hexagonal  a=p=y=90°  a = b±c a=p= 90° y= 120°  Trigonal / Rhombohedral Orthorhombic  a=b =c « = a±b±c  Monoclinic  a±b±c  Triclinic  a±b±c  90°  a=  p=y=90°  a = p= 90° ^ 9 0 ° a^p^y^90°  (1.61)  P,I P R P, C, I, F P,C P  Table 1.2 The Bravais Lattices. Unit cell parameters are defined with respect to the crystallographic axis system (Figure 1.28). P; Primitive - one lattice point per unit cell with fractional co-ordinates (x,y,z). I; Body centred - 2 lattice points per unit cell at (x,y,z) and (x+Viy+Vi^z+Vi). C (or A or B); Face centred - 2 lattice points per unit cell at (x,y,z) and (x+'Ajrt- Aj) (or (xy+Vi£+ Vi) or (x+ Viyj+Vi)). F; Face centred - 4 lattice points per unit cell at (x,y,z), {x+ A,y+ 'A^), (x y+ / z+ V ) and (x+ 'Ay^+'A). R; Face centred - 4 lattice point per unit cell at (x,y,z), {x+ Ay (x,y+'A^) and (x^^+ A). l  l  1  l  l  2)  2  l  63  c  a Figure 1.28 The crystallographic axis system.  The distance between lattice planes hkl, denoted duk is trigonometrically related to the unit cell dimensions. The simplest cases, where a, J3 and /are all 90° yield the relationships:  a (h +k +l ) 2  2  h +k 2  c  2  1  h  k  I  d hkl  a  b  c  2  (1.62a)  Tetragonal  (1.62b)  Orthorhombic  (1.62c)  2  a  hkl  l  Cubic  2  2  2  Bragg's Law (Eqn 1.60) relates dhu to the position of a peak in the X R D pattern and therefore i f sufficient peaks can be indexed, i.e. have Miller indices assigned to them, then the unit cell parameters can be extracted by solution of equations such as Eqn 1.62(a-c). The indexing of powder patterns can be difficult because of the superposition of diffraction patterns from the 3D array of lattice points onto one angle dimension, 6? and is usually computed from experimentally observed reflections and a known space group. Data may be consistent with many sets of unit cell parameters, but the number of structurally reasonable possibilities diminishes as the number of indexed reflections increases.  64  1.2.2(d)  Experimental Apparatus.  The efficiency of coherent scattering is proportional to A and therefore X R D experiments 3  utilise relatively low frequency X-ray sources. For powder diffraction studies, by far the most common is Cu K a radiation, generated by the selective ionization of core electrons followed by the subsequent emission of radiation as the ion relaxes to its ground state. Absorption of the radiation may necessitate the use of another metal (e.g. molybdenum) for the generation of the X-rays. For example, many nickel-containing materials absorb in the frequency range covered by copper emissions. Every method of X-ray generation requires some form of monochromation and collimation of the X-rays directed at the sample in order to observe high resolution X-ray scattering. Aside from the alternate approaches to generating the X-ray source, there are a number of different methods of collecting X R D data, each with its own advantages and disadvantages. A description of the apparatus used to generate the X R P D patterns presented in this thesis is given in Section 2.2.  1.2.2(e)  XRD and NMR Equivalence.  The asymmetric unit contains all of the unique atoms in a structure. The members of the asymmetric unit associated with an adjacent lattice point are related by the symmetry elements of the lattice's space group but there is no general relationship between members of the basis. Therefore the lattice planes of all of the atoms in the asymmetric unit are independent and each atom in the asymmetric unit is considered crystallographically inequivalent. Crystallographic inequivalence is a precisely defined concept relating to the mathematical symmetry properties of the crystal structure and can apply to chemically indistinguishable atoms. N M R inequivalence, by contrast, is a terms loosely applied to describe whether or not two signals are experimentally distinguishable. This is an important consideration when comparing N M R and crystallographic information. For example, a crystal structure may predict four inequivalent atoms of a given element. A quantitatively acquired N M R spectrum of that element is consistent with the crystal structure i f  65  it shows between one and four signals, providing that the intensity can be broken into four units and distributed amongst the observed signals. However, an N M R spectrum showing more than four signals is not consistent with the crystallographic information, and might indicate the presence of an impurity or suggest a structure with a lower space group symmetry.  => Atoms that are inequivalent in N M R spectroscopy must be crystallographically inequivalent. => Crystallographically inequivalent atoms may be N M R inequivalent.  1.3 Materials Studied In this thesis are presented the results of solid state N M R structural investigations of a variety of aluminium containing materials, supported by other experimental data. The aim of this work is to demonstrate the applicability of multiple field and quantitative high field solid state A l M A S N M R spectroscopy to materials consisting of both crystalline and amorphous 2 7  phases. A n overview of the structures and applications of these materials is presented below.  1.3.1  Z E O L I T E M O L E C U L A R SIEVES  '  Naturally occurring zeolites were known only as a class of low density, hydrated aluminosilicate minerals from their discovery in the 18 century until the first observation of th  their "molecular sieving" properties in the middle of the 20 century. The huge commercial 81  th  interest in zeolites followed the first synthesis of acid zeolites, since it was soon apparent that 82  these acid zeolites had applications as active Bransted acid catalysts. Acid zeolites show high thermal stability and their use avoids many of the problems associated with the handling and disposal of conventional mineral acids and their chemical by-products, whilst providing considerably higher activity than alternative heterogeneous catalysts " such as amorphous 83  85  alumina. These properties, together with the molecular sieving qualities of zeolites account for their widespread use throughout the petroleum industry as cracking and reforming catalysts. ' ' 86 8  66  Interest in this class of material has grown along with the range of known structures,  75  particularly since it became apparent that many structures not found in nature could be synthesised. The general approach to zeolite synthesis (autoclaving alumina and silica gels at 88  high temperatures and pressures for prolonged periods under basic conditions) mimics the conditions of their formation in nature. In addition, use of an organic template, usually a quaternary ammonium ion, facilitates the formation of a far wider range of structural types and with higher framework silicon to aluminium ratios (Si:Al) than are found in nature. '  89 90  Zeolites have the general chemical formula:  Mr [(Si0 ),_„ (A10 )„ ] - . y H 0 2  2  (1.63)  2  x  The anionic aluminosilicate framework is composed of edge and corner sharing, neutral S i 0 and negatively charged 4  AIO4  tetrahedra that form the walls of pores and channels which  are of molecular dimensions (most commonly in the range of 4-13 A ) .  75  The locations occupied  by Si and A l atoms are collectively known as "T-sites" and every aluminium T-site must be balanced by the charge of extra-framework cationic species, M . Although the framework x+  structures are highly crystalline, the distribution of aluminium T-sites is disordered, with the exception that i f the framework Si:Al = 1, such as for zeolite-A (Figure 1.29(c)), then there is a perfect alternation of silicon and aluminium T-sites. There have been proposals for a medium range ordering of aluminium T-sites when the framework Si:Al > l ,  9 1  but in general any  preferred site occupation of aluminium T-sites is insufficiently ordered to be studied by diffraction techniques. The charge balancing cations can be exchanged with other cations and water of hydration can be reversibly removed from the structure. Organic template molecules left from the synthesis may initially be trapped in the structures of synthetic zeolites, but these can also in general be removed by calcination (i.e. heating to ca. 500°C to induce thermal decomposition). Important characteristics of a given zeolite structure are the dimensionality of and the size of the pore and channel system. The size of the channels of a zeolite framework determines the size selectivity towards probe molecules. For example, naturally occurring mordenite has two systems of one dimensional (001) channels. Its structure contains channels that run parallel to 92  67  one another along the c-axis of the unit cell and do not intersect. The largest channels have a circumference of 12 T-sites, termed a "12-ring". This description is often more representative of the pore capacity than the equilibrium pore diameters (7.0 A and 6.5 A in this elliptical case) since diffusion through the pores is facilitated by expansion and contraction of the pores via T-O-T bond angle and length vibrations. Stillbite, one the earliest known zeolites, is an example of a structure with a 2D intersecting channel network and faujasite has a structure with a 3D 93  network of channels (Figure 1.29(b)). The pore size and the geometry determines which materials may enter the pores. The cavity size places a restriction on the maximum size of reaction intermediates and the dimensionality plays a role in the rate of diffusion through the structure and how susceptible the structure is to pore blocking processes. It can be instructive to describe zeolite structures in terms of "secondary building units" (SBUs) in order to focus on regions of chemical interest, to highlight symmetry properties of the structure, or to aid the conceptualisation of complex structures. The sodalite cage (Figure 1.29(a)) is a S B U of many simple zeolite structures. The T-sites of a sodalite cage are arranged in the form of a dodecahedron, each vertex of which represents an outward pointing (T-0-) bond that connects to another T-site in the complete zeolite structure. The structure of the parent material, sodalite, is made up of cages directly connected at the 4-ring faces such that adjacent cages share 4 T-sites and form a cubic (I) array. The complete structure has the smallest pores 95  of any zeolite; the ~ 3 A diameter 6-rings of the sodalite cages themselves. If the same SBUs are joined by a "double 6-ring" the faujasite structure is formed (Figure 1.29(b)). This is representative of the family of materials that includes zeolite-Y, the acid forms of which are 96  used in greater quantities than any other zeolite in the petrochemical industry and zeolite-X, 97  98  a synthetic form with framework Si:Al lower than that found in the natural mineral (Si:Al ~ 2.5). Another industrially important zeolite and the first zeolite synthesized with a structure not found in nature, zeolite-A,  82,99  is composed of sodalite cages joined by double 4-rings. This  material is ubiquitous in synthetic chemistry as a drying agent because of its unusually small diameter pore system (4 A). Other prominent uses of zeolites utilize their ion exchange properties (zeolites are commonly found in water softener systems, detergents and even in kidney dialysis equipment) and the spatial and dimensional limitations of the chemistry which can take place in the cavities. For example, small silver clusters can be assembled inside sodalite cages to produce materials  68  with photochemical and thermochromic properties applicable to printing technology. ' is also interest in the semiconducting properties of zeolites. Cationic conduction  101  There  is not unique  to zeolites, but the pore structure of zeolites gives their ionic conductivity unusually small temperature dependence making them potentially useful solid electrolytes in batteries.  102  There  is interest in research into the use of zeolites in redox chemistry. Their ionic conductivity facilitates their use as electrodes for electrochemical and redox reactions, '  103 104  and zeolites can  act either as supports for transition metal catalysts (most importantly platinum), transition metal complexes '  106 107  or ions,  108  105  hosts to  and as transition metal redox catalysts themselves  where T-sites have been doped with transition metals during the zeolite synthesis. '  109 110  Zeolites  have been described as "mineral enzymes" because of the direction that the pores and channels impose upon the reactions.  111  The majority of zeolite structure information has been obtained from X R D studies. Only a limited number of zeolites can be synthesised or have been found in nature with crystals large enough for single crystal X-ray diffraction studies. '  112 113  The Rietveld refinement technique  applied to X R P D data can often yield average framework structures, but neither single crystal nor powder X R D are sensitive to any disorder within the structure. Therefore the distribution of aluminium or dopant atoms amongst the T-sites cannot be investigated. This is also true of the locations of counter ions, hydrating water and in general, template molecules. X R P D studies are also quite insensitive to the presence of amorphous material and are not generally successful in locating the preferred positions of probe molecules introduced into the channels. A considerable amount of the chemical functionality of zeolites is associated with the positions of the non-silicon T-sites and extra framework material: Heteroatoms in T-sites positions are associated with electrostatic charge and the framework, or species co-ordinated to the framework, due to the charges are therefore more susceptible to attack by, or reaction with, adsorbed molecules or ions. For example, the Bransted acidity of H-zeolites is associated with charge balancing protons attached to framework oxygen atoms adjacent to an aluminium T-site. Catalysis of cracking and reforming reactions of organic molecules takes place at these sites. In addition, hydrothermal treatment of such zeolites leads to framework dealumination (i.e. cleavage of framework T - 0 bonds selectively at aluminium T-sites) and the formation of extra framework material thought to be responsible for additional catalytic activity.  69  Figure 1.29 Example of the relationship between zeolite structures and secondary building units, (a) Sodalite cage secondary building unit, arranged to form (b) the structure of Faujasite and (c) the structure of zeolite-  A. T-sites are located at vertices and T-O-T linkages are represented by straight lines for clarity, (a) The T-sites of a sodalite cage are arranged in the form of a dodecahedron, (b) The faujasite structure is composed of sodalite cages joined by double 6-rings in a diamond-like array (i.e. related to each other tetrahedrally) such that the centre of the cages form a cubic (R) array. The 3 dimensional channel system is composed of mutually orthogonal 12-ring channels, aligned parallel to the unit cell axes, (c) The zeolite-A structure is composed of sodalite cages joined by double 4-rings in a cubic (P) array. The 3D channel system is composed of 8-ring channels aligned with the unit cell axes, cf. (b).  IR spectroscopy can be used to investigate specific bond types  114  (e.g. hydroxyl groups) but  solid state N M R spectroscopy has become the technique most applicable to the investigation of local atomic environments in zeolite structures to which diffraction studies are insensitive. In this thesis the results of extensive studies of the aluminium environments in acid zeolite-Y materials containing substantial amounts of amorphous extra framework material are presented.  70  The aluminium in these systems is intimately associated with their catalytic activity and the disordered distribution of framework and extra-framework aluminium can only be directly *f*J  characterised using solid state A l N M R spectroscopy.  1.3.2  ALUMINOPHOSPHATE M O L E C U L A R SIEVES  Aluminophosphate molecular sieves  are an entirely synthetic class of materials  (AIPO4S)  closely related to zeolites that have been developed in the last twenty years.  115  Instead of a  distribution of silicon and aluminium T-sites, since their framework A1:P ratio is equal to one, AIPO4S  have a perfect alternation of aluminium and phosphorous T-sites. Many  AIPO4S  zeolites share the same framework topology (e.g. SAPO-37 has the faujasite structure AIPO4-2O  has the sodalite structure),  structures than zeolites; 18,  118  75  AIPO4-36  but  115  1 1 7  AIPO4S  116  and and  can be synthesized with a wider range of  and the material studied in the present work, AIPO4-  are examples of this.  AIPO4S  have the general chemical formula:  [(A10 XP0 )}«H 0 4  4  (1.65)  2  Their frameworks are composed of equal numbers of P04 and AIO4" tetrahedra. Therefore, +  although they share the molecular sieving properties of zeolites, the frameworks are neutral, precluding ion exchange or catalytic activity. In addition, the frameworks are generally less stable than those of zeolites because the local alternation of positive charges (phosphorous Tsites) and negative charges (aluminium T-sites) are susceptible to chemical attack. As a result, AIPO4S have yet to find any large scale industrial applications. However, there is considerable interest in the potential uses of these materials. A variety of elements can be substituted into the T-sites  116  more easily than for zeolites, thereby introducing  charge and catalytic potential to the framework. AIPO4S,  are called variously T A P O s ,  120  119  VAPOs,  121  The resulting materials, although strictly SAPOs  116,122  and M A P O s  1 2 3 , 1 2 4  (named  after titanium, vanadium, silicon and magnesium atoms substituted into the frameworks) to name just four. These materials can be used as "redox molecular sieves" 71  119,125  for a variety of  chemical processes, including conversions of alcohols to hydrocarbons Although substituted  AIPO4S  126  and vice versa.  127  typically suffer from thermal stability and coking problems, there  is huge potential for their use due to their complimentary range of size, shape and chemical selectivity compared to zeolites. The issues relating to the structural analysis of  AIPO4S  are largely the same as those  associated with zeolites. The X-ray scattering from aluminium and phosphorous is sufficiently similar that it is often only possible to determine average  AIPO4  structures, as opposed to a  complete assignment of the A l and P T-sites. The crystal structure topology may only be consistent with one assignment, but this is not always the case. In addition, the organic template locations and hydration characteristics of Unlike zeolites, the T-sites of  AIPO4S  AIPO4S  are more significant than those for zeolites.  are frequently directly hydrated or co-ordinated by  occluded molecules, a fact which has the potential to give A I P O 4 S a range of functionalities 128  towards adsorbed molecules that are not found with zeolites. The factors governing these interactions are not well understood, but solid state N M R techniques have the potential to probe the local spatial relationship between framework and non-framework nuclei in these systems. In the present work are presented the results of a multiple field M A S N M R study of the hydration characteristics of  AIPO4-I8;  a material for which both the framework connectivities and the  structure of the dehydrated material, respectively, are known from X R P D .  1.3.3  ALUMINOPHOSPHATE CERAMIC MIXTURES  A range of dense crystalline phases can be produced from aluminophosphate solutions (e.g. tertiary alkoxy aluminium salts and phosphoric acid) under conditions similar to those used to synthesize A I P O 4 S , but in the absence of organic templates. These materials can be collectively classified, by their physical properties (primarily hardness, thermal stability and chemical \  -129  inertness), as ceramics. These materials have been found to bind strongly to surfaces during in situ synthesis, providing a mechanically robust protective layer that is chemically inert (with the exception of conditions similar to their synthesis). Although by no means unique in these properties, aluminophosphate ceramics can be synthesized at lower temperatures than alternative materials; 72  a fact that considerably broadens the range o f possible applications. For example, aluminophosphate ceramics have been used to encapsulate radioactive waste for permanent storage.  130  Ceramic layers around radioactive metal particles are impervious to environmental  degradation and the relatively low temperature treatment conditions circumvent many o f the risks associated with alternative "glassification" o f radioactive waste. These materials have also been reportedly used as robust supports f o r ,  131  or components o f ,  been shown to be useful for the coating of metal surfaces.  132  industrial catalysts and have  133  Factors such as the homogeneity and average crystal size, as well as the identity and phase composition o f the ceramic mixtures are crucial factors governing the physical properties of the ceramics. For example, an alternative use of aluminophosphate ceramics is as a component in ceramic-epoxy composite materials.  134  These materials require the synthesis o f a porous ceramic  component into which the epoxy can penetrate, completely unsuitable for waste management 135  and metal surface coating applications,  but similar to the high surface area requirements of  heterogeneous catalysts. Therefore the complete characterisation o f sometimes complex mixtures o f crystalline and amorphous phases is required to further understand the factors required for the design o f these materials. X R D can be an effective means for the quantitative detection and identification of the crystalline components, but in the present work it w i l l be shown that only i n conjunction with solid state N M R can a quantitative analysis of all phases be achieved. In this thesis the results o f the quantitative assignment of the phases present in a series o f aluminophosphate ceramic mixtures are presented which show how the extent of conversion o f the starting materials and the proportions o f phases comprising the ceramic mixtures vary with the type o f alumina source.  73  1.4 References for Chapter 1. 1. Atkins, P.W. Nuclear Magnetic Resonance. In Physical Chemistry - 4 ed.; Oxford th  University Press: Oxford, 1992; pp536-548. 2. Harris, R.K. Nuclear Magnetic resonance Spectroscopy; Wiley: New York, 1986. 3. Homans, S.W. A Dictionary of Concepts in NMR; Clarendon Press: Oxford, 1989. 4. Derome, A . E . Describing Pulse N M R . In Modem NMR Techniques for Chemistry Research; Pergamon Press: Oxford, 1987; pp63-96. 5. Sclichter, C P . Basic Theory. In Principles of Magnetic Resonance - 3 ed.; Springerrd  Verlag: New York, 1990; pp 11-65. 6. Ernst, R.R.; Bodenhausen, G.; Wokaun, A . Time and frequency domains. In Principles of Nuclear Magnetic Resonance in One and Two Dimensions; Clarendon Press: Oxford, 1987; pp96-97. 7. Atkins, P.W. The uncertainty principle. In Molecular Quantum Mechanics - 2  nd  ed.; Oxford  University Press: Oxford, 1993; ppl7-19. 8. Stephenson, G. Mathematical Methods for Science Students - 2  nd  ed.; Longman: London,  1984. 9. Sanders, J.K.M.; Hunter, B.K. Solids. In Modern NMR Spectroscopy - 2  nd  ed.; Oxford  University Press: Oxford, 1994; pp260-287. 10. Mehring, M . Nuclear Spin Interactions in Solids. In High Resolution NMR Spectroscopy in Solids; Springer-Verlag: New York, 1976; pp6-30. 11. Ernst, R.R.; Bodenhausen, G.; Wokaun, A . The nuclear spin Hamiltonian. In Principles of Nuclear Magnetic Resonance in One and Two Dimensions; Clarendon Press: Oxford, 1987; pp44-49. 74  12. Atkins, P.W. Techniques of Approximation. In Molecular Quantum Mechanics - 2 ed.; Oxford University Press: Oxford, 1993; ppl71-180. 13. Vega, A.J. Magnetic Resonance of Half-Integer Quadrupolar Nuclei. In New Developments and Applications of NMR and ESR Spectroscopy; Bat-Sheva Workshop: Israel, 1990. 14. Atkins, P.W. Physical Chemistry - 4 ed.; Oxford University Press: Oxford, 1992; pp332th  344. 15. Atkins, P.W. Angular momentum. In Molecular Quantum Mechanics - 2 ed.; Oxford nd  University Press: Oxford, 1993; ppl03-131. 16. Massiot, D. Quadrupolar Nuclei, notes from Bruker Symposium, Billercia, MA 1997. 17. Slichter, C P . Electric Quadruple Effects. In Principles of Magnetic Resonance - 3 ed.; rd  Springer-Verlag: New York, 1990, pp485-503. 18. Skibsted, J. Ph.D. Thesis; Aarhus Universitet, 1994. 19. Mueller, K.T. Ph.D. Thesis; Lawrence Berkley Laboratory: University of California, 1991. 20. Bloembergen, N . ; Rowland, T.J. Acta Metall, 1953,1, 731-746. 21. Mehring, M . Appendix A . Irreducible Tensor Representation of Spin Interactions. In High Resolution NMR Spectroscopy of Solids; Springer-Verlag: New York, 1976; p213. 22. Slichter, C P . Appendix I. Powder Patterns. In Principles of Magnetic Resonance - 3 ed.; rd  Springer-Verlag: New York, 1990, pp605-616. 23. Pake, G.E. J. Chem. Phys., 1948,16, 327-336. 24. Lowe, I.J. Phys. Rev. Lett., 1959, 2, 285-287. 25. Andrew, E.R.; Bradbury, A.; Eades, R.G.; Nature, 1958,182, 1659. 26. Hon, J.F.; Bray, P.J. Phys. Rev., 1958,110, 624-629. 27. O'Reilly, D.E.J. Chem. Phys., 1958, 28, 1262-1264.  75  28. Slichter, C P . Advanced Concepts in Pulsed Magnetic Resonance. In Principles of Magnetic Resonance - 3 ed.; Springer-Verlag: New York, 1990, pp367-423. rd  29. Stejkal, E.O.; Schaefer, J.; McKay, R.A. J. Magn. Reson., 1977, 25, 569-573. 30. Cohn, M . ; Kowalsky, A.; Leigh, J.S.; Maricic, S. Magn. Reson. Biol. Syst., 1967, 45-48. 31. Bonera, G.; Galimerti, M . Solid State Commum., 1966, 4, 589-591. 32. Hahn, E.L. Phys. Rev., 1950, 80, 580-581. 33. Carr, H.Y.; Purcell, E . M . Phys. Rev., 1954, 94, 630-638. 34. Meiboom, S.; Gill, D. Rev. Sci. Instr., 1958, 29, 688-691. 35. Derome, A . E . Polarisation Transfer and Spectrum Editing. In Modern NMR Techniques for Chemistry Research; Pergamon Press: Oxford, 1987; ppl29-141. 36. Sanders, J . K . M . ; Hunter, B.K. Multiple Pulse Experiments. In Modern NMR Spectroscopy 2  nd  ed.; Oxford University Press: Oxford, 1994; pp63-97.  37. Morris, G.A.; Freeman, R. J. Am. Chem. Soc, 1979,101, 760-762. 38. Morris, G.A. J. Magn. Reson., 1980, 41, 185-188. 39. Fyfe, C.A.; Wong-Moon, K.C.; Huang, Y . ; Grondey, H . J. Am. Chem. Soc, 1995,117, 10397-10398. 40. Burum, D.P.; Ernst, R.R. J. Magn. Reson., 1980, 39, 163-168. 41. Jeneer, J. Lectures to the Ampere International Summer School, Basko Polje, Yugoslavia, 1971. 42. Bodenhausen, G.; Freeman, R.; Niedermeyer, R.; Turner, D.L. J. Magn. Reson., 1977, 26, 133-164. 43. Morris, G.A.; Freeman, R. Bull. Magn. Reson., 1979, / , 5-26. 44. Bodenhausen, G.; Freeman, R. J. Magn. Reson., 1977, 28, 471-476. 45. Maudsley, A . A . ; Miiller, L.; Ernst, R.R. J. Magn. Reson., 1977, 28, 463-469. 76  46. Frydman, L.; Harwood, J.S.; J. Am. Chem. Soc., 1995,117, 5387-5368. 47. Samoson, A.; Lippmaa, E.; Pines, A . Moi. Phys., 1988, 65, 1013-1018. 48. Llor. A.; Virlet, J. Chem. Phys. Lett, 1988,152, 248-253. 49. Mueller, K.T.; Sun, B.Q.; Chingas, G . C ; Zwanziger, J.W.; Terao, T.; Pines, A . J, Mag. Res., 1990, 86, 470-487. 50. Youngman, R.E.; Werner-Zwanziger, U . ; Zwanziger, J.W.; Z. Naturforsch., 1996, 51a, 321329. 51. Wu, G; Rovyank, D; Sun, B.; Griffin, R.G.; Chem. Phys. Lett., 1996, 249, 210-217. 52. Bodart, P.R. J. Magn. Reson., 1998,133, 207-209. 53. Medek, A.; Marinelli, L.; Frydman, L . Multiple Quantum Magic-Angle Spinning N M R of Half-Integer Quadrupolar Nuclei. In Modeling NMR chemical shifts : gaining insights into structure and environment.; Facelli, J . C ; de Dios, A . C . Eds.; A.C.S. Symposium Series 732: Washington, DC, 1998; ppl36-155. 54. Amoureux, J.P.; Fernandez, C ; Frydman, L . Chem. Phys. Lett., 1996, 259, 347-355. 55. Wang, S.H.; X u , Z.; Baltisberger, J.H.; Bull, L . M . ; Stebbins, J.F.; Pines, A . Solid State Nucl. Magn. Reson., 1997, 8, 1-16. 56. Massiot, D.; Touzo, B.; Trumeau, D.; Coutures, J.P.; Virlet, J.; Florian, P.; Grandinetti, P.J., Solid State Nucl. Magn. Reson., 1996, 6, 73-83. 57. Fyfe, C.A.; Skibsted, J.; Grondey, H.; Meyer zu Altenschildesche, H . Chem. Phys. Lett., 1997, 281, 44-48. 58. Ding, S.; McDowell, C A . Chem. Phys. Lett., 1997, 270, 81-86. 59. Massiot, D. J. Magn. Reson. A, 1996,122, 240-244. 60. Baltisberger, J.H.; X u , Z.; Stebbins, J.F.; Wang, H.; Pines, A . J. Am. Chem. Soc, 1996,118, 7209-7214. 77  61. Vega, A.J. J. Magn. Reson., 1992, 96, 50-68. 62. Hayashi, S. Solid State Nucl. Magn. Reson., 1994, 3, 91-103. 63. Grey, C.P.; Vega, A.J. J. Am. Chem. Soc., 1995, 117, 8232-8242. 64. Jeschke, G. J. Chem. Phys., 1998,108, 907-917. 65. Kwak, H.-T.; Prasad, S.; Yao, Z.; Grandinetti, P.J.; Sachleben, J.R.; Emsley, L. J. Magn. Reson., 2001,150, 71-80.  66. Kwang, H.L.; Charpentier, T.; Pines, A. J. Magn. Reson., 2002,154, 196-204. 67. Bretherton, J.L., 1997: 3QMAS experiments were performed at U.B.C. on a sample of VPI5 using PFG-3QMAS (see Ref. 60) with and without the nil preparation pulse using a Bruker MSL400 spectrometer. 68. Mildner, T.; Smith, M.E.; Dupree. R. Chem. Phys. Lett., 1999, 301, 389-394. 69. Amoureux, J.P.; Fernandez, C.; Steuernagel, S. J. Magn. Reson. A, 1996,123, 116-118. 70. Battle, P.D. Public lecture series on the subject of Crystallography; Inorganic Chemistry Laboratory: Oxford, England, 1993. 71. Woolfson, M . M . An Introduction to X-Ray Crystallography - 2  nd  ed.; Cambridge University  Press: Cambridge, 1997.; pp32-42. 72. Jenkins, R.; Snyder, R.J. Introduction to X-Ray Powder Diffractometry; Winefordner, J.D. Ed.; Chemical Analysis, volume 138; Wiley: New York, 1996; ppl4-43. 73. Bragg, W.H.; Bragg. W.L. X-Rays and Crystal Structure; Bell and Sons, London, 1915. 74. Atkins, P.W. Diffraction Methods. In Physical Chemistry - 4 ed.; Oxford University Press: th  Oxford, 1992; pp617-641. 75. Ch. Baerlocher and L . B . McCusker, Database of Zeolite Structures: http://www.izastructure.org/.  78  76. Inorganic Crystal Structure Database: FIZ, Karlsrhue, Germany. 77. Jenkins, R.; Snyder, R.J. The Intensity Equation. In Introduction to X-Ray Powder Diffractometry; Winefordner, J.D. Ed.; Chemical Analysis, volume 138; Wiley: New York, 1996; pp80-82. 78. International Tables for X-Ray Crystallography- Volume A.; Hahn, T. Ed.; Kynoch Press: Dordrecht, 1993. 79. Ozin, G.A.; Kuperman, A.; Stein, A . Angew. Chem. Int. Ed., 1989, 28, 359-376. 80. Wong-Moon, K. Ph.D. Thesis, University of British Columbia, 1995. 81. McBain, J.W. Sorbtion of Gasses and Vapours by Solids; Rutlidge and sons, 1932. 82. Barrer, R . M . Nature, 1949,164, 112-113. 83. Arends, I.W.C.E.; Sheldon, R.A.; Wallau, M . ; Schuchardt, U . Angew. Chem. Int. Ed., 1997, 36, 1144-1163. 84. Mobil Oil Co., Inc.. Aluminosilicate catalysts for reforming hydrocarbons. N L Patent 6405402, Chem. Abstr., 1964, 62, 65692. 85. Plank, C.J.; Rosinski, E.J.; Leaman, W.K. High activity catalysts for hydrocarbon cracking. Fr. Patent 1398236, Chem. Abstr. 1965, 64, 34761. 86. Perot, G.; Guinset, H. J. Moi. Catal, 1990, 61, 173-196. 87. Pujado, P.R.; Rabo, J.A.; Antos, G.J.; Gembicki, S A . Catalysis Today, 1992,13, 113-141. 88. Breck, D.W.; Eversole, W.G.; Milton, R . M . ; Reed, T.B.; Thomas, T.L. J. Am. Chem. Soc, 1956, 78, 5963-5971. 89. Breck, D.W. Zeolite Molecular Sieves.; Wiley: New York, 1974. 90. Gilson, J.P. Organic and Inorganic Agents in the Synthesis of Molecular Sieves. In Zeolite Microporous Solids: Synthesis, Structure and Reactivity., Derouane, E.G.; Lemos, F.;  79  Naccache, C ; Ribeiro, F.R. Eds.; N A T O ASI Series - Vol. 352; Kluwer Academic Publishers: Boston, 1992.; pp3-19. 91. Melchior, M.T.; Vaughan, D.E.W.; Pictroski, C.F. J. Phys. Chem., 1995, 99, 6128-6144. 92. Meier, W . M . Z. Kristallogr., 1961,115, 439-450. 93. Galli, E.; Gottardi, G. Miner. Petrogr. Acta, 1966,12, 1-10. 94. Bergerhoff, G.; Baur, W.H.; Nowaki, W. N. Jb. Miner. Mk, 1958, 193- 200. 95. Pauling, L. Z. Kristallogr., 1930, 74,213-225. 96. Breck, D.W. U.S. Patent 3130007, 1964. 97. Elliot, K . M . ; Eastwood, S.C. Oil GasJ., 1962, 60, 142-144. 98. Milton, R . M . US Patent 2882244, 1959. 99. Reed, T.B.; Breck, D.W. J. Am. Chem. Soc, 1956, 78, 5972-5977. 100.  Ozin, G.A.; Stein, A.; Stuckey, G.D.; Goober, J.P. Silver Sodalites: Optically responsive  nanocomposites. In Inclusion Phenomena Conference Review; Atwood, J. ed.; Plenum: New York, 1990; pp379-393. 101.  Petranovic, N . ; Susie, M . Zeolites, 1983, 3, 271-273.  102.  Freeman Jnr., D.C. US Patent 3186875, 1965.  103.  L i , Z.; Mallouk, T.E. J. Phys. Chem., 1987, 91, 643-648.  104.  De Vismes, B.; Bedioui, F.; Devynck, J.; Bied-Charreton, C. J. Electroanal. Chem.,  1985,187, 197-202. 105.  Rolison, D.R.; Nowak, R.J.; Welsh, T,A.; Murray C.G. Tantala, 1991, 38, 27-35.  106.  Montes de Correa, C.; Corodoba, F.; Bustamante, F. Microporous and Mesoporous  Mater., 2000, 40, 149-157. 107.  Adam, W.; Mitchell, C . M . ; Saha-Moller, C.R.; Selvam, T.; Weichold, O. J. Moi. Catal.  A, 2000,154, 251-255. 80  108.  Yokoyama, C ; Misono, M . J. Catal, 1994,150, 9-17.  109.  Barker, C M . ; Kaltosyannis, N . ; Catlow, C.R.A. In Stud. Surf. Sci. Cata., 2001,135,  2580-2587. 110.  Benazzi, E.; Tivdar, C ; Marchal, G.N.; Marion, P.; Gueret, C ; Kasztelan, S. Fr. Patent  2792946, 2000. 111.  Parton, R.F.; Vankelcom, I.F.; Casselman, M.J.; Bezoukhanova, C P . ; Uytterhoeven  J.B.; Jacobs, P.A. Nature, 1994, 370, 541-544. 112.  Charnell, J.F. J. Crystal Growth, 1971, 8, 291-294.  113.  Qiu, S.; Y u , J.; Shu, G.; Terasaki, O.; Nozue, Y . ; Pang, W.; X u , R. Micropor. Mesopor.  Mater., 1998, 27,245-251. 114.  Baker, M.D.; Ozin, G.A.; Godber, J. Catal. Rev. Sci. Eng., 1985, 27, 591-651.  115.  Wilson, S.T.; Lok, B . M . ; Messina, C.A.; Cannan, T.R.; Flanigen, E . M . J. Am. Chem.  Soc, 1982,104, 1146-1147. 116.  Lok, B . M . ; Messina, C.A.; Patton, R.L.; Gajek, R.T.; Cannan, T.R.; Flanigen, E . M . J.  Am. Chem. Soc, 1984,106, 6092-6093. 117.  Zahedi-Niaki, M . H . ; X u , G.Y.; Meyer, H.; Fyfe, C A . ; Kaliaguine, S. Micropor.  Mesopor. Mater., 1999, 32, 241-250. 118.  Simmen, A.; McCusker, L.B.; Baerlocher, Ch.; Meier, W . M . Zeolites, 1991,11, 654-  661. 119.  Hartman, M . ; Kevan, L. Chemical Reviews, 1999, 3, 635-664.  120.  Zahedi-Niaki, M . H . ; Joshi, P.N.; Kaliaguine, S. Chem. Comm., 1996,1, 47-48.  121.  Zahedi-Niaki, M . H . ; Zaidi, S.M.J.; Kaliaguine, S. Appl, Catal. A, 2000,196, 9-24.  81  122.  Chen, J.; Wright, P. A.; Thomas, J. M . ; Natarajan, S.; Marchese, L.; Bradley, S.M.;  Sankar, G.; Catlow, C. R. A.; Gai-Boyes, P.L.; Townsend, R. A.; Lok, C. M . J. Phys. Chem. 1994, 98, 10216-10224. 123.  Akolekar, D.B.; Howe, R.F. J. Chem. Soc. Faraday Trans., 1997, 93, 3263-3268.  124.  Smith, J.V.; Pluth, J.J.; Andries, K.J. Zeolites, 1993,13, 166-169.  125.  Sheldon, J.D.; Chen. J.; Dakka, E.; Neeleman, E. Stud. Surf. Sci. Catal, 1994, 82, 515-  526. 126.  Notari, B. Stud. Surf. Sci. Catal, 1991, 60, 343-352.  127.  Chen, J.D.; Sheldon, R.A. J. Catal, 1995,153, 1-8.  128.  McCusker, L.B.; Baerlocher, Ch.; Jahn, E.; Bulow, M . Zeolites, 1991,11, 308-313.  129.  The American Ceramic Society, http://www.acers.org, 2002.  130.  Singh, D; Wagh, A.S.; Tlustochowicz, M.S.; Jeong Y . Waste Management, 1998,18,  135-143. 131.  Wijzen, F.; Koch, B.; Rocha, J.; Esculcas, A ; Rulmont, A.; Mater. Res. Bull, 1998, 33,  1667-1680. 132.  Bautista, F . M . ; Campelo, J.M.; Garcia, A.; Luna, D.; Marinas, J.M.; Romero, A . A . Appl.  Catal. A, 1993, 96, 175-199. 133.  Moorlag, C. M.Sc. Thesis, University of British Columbia, 2000.  134.  Craig, B.D.; Francis, L.F. J. Am. Ceram. Soc, 1998, 81, 3109-3116.  135.  Bothe Jr., J.V.; Brown, P.W. J. Am. Ceram. Soc, 1993, 76, 2353-2358.  82  Chapter 2. Experimental  2.1 Solid State NMR Spectroscopy  2.1.1  SPECTROMETERS  Solid state N M R spectra were obtained at a field of 9.4 T (400 M H z proton frequency) using a Bruker DMX400 solid state spectrometer with an Oxford Instruments 400 M H z wide bore (89 mm) superconducting magnet and a SGI Unix interface running Bruker X W I N N M R v.2.6 software. Solid state N M R spectra were obtained at a field of 14.1 T (600 M H z proton frequency) using a Bruker AMX600 spectrometer with an Oxford Instruments 600 M H z narrow bore (54 mm) superconducting magnet and a Unix32 based Bruker interface running Bruker U X N M R software and located at Simon Fraser University, Burnaby, British Columbia. Solid state N M R spectra were also obtained at fields of 17.6 T (750 M H z proton frequency) and 18.8 T (800 M H z proton frequency) using Varian Inova spectrometers with solids capabilities interfaced to Sun Microsystems workstations running Varian V N M R software. These instruments used Oxford Instruments 750 M H z narrow bore (51 mm) and Oxford Instruments 800 M H z medium bore (63 mm) superconducting magnets, respectively. Research using these instruments was performed in the Environmental Molecular Sciences Laboratory (a national scientific user facility sponsored by the U.S. DoE Office of Biological and Environmental Research) located at the Pacific Northwest National Laboratory, operated by Battelle for the DoE.  83  2.1.1(a)  General spectrometer layout.  1  All of the spectrometers used for the present research possess the basic components shown in Figure 2.1. All pulse sequence commands are controlled by the computer interface. The transmitter frequency is generated by the synthesizer and the signal divided into rf. pulses of a particular phase and duration by the modulator.  •  Cryostat  Synthesizer  Modulator  Router Liq. helium  Amplifier 1 low power  Liq. nitrogen  Amplifier 2 high power  Vacuum Superconducting magnet coils Superconducting cryoshims  Filter Receiver  Room temperature shims  Pre-Amplifier z Digitizer  Interface Figure 2.1 Component parts of a solid state N M R spectrometer and simplified cross-sectional diagram of a  superconducting N M R magnet. Refer to text for details.  Pulse amplitude is set at the first of a two stage amplification process. The second stage high power amplification of this signal is a particular feature of solid state spectrometers 84  (Section 1.1.4). The AMX600 spectrometer used in this study was not equipped with high power amplification and was limited to a maximum rf. power output at the amplifiers of =50 W. Depending on the experimental setup, the other spectrometers used in this study were equipped with 300 W or 1 kW amplifiers on all channels. Bruker spectrometers use "gated" low power amplifiers that must be driven by a minimum amplitude of signal before activating. By this method amplified random noise of all frequencies is prevented from reaching the probe. Varian spectrometers use un-gated amplifiers and gating is applied by the receiver only. Experimental noise at the observation frequency is minimized by the use of an additional carrier frequency (10 M H z in the case of the 750 M H z instrument and 20 M H z in the case of the 800 M H z instrument) which is subtracted by a "step down unit" within the pre-amplifier prior to delivery of the pulses to the probe. The Bruker approach minimizes experimental noise. The Varian approach facilitates the generation of square pulses by reducing the amplifier rise time, thereby allowing more accurate control over pulse shapes. For the majority of experiments employed in this work, the two methods of pulse generation are equivalent. Where un-gated amplification presented a problem, the Varian instruments could be configured to operate in a "pulsed" or gated mode. The amplified signal is passed through analogue filters to the probe, where it is delivered to the sample. The observation frequency must be routed through the pre-amplifier where it is internally routed to the probe or to the receiver. Any other frequencies are passed directly from the amplifiers to the probe and filters are inserted in-line at the probe. At the conclusion of the pulse sequence the signal generated by the sample is amplified at the pre-amplifier (i.e. pre-receiver amplifier) and sent to the receiver. The transmitter frequency, which is directed by the router to the receiver during the acquisition time, is subtracted at the receiver. The digitizer (sometimes called an "analogue to digital converter") samples discrete data points and the digital FID in the rotating frame is sent to the interface for processing. In the latest generation of spectrometers, the transmitter frequency is generated digitally, the modulator controls the phase of the transmitter digitally and the can signal is digitally filtered at the receiver. The first stage amplifiers perform a digital to analogue conversion. With older spectrometers (the AMX600 being an example of this) the synthesizers generate an analogue transmitter frequency and the filtering of the signal is performed by one of a set of band pass filters within the receiver.  85  The router and modulator additionally perform phase cycling. Quadrature detection is achieved by sending transmitter frequencies with orthogonal phases to the receiver, where they are subtracted from duplicates of the signal. This is simultaneous quadrature detection. The AMX600 cannot simultaneously handle two signals and quadrature detection is achieved by subtracting orthogonal transmitter frequencies sequentially. For experiments utilizing two of more frequencies, or "channels", additional synthesizers, modulators and amplifiers are required.  2.1.1 (b)  Experimental Setup.  A variety of commercial probes equipped with M A S stators were used to acquire the 31  2 7  Al,  P and S i M A S N M R spectra presented in this work. Depending on the tuning range of the 2 9  probe, the magic angle was experimentally optimized by maximizing the amplitude of the rotational echoes of the FID of solid K B r , N a N 0 or K 79  2 3  3  1 2 7  I (100.2 M H z , 105.8 M H z and 80.0  MHz, respectively, at 9.4 T). A l l shimming was performed on either the H signal of water or ]  the  77  77  A l signal of aluminium nitrate solution, to a value of A v « 0.07 - 0.15 ppm. A l N M R L  2  spectra are referenced to 1.0 M AlCNO^aq) = 0 ppm.  H3PO4  2 9  = 0 ppm. P N M R spectra are referenced to 85% 31  S i N M R spectra are referenced to the S i signal from zeolite NaA at -89.7 2 9  ppm from tetramethylsilane (TMS). Pulse length measurements were made directly on the reference samples a n d A l solid pulse lengths converted from the value measured for 1.0 M 27  Al(N0 ) (a?). 3  3  2.1.1(c)  M A S Probes.  M A S N M R spectroscopy was performed using a variety of home built and commercially manufactured probes: Commercial Bruker triple tuned H X Y M A S probe designed for the DMX400 spectrometer, equipped with a silicon nitride stator using 4 mm diameter zirconia sample rotors and Kel-F caps, capable of a maximum spinning rate of 15 kHz.  86  Home built double-tuned P / A 1 probe designed for the DMX400 spectrometer 31  27  equipped with a supersonic Doty spinning system composed of silicon nitride, using 5 mm silicon nitride rotors and Aurum caps, with a maximum spinning rate of 17 kHz. Home built double-tuned P / A 1 probe designed for the Varian Unity 18.8 T 31  27  spectrometer equipped with a supersonic Doty spinning system composed of silicon nitride, using 5 mm silicon nitride rotors and Aurum caps, with a maximum spinning rate of 17 kHz. Home built single frequency probe designed for narrow bore magnets equipped with a supersonic Doty spinning system composed of zirconia with a silicon nitride coil insert, also capable of a maximum spinning rate of 17 kHz. This probe was modified to tune to the A1 frequency at 14.4 T (156.4 MHz), 17.6 T (195.4 MHz) and 18.8 T 27  (208.4 MHz) at various times. Home built double-tuned H / S i probe designed for the DMX400 spectrometer 1  29  equipped with a standard speed Doty spinning system composed of Kel-F, using 7 mm zirconia rotors and Kel-F caps, with a maximum spinning rate of 5 kHz. Home built single frequency S i probe designed for the DMX400 spectrometer 2 9  equipped with a supersonic Doty spinning system composed of zirconia with both zirconia and silicon nitride 10 mm rotors and Aurum caps, capable of a maximum spinning rate of 6 kHz.  2.1.2  SPECTRAL PROCESSING.  2.1.2(a)  Zero Filling and Apodization.  2  Zero filling and apodization are routinely applied to the FID to enhance the appearance of the spectrum after the Fourier transform. No additional information is contained in spectra following apodization or from zero filling to more than In points (where n is the number of experimental data points), but that information can be made easier to interpret.  87  For a given spectral sweep width, the spacing between data points in the frequency domain is inversely proportional to the total number of points in the time domain. A decrease in the number of Hz per point can be achieved by lengthening the acquisition time. Assuming that the FID has already decayed to zero, the same result is achieved by the addition of a string of data points of zero amplitude to the experimental FID, without the disadvantage of the addition of further experimental noise. This is known as "zero filling".  (b)i)  (a)i)  (c)i)  FT  FT  f (b)ii)  (a)ii)  (c)ii)  Figure 2.2 Fourier transformation of a truncated FID. (a)i) A FID that does not decay fully to zero can be represented by the product of (b)i) a step function and (c)i) a full FID. These Fourier transform into (b)ii) a sine function and (c)ii) a Lorentzian function and the Fourier transformation of the truncated FID (a)ii) is therefore the product of these. The characteristic spectral distortions associated with truncation are known as "sine wiggles".  88  The N M R experiments in this work that were performed on Bruker instruments were acquired with an excess of time domain data points. During processing, the time domain was truncated at the point where the FID was no longer apparent above the experimental noise and zero filled to a number 2" such that 2" is typically between two and four multiples of the number of remaining experimental time domain points. The Varian V N M R interface does not permit post-acquisition truncation of the FID before zero filling. Where necessary, these data were processed using X W I N N M R . For the indirectly detected dimension of 2D experiments that give low S:N it may not always be practically possible to obtain enough time domain data points for the signal in t to x  fully decay. Fourier transform of a truncated FID is equivalent to the product of the Fourier transformation of the full FID and a square step function (Figure 2.2). The effect of truncation can be reduced or eliminated by "apodization". If the experimental data is multiplied by an exponential decay function, exp(- t/T), the truncated FID is equivalent to the product of the full FID, the step function and the exponential apodization function. If the exponential function is close to zero at the step (Figure 2.2(b)i)) then the product of the exponential decay function and the step function is approximately equivalent to the exponential decay alone. The Fourier transform of the resulting apodized FID results in a Lorentzian signal (cf. Figure 2.2(c)ii)) that has been broadened by an amount T' Hz. i.e. Exponential apodization x  has the same effect as an exponential relaxation parameter. It is for this reason that exponential apodization is usually expressed in units of Hz. A secondary effect of exponential multiplication is to enhance S:N at the expense of resolution. The experimental data are multiplied by unity where there is the largest amplitude of signal (t = 0) and by exp(- t/T) where signal amplitude is smaller. Random noise is constant throughout the acquisition time and therefore the signal is selectively enhanced. Apodization can also be applied to enhance resolution by the application of an exp(+ t/T) function at the expense of S:N. In solid state N M R this is rarely beneficial. A variety of other mathematical functions can be applied to the FID, such as sin(a + t/T) or sin (a + t/T). Each apodization function produces a characteristic distortion of the line shape. For the spectra presented in this work, exponential multiplication, or no apodization is used almost exclusively. Where relevant, the exponential broadening is quoted in the figure captions.  89  2.1.2(b)  2D Spectroscopy.  2.1.2(b)i) Phase sensitivity in the indirectly detected dimension. The general considerations for spectral processing in the second, indirectly detected dimension are the same as for the first, directly detected dimension. The important difference relates to quadrature detection. Two methods were employed for the phase sensitive, quadrature detection in the indirectly detected dimension of the 2D spectra presented in this work; hypercomplex acquisition and Time Proportional Phase incrementation (TPPI). 3  4  As discussed in Section 1.1.5, quadrature detection requires the linear combination of orthogonal data sets. Use of the hypercomplex method effectively requires that two full 2D data sets are acquired that differ only by the phase of either the receiver or a pulse that governs the phase of the signal, i.e. For each value of t two FIDs are acquired where the phase of the u  receiver and the FID differ by 7t/2. The means by which this is achieved does not affect the result, providing that the same method is used throughout. Following the incrementation of t , x  the phases are re-set such that the phase of each pair of FIDs are generated by the same pulse and receiver phases. The Fourier transform and phasing in F l is then analogous to the simultaneous quadrature detection in F2. TPPI is the incrementation of the phase of either the receiver, or a pulse that governs the phase of the signal, by TC/2 with each increment of t . The pulse and receiver phases return to the x  starting point every four values of t . The Fourier transform and phasing in F l is then analogous x  to sequential quadrature detection in F2. These two methods are entirely equivalent. TPPI apparently employs more thorough phase cycling. However, all of the phase cycling necessary for the minimization of instrumental artefacts is carried out within each FID in F2. The choice of TPPI or hypercomplex F l quadrature detection and the means by which the phases are alternated is determined by considerations specific to each experiment and is primarily dictated by which method gives the best observed results.  90  2.1.2(b)ii)  3QMAS experiments.  The three variations of the 3QMAS experiment used in this work are 2D experiments in which the indirectly detected time domain is an incremented evolution time prior to the final pulse and the directly detected time domain is the collection of an echo. If the increment in the F l time domain is 8t then the increment in the time at which the echo reaches maximum x  amplitude is S\k\t . The value of k (Table 1.1) is discussed in Section 1.1.10(b). x  The experiment can account for this in two ways: The acquisition in t can be shifted in time 2  to always begin at the echo maximum, i.e. t = 0 moves away from the second pulse at the same 2  rate as the echo. The sweep width in F l is then equal to \/St (l + \k\). Following the Fourier x  transformation of data acquired in this way, the F l and F2 dimensions of the 3QMAS spectrum are the "isotropic" and M A S dimensions described in Section 1.1.10(d). However, quadrature detection in F l is not straightforward using this approach, since control of the relative phases of FIDs taken at successive values of t (an F l sweep width dependent property) is problematic. x  The t data of the 3QMAS experiments presented here were acquired beginning after the 2  ringdown delay following the final pulse. Therefore the echo becomes more distant in the F2 time domain as a function of U by the amount S\k\t and the sweep width in F l is \/St . The 2D x  x  Fourier transformation of the resulting data produces a spectrum in which the "isotropic dimension" and the M A S dimension are no longer aligned with F l and F2 respectively (Figure 2.3(a)). However, the pulses and the magnetization are identical in both experiments, and therefore the 3QMAS spectra must also be identical. This can be explained and corrected for by using the Shift Theorem. A result of the Fourier transform is simply that a time shift in the time domain 5  produces a frequency shift in the frequency domain. If tis the "shift" (in time) then this can be expressed as:  / ( f - r ) = exp(tor)/W  91  (2.1)  (b)  _l  60  •  r  40  20  0  -20  -40  ppm  Figure 2.3 (a) Unsheared and (b) sheared z-filtered A l 3 Q M A S spectrum of calcined, hydrated A I P O 4 - I 8 . 2 7  Spinning sidebands are denoted Skyline projections are shown. Calcined, hydrated A1P0 -18 contains six aluminium environments, marked in the figure. In the unsheared spectrum (a) the isotropic dimension and the MAS dimension are orthogonal to each other but do not lie parallel to the F l and F2 axes. The F l projection is therefore broadened compared to a projection along the isotropic dimension. In the sheared spectrum (b) the isotropic dimension and the MAS dimension are aligned with Fl and F2 respectively. The F l projection shows six sharp peaks and spinning sideband peaks. The scaling of the F l ppm scale is also different to that in (a) such that the F2 projections of (a) and (b) are equivalent. 4  92  The time shift in t can be corrected for by a "shearing transformation" analogous to that x  first proposed for the D A S experiment. The shearing transformation is the application of a time 6  (ti) dependent first order phase correction (i.e. a phase correction proportional to 6%) following the first Fourier transform. The steps in the 2D processing of 3QMAS spectra are:  S„ («<*, ,t )= S* (ndt, ,t )+ iS* (*«*,, t )  (2.2)  S («<*,, t ) — ^  (2.3)  2  n  2  2  2  S'„ (ndt, ,co ) 2  S (nSt ,eo )= e x p ^ A : ^ , ^ )s' (n&,, <o )  (2.4)  t,S"M a> )-^f(a> iD )  (2.5)  n  n  x  n  2  lt  2  lt  2  2  n=0  Where: S (n5t , t ) n  x  Raw «th data set of the 2D experiment.  2  Real and imaginary components acquired at the same value of ti using the hypercomplex method. 5^ (nSt , 0) )  nth data set in the mixed time domain/frequency domain  2  t  interferogram following first Fourier transform. S"(nSt , co ) {  nth data set in the sheared, mixed time domain/frequency  2  domain interferogram. ^ S" (n8t , co ) n  x  2  Sheared, mixed time domain/frequency domain  71=0  interferogram with ni increments in t . x  f(co , co ) x  2  Frequency domain 2D spectrum.  The shearing transformation itself (Eqn 2.4) is performed in X W I N N M R using the programme mqxfb (Appendix III.A(xi)), a modified version of xj"shear. It is performed in 7  V N M R using the programme daslp that is a component part of that software. 93  The steps between Eqns 2.2-2.5 represent the most convenient method of applying the shearing transformation within the spectrometer processing software. However, if the shearing transformation were applied prior to Fourier transformation it would have the additional effect of changing the sweep width in F l to \/St (l + T h i s does not occur if the shearing l  transformation is applied following the first Fourier transformation and therefore the F l frequency scale must be modified in order that the F l shifts correspond to Eqn 1.69. mqxfb has been adapted from xfshear in order that the F l chemical shift scale (ppm) follows general 8 11  conventions. " Spectra acquired on Varian instruments were referenced manually with respect to the transmitter frequency in F2 according to the same conventions, since daslp cannot correctly scale the F l chemical shift.  2.2  Powder X-Ray Diffraction X R P D patterns were acquired using a Rikagu Rotorflex diffractometer with pseudo-  focusing Bragg-Brentano geometry (Figure 2.4). The instrument was equipped with a water 12  cooled rotating copper anode X-ray source, rotating at 2500 rpm and operated at 7.5 kW (50 13  k V and 150 mA) collimated with 0.3 mm dispersion and 0.6 mm Soller slits in both the incident and diffracted beam and monochromated before the detector. B y this method reflections of Cu Koc radiation (X = 1.5418 A) are measured, but there is no resolution of Cu K a i and Cu K a  2  radiation. Samples were pressed into cavities in the faces of aluminium plates which were then mounted vertically. Both the sample and goniometer (the apparatus containing the detector, a monochromator and a pair of slits) rotate in order to maintain the geometric relationship indicated in Figure 2.4. This arrangement produces a pseudo focusing effect to compensate for differing angles of incidence at different parts of the sample. A true refocusing effect would be produced by a concave sample, but the errors associated with a flat sample are small. Fingerprint diffraction patterns were typically recorded between angles of 20 = 4° and 40° using a step size of 0.02° at 2°min" . Acquisition parameters are presented in figure captions 1  accompanying the data.  94  Detector  Sample Figure 2.4 Plan view of the Bragg-Brentano geometry of the Rikagu Rotorflex diffractometer. The sample  rotates at half the rate of the goniometer, maintaining the 0-20 ratio as shown. With the detector thus placed, this arrangement produces a pseudo-focusing effect. DS and SS denote divergence and Soller slits, respectively.  X R P D patterns were also acquired on two Siemens D5000 diffractometers. The geometry of this equipment is analogous to Figure 2.4, except the sample is loaded into a vertical glass capillary that is rotated around its axis during the experiment to minimise the effects of any preferred crystallite orientation. Fingerprint patterns were collected with C u K a radiation between 26 = 5° and 70° at a rate of 0.0257sec on an instrument located at the Department of Mineralogy, University of British Columbia. High resolution patterns were collected by the research group of Professor Hermann Gies at Institut fur Minerologie, Ruhr Universitat Bochum, Germany. The scan range was 20 = 5°- 40° in steps of 0.0077°. Each scan took approximately 4 hours with 4 scans acquired for each sample.  95  2.3  Thermogravimetric Analysis. Thermogravimetric analyses (TGA) were carried out using a T A Instruments TG51  thermogravimetric analyser under a flow of 80 cc/min of dry nitrogen gas. The platinum sample pan and quartz balance arm, balance assembly and thermocouple were housed in a sliding glass dewar and connected to the nitrogen supply by a needle valve and vented to a fume hood. The sample chamber was inserted into the tube furnace and the temperature was ramped from room temperature to 800°C at a rate of 10°C/min and maintained at this temperature for 60 minutes to ensure constant mass. A starting sample mass of between 15 mg 25 mg was used for each analysis.  2.4 Phvsisorption Measurements. Physisorption measurements were made using a Micrometrics A S A P 2010 instrument using nitrogen as the analysis gas. Samples were degassed overnight at 350°C under vacuum prior to analysis. Experiments were performed at 77.3 K with the sample vessel immersed in a liquid nitrogen bath. Surface area measurements were fit to a Langmuir isotherm ' and porosity 14 15  measurements were obtained from each isotherm using a Saito-Foley modified HovarthKawazoe analysis. The Langmuir isotherm (Eqn 2.7) is based on a physical model that is appropriate for the measurement of the internal surface area of molecular sieve materials.  e = -^— 1+  (2.7)  Kp'  k Where the equilibrium constant K = — and is calculated from the adsorption rate constant (k ) and the desorption rate constant (k ) at a specified temperature. Partial surface coverage 9 a  b  (where 0 < 9 < 1) at that temperature is therefore a function of the adsorbent pressure partial pressure, p'.  96  16  The derivation of Eqn 2.7 assumes that the surface consists of an array of equivalent potential adsorbent sites and that adsorbent-adsorbent interactions are negligible. The model also assumes that a maximum of a complete monolayer of adsorbent molecules may adsorb to the surface and does not allow for differences in adsorption and desorption rates associated with different surface environments (such as those associated with defect sites, for example). For this reason, the extremely low pressure regions of the acquired nitrogen isotherms (p/p < 10" ), 5  0  where a clear deviation from the mathematical form of Eqn 2.7 is evident, were not considered reliable. Alternative isotherms that are based upon a physical model that permits further layers of physisorption cannot accommodate the restrictions imposed by the zeolite pores and channels and therefore the most representative surface area measurements are those taken from the low pressure (p/p ^ 10" ) region of the nitrogen adsorption isotherm. 3  17  0  The surface area (S, m g ) is calculated (Eqn 2.8) from the physical constants A (cross 2  _1  sectional area, m ) and M (molecular mass, gmol") of the adsorbent molecule and from the weight of the monolayer (W, g). Wis calculated from the sum of the pressure differences following fixed doses of sample gas up to the pressure required for the formation of a monolayer  (9= 1) determined from Eqn 2.7.  S=  WN A A  M  (2.8)  Values of S calculated in this way are known as Langmuir surface areas. The Horvath-Kawazoe pore size analysis was developed to measure pore size 18  distributions in layered carbon materials, where the distances between layers is of the same order of magnitude as the pore channel diameters of zeolites. The Saito-Foley modification '  19 20  uses a tubular pore model of the structure. The mathematical relationships between partial pressure and pore diameter arise from the summation over the appropriate geometry of LenardJones (i.e. potential energy, U = -Ar'  6  \PoJ  *-o  +Br~ ) interactions. n  1  21  2k + l  32  97  (2.9)  Where a and B are ^-dependent constants, d is the equilibrium separation between the 0  surface and the adsorbed molecules and r is the pore radius. Eqn 2.9 does not yield realistic physical constants for the surface potential for molecular sieve systems and a and B should be 21  treated as empirical parameters. However, the analysis is useful for measurements of relative pore size distributions and was used in this fashion in the present work. Mesopore areas were calculated using the method of Barrett, Joyner and Halenda (the B JH method). The calculation assumes that at the highest pressures of the isotherm close to p/p = 22  0  1, all of the mesopores in the sample are cylindrical pores that are fully filled with condensed adsorbate molecules. The "adsorbed layer" is the first layer of adsorbate molecules covering the walls of the mesopores. The "core" are those adsorbate molecules condensed within the pore and within the adsorbate layer. The BJH method treats isotherm data as a desorption record and assumes that the core of a mesopore evaporates simultaneously at a critical pressure that is a function of the radius of that pore, as given by the Kelvin equation (Eqn 2.10) and that the adsorbate layer desorbs gradually.  23  Jp] KPo)  =  .W * r  R  ( 2 1 0 T  Where V is the molar volume of the adsorbate and r is the pore radius. This pore radius can be related to the volume of desorbed gas and the length of the pore and following a complex series of corrections for changes in the thickness of the adsorbed layer the incremental surface areas associated with each pressure change can be summed to give a total surface area for different ranges of r.  98  2.5 fl-Hexane Catalytic Cracking Activity Measurements. Catalytic n-hexane cracking activity measurements were made using a quartz reactor system within a tube furnace coupled to a Hewlett Packard 5890A Gas Chromatograph. The gas chromatograph (GC) was equipped with a pneumatically actuated injector valve, a packed column (using Chromosorb 101) and a hydrogen flame ionization detector. For each measurement, 150 mg of sample was pelletised to ensure a constant rate of diffusion to the sample surface. The sample temperature was ramped to 540°C at 10°C/min in a 50 cm /min 3  flow of helium gas prior to a 30 minute activation period in a 10 cm /min flow of compressed 3  air. A 50 cm /min flow of the helium carrier gas containing a saturated vapour pressure of nhexane (>99%, Aldrich) was passed over the activated sample for 4 minutes a sample then taken and the products then analyzed at this time. The G C oven temperature was ramped from 30°C to 175°C at a rate of 10°C/min and the C1-C4 product distributions measured using a Hewlett Packard 3396 integrator. Sample size, flow rate and reaction time were chosen so as to avoid excessive coking of the samples, using the experimental conditions specified for the Mobil alpha test.  24  The apparatus was found to be sensitive to the ambient conditions and the levels of hexane in the bubblers (Figure 2.5) and therefore each measurement was accompanied by a measurement of the activity of a commercial sample of ultra-stable acid Y (USY), LZY-84, purchased from UOP. The GC response and integrations of the resolved product isomers were calibrated using a flow of a gas mixture of known composition (2.99% methane, 3.98% ethane, 2.98% ethene, 10.94% propane, 3.70% butane, 79.39% helium) directed through the apparatus under analysis conditions. Experiments performed without a sample present showed negligible background hexane conversion (ca. 0.05 %).  99  Air Flow meter Needle valve Tube furnace 3-way valve  He  -Sample • Quartz wool  Temperature equilibration water  Key:  x=^^J/  - Quartz tube  >=!^^/  'Vent n-Hexane bubblers  Compressed air  !•  Gas Chromatograph and integrator  Dry helium «-Hexane loaded helium  The apparatus has three gas circuits; dry helium, a saturated vapour pressure of w-hexane over the helium carrier gas (generated by passing the helium supply through two water-jacket equilibrated bubblers) and a sample activation air supply. Three-way valves and needle valves control the direction and flow of gas during each phase of the experiment. A vent at the base of the quartz reactor tube must be opened during sample activation, during which timefirstdry, then hexane-loaded helium, bypasses the GC to a fume hood. Prior to analysis, the vent is closed and the hexane-loaded helium is directed through the sample and after 4 minutes a sample of the flow of hexane and product-loaded helium is injected into the GC. F i g u r e 2.5 F l o w d i a g r a m o f n - h e x a n e c r a c k i n g u n i t .  100  2.6 References for Chapter 2. 1. Derome, A . E . Practical Implications of Pulse N M R . In Modem NMR Techniques for Chemistry Research; Pergamon Press: Oxford, 1987; ppl4-18. 2. Sanders, J . K . M . ; Hunter, B.K. Sensitivity and resolution enhancement. In Modern NMR Spectroscopy - 2 ed.; Oxford University Press: Oxford, 1994; pp30-33. nd  Magn. Reson., 1982, 48, 286-292.  3. States D.J.; Haberkorn, R.A.; Ruben, D.J.  4. Marion, D.; Wuthrich, K. Biochem. Biophys. Res. Commun., 1983,113, 967-974. 5. Massiot, D. Quadrupolar Nuclei, notes from Bruker Symposium, Billercia, MA 1997. 6. Grandinetti, P.; Baltisberger, J.H.; Llor, A.; Lee. Y . K . ; Werner, U . ; Eastman, M . A . ; Pines, A. J. Magn. Reson., 1993,103, 72-81. 7. Fernandez, C.; Steuernagel, S. xfshear processing macro for X W I N N M R v.2.6, Bruker Analytik, GmbH., 2000. 8. Youngman, R.E.; Werner-Zwanziger, U . ; Zwanziger, J.W.; Z. Naturforsch., 1996, 51a, 321329. 9. Massiot, D.; Touzo, B.; Trumeau, D.; Coutures, J.P.; Virlet, J.; Florian, P.; Grandinetti, P.J., Solid State Nucl. Magn. Reson., 1996, 6, 73-83. 10. Medek, A.; Harwood, J.S.; Frydman, L. J.Am. Chem. Soc. 1995,117, 12779-12787. 11. Skibsted, J. Personal Communication, 2000. 12. Snyder, R.L.; Jenkins, R. The Bragg-Brentano Diffractometer. In Introduction to X-Ray Powder Diffractometry; John Wiley and Sons: New York, 1996; ppl80-187. 13. Woolfson, M . M . X-Ray Sources. In An introduction to X-Ray crystallography - 2 ed.; nd  Cambridge University Press: Cambridge, 1997; ppl43-151. 14. Langmuir, I. Phys. Rev., 1915, 6, 79-80. 15. Langmuir, I. Phys. Rev., 1916, 8, 149-176. 16. Atkins, P.W. Adsorption isotherms. In Physical Chemistry - 4 ed.; Oxford University th  Press: Oxford, 1992; pp885-887. 17. Sawada, J.A. Ph.D. Thesis; University of British Columbia, 2000. 18. Horvath, G.; Kawazoe, K . J. Chem. Eng. Jpn. 1983,16, 470-475. 19. Saito, A.; Foley, H.C. AIchEJ. 1991,37, 429-436. 20. Saito, A.; Foley, H.C. Microporous Mater. 1995, 3, 531-542.  101  21. Rakiewicz, E.F.; Mueller, K.T.; Jarvie, T.P.; Sutovich, K.J.; Roberie, T.G.; Peters, A.W. Microporous Mater. 1996, 7, 81-88. 22. Barrett, E.P.; Joyner, L.S.; Halenda, P.P. J. Am. Chem. Soc., 1951, 73, 373-380. 23. Micrometrics ASAP 2010 Accelerated Surface Area and Porosimetry System Operator's Manual vl.02,1994, cl3-c22. 24. Miale, J.N.; Chen, N . Y . ; Weisz, P.B. J. Catal, 1966, 6, 278-287.  102  Chapter 3. Multiple Field A1 MAS and 30MAS Z/  NMR Characterization and Quantification of the Aluminium Environments in Ultrastable Zeolite Y.  3.1  Introduction  Acid zeolites are formed by exchanging N H  + 4  for the charge-balancing cations such as N a  +  present after synthesis. The ammonium ions decompose above 350°C, liberating ammonia and leaving behind " H " . Steam treatment of the resulting acid zeolites at elevated temperatures +  (generally at or above 500°C) removes aluminium from the frameworks and yields catalytically active "ultrastable" catalysts that show a high degree of thermal stability in anhydrous conditions.  1  Early S i M A S N M R studies " of the most widely used of these catalysis, U S Y , clearly 2 9  2  5  documented the increase in the framework silicon to aluminium ratio (Si:Al) which the steaming produced, but in these studies the nature of the aluminium species present was much less well described due to the poor resolution of t h e A l M A S N M R spectra. The relatively low magnetic 27  fields (< 9.4 T) and spinning rates (~ 4 kHz) which were available at that time gave A l M A S 2 7  N M R spectra with overlapping quadrupolar broadened resonances covering a frequency range greater than the spinning rate. However, it was clear that six co-ordinate aluminium (Al° ) was produced even under the ct  mild conditions employed for calcination (ca. 500°C) and that steaming generated substantial amounts of material with a distribution of distorted aluminium environments, characterised by a broad resonance between the signals of the tetrahedral framework aluminium ( A l Al  0 c t  F w k T e t  ) and  at approximately 60 ppm and 0 ppm, respectively. " It was generally assumed that this 2  9  material was amorphous in nature, but there were numerous, contradictory assignments made of  103  its co-ordination. Various authors have demonstrated that the maximum in the chemical shift for this resonance was similar to that of five co-ordinate aluminium ( A l resonance to this species. ' '  Pent  ) and assigned the  However, other authors proposed that it was due to aluminium  7 8 11,12  in a more distorted tetrahedral environment ' '  5,6 10 13  (Al  BrTet  ).  Quantitation of A l M A S N M R spectra has also been a general problem, with 2 7  approximately 30% of the total aluminium known to be present not being observed even when very small pulse angles were used; the so-called "invisible aluminium". ' ' " ' This in turn has 7  2 5 8  11 13  lead to the interpretation of the physical and chemical properties of U S Y in terms of the proposed nature of the invisible aluminium in the absence of direct evidence. More recently, the availability of higher field strengths (11.8 T, 500 MHz for protons) and the use of nutation dependent spectroscopy to distinguish aluminium resonances by the quadupolar coupling 12  dependence of the nutation rate have provided further information on the multiple aluminium environments present. A n excellent summary of this work has been given by Fripiat et al} who 1  also drew parallels between the A l N M R spectra of the U S Y materials and those of amorphous 27  alumina gels containing four, five and six co-ordinate aluminium ' and interpreted the 14 15  N M R spectra of U S Y in terms of A l  F w k T e t  , Al  0 c t  and either of A l  P e n t  or A l  B r T e t  2 7  Al  .  There is a small improvement in the resolution of the A l M A S N M R spectra when the 27  satellite spectra are observed, but again the interaction is only reduced and not eliminated. In 16  an application of the DOR technique to USY, Ray and Samoson obtained only a marginal gain 7  in A1 spectral resolution because, apart from problems with overlapping sidebands, DOR is not 27  able to separate signals from distributions of chemical shifts and quadrupolar couplings such as those that are present in amorphous systems. DAS experiments provide an additional spectral 17  dimension that can potentially provide this resolution, but are not well suited to systems with very rapid T relaxation and no A l DAS N M R studies of ultrastable zeolites have been 2 7  x  published. The recent availability of very high magnetic field strengths (up to 18.8 T) and the development of the M Q M A S experiment (i.e. a 2D quadrupolar line-narrowing experiment with no additional mechanical limitations that is also suitable for the study of systems with fast T ) x  provide conditions that are far more favourable for the study of the aluminium environments in USY. Results presented in this chapter show the quantitative detection of all of the aluminium environments present in an ultrastable zeolite Y (USY) material similar in composition to the  104  U S Y catalysts used in the petrochemical industry as acid cracking and reforming catalysts. A l M A S N M R spectra acquired at various magnetic field strengths (18.8 T, 17.6 T, 14.4 T and 9.4 T) have been simulated using N M R lineshape parameters that are consistent with A1 3QMAS 27  N M R spectra. These parameters are compared to those obtained from the simulation of similar spectra of an amorphous alumina material reported to be analogous to the extra-framework material present in U S Y . ' 1 1  1 4  3.2 Materials and Methods  3.2.1  SYNTHESIS OF AMORPHOUS ALUMINA AND U S Y  A reference amorphous alumina material was prepared by the method of Coster and Fripiat. ' Glacial acetic acid was added to a 0.4 M solution of tri-2-butoxyaluminium in 211 15  butanol to give a solution containing an approximately 0.7:1 H20:A1 ratio. The solution was stirred at 85°C for 3 days and then NH4CO3 added to cause the precipitation of an alumina gel. After washing with 2-butanol, the gel was calcined at 550°C for 24 h. The U S Y sample was prepared by Dr. L . Y . Lam, Petrobras CENPES Research Centre, Rio de Janeiro, Brazil. The sodium zeolite Y (NaY) starting material was ammonium exchanged and steam calcined at 600°C, ammonium re-exchanged and again steam calcined at 600°C. At the Petrobras CENPES Research Centre, the framework Si: A l was determined by IR spectroscopy to be 9.1 and the "crystallinity" was determined by X R P D to be 92% compared to the NaY starting material.  3.2.2  EXPERIMENTAL APPARATUS  X R P D patterns were obtained between 20 = 5° and 70° at a rate of 0.025°/s using a Siemens D5000 powder diffractometer and Cu Koc radiation. Thermogravimetric analysis (TGA) was carried out using a T A Instruments TG51 thermogravimetric analyser under a flow  105  of 80 cc/min of dry nitrogen gas. The temperature was ramped from room temperature to 800°C at a rate of 10°C/min and maintained at this temperature for 60 minutes. Single pulse S i M A S N M R spectra were obtained at 9.4 T (79.5 MHz) using a Bruker 2 9  MSL400 spectrometer and a home built probe equipped with a standard speed Doty 7 mm stator with a 90° pulse width of 8.0 us. Single pulse and 90°-180° echo A1 M A S N M R spectra were 27  obtained at 9.4 T (104.3 MHz) using a Bruker MSL400 spectrometer and a home built probe equipped with a Doty 5 mm supersonic stator with a liquid state 90° pulse width of 4.8 us. Nutation A l 3QMAS spectroscopy was performed at this field using the same probe with a 2 7  Bruker DMX400 spectrometer. A l N M R M A S and 3QMAS N M R spectra were also acquired 2 7  at 14.4 T (156.4 MHz) using a Bruker A M X 600 spectrometer and at 17.6 T (195.4 MHz) and 18.8 T (208.4 MHz) using Varian Inova spectrometers (see Acknowledgements). A l l A l N M R 2 7  spectra were obtained using home built M A S probes equipped with Doty 5 mm supersonic stators. The liquid state 90° pulse lengths were 6.0 us (156.4 MHz) and 5.1 us (208.4 MHz). At 27  all fields  A l M A S N M R experiments were conventional and echo experiments were rotor  synchronised. Pulse lengths of 1.0 us were used for the single pulse experiments and for the first pulse of the echo experiments. Referencing and processing are discussed in Chapter 2 and the pulse sequences used are listed in Appendix III.  3.3 Results and Discussion  3.3.1  CHANGES IN THE FRAMEWORK  The changes in the S i M A S N M R spectra between the NaY starting material (Figure 29  3.1(a)) and the U S Y material (Figure 3.1(b)) are consistent with literature data. " There is a 2  5  clear increase of the upfield peak intensities in the spectrum of U S Y indicating an increase in framework Si:Al. Using the peak assignments indicated in the figure, ' where Si[«Al] is the 19 20  signal from a silicon T-atom surrounded by n aluminium atoms in the adjacent T-sites, the framework Si:Al can be calculated.  106  Si[2Al] (a)  (b)  Figure 3.1  Si M A S N M R spectra of (a) N a Y (b) the U S Y material recorded at 9.4 T . Spectra were acquired  using a pulse angle of 22°, a recycle delay of 10s and a spinning rate of (a) 3.6 kHz and (b) 4.1 kHz. Spectra were simulated using the programme DmFit using isotropic peaks with mixed Lorentzian and Gaussian character. Deconvolutions are displayed under each spectrum. Two peaks were used to simulate the Si[l Al] peak to account for the two possible relative aluminium atom positions of "next nearest neighbour" Si[lAl] atoms. This is the only resolvable next nearest neighbour effect, (b) Two peaks are also used to simulate the SifOAl] peak of USY in order to accommodate the asymmetry of this signal. A broad peak is also included in the simulation of this spectrum to account for the presence of any amorphous silicon-containing material. The parameters used for this peak were obtained from the simulation of the Si MAS NMR spectrum of a USY sample treated under very harsh conditions (steam calcination at 850°C, not shown) that contained a substantial amorphous component. 18  2  107  Although disordered, the distribution of aluminium atoms in zeolites obeys Lowenstein's 21  rule,  which states that A l - O - A l bridges are structurally unstable and are avoided where  possible, and framework Si: A l values can be calculated using the formula:  Si_ Al  4]T/(„A1) n=0  (3-1)  22  JV(/iAl) n=0  Where 7(«A1) is the intensity of the Si[«Al] peak. The integrals of the peaks used to simulate the S i M A S N M R spectra were inserted into Eqn 3.1 and indicate that the framework 2 9  Si:Al has increased from a value of 2.6 ± 0.1 for NaY to a value of 10.7 ± 0 . 1 for U S Y , in good agreement with the value obtained using IR spectroscopy (Section 3.2.1). X R P D patterns (Figure 3.2) are consistent with the changes observed in the S i M A S N M R 29  spectra. Both materials show a high degree of crystallinity, but the pattern from U S Y (Figure 3.2(b)) shows an increase in the baseline intensity between 20 ~ 18-35°. This is partly due to a general broadening of the peaks resulting from a decrease in sample crystallinity, but there is also a broad underlying signal indicative of diffuse scattering from an amorphous component of unknown composition. However, X R P D data cannot be used to quantify or characterise the amorphous material and S i N M R does not detect the presence of any amorphous material that 2 9  does not contain silicon.  3.3.2  L o w FIELD A1 N M R INVESTIGATION OF U S Y 27  The A l M A S spectra of the U S Y material (Figure 3.3) contain broad resonances with 2 7  general features similar to previous literature data at 9.4 T, indicating that the nature of the aluminium in the present sample is typical of this type of catalyst. At the high spinning rate employed, there is no overlap of spinning sidebands with any of the isotropic peaks, a problem with earlier work. Three maxima are observed at 58, 30 and -3 ppm which previous authors have assigned to 4, 5 and 6 co-ordinated aluminium species, as discussed above.  108  (a)  (b)  "1  10  I  I  I  1  1  1  20  30  40  50  60  70  20 (degrees) Figure 3.2 Powder X R D patterns of (a) NaY and (b) USY. Acquired on a Seimens D5000 powder diffractometer with Cu Kct radiation between 20 = 5° and 20 = 70° in steps of 0.0257s.  However, even at 9.4 T there are indications that the situation is more complex than the interpretations proposed in the literature " ' ' . There are significant differences between the 5  7 10 13  intensity differences of the A l single pulse (Figure 3.3(a)) and echo (Figure 3.3(b)) M A S N M R 2 7  spectra. The maxima in the spectra correspond well with each other but there is an indication from the spinning sideband pattern of the echo spectrum that there is a major contribution from one or more broad resonances, centred at approximately 30 ppm, which is greatly attenuated in the single pulse experiment.  109  (a)  ] i i i i | i i i i | i i i i | i i 300 200 100 0 Figure 3.3  2 7  i—i—|—i—i—i—i—|—i—r -100 -200 (ppm)  A l M A S N M R (a) single pulse (b) 90°-180° echo spectra of U S Y at 9.4 T . Spectra were acquired at  a spinning rate of 9.8 kHz with a 200ms recycle delay.  The total integrated intensities of both the  27  A l single pulse and echo M A S N M R  weighed samples were calibrated at all fields against spectra of weighed NaY and NaA  spectra of samples  acquired under identical conditions (Table 3.1). Sample hydration was measured by TGA and corrected for. The composition of the extra-framework aluminium-containing material is not known. Calculations were performed assuming compositional extremes of (Si0 )x(HA102)/. 2  x  and (Si02)x(HA102)^(Al203)c;.^/2 and an error or ~1 % is thus introduced. In line with previous studies  '  4,5,9 11  at low fields, the 9.4 T single pulse experiment shows intensity from only 65% of  110  the aluminium present, whereas the echo experiment shows intensity from 85% of the aluminium present, suggesting that the broad resonance that is discriminated against in the single pulse experiment may account for the "invisible aluminium". ' ' " ' 2 5 9  Field  11 15  9.4 T  14.4 T  18.8 T  (± 5 %)  (± 10 %)  (± 10 %)  Single pulse  65  87  95  Echo  85  -  98  Table 3.1 Integrated total intensities o f A l single pulse and echo M A S spectra. Values are expressed as an 2 7  averaged percentage of the integrated intensities of Al MAS spectra of weighed samples of NaY and NaA acquired under identical conditions, adjusted for hydration (measured by TGA) and composition (refer to text). 27  Figure 3.4 shows the A l 3QMAS N M R spectrum of the U S Y material at 9.4 T. It is clear 2 7  that two distinct tetrahedral environments are present; the maximum at 59 ppm in F2, corresponding to the maximum in the M A S spectrum that is assigned to tetrahedral framework aluminium, and another with a much larger quadrupolar coupling corresponding to the chemical shifts of the unresolved broad resonances in the M A S spectrum whose intensity is enhanced in the echo spectrum. The main difficulties in the further interpretation of these data arise from the limited resolution of the M A S spectra obtainable at 9.4 T, even at these high spinning rates, and the 97  discrimination of both the  A l M A S and 3QMAS N M R spectra against the broad resonances.  To some extent this is due to the largely amorphous nature of the aluminium containing species (where distributions of chemical and quadrupolar shifts are to be expected) but a major contribution is from the reduced effect of quadrupolar coupling. This is inversely field dependent and both the interaction and the distribution of the interaction, introduced by distributions in local environments, should be reduced at higher magnetic field strengths.  Ill  Figure 3.4 Nutation A13QMAS NMR spectrum of USY acquired at 9.4 T. Acquired using a Bruker 4 mm MAS probe at a rf. power of 52 kHz and a spinning rate of 12.0 kHz and using a 200 ms recycle delay. Exponential line broadenings of 110 Hz and 80 Hz were applied to F2 and F l respectively prior to the double Fourier transform and shearing transformation. Skyline projections are shown. '*' denotes a spinning sideband. 27  112  3.3.3  H I G H FIELD  2 7  Al  N M R INVESTIGATION OF U S Y  Figure 3.5 shows the single pulse and echo^'Al M A S spectra 18.8 T and can be compared directly with those at 9.4 T (Figure 3.3). At 18.8 T the single pulse and spin-echo spectra are far more consistent with each other and both show intensity from > 95 % of the aluminium present (i.e. quantitative within the error limits of these data) indicating little discrimination between the resonances; quite different from the situation at 9.4 T. The small spinning sidebands observed are due to the approximate doubling of the chemical shift contribution in frequency units while the spinning rates are ca. 10 kHz at both fields. The general resolution of the spectra is greatly improved and the intensity in the centre of the spectrum is greatly reduced. Three resonances are now clearly resolved at 61, 30, and 1 ppm with an additional broad resonance which appears as a distinct shoulder to high field of the tetrahedral resonance, at ~ 54 ppm. This signal is considered to be due to the second tetrahedral aluminium species revealed by the 3QMAS experiment at 9.4 T. Further investigations of these four aluminium environments were carried out by A1 3QMAS N M R at 18.8 T (Figure 3.6). 27  The  2 7  A l 3QMAS N M R spectrum of the U S Y material at 18.8 T shows four clearly  resolved signals, assigned to aluminium in tetrahedral ( A l ) , broad tetrahedral ( A l Tet  co-ordinate ( A l  Pent  B r T e t  ) , five  ) and octahedral (Al° ) environments. The signal intensity of the five coct  ordinate environment is discriminated against but the signal in the 2D plot is clearly evident at 20 ppm (Fl) and 31 ppm (F2) and corresponds to the 30 ppm resonance of the M A S spectra (Figure 3.5). 77  These data facilitate further interpretation of the 9.4 T experiments. The A l 3QMAS N M R spectrum of U S Y at 9.4 T (Figure 3.4) has a small feature at 23 ppm ( F l ) and approximately 30 ppm (F2) which can be confirmed, in light of the high field data, as the A l addition, the M A S (F2) shifts of A l  B r T e t  and A l  P e n t  P e n t  resonance. In  can be seen to overlap.  However, this assignment of the low field data is not possible without access to high field data, since at low field the broad resonances are discriminated against and overlap to such an extent that the assignment of spectra is at best ambiguous.  113  I  I  240  I  I  I  160  1  1  1  0  80  1  1  1 —  -160  -80  (ppm)  I  240  I  !  I  !  160  !  I  0  80  |  !  -80  !  !  -160  (ppm)  Figure 3.5 (a) Single pulse and (b) e c h o A l M A S N M R spectra of U S Y acquired at 18.8 T . Recorded using (a) 27  100 ms recycle delay at spinning rate of 10.2 kHz, (b) 500 ms recycle delay at a spinning rate of 11.9 kHz.  114  90 80 70 60 50 40 30 20 10 0 -10 -20 -30 F2 (ppm) Figure 3.6 A 1 R I A C T I I 3 Q M A S spectrum of U S Y at 208.43 M H z . Recorded at a spinning rate of 10.2 kHz 27  using a recycle delay of 100 ms. Exponential line broadenings of 100 Hz and 50 Hz were used in the F2 and Fl dimensions, respectively. Skyline projections along both axes are shown. '*' indicate spinning sidebands. Peak assignments are discussed in the text.  115  3.3.4  SIMULATION, DECONVOLUTION AND PARAMETERISATION OF A I N M R D A T A 2 7  It is now possible to extract the spectral parameters of all four sites, beginning with an interpretation of the A l M A S N M R spectra recorded at 18.8 T. The spectra were simulated 2 7  18  using non-integer spin, quadrupolar M A S lineshapes of the central transition for each of the four aluminium environments. A n additional symmetrical peak is used to simulate A l  0 c t  to account  for a shoulder in the octahedral region at ~ 0 ppm in the 3QMAS spectra at both high and low field, " A l ° " . Narrow signals at this chemical shift are also observed for samples treated Iso  ct  under extremely mild conditions and can be assigned to isolated "A1(H.20)6 " or closely related 3+  species in which the aluminium environment is close to perfectly octahedral. Spinning sideband and satellite transition intensities were simulated using peaks with mixed Lorentzian and Gaussian character with intensities estimated from the first order spinning sideband regions of the spectra that did not overlap with the zeroth order region. Distributions of chemical and quadrupolar shifts and quadrupolar coupling were accounted for by an exponential broadening function. B y this method, average lineshape parameters were obtained and the results are shown in Figure 3.7 with the fitting parameters given in Table 3.2. More detailed notes relating to the simulations are given in the table caption. The A l  T e t  resonance accounts for 33% of the total intensity. Framework Si: A l of the NaY  starting material (measured by S i M A S N M R ) and of the U S Y material (measured both by S i 2 9  2 9  N M R and IR spectroscopy) also indicate that ca. 33% of the aluminium remains in the U S Y framework following steaming. This suggests that the vast majority of the intensity from the other aluminium species are extra-framework in nature and that the A l assigned as framework tetrahedral aluminium ( A l  116  F w k T e t  ).  T e t  resonance can be  T  1  120  1  1  80  1  1  1  1  0  40  1  1—  -40  (ppm)  Figure 3.7 Experimental, simulated and deconvolved single pulse A1 M A S N M R spectrum of U S Y . Acquired at 18.8 T using a spinning rate of 10.2 kHz. 27  117  Confirmation of the accuracy of the analysis is provided by the excellent agreement between the predicted and experimentally observed isotropic chemical shifts of the corresponding 3QMAS spectra at both 18.8 T and 9.4 T as shown in Table 3.3.  Average v  % of total intensity''  Q  (ppm) Al A1  fl  (kHz)*  60.7  360  0.5  33  B,Te.  60.4  940  0.1  21  Pent  32.2  578  0.1  c  20  3.2  492  0.1  c  26  A1  T e t  Al°  ct  Table 3.2 Summary of the spectral simulation parameters for the single pulse and echo A l M A S 2 7  NMR  spectra of USY at 18.8 T. The spectra were simulated using the programme DmFit. is the chemical shift contribution to the isotropic shift. DmFit does not account for the contribution to the isotropic shift from the portion of the exponential broadening function attributable to a distribution of quadrupolar couplings. Therefore these values include a correction that is inversely proportional to the magnetic field, calculated from differences in the b 2 isotropic chemical shift values used in DmFit. vQ- 3e qQ/2I(2I -1) Signals are simulated using an average 18  value of v , with distributions accounted for using an exponential broadening function (E ) with a chemical shift component proportional to the magnetic field strength and a quadrupolar coupling component inversely proportional to the magnetic field. Values are estimated to be accurate to ± 10 %. The simulation is relatively insensitive to this variable and values are only assumed to be accurate to ± 0.5. ''Values are calculated to ± 3% for 18.8 T spectrum using spinning sideband intensities simulated by peaks with mixed Lorentzian and Gaussian character. Q  m  A demanding test of the lineshape parameters extracted from the 18.8 T A l M A S  NMR  spectra is to check that they correctly simulate the lower field M A S spectra. The average parameters of the four aluminium environments used to simulate t h e A l M A S N M R spectrum 27  at 18.8 T have been used to accurately simulate the A l M A S N M R spectra of U S Y at both 9.4 2 7  T and 14.4 T (Figure 3.1). In these simulations, only the magnitudes of the four resonances were allowed to vary (because the lower field spectra are not quantitative) and the line shapes and positions were determined from the parameters at 18.8 T and the known field dependencies of chemical shielding frequencies and second order quadrupolar interactions. Spinning sideband and satellite transition intensities were treated in the same manner as for the 18.8 T spectra.  118  There is an excellent reproduction of all of the features of the spectra. The total intensities of the spectra at 18.8 T and 14.4 T are quite close (Table 3.1), as are the relative intensities of the peaks used to simulate the spectra. The values at 9.4 T are quite different, indicating that A1 27  M A S N M R measurements on these systems should be made at fields of 14 T or greater to ensure quantitative reliability. A similar conclusion has recently been drawn by Fitzgerald for 27  2S 26  23  A l in various aluminas  and by several workers on faujasite systems. '  9.4 T Calculated  Experimental  Calculated  Experimental  34.2  35.1  33.6  33.9  Br.Tet  40.2  40.5  35.0  35.8  Pent  20.4  22.8  18.5  19.7  3.8  4.3  2.5  3.8  ^jFwk.Tet  A1  18.8 T  A1  A 1  oc  Table 3.3 Calculated and experimental 3 Q M A S isotropic (Fl) shifts of U S Y at 9.4 T and 18.8 T . Values are calculated using the parameters given in Table 3.2 using Eqn 1.59 (Section 1.1.10(d)).  Further general insight into the past difficulties in the interpretation of the lower field A l M A S N M R spectra can be gained from an examination of the contributions of the individual resonances to the simulation of the 9.4 T A l M A S spectrum. At this field it can be seen that the 2 7  Al  B r T e t  resonance has developed a clearly defined M A S quadrupolar line shape. The upfield  portion of this resonance overlaps with the A l  P e n t  resonance (which is now centred at 25 ppm)  giving a composite 'signal' observed at approximately 30 ppm, while the low field portion of the line shape is overlapped with and obscured by the A l  119  F w k T e t  resonance at 59 ppm.  (a) 14.4 T  i  1  120  1  1  80  1  1  1  1  0  40  r  -40  (ppm)  Figure 3.8 Experimental, simulated and deconvolved single pulse A 1 M A S N M R spectra of U S Y acquired at 27  (a) 14.4 T and (b) 9.4 T. Spinning rates were (a) 13.0 kHz (b) 9.8 kHz. Simulations generated using parameters given in Table 3.2.  120  3.3.5  27  A1 N M R  INVESTIGATIONS OF A M O R P H O U S A L U M I N A  This material was chosen because of the previously proposed similarities between such materials and the extra-framework aluminium species in U S Y materials. The A l M A S 11  2 7  NMR  spectrum of the amorphous alumina gel (Figure 3.9(a)) at 17.6 T shows well resolved broad resonances with maxima at 67, 35 and 6 ppm from four, five and six co-ordinate aluminium species. The resonances have slightly different line widths and there are small contributions from spinning sidebands. These are clearly reflected in the A l 3QMAS N M R spectrum (Figure 2 7  3.9(b)), but while the four and six co-ordinate resonances are observed with high S:N and in approximately the correct proportions, the contribution from the five co-ordinate aluminium is reduced. The same discrimination against five co-ordinate aluminium was observed in the  2 7  Al  3QMAS N M R spectra of the U S Y material and this has been found to be a general characteristic of 3QMAS experiments on materials of these types. The parameters derived from the fitting of the M A S spectra at 14.4 T and 18.8 T are presented in Table 3.4.  Aluminium coordination  8 s (ppm)"  Average v (kHz)  _ c  4  65.2  655  0.3  5  36.8  452  0.7  6  1.9  535  0.0  C  Q  %  Table 3.4 Summary of the spectral simulation parameters for the 14.4 T and 18.8 T single p u l s e A l M A S 27  spectra of amorphous alumina gel. The spectra were simulated using the programme DmFit using the methods described in the caption for Table 3.2 "S s is the chemical shift contribution to the isotropic shift. v = 3 e 9 Q / 2 / ( 2 / - l ) Signals are simulated using an average value of v , with distributions accounted for 18  C  b  2  Q  Q  using a scaled exponential broadening function. Values are estimated to be accurate to ± 10%. The simulation is insensitive to this variable and values are only assumed to be accurate to ± 0.5.  121  T  I  I  |  I  I  I  120  I  j  I  80  I  I  I  |  I  40  I  I  I  0  |  I  I  I  -40  1  j  1  1  1  1  [•  -80  F2 (ppm) Figure 3.9 A1 (a) 9 0 ° - 1 8 0 ° echo M A S and (b) nutation 3 Q M A S N M R spectra of amorphous alumina 27  acquired at 17.6 T . Recorded using a recycle delay of 200 ms at a spinning rate of (a) 13.7 kHz and (b) 15.1 kHz. Skyline projections are shown.  The amorphous alumina is quite different to the U S Y both in the distributions and average quadrupolar coupling and chemical shift parameters of the aluminium environments, particularly the broad four co-ordinate aluminium (Table 3.2 and 3.4). Such a situation might be  122  expected since the pore and channel system dimensions of the zeolite framework may well place constraints on the geometries and sizes of any extra-framework species within the lattice and they may well be less random or amorphous than the gel materials and contain quite anisotropic local environments. The gels have been formed under conditions where there are comparatively greater opportunities for the "relaxation" of the local aluminium geometries.  3.4  Conclusions  Simulations of the A1 M A S N M R spectra and calculated A 1 3QMAS N M R isotropic 27  27  shifts at the different magnetic field strengths reveal a consistent pattern of four clearly defined aluminium environments for this U S Y material (Table 3.2 and 3.3) and this is considered to be representative of this class of catalyst. Further, there are differences between the behaviours of these signals and those in the spectra of the amorphous alumina sample (Table 3.4), indicating that amorphous materials of this type may be more limited as models for extra-framework alumina than originally proposed. This may be due to constraints imposed by the limited 15  volumes available in the channel systems, which could restrict both the sizes and geometric arrangements in the oligomeric species produced during the framework dealumination resulting from steam treatment. However, this study shows that the experimental protocol of T G A ,  Si M A S N M R and  multiple field A1 M A S and 3QMAS N M R spectroscopy is an effective means of quantifying 27  and characterising the aluminium environments in acid zeolites that have undergone steam dealumination. In the following chapters, these techniques will be applied to a range of U S Y materials in order to investigate the relationship between the aluminium and the chemical and physical properties of the systems.  123  3.5 1.  References for Chapter 3 Occelli, M . L . Fluid Catalytic Cracking II- Concepts in Catalytic Design; A C S Symposium Series 452: Washington, DC, 1991; pp27-44.  2.  Klinowski, J.; Thomas, J.M.; Fyfe, C.A.; Gobbi, G.C. Nature, 1982, 296, 533-536.  3.  Scherzer, J. A.C.S. Symp. Ser., 1984, 248, 157-200.  4.  Klinowski, J . ; Fyfe, C.A.; Gobbi, G.C. J. Chem. Soc, Faraday Trans. 1,1985, 81, 30033019.  5.  Freude, D.; Brunner, E.; Pfeifer, H.; Prager, D.; Jerschkewitz, H.-G.; Lohse, U . ; Oehlmann, G. Chem. Phys. Lett., 1987,139, 325-330.  6.  Samoson, A.; Lippmaa, E.; Engelhardt, G.; Lohse, U . ; Jerschkewitz, H.-G. Chem. Phys. Lett., 1987,134, 589-592.  7.  Ray, G.J.; Samoson, A.; Zeolites, 1993,13, 410-413.  8.  Gilson, J.-P.; Edwards, G . C ; Peters, A.W.; Rajagopalan, K.; Wormsbecher, R.F.; Roberie, T.G.; Shatlock, M.P. Chem. Commun.,1981, 91-92.  9.  Shertukde, P.V.; Hall, W.K.; Dereppe, J.-M.; Marcelin, G. J. Catal., 1993,139, 468-481.  10.  Remy, M.J.; Stanica, D.; Poncelet, G.; Feijen, E.J.P.; Grobet, P.J.; Martens, A.J.; Jacobs, P.A. J. Phys. Chem., 1996,100, 12440-12447.  11.  Coster, D.; Blumenfield, A . L . ; Fripiat, J.J. J. Phys. Chem.,1994, 98, 6201-6211.  12.  Blumenfield, A . L . ; Fripiat, J.J. Topics in Catalysis, 1997, 4, 119-129.  13.  Grobet, P.J.; Geerts, H.; Tielen, M . ; Martens, J.A.; Jacobs, P.A. Stud. Surf. Sci. Catal.; Elsevier: Amsterdam, 1989, 46, 721-734.  14.  Wood, T.E.; Siedle, A.R.; Hill, J.R.; Skarjune, R.P.; Goodbrake, C.J.; Mat. Res. Symp. Proc, 1990,180,97-115. 124  15.  Coster, D.; Fripiat, J.J.; Chem. Mater.,1993, 5, 1204-1210.  16.  Kunath-Fandrei, G.; Bastow, T.J.; Hall, J.S.; Jager, C ; Smith, G.E. J. Phys. Chem., 1995, 99, 15138-15141.  17.  Llor. A.; Virlet, J. Chem. Phys. Lett, 1988,152, 248-253.  18.  Massiot, D.; Fayon, F.; Capron, M . ; King, I.; Le Calve, S.; Alonso, B., Durand, J.-O.; Bujoli, B.; Gan, Z.; Hoatson, G. Magnetic Resonance in Chemistry 2002, 40, 70-76.  19.  Lipmaa, E.; Magi, M . ; Samoson, A.; Grimmer, A.R.; Englehardt, G. J. Am. Chem. Soc, 1986,108, 1730-1735.  20.  Fyfe, C.A.; Thomas, J.M.; Klinowski, J.; Gobbi, G.C. Angew. Chem. Int. Ed. Engl., 1983, 22, 257-273.  21.  Lowenstein, W.Am. Mineral., 1954,59,92-96.  22.  Fyfe, C A . Solid State NMR for Chemists, C.F.C. Press: Guelph, 1984.  23.  ACS Symposium Series 717: Solid State Spectroscopy of Inorganic Materials, Fitzgerald, J.J. Ed; American chemical Society: New York, 1998; ppl82-226.  24.  Fyfe, C.A.; Bremerton, J.L.; Lam, L . Y . Chem. Commun., 2000,17, 1575-1576.  25.  Fyfe, C.A.; Bretherton, J.L.; Lam, L . Y . J. Am. Chem. Soc. 2001, 123, 5285-5291.  26.  van Bokhoven, J.A.; Roest, A . L . ; Koningsberger, D . C ; Miller, J.T.; Nachtegaal, G.H.; Kentgens, A . P . M . J. Phys. Chem. B, 2000,104, 6743-6754.  125  Chapter 4. Solid State NMR and Nitrogen Physisorption Investigations of the Effect of Acid Washing of USY Materials.  4.1 Introduction In Chapter 3, the experimental protocol for the characterisation of the aluminium environments in U S Y was shown to be effective in the study of a U S Y material with a moderately high framework Si: A l and high crystallinity. Presented in this chapter are the results of a study of a series of U S Y materials that have been subjected to a sequence of steam dealumination and acid washing treatments which selectively remove framework and extraframework aluminium, respectively. '  1 2  Since the role of the various aluminium species in ultrastable zeolite systems is not well understood, the selective removal of extra-framework aluminium containing material provides the opportunity to investigate the chemical and physical properties of this extra-framework component of U S Y . Additionally, acid washing is a means of tuning the properties of U S Y , which is of particular interest because of the role that extra-framework material is thought to play in catalytic activity. " Various studies have shown that the removal of the extra-framework material reduces the activity of U S Y catalysts towards the catalytic cracking of small organic molecules " whereas other studies have suggested that this extra-framework material has a 6  8  detrimental effect on the cracking activity of larger molecules because of the steric hindrance that it provides. " 9  11  Control of the proportions of framework and extra-framework aluminium can be achieved in various ways. The framework Si:Al of the starting material can be determined during synthesis such that a given amount of hydrothermally generated extra-framework material can  126  be produced at various framework Si: Als. Washing with mineral acids, E D T A or hexafluorosilicate salts is aimed at selectively removing extra-framework material without changing the aluminium content of the framework. The effects of washing are dependent on 12  the conditions and reagent chosen and are a matter of some debate. ' ' Hexafluorosilicates are 1 2 9  generally considered to effectively remove extra-framework material from ultrastable zeolites whilst also further dealuminating the framework and otherwise facilitating framework healing. A high degree of dealumination can thus be achieved without the generation of extra framework material. ' Mineral acid washing is reported to act similarly, but with a far greater degree of 13 14  12  defect formation and some ambiguity regarding the removal of framework aluminium. It has also been reported that a substantial amount of extra-framework material cannot be removed by washing with mineral acids.  16  However, these conclusions are often based on indirect evidence, such as the evaluation of the amount of framework and extra-framework material by the concentration of Lewis or Bransted acid sites, assigned exclusively to the extra-framework and framework aluminium respectively. ' ' " U S Y systems may be too complex for such simple assignments and there is 1 6 17  19  some suggestion that the general nature of extra-framework aluminium varies between studies.  1  Investigations have been hampered in the past by the inability to directly quantitatively observe and assign the aluminium present in ultrastable zeolite systems using solid state Al 27  N M R . The series of U S Y materials studied in this chapter have been chosen as being representative of extreme changes associated with hydrothermal dealumination and acid washing. This study also provides the opportunity to demonstrate the general utility of the experimental protocol described in Chapter 3 to quantitatively measure the amount of each type of aluminium species present and to follow the changes in these values as a result of the steam and acid treatments.  127  4.2  Materials and Experimental Methods  4.2.1  MATERIALS  The series of U S Y samples were prepared by Dr. L . Y . Lam, Petrobras CENPES Research 9ft  Centre, Rio de Janeiro, Brazil.  The NaY starting material was ammonium exchanged and  steam calcined at 650°C for 90 minutes to create sample U l . This was then washed in dilute sulphuric acid at 80°C for 60 minutes, during which time acid was added in order to maintain a constant ph = 2.5, to create sample U l w . Sample U2 resulted from two further steam treatments separated by an acid washing and sample U2w resulted from a further acid washing of sample U2. U2w has therefore undergone a total of three cycles of steam treatments and acid washings.  4.2.2  E X P E R I M E N T A L APPARATUS  Experiments were performed on the U S Y samples (Section 4.2.1) using the same methods and conditions as those discussed in Chapter 3. Single pulse S i M A S N M R spectra were obtained at 9.4 T (79.5 MHz) using a Bruker 2 9  DMX400 spectrometer and a home built probe equipped with a standard speed Doty 7 mm stator delivering a 90° pulse width of 8.0 us. Single pulse and 90°-180° echo A l M A S N M R 2 7  and 3QMAS spectra were obtained at 9.4 T (104.3 MHz) using a Bruker DMX400 spectrometer and a Bruker 4 mm M A S probe at a rf. power of 56 kHz. Pulse lengths of 0.8 us were used for the single pulse experiments and for the first pulse of the echo experiments. 2 7  A l N M R M A S and 3QMAS N M R spectra were also acquired at 17.6 T (195.4 MHz)  using a Varian Inova spectrometer. A l l A l spectra were obtained using home-built M A S 2 7  probes equipped with a Doty 5 mm supersonic stator at a rf. power of 56 kHz. Pulse lengths of 1.0 us were used for the single pulse experiments and for the first pulse of the echo experiments. Referencing and processing is discussed in Chapter 2 and the pulse sequences used are listed in Appendix III.  128  T G A was carried out using a T A Instruments TG51 thermogravimetric analyser under a flow of 80cc/min of dry nitrogen gas. The temperature was ramped from room temperature to 800°C at a rate of 10°C/min and maintained at this temperature for 60 minutes. Physisorption measurements were taken using a Micrometrics A S A P 2010 instrument using nitrogen as the analysis and saturation gas. Experiments were performed at 77.3 K with the sample vessel immersed in a liquid nitrogen bath. Surface area measurements were fit to a Langmuir isotherm, porosity measurements were obtained from each isotherm using a SaitoFoley modified Hovarth-Kawazoe analysis and mesopore areas for pore sizes in excess of 15 A were obtained from a cumulative BJH analysis. Refer to Section 2.4 for details.  4.3 Results and Discussion  4.3.1  2 9  C H A N G E S IN T H E F R A M E W O R K  S i M A S N M R spectra (Figure 4.1) were simulated and the integrals of the 22  deconvolutions used to calculate the framework Si:Al (Table 4.1, S i : A l ) according to Eqn 3.1. 29  The spectra from the pairs of samples differing by an acid washing only ( U l , U l w and U2, U2w) are not significantly different. Therefore, differences in the calculated Si:Al values between the members of each pair of samples reflect the experimental error. The results of the analyses of N2 adsorption isotherms for all of these U S Y materials are summarized in Table 4.2. These data are consistent with physical changes associated with steam dealumination and washing with mineral acids " but the overall changes in this series of 9  16  samples are relatively small. Both treatments are associated in the literature with mesopore formation and these data show that the median pore diameters of the U S Y materials do increase with successive treatments. The Langmuir surface areas show the effects of pore blockage by extra-framework material. There is an increase in surface area following acid washing between samples U l and U l w and more markedly between samples U2 and U2w. By contrast, there is a decrease in surface area between samples U l w and U2 which are separated by an acid washing and two steam treatments. It might be argued that the effect of acid washing on total and mesopore surface area between U2 and U2w is indicative of the framework healing that has  129  been reported elsewhere ' ' primarily by X R P D and physisorption measurements. Powder X R D 1 2 12  studies of the samples used in the present work (not shown) show a gradual increase in 20  crystallinity with successive treatments (as described in Section 4.2.1). The samples U l and U l w and the samples U2 and U2w differ by an acid washing only. S i M A S N M R data show 2 9  that this acid washing does not change the framework S i A l (Table 4.1), but the X R P D data indicates that the framework is modified.  29  Si:Al  a  6  Si:Al  IR  c  Bulk Si:Al  ±  Si:Al -  rf  Fwk Tet  e  ±  Si:Al  T o t a l T e t  ±  ±  Ul  6.1  0.4  5.6  2.6  0.6  13.2  4.4  3.8  1.3  Ulw  6.1  0.4  5.6  3.9  0.9  15.2  3.6  5.5  2.2  U2  24.3  0.9  11.8  5.1  1.3  38.8  13.4  11.2  3.8  U2w  24.3  0.9  11.9  8.7  2.0  36.1  11.3  16.8  5.1  Table 4.1 Si:Al values calculated for the series of U S Y samples. "Calculated from Si MAS NMR spectra recorded at 9.4 T, according to Eqn 3.1. Values obtained from the integration of FT IR spectra. Taken from reference 18, Table 1. Values above SiiAl^a value of approximately 10 are not considered to be reliable. 'Molar percentages of Si and Al in the bulk material are derived from TGA measurements, assuming a bulk stoichiometry of the dry material of (Si02) (HA102)(i- )/2, since the molar mass of this stoichiometry is independent of the unknown bulk Si.Al and from the total integrals o f A l MAS NMR spectra. Analogous calculations on systems where bulk Si:Al is identical to the starting material (Chapters 3 and 5) facilitate the estimation of uncertainties in the bulk stoichiometry in this case. These molar percentages are used to calculate the bulk Si:Al. S i : A l are Si: A l then calculated by assuming that the proportion of the total aluminium in the framework is equal to the integral of A l . and ( A l + Al ) respectively of the peaks used to simulate the A1 MAS NMR spectra recorded at 17.6 T. 29  21  x  X  27  d  c  Total T e t  F w l c T e t  F w k T e t  BrTet  27  130  FwkTet  Figure 4.1  Si M A S N M R spectra of (a) U l and U l w (b) U2 and U2w, recorded at 9.4 T . Spectra were  acquired using a pulse angle of 28°, a recycle delay of 30s and at spinning rates of ca.2 kHz. Spectra were simulated using the programme DmFit using isotropic peaks with mixed Lorentzian and Gaussian character. Simulations and deconvolutions are displayed under each spectrum. Two peaks were used to simulate the Si[lAl] peak of samples U l and Ulw to account for the two possible relative aluminium atom positions of "next nearest neighbour" Si[lAl] atoms. 22  "Langmuir Surface Area (m g )  *Median Pore Diameter (A)  "BJH Mesopore Area (m g )  Mesopore Area as % of Total  NaY  890.2  8.39  22.0  2.5  Ul  768.8  9.72  28.6  3.7  Ulw  775.0  9.98  25.3  3.3  U2  695.4  10.42  38.2  5.5  U2w  780.4  10.67  69.2  8.9  2  c  _1  2  _1  Table 4.2 Physisorption data obtained from N adsorption isotherm. "± 2%. *Values are accurate to ± 0.1 A . 2  Estimated error limits are based upon repeat experiments. Starting material.  131  4.3.2  H I G H AND L O W F I E L D A 1 M A S N M R D A T A 2 7  Single pulse and echo  A l M A S N M R spectra of the U S Y materials were recorded at both  17.6 T and 9.4 T and calibrated against spectra of NaY (both fields) and NaA (17.6 T only) acquired under identical conditions. In general terms, the single pulse A l M A S N M R spectra 2 7  of samples U l and U2 (Figure 4.2) are similar to the spectra of the U S Y material studied in Chapter 3 but with a lower intensity between ~ 15-55 ppm (the chemical shift range corresponding to A l  B r T e t  and A l  P e n t  ) relative to the A l  F w k T e t  and A l ° resonances. The effect of ct  acid washing of both of these samples can be seen in t h e A l M A S N M R spectra of U l w and 27  U2w by the reduced the intensity of the broad resonances (particularly noticeable at 9.4 T, Figure 4.2(b)) and the appearance of fine structure in the A l °  ct  resonance.  Further indication that the "invisible aluminium" at lower magnetic field strengths, discussed in Chapter 3, comprises these broad signals is given by the comparison of the total integrated intensities of the A1 M A S N M R spectra at high and low field (Figure 4.3). The 27  spectra acquired at 17.6 T can be considered to be quantitative and therefore the amount of invisible aluminium at 9.4 T is at a minimum directly following an acid washing, i.e. for samples U l w and U2w, the A l M A S N M R spectra of which have the lowest contributions 2 7  from the broad resonances. The integrals of the high field spectra (Table 4.3) show that each cycle of steaming and acid washing results in a reduction of ca. 25 % of the total aluminium content.  17.6 T  9.4 T  Single pulse  Echo  ±  Single pulse ±  Echo  -±  a  ±  Ul  93  11  106  9  67  10  76  11  Ulw  75  9  70  6  64  10  70  10  U2  62  8  55  5  39  6  48  7  U2w  38  5  35  4  33  5  29  4  Table 4.3 Integrated total intensities of single pulse and e c h o A l M A S N M R spectra at 17.6 T and 9.4 T . 27  Values are expressed as an averaged percentage of the integrated intensities o f A l MAS spectra of weighed samples of NaY acquired under identical conditions, adjusted for hydration (measured by TGA) and composition. "The error of ± 15% of each value is based upon repeat experiments. 27  132  — i  1  120  1  80  1  1  i  40  1  1  0  1  1  r —  -40  1  i  120  1  80  (ppm)  1  1  1  40  1  0  1  1—  -40  (ppm)  Figure 4.2 Single pulse A1 M A S N M R spectra of U S Y samples acquired at (a) 17.6 T and (b) 9.4 T . 27  Acquired using (a) a 60° pulse width and a 500 ms recycle delay at spinning rates of 10.5-13.3 kHz, (b) a 45° pulse width and a 100 ms recycle delay at a spinning rate of 12.0 kHz.  T h e A l 3QMAS N M R spectra of U l (Figure 4.4) can be assigned analogously to the 27  spectra of the U S Y material studied in Chapter 3. The general position and shape of the resonances is also similar to the U S Y studied previously, showing a characteristic overlap between A l Al  B r T e t  B r T e t  and A l  and both of A l  F w k T e t  F w k T e t  resonances at 17.6 T and an overlap between the resonances of and A l  P e n t  at 9.4 T. The corresponding spectra of U2w (Figure 4.5)  show a quite different situation. The A l  0 c t  region has a sharp resonance at ca. 0 ppm in both  dimensions from a symmetrical aluminium species, A l  I s o 0 c t  , most probably A1(H20)6 or a 3+  closely related species. In addition, there is a suggestion of resolution between broad octahedral resonances, which is also seen in the spectra of U l w (not shown). These new features of the  133  octahedral region of the A l 3QMAS N M R spectra are a result of selective acid leaching. 2 7  Similarly, the broad tetrahedral signal in the A1 3QMAS spectra of U2w is significantly 27  different from the corresponding spectra of U l , U l w and the spectra presented in Chapter 3. There is little resolution between A l  F w k T e t  and A l  B r T e t  resonances and the width of the A l  resonance is considerably reduced; manifested in a decrease in the F l shift of the A l  B r T e t  B r T e t  resonance. Again, these changes are due to the selective removal of the broadest of the A l  B r T e t  environments by repeated acid washing.  120 i  100  So  80  <D * * C  I—I  U d>  t  „  -a a-  O  2  *  60  40  20 Ul  Ulw  U2  U2w  Figure 4.3 Total integrated intensity o f A l M A S spectra of U S Y samples. Values averaged for single pulse and 2 7  echo experiments, o 17.6 T data, • 9.4 T data. The average 17.6 T value of 99.5% for sample U l indicates that no dealumination occurs following the first steam treatment and that the 17.6 T data are quantitative within the estimated error limits.  134  F2 (ppm) Figure 4.4 AI3QMAS NMR spectra of U l recorded at (a) 17.6 T and (b) 9.4 T. (a) Acquired using a two pulse nutation pulse sequence at a spinning rate of 13.6 kHz using a 100 ms recycle delay. Exponential line broadenings of 90 Hz and 200 Hz were applied to F2 and Fl, respectively, prior to the double Fourier transform and shearing transformation, (b) Acquired using three pulse z-filtered pulse sequence with a z-filter rf. power of 2.1 kHz, at a spinning rate of 12.0 kHz and using a 100 ms recycle delay. Skyline projections are shown. '*' denotes a spinning sideband. J7  135  ^jlso.Oct  Al0Ct  10  \  ^&  Fl 20 (ppm) 30  ^jFwk.Tet  £ Al  B r T e t  40 100  100  80  60  40 20 F2 (ppm)  80  60  40 20 F2 (ppm)  0  -20  6 3o  Figure 4.5 AI3QMAS NMR spectra of U2w recorded at (a) 17.6 T and (b) 9.4 T. (a) Acquired using a three pulse RIACTII pulse sequence with coherence generation by spin locking pulses of l/4r at a spinning rate of 13.5 kHz using a 100 ms recycle delay. Exponential line broadenings of 90 Hz and 200 Hz were applied to F2 and Fl,respectively, prior to the double Fourier transform and shearing transformation, (b) Acquired using three pulse z-filtered pulse sequence with a z-filter rf. power of 2.1 kHz, at a spinning rate of 12.0 kHz and using a 100 ms recycle delay. Skyline projections are shown. '*' denotes a spinning sideband. 27  R0T  136  In order to obtain a more detailed picture of the changes taking place, all A l single pulse and echo M A S N M R spectra were simulated using the programme DmFit. The M A S 22  quadrupolar lineshapes used to simulate the four main resonances were consistent in width and position with the features of the 3QMAS spectra, discussed above. Two broad peaks, in addition to a sharp isotropic peak for A l ° , were used to simulate the octahedral regions of all samples. Iso  ct  In the case of U l and U2, where there is no resolution between A l ° environments in the F l ct  dimension of the A l 3QMAS N M R spectra, this treatment should be considered only as a 2 7  means of simulating a particularly broad distribution of quadrupolar couplings more accurately than is possible using a single peak. The same chemical shift and quadrupolar coupling parameters were used to simulate the spectra at both high and low fields and calculated 3QMAS F l shifts closely matched experimental values. These data are summarised in Table 4.4.  arisow ,  (a)Ul  .  (ppm)  t »  Isotropic 3 Q M A S Shift (ppm)  rj  Vq  Q  9.4 T  s„  ±  9.4 T  17.6 T  (kHz)  3.5  0.1  -0.7  -0.2  6.0  0.1  -12.5  32.2  0.1  60.8 59.8  c  Al°  Pe„t  A 1  A  ct  jFwk.Tet  A1  Br.Tet  17.6 T  ±0.3  Cak.  ±  Expt.  Calc.  ±  Expt.  425.3  0.1  3.4  0.5  *4.3  2.3  0.2  **2.7  -3.6  839.7  0.1  8.9  1.6  *7.4  4.9  0.5  **4.2  -5.2  -1.5  559.2  0.5  20.4  1.0  22.0  18.4  0.3  18.9  0.1  -0.7  -0.2  259.7  0.1  33.9  0.2  35.3  33.5  0.1  34.0  0.2  -15.9  -4.5  870.2  0.5  39.3  2.4  40.8  34.7  0.7  35.9  d  Single resonance, = -1—>+9ppm. **Single resonance, ~ +l-++7ppm.  ac-isod)  (b) U1W  OyXIy (ppm) 7 2  A 1  A 1  A  Oct  Pe„t  |Fwk.Tet  Al  B r T e t  7 2  by  ±  9.4 T  17.6 T  (kHz)  6.3  0.1  -2.8  -0.8  423.6  4.4  0.1  -9.6  -2.7  31.2  0.1  -3.4  59.8  0.1  59.7  0.1  Isotropic 3 Q M A S Shift (ppm)  fj  u  9.4 T  K  17.6 T  Calc.  ±  Expt.  Calc.  ±  Expt.  0.7  5.1  0.7  4.4  3.9  0.2  *3.9  749.8  0.8  7.8  1.8  6.6  3.9  0.5  -  -1.0  541.4  0.5  19.6  0.9  21.2  17.8  0.3  19.1  0.1  0.0  273.4  0.1  33.4  0.2  35.4  33.0  0.1  33.8  -15.9  -4.5  868.1  0.5  39.2  2.3  41.1  34.6  0.7  35.5  ""Single resonance = +1—H-6ppm.  137  ±0.3  d  M  A1  A  Pent  jFwk.Tet A 1  B,Tet  9.4 T  17.6 T  ±  9.4 T  17.6 T  (kHz)  '±0.3  Calc.  ±  Expt.  Calc.  ±  Expt.  1A  0.1  -3.6  -1.0  495.2  0.7  6.3  0.8  *5.0  4.7  0.2  **3.2  6.4  0.1  -13.2  -3.7  788.7  0.8  9.5  2.0  *8.0  5.2  0.6  31.7  0.1  -2.2  -0.6  568.3  0.5  20.1  1.0  22.1  18.2  0.3  20.0  61.0  0.1  -0.6  -0.2  267.9  0.1  34.0  0.2  35.3  33.6  0.1  *33.7  55.8  0.1  -1.4  -0.4  568.3  0.5  33.4  1.0  32.8  31.4  0.3  -  a  ct  b  5  c  Al°  Isotropic 3 Q M A S Shift (ppm)  "dffx (PP ) V„  (c)U2  * Single resonance, = 0—+ 1 lppm.**Single resonance, ~ 1—<•+7ppm. *Singl e resonance, = +30->+38ppm. #  (d)  U2w  A1  A  Pent  jFwk.Tet A1  Br.Tet  "Isotropic 3 Q M A S Shift (ppm)  %  ±  9.4 T  17.6 T  (kHz)  6.2  0.1  -2.0  -0.6  4.1  0.1  -10.5  31.4  0.1  60.4 55.5  a  ct  (ppm)  5  c  Al°  a ciso(2)  9.4 T  17.6 T  ±0.3  Calc.  ±  Expt.  Calc.  ±  Expt  476.7  0.4  5.3  0.7  3.8  3.9  0.2  **3.8  -3.0  814.8  0.7  8.3  2.4  7.1  4.0  0.7  **4.7  -4.5  -1.3  568.3  0.5  20.0  1.0  -  18.0  0.3  -  0.1  -0.2  -0.1  267.9  0.1  33.7  0.2  *34.1  33.3  0.1  '33.9  0.1  -0.8  -0.2  552.3  0.5  33.1  0.9  -  31.2  0.3  -  d  •Single resonance, « +30—H-38ppm. **Single resonance, ~ +2—>+7ppm. Table 4.4 Quadrupolar lineshape parameters for the simulation o f A l M A S N M R spectra of (a) U l , (b) U l w , (c) U2 and (d) U2w at 9.4 T and 17.6 T . "(Eqn 1.44) *(Eqn 1.38) Signals are simulated using an average value of VQ, with distributions accounted for using an exponential broadening function (E ) with chemical shift contributions proportional to the magnetic field and quadrupolar coupling contributions inversely proportional to the magnetic field. Values are estimated to be accurate to ±10 %. TDmFit does not account for the contribution to the chemical shift from the portion of E attributable to a distribution of quadrupolar couplings. Therefore these values include a correction that is inversely proportional to the magnetic field, calculated from differences in the isotropic chemical shift values used in DmFit used at each field. The quality of the simulations are insensitive to this parameter. S™ (Eqn 1.69). 2 7  m  m  d  e  4.3.3  DISCUSSION  Integration of the 17.6 T A l single pulse and echo M A S N M R spectral simulations (Figure 2 7  4.6) and the total integrals of the experimental spectra with respect to N a Y and NaA calibrants (Table 4.3) give the percentages of each aluminium environment present in each sample on the same scale (Table 4.5 and Figure 4.7(a)).The steps U l — • U l w and U2—>U2w show that acid 138  washing removes aluminium contributing to all of the broad resonances but does not remove Al  . Further, the relative percentages of aluminium environments in each sample (Figure  F w k T e t  4.7(b)) show that the susceptibility of each environment to acid leaching is in the order A 1  Pent  Al  > A 1  Br.Tet  > A 1  Oct  T  h  g  S  T  E  P  TJI _>TJ2 W  highlights a distinction between the behaviour of  and aluminium with higher co-ordination. The generation of A l  B r T e t  P e n t  and A l ° during the ct  two steam treatments outweighs their depletion by acid leaching whereas the percentage of Al  falls steadily across the series.  B r T e t  Si: A l values can be calculated using the percentages of A l Figure 4.1). The value of S i : A l 29  F w k T e t  F w k T e t  and A l  B r T e t  (Table 4.1 and  follows the same trend as, but is always greater than  S i : A l (the framework Si:Al calculated from S i M A S N M R data), suggesting that 100% of the 2 9  aluminium contributing to the A l  F w k T e t  resonance is indeed in the framework and that the  assignment, first made in Chapter 3, is also correct in this more general study. However, in contrast to the material studied in Chapter 3, these data show that a variable proportion of Al  B r T e t  is part of the framework. This proportion is greatest following an acid washing,  confirming the conclusion from S i M A S N M R that only exfr-a-framework material is removed 29  by acid washing.  T  120  1  1  1  1  1  40 (ppm)  0  I  80  1  1  1  1  -40  120  1  1  80  1  1  1  40 (ppm)  1  0  1  1  r  -40  Figure 4.6 Single p u l s e A l M A S N M R spectra, simulation and deconvolution of U l acquired at (a) 17.6 T 27  and (b) 9.4 T . Parameters used for these simulations are tabulated above. Simulations of similar quality were obtained for the A1 MAS NMR spectra of all samples. 27  139  A  |Fwk.Tet  A  ,Br.Te,  A1  Pent  Al°  ct  Ul  19.6  ±2.9  49.1  ±7.2  6.6  ± 1.0  24.1  ±3.5  Ulw  18.2  ±2.7  33.0  ±4.8  0.5  ±0.1  21.1  ±3.1  U2  7.7  ± 1.1  18.8  ±2.8  5.7  ±0.8  26.1  ±3.8  U2w  8.7  ± 1.4  10.1  ± 1.7  1.8  ±0.3  15.7  ±2.6  Table 4.5 Percentage abundance of aluminium environments in USY samples. Values are obtained from the  integration of peaks used to simulate single pulse and echo A1 MAS NMR spectra acquired at 17.6 T, scaled with respect to NaY and NaA framework tetrahedral signals. 27  140  4.4  Conclusions  This study of a series of steamed and acid washed U S Y materials shows that the experimental techniques introduced in Chapter 3 are suitable for the quantitative analysis of a range of U S Y materials. These data show that while acid washing of U S Y selectively removes some but not all of the extra-framework aluminium, it does not leach aluminium from the zeolite framework. The 16  susceptibility of the aluminium species to acid washing follows the order A l  P e n t  >Al  B r T e t  >Al  0 c t  .  The tetrahedral aluminium in the most distorted environments is preferentially removed, evidenced by a downfield A l 3QMAS F l shift and reduction in width in the A l M A S 2 7  spectra of the A l  B r T e t  NMR  2 7  resonance. Such a selective removal of the most distorted environments  cannot be verified for pentahedrally co-ordinated aluminium, as the almost complete removal of Al  P e n t  by acid washing prevents accurate parameterisation. The removal of A l ° environments ct  by acid washing shows a more complex pattern of selectivity. These conclusions are supported by S i M A S N M R data and literature IR data, which 2 9  20  show that the framework Si A l does not change following acid washing, and by physisorption data, which show an increase in surface area following acid washing that can be attributed to the removal of pore-blocking extra-framework material. There is some suggestion from the physisorption data that acid washing facilitates framework healing. It is also clear from these results that the A l  F w k T e t  12  resonance cannot account for 100% of  the framework aluminium and that the assignment of A l  B r T e t  as extra-framework aluminium,  made in Chapter 3, is not necessarily true in all cases. Indeed, the proportion of A l  B r T e t  that is  part of the framework rises as a result of acid washing from ~ 45% for sample U l to ~ 85% for sample U l w and from ~ 25% for sample U2 to ~ 45% for sample U2w as a result of acid washing. Information is also contained in these data on the effect of steam dealumination, exemplified by the differing behaviour of A l  B r T e t  to the distorted aluminium environments of  higher coordination. Firm conclusions relating to the effects of hydrothermal dealumination are not possible from this study, however, since this series of samples does not isolate the changes due to steam treatment alone. A series of samples related to each other only by hydrothermal treatments are investigated in the following chapter in order to study structural changes  141  associated with hydrothermal dealumination, as part of a preliminary investigation of the relationships between U S Y composition and catalytic cracking activity.  4.5 References for Chapter 4. 1. Corma, A.; Martinez, A.; Martinez, C. Appl. Catal. A, 1996,134, 169-182. 2. Giudici, R.; Kouwenhoven, H.W.; Prins, R. Appl. Catal. A, 2000, 203, 101-110. 3. Cerqueira, H.S.; Ayrault, P.; Datka, J.; Guinset, M . Micropor. Mesopor. Mater., 2000, 38, 197-205. 4. Santos, R.F.; Urquieta-Gonzalez, E.A. Proceedings of the International Zeolite Conference, 12 ; Materials Research Society: Pensylvania, 1998. th  5. Ward, J.W.; Carlson, T.L. US Patent 4517073, 1985. 6. Beyerlein, R.A.; McVicker, G.B.; Yacullo, L . N . ; Zeimak, J.J. J. Phys. Chem., 1998, 92, 1967-1970. 7. Biaglow, A.I.; Parrillo, D.J.; Kokotailo, G.T.; Gorte, R.J. J. Catal, 1994,148, 213-223. 8. Narbeschuber, T.F.; Brait, A.; Seshan, K.; Lercher, J.A. Appl. Catal. A, 1996,146, 119-129. 9. Rhodes, N.P.; Rudham, R.; Standbridge, N.J.H. J. Chem. Soc. Faraday Trans., 1996, 92, 2817-2823. 10. Bamwenda, G.R.; Zhao, Y . X . ; Wojciechowski, B.W. J. Catal, 1994,150,243-253. 11. Bamwenda, G.R.; Zhao, Y . X . ; Groten, W.A.; Wojciechowski, B.W. J. Catal, 1995,157, 209-221. 12. Hudek, P.; Jorik, V . ; Sneiskova, A . ; Zidek, Z. React. Kinet. Catal. Lett, 1997, 60, 15-19. 13. Sulikowski, B.; Klinowski, J. / . Chem. Soc. Faraday Trans., 1990, 86, 199-204. 14. Fejes, P.; Kirisci, I.; Hannus, I. Acta. Phys. Chem., 1982, 28, 173-180. 15. Banazzi, E.; Lynch, J.; Gola, A.; Lacombe, S.; Marcilly, C. Proc. Int. Zeolite Conf, 1999, 12, 2735-2742. 16. Gola, A.; Rebours, B.; Milazzo, E.; Lynch, J.; Benazzi, E.; Lacombe, S.; Delevoye, L.; Fernandez, C. Micropor. Mater., 2000, 40, 73-83. 17. Stockenhuber, M . ; Lercher, J.A. Micropor. Mater., 1995, 3, 457-465.  142  18. Jolly, S.; Saussey, J.; Lavalley, J . C ; Zanier, N . ; Benazzi, E.; Joly, J.F. Der Bunsen Chem. Phys., 1993, 97, 313-315. 19. Mariey, L.; Khabtou, S.; Marzin, M . ; Lavalley, J . C ; Chambellan, A . ; Chevereau, T. Stud. Surf. Sci. Catal., 1995, 97, 501-506. 20. Menezes, S . M . C ; Camorim, V . L . ; Lam, Y . L . ; San G i l , R.A.S.; Bailly, A . ; Amoureux, J.P. App. Catal. A, 2001, 207, 367-377. 21. Lam, L . Y . Personal communication, 2001. 22. Massiot, D.; Fayon, F.; Capron, M . ; King, I.; Le Calve, S.; Alonso, B., Durand, J.-O.; Bujoli, B.; Gan, Z.; Hoatson, G. Magnetic Resonance in Chemistry 2002, 40, 70-76.  143  Chapter 5. Multiple Field A1 NMR Study to Probe 27  the Effects of Calcination on the Structure and fi-Hexane Cracking Activity of USY.  5.1 Introduction A particular focus of research on U S Y has been the role played by aluminium and the Bransted and Lewis acid sites associated with it. ' Following steam dealumination, the acidity 1 2  of the framework is modified, quantities of extra-framework aluminium containing material are generated and a degree of framework degradation occurs ' (refer to Section 3.1 for discussion 3 4  and further references). Several studies show that both the framework and extra framework material play an important role in acid catalysis (refer to Section 4.1 for discussion and references) prompting research into the dealumination process and the aluminium species that it generates. A great deal of accepted understanding of the catalytic properties of U S Y was based upon the proposed nature of the "invisible aluminium" (Section 3.1). For example, it has been suggested that the enhancement of catalytic activity provided by the extra-framework material derives from "superacid sites". It is reasoned that any aluminium that cannot be observed by 5  27  A l N M R must be in an exceptionally distorted environment and therefore responsible for highly Lewis-acidic sites. Pyridine and ammonia adsorption experiments and IR spectra confirmed that steaming generated strong Bnansted and Lewis acid sites compared to the parent materials. The mode of "superacidity" was proposed to be from the new Brensted acid sites, from a synergic interaction between neighbouring Brensted and Lewis acid sites ' ' ' or from the 2 5 8 9  Lewis acid sites alone. It has also been suggested, in part due to the inability to generate 10  analogues of the extra framework material with catalytic properties that support these postulates  144  (refer to Section 3.3 for discussion and references) that there are no special catalytic sites generated in U S Y by steaming and that changes in the catalytic activity are in fact due to the diffusion effects of mesopores formed by framework decomposition. '  11 12  Latterly, as access to high magnetic fields has increased, it has been possible to quantitatively observe 100% of the aluminium present in ultrastable zeolites. ' Further, the 13 14  development of the two dimensional M Q M A S N M R experiment has lead to an unambiguous assignment of all of the signals in the A l M A S N M R spectra " that had not been possible 2 7  13  15  using earlier line-narrowing techniques for quadrupolar nuclei, such as DOR and D A S . Equipped for the first time with an understanding of the signals that comprise the A l N M R 2 7  spectra, it is now possible to characterize all of the aluminium species present and to re-examine the properties of U S Y materials. In this chapter the results of a study to investigate the progressive structural changes brought about by repeated ammonium exchange and calcination on two series of U S Y samples are presented. Using the protocol presented in Chapters 3 and 4, the changes in the amounts of each type of aluminium as a result of successive calcination are measured. These data are correlated with physisorption data and measurements of high temperature (540°C) H-hexane catalytic cracking activity to investigate the relationships of surface area, porosity and mesopore surface area to the changes in the proportions of the aluminium species present and to provide a preliminary interpretation of the factors influencing the catalytic activity.  5.2  5.2.1  Materials and Methods  MATERIALS 9Q  NaY (LZY-52, Union Carbide Corporation) that was determined by  Si M A S N M R to have  a framework Si:Al of 2.6, was stirred at room temperature for 1 hour in ammonium nitrate solution and then rinsed with distilled water during suction filtration. A total of four such ammonium exchanges were performed in ammonium nitrate solution containing a 10 times molar excess of N H / over N a (based on the total sodium present in the parent material) and a +  145  further three ammonium exchanges were performed in an ammonium nitrate solution containing a 50 times molar excess of NFLi , to produce an N H 4 Y material to act as the source for all U S Y +  samples. The N H 4 Y was dried and then calcined at 550°C for 4 hours and allowed to cool in air. The sample temperature was ramped from room temperature to 550°C at a rate of 9°C/min. A portion of the resulting H Y sample was set aside and the remainder was re-washed twice in ammonium nitrate solution containing an approximately 50 times molar excess of N H  + 4  over H  +  (based on a complete ion exchange of the sodium present in the NaY parent material), rinsed in distilled water, dried and re-calcined for 4 hours at 550°C. This procedure was repeated a further four times to produce a series of six samples. A further series of six samples was prepared from the NH4Y starting material using a calcination temperature of 650°C.  5.2.2  EXPERIMENTAL APPARATUS  Solid state S i M A S N M R experiments were performed at 9.4 T (79.5 MHz) on a Bruker 2 9  DMX400 spectrometer using a home-built probe incorporating a 10 mm supersonic Doty stator. Single pulse M A S spectra were obtained at spinning speeds of 1.5-2.5 kHz using a 25° pulse width (4.5 ps) at a rf. power of 15 kHz and a 120 s recycle delay. Spectra were referenced to the 2 9  S i M A S signal of NaA, at -89.7 ppm from TMS. Solid state Al M A S N M R experiments were performed at 9.4 T (104.3 MHz) on a Bruker 27  DMX400 spectrometer and at 17.6 T (195.4 MHz) on a Varian Inova spectrometer (see Acknowledgements) using home built probes incorporating supersonic 5mm Doty stators. Single pulse A l M A S N M R spectra at 9.4 T were obtained at spinning speeds of 12.0-14.1 2 7  kHz using 1.0 u.s pulse widths at a rf. power of 18 kHz. A1 M A S and 3QMAS spectra at 17.6 27  T were obtained at spinning speeds of 12.2-13.8 kHz using a 1.0 u.s pulse length at a rf. power of 56 kHz. 90°-180° echo experiments at both fields were acquired using a 1.0 u.s first pulse and a 180° refocusing pulse after an echo delay of one rotor cycle. A l l  2 7  A l chemical shifts were  referenced to 1M A1(N03)3 (aq). A recycle delay of 0.5 s was used for a l l A l M A S N M R 2 7  experiments. Spectra acquired using a recycle delay of 200 ms were of equal intensity, confirming that no saturation of the signal occurred. 146  Referencing and processing were as discussed in Chapter 2 and the pulse sequences used are listed in Appendix III. X R P D patterns were obtained using a Rikagu Rotating Anode Diffractometer generating C u K a radiation (1=1.5418 A) using a 20 step size of 0.02° at a rate of 2°/min. (Refer to Section 2.2 for further details). T G A was carried out using a T A Instruments TG51 thermogravimetric analyser under a flow of 80 cc/min of dry nitrogen gas. The temperature was ramped from room temperature to 800°C at a rate of 10°C/min and maintained at this temperature for 60 minutes to ensure constant mass. Physisorption measurements were carried out at 77.3 K using a Micrometrics A S A P 2010 instrument using nitrogen as the analysis and saturation gas. Samples were degassed overnight at 350°C under vacuum prior to analysis. Surface area measurements were fit to a Langmuir isotherm, and porosity measurements were obtained from each isotherm using a Saito-Foley modified Hovarth-Kawazoe analysis. Mesopore areas for pore sizes in excess of 15 A were obtained from a cumulative BJH analysis. (Refer to Section 2.4 for further details). Catalytic n-hexane cracking activity measurements were carried out using a quartz reactor system coupled to a Hewlett Packard 5890A Gas Chromatograph. The sample temperature was ramped to 540°C at 10°C/min in a 50 cm /min flow of helium gas prior to a 30 minute 3  activation period in a 10 cm /min flow of compressed air. A 50 cm /min flow of the helium carrier gas containing a saturated vapour pressure of n-hexane (>99%, Aldrich) was passed over the activated sample for 4 minutes and at this time the products were analyzed. The GC oven temperature was ramped from 30°C to 175°C at a rate of 10°C/min and the C1-C4 product distributions measured using a Hewlett Packard 3396 integrator. (Refer to Section 2.5 for details). For the catalytic testing in this study, the sample preparations included the step of pelletizing the samples (using a simple hand operated press to form 1 mm diameter pellets). This procedure, combined with the use of a constant flow rate of hexane through the reactor for all samples keeps the effect of diffusion across macroscopic intercrystallite and "inter-particle" distances (many orders of magnitude greater than those influenced by the microscopic structure of the samples) constant. Therefore these bulk diffusion effects should not be observed in these data.  147  5.3 Results 5.3.1  29  S i MAS, A1 MAS AND 3QMAS NMR SPECTROSCOPY 27  Quantitative Si a n d A l M A S N M R experiments were performed as described in Chapters 29  27  3 and 4 in order to obtain average lineshape parameters and the percentage abundance of all aluminium species as a function of the number of calcinations. Figure 5.1 shows the S i M A S N M R spectra at 9.4 T for the series of samples calcined at 2 9  550°C, the "550-series" and the series of samples calcined at 650°C, the "650-series". Figure 5.2 shows the corresponding single pulse A1 M A S N M R spectra at 17.6 T. Across both series the 27  2 9  S i M A S N M R spectra show a gradual increase in the framework Si:Al (Table 5.1) a n d A l 27  M A S N M R spectra show a general increase in the intensity of broad signals and a gradual broadening of all resonances with successive calcinations (Figure 5.2). The changes across the 650-series are greater than the changes across the 550-series. These changes are similar to those produced by "steaming", i.e. steam calcination, ' " and indicate that U S Y undergoes "self 14 16  26  steaming" ' ' as it dehydrates. 5 9 12  Wang et al. found that the steam calcination reaction is observed to be significantly faster 5  in the initial stages of treatment. In the present work, the hydrothermal conditions that are largely responsible for the dealumination reactions are only present for a short time and therefore, although it was found that the spectroscopic changes of these systems are extremely sensitive to the precise hydration of the samples prior to calcination and to the rate of temperature increase towards the final calcination temperature, the structural changes observed in this study reflect the extent of reaction occurring during a short time period early in the calcination process. A l l *Si M A S N M R spectra were simulated in order to obtain framework z  Si:Al values  (Table 5.1) using Eqn 3.1. In addition to the peaks representing Si[nAl], each simulation included a broad peak to account for an amorphous silica-containing component, most obvious in the simulations of the S i M A S N M R spectra of the 650-series after successive calcinations. 2 9  The  2 9  S i M A S N M R spectrum of a U S Y sample treated under extremely harsh conditions  (steam calcination at 850°C) shows an amorphous silicon signal of much greater intensity relative to the framework signals and the peak parameters obtained from the deconvolution of  148  this spectrum were used in the simulation of the  Si M A S N M R spectra of the 550- and 650-  series.  (a) U S Y materials calcined at 550°C.  Number of  2 9  Si:Al  a  6  Calcinations  Si:Al  F w k  "  T e t  ±  c  Si:Al  T o t a l T e t  -  ±  ±  1  3.4  0.2  8.3  1.9  3.1  0.7  2  3.8  0.2  8.8  2.6  3.6  1.0  3  4.3  0.2  9.5  2.1  3.4  0.8  4  4.6  0.3  10.6  2.0  3.6  0.7  5  4.9  0.3  11.1  2.1  3.6  0.7  6  5.2  0.3  11.6  2.3  3.5  0.7  b) U S Y materials calcined at 650°C.  Number of  2 9  Si:Al  *Si:Al  a  Calcinations  ±  F w k  "  T e t  c  Si:Al  T o t a l T e t  ±  -  ±  1  3.7  0.2  9.5  1.7  3.3  0.6  2  4.7  0.2  11.7  3.7  4.3  1.4  3  5.2  0.2  11.9  3.9  5.1  1.7  4  5.8  0.3  16.8  4.7  5.8  1.6  5  6.5  0.3  36.6  9.2  6.4  1.6  6  8.5  0.3  40.4  7.9  IA  1.4  Table 5.1 Si:Al values calculated for the series of U S Y materials calcined at (a) 5 5 0 ° C and (b) 6 5 0 ° C .  "Calculated from Si MAS NMR spectra recorded at 9.4 T, according to Eqn 3.1. Tvlolar percentages of Si and Al in the bulk material are derived from TGA measurements, assuming a bulk stoichiometry of (SiC^^A^C^)^.^ where 2xl(\ -x)- 2.6. Values of the bulk Si:Al calculated from these molar percentages were found to 2.60 ± 0.02 for all samples of both series. S i : A l and Si:Al were then calculated by assuming that the proportion of the total aluminium in the framework is equal to the integral of A l and ( A l + Al ) respectively of the peaks used to simulate the Al MAS NMR spectra recorded at 17.6 T. 29  FwkTet  c  Total T e t  F w k T e t  F w k T e t  BrTet  27  It has been shown that  Si:Al values for U S Y systems may underestimate the true  framework Si:Al ratio due to contribution to Si[«Al] signals from Si[«OH] silanol groups lying at similar chemical shifts. However, our studies of U S Y systems of comparable crystallinity  149  and framework S i : A l ' have shown good agreement between S i : A l values and those obtained 6  14  29  by IR spectroscopy, suggesting that the framework S i : A l values of samples treated under these 29  relatively mild conditions are reliable.  i  -80  1 — i  -90  1  1  1 — i — i  i  1  -100 -110 -120 -80 (ppm)  1  1  -90  1  1  1  1  1  1  -100 -110 -120 (ppm)  Figure 5.1 " S i M A S N M R spectra of the series of USY materials calcined at (a) 5 5 0 ° C and (b) 6 5 0 ° C ,  recorded at 9.4 T. Samples were successively ammonium exchanged and calcined at each temperature and the spectra are labelled according to the number of treatments. All spectra were simulated using the program DmFit and a simulation and deconvolution of thefirstmember of each series are shown as examples. Peak assignments common to all spectra and simulations are shown above the deconvolutions. Spectra were obtained at spinning speeds of 1.5-2.5 kHz using a 4.5 us pulse width and a 120 s recycle delay.  27  150  Figure 5.2 Single pulse A1 M A S N M R spectra of the series of U S Y materials calcined at 5 5 0 ° C and 6 5 0 ° C , recorded at 17.6 T labeled according to the temperature of and number of calcinations. Spectra were obtained 27  at spinning speeds of 12.2-13.8 kHz using a pulse length of 1.0 us and a recycle delay of 0.5 s. Asterisks indicate spectra whose simulations are shown in *Figure 5.5(b) and **Figure 5.6(b).  151  A l 3QMAS N M R spectra of selected members of each series were also recorded at both field strengths. The spectra of samples calcined once and six times at 550°C and 650°C recorded at 17.6 T are shown in Figure 5.3 and Figure 5.4, respectively. A l l spectra were qualitatively similar to samples studied in Chapters 3 and 4. In all members of the 550-series and the first member of the 650-series, a narrow, isotropic signal ( A l ° ) can be resolved in the octahedral Iso  regions of both the  ct  A l M A S (Figure 5.2) and 3QMAS spectra and is assigned analogously to  the resonance seen in the A l M A S and 3QMAS spectra of samples U l w and U2w (Chapter 4). 2 7  The total integrals of t h e A l M A S N M R spectra were calibrated against spectra of NaY 27  and NaA as described previously and the results are summarized in Table 5.2.  (a) Series calcined at 550°C. Number of Calcinations  17.6 T Single Pulse  9.4 T  ±  Echo  ±  92 111  13 15  ±  Echo  ±  59 62  9 9  73 72  11 11  Single Pulse  1 2  102 105  5 6  3  96  5  105  15  60  9  63  9  4  104  5  101  14  56  8  60  9  5  102  5  102  14  57  8  56  8  6  93  5  100  14  61  9  47  7  ±  Echo  ±  (b) Series calcined at 650°C. Number of Calcinations  17.6 T Single Pulse  9.4 T  ±  Echo  ±  Single Pulse  1  97  8  98  11  57  9  77  12  2  111  9  111  13  64  10  65  10  3  108  9  115  13  60  9  66  10  4  107  9  110  13  61  9  63  9  5  108  9  106  12  60  9  54  9  6  97  8  96  11  59  9  51  8  Table 5.2 Percentage of total aluminium in U S Y materials calcined at (a) 5 5 0 ° C and (b) 6 5 0 ° C , observed using A1 M A S N M R . J7  152  100  80  60  40 20 F2 (ppm)  0 -20  Figure 5.3 A 1 3 Q M A S N M R spectra of U S Y materials ammonium exchanged and calcined at 5 5 0 ° C : 7  (a) once (b) six times, recorded at 17.6 T. Spectra were recorded using a 2-pulse nutation pulse sequence at spinning rates of (a) 14.1 kHz (b) 14.2 kHz. Skyline projections are shown. '*' denote spinning sidebands. A diagonal offset artefact is visible in (b).  153  100  80  60  40 20 F2 (ppm)  0 -20  Figure 5 . 4 A l 3 Q M A S N M R spectra of USY materials ammonium exchanged and calcined at 6 5 0 ° C 2 7  (a) once (b) six times, recorded at 17.6 T . Spectra were recorded using a 2-pulse nutation pulse sequence at  spinning rates of (a) 14.5 kHz (b) 14.6 kHz. Skyline projections are shown. '*' denote spinning sidebands.  154  As described previously (Chapters 3 and 4), the spectra at 17.6 T (Figure 5.2(a)) show intensity from 100% of the aluminium present, within experimental error. The single pulse experiments at 9.4 T (Figure 5.2(b)) consistently show intensity from only approximately 60% of the aluminium present and the echo experiments show approximately 75% of the expected intensity at the beginning of each series, dropping to approximately 50% at the end of each series. This clearly indicates that the broad aluminium signals are responsible for the "invisible aluminium" (refer to Section 3.1 for discussion and references) since the intensities of these broad resonances increase along both series. 77  77  A l l 'A1 M A S N M R spectra were simulated using the programme DmFit Z  and examples of  the simulations of high and low field single pulse spectra are shown in Figure 5.5 and Figure 5.6. A summary of the simulation parameters is given in Appendix TV and the integrals of each simulation are presented in Table 5.3. These integrals are used to calculate the values S i : A l and S i : A l  T o t a l T e t  F w k T e t  (Table 5.1). Each of the aluminium resonances was simulated by a second order  quadrupolar M A S line multiplied by an exponential broadening function in order to account for distributions in chemical shift and quadrupolar coupling. This method does not fully reproduce the asymmetry caused by distributions in quadrupolar coupling, hence the intensity maxima of the experimental spectra and their simulations are slightly offset. In the octahedral region, in addition to a narrow isotropic peak close to zero ppm, the spectra were simulated using two 77  peaks. The  A l 3QMAS N M R spectra show a distribution of environments in the octahedral  region and therefore this feature of the simulations is not intended to infer the existence of two distinct broad octahedral environments, but rather to account for one particularly broad resonance. Additional peaks with mixed Lorentzian and Gaussian character are used to simulate spinning sideband and satellite transition intensity, as described in Section 3.3.4.  155  I  T  I  I  I  1  1  1  1  1  1  1  I  I  I  I  I  I  I  I  I  I  I  150  100  0  50  -50  -100  (ppm) Figure 5.5 " A l single pulse MAS NMR spectra and spectral simulations of USY ammonium exchanged and  calcined four times at 5 5 0 ° C , recorded at (a) 9.4 T and (b) 17.6 T . Spectra were simulated using the program DmFit. A l resonances were simulated using second order quadrupolar MAS lineshapes with the average VQ values given in Table 5.3 and afielddependent exponential broadening to account for distributions in chemical shift and quadrupolar coupling. Intensity from A l , spinning sidebands and satellite transitions are simulated using isotropic peaks with mixed Gaussian and Lorentzian character. 27 2 7  lso 0 c t  156  Figure 5.6 A I single pulse M A S N M R spectra and spectral simulations of U S Y ammonium exchanged and calcined five times at 6 5 0 ° C , recorded at (a) 9.4 T and (b) 17.6 T . Spectra were simulated using the program DmFit as described in Figure 5.5. 27  27  157  (a) Series calcined at 550°C.  Number of Calcinations  A  jFwk.Tet  A 1  B,Tet  ±  A  jTotal Tet  ±  A 1  Pe„t  ±  A 1  Oct  ±  ±  1  31.7  7.0  51.7  11.4  83.4  18.4  2.2  0.5  14.5  3.2  2  29.8  8.5  43.7  12.5  73.4  20.9  2.6  0.7  24.0  6.9  3  27.3  5.8  42.3  9.0  69.6  14.9  5.7  1.2  24.8  5.3  4  24.6  4.4  48.6  8.6  73.2  13.0  5.8  1.0  21.1  3.7  5  23.5  4.3  48.8  8.9  72.3  13.2  6.5  1.2  21.2  3.9  6  22.3  4.2  51.0  9.5  73.3  13.7  7.4  1.4  19.3  3.6  (b) Series calcined at 650°C.  Number of Calcinations  A  jFwk.Tet  A1  Br.Tet  ±  A  jTotal Tet  ±  A 1  p „t  ±  e  A  ,Oct  ±  ±  1  27.4  4.7  52.3  9.0  79.7  13.7  4.7  0.8  15.6  2.7  2  22.2  6.8  37.7  11.5  59.9  18.3  18.5  5.7  21.6  6.6  3  21.9  7.2  28.8  9.5  50.6  16.7  25.0  8.2  24.3  8.0  4  15.4  4.2  29.7  8.1  45.1  12.4  30.4  8.3  24.5  6.7  5  7.9  1.9  36.1  8.8  44.0  10.7  34.8  8.5  21.2  5.2  6  6.9  1.3  30.6  5.6  37.5  6.9  48.1  8.8  14.3  2.6  Table 5.3 Percentage abundance" of aluminium species in U S Y materials calcined at (a) 5 5 0 ° C and (b)  6 5 0 ° C . "Percentages are calculated from the integration of peaks used to simulate single pulse and echo Al MAS NMR spectra recorded at 17.6 T (Figure 5.2), the total integrals of which are assumed to represent 100% of the aluminium present. 17.6 T echo Al MAS NMR experiments were simulated using peak parameters obtained from the simulation of single pulse experiments performed at 17.6 T and 9.4 T (Appendix IV). Error limits are estimated from deviations of the total spectral integrals from 100% (Table 5.2) and from intensity differences between simulations of single pulse and echo experiments. 27  27  158  5.3.2  PHYSISORPTION AND P O W D E R X R D D A T A  The nitrogen physisorption data are summarized in Table 5.4. The analyses assume that there is uniform interaction between the sorbate molecules (N2) and the sorbent (USY) and that these interactions are large compared to sorbate-sorbate interactions. U S Y potentially contains significantly different surface interactions associated with the extra framework material and the zeolite framework and localized charges associated with the aluminium that may interact significantly with the N2 molecular quadrupole, thereby complicating the interpretation of nitrogen physisorption data for these systems. ' For these reasons, Langmuir surface area 29 30  values are not taken from regions of the isotherm at extremely low pressure or at pressures higher than P/Po ~ 10" so as to avoid data that does not conform to the form of a Langmuir isotherm. This method gave a surface area measurement of a commercially available U S Y material (LZY-84, U.O.P.) of 800 m ^ that was in good agreement with the published data 2  1  sheet which quoted a surface area of 820 m g . Large differences in surface area values 2  _1  extracted from isotherm data can be brought about by changing the range of data analyses. In this work, the P/Po data range was constant for all Langmuir surface area analyses. The Saito-Foley modified Hovarth-Kawazoe porosity analyses of these data (Figure 5.7) show a bimodal pore size distribution for all samples, at approximately 8 A and 10 A. However, it is unlikely that this represents a true picture of the structure. The peak at approximately 8 A is often composed of a single data point and is always derived from the very low relative pressure data that has been ignored for the purpose of the other analyses. In a study of mixed lanthanumacid Y materials (H^La^Y), de la Puente and Sedran attribute such data to preferred site 29  occupation by the nitrogen rather than to a true bimodal distribution, highlighting peaks defined by the first one or two data points as an indication of this. Further, de la Puente and Sedran did not see a bimodal pore size distribution in argon physisorption measurements. For this reason, the median pore diameter values (which derive from the same regions of data as the surface area values) may be regarded as reliable and the observation of two distinct pore sizes is assumed to be an experimental artefact. Powder X R D patterns for both series (Figure 5.8) are consistent with the physisorption data, showing a dramatic increase in line widths, a decrease in peak intensities and an increasing  159  amorphous background signal (not quantitative) in the 650-series and analogous but far less extensive changes in the 550-series.  (a) Series calcined at 550°C.  "Langmuir Surface Area  Number of Calcinations  (mV)  *BJH Mesopore Area (m g")  Mesopore Area as % of Total  ''Median Pore Diameter (A)  2  !  NaY  890.2  22.0  2.5  8.39  1  809.5  22.3  2.8  9.41  2  801.3  31.7  4.0  9.88  3  774.5  36.6  4.7  10.11  4  769.0  49.0  6.4  10.31  5  772.2  69.1  9.0  10.36  6  755.2*  39.6  5.3  10.50  rf  (b) Series calcined at 650°C.  Number of Calcinations  "Lagmuir Surface Area (m g )  *BJH Mesopore Area (m g )  Mesopore Area as % of Total  ^Median Pore Diameter (A)  1  787.6  49.8  6.3  9.41  2  746.3  44.2  5.9  10.15  3  670.6  63.4  10.0  10.51  4  585.1  81.7  14.0  10.54  5  515.7  107.8  20.9  10.67  6  457.3  155.7  34.0  11.10  2  _1  2  _1  Table 5.4 Physisorption data from N isotherm analysis of USY materials calcined at (a) 5 5 0 ° C and (b) 2  6 5 0 ° C . "± 2%. ± 7% ± 14% Values obtained from a Saito-Foley modified Horvath-Kawazoe analysis of N isotherm (refer to Section 2.4). Accurate to ± 0.1 A . Error limits are estimated from the distribution of values obtained from repeated experiments on the reference sample LZY-84. b  c  d  2  160  5  10  15  20 5  Pore Size (A)  10  15  20  Pore Size (A)  Figure 5.7. Differential pore size distributions based on nitrogen adsorption isotherms of U S Y materials.  dVacis I dD (arbitrary units) is plotted against D, where V is the volume of adsorbed gas at s.t.p. and D is the pore diameter. Plots are labelled according to calcination temperature and the number of ammonium exchange and calcination treatments. aa!s  161  (a)  4  8  12  16  20 24 20 (degrees)  28  32  36  40  4  8  12  16  20 24 20 (degrees)  28  32  36  40  Figure 5.8 Selected X R P D patters of U S Y materials calcined at (a) 5 5 0 ° C (b) 6 5 0 ° C . The number of  ammonium exchange and calcination treatments is indicated next to each powder pattern. Diffractograms obtained using CuKot radiation (k=l .5418 A) using a 20 step size of 0.02° at a scan rate of 27min.  162  5.3.3  »-HEXANE CATALYTIC CRACKING DATA  Table 5.5 summarizes the n-hexane catalytic cracking data acquired for the two series of U S Y samples. Values have been corrected for hydration and scaled with respect to the activity of the standard sample LZY-84 (see caption, Table 5.5).  (a) Series calcined at 550°C.  Number of Calcinations  Methane  Ethene  Ethane  ±  ±  ±  ±  ±  ±  ±  Total  C4  C3  C2  1  0.09  0.1  0.12  0.1  0.11  0.1  0.22  0.1  0.83  0.1  0.35  0.1  1.49  0.2  2  0.17  0.1  0.37  0.1  0.20  0.1  0.57  0.1  3.59  0.4  1.30  0.2  5.62  0.6  3  0.21  0.1  0.45  0.1  0.23  0.1  0.68  0.1  4.32  0.5  1.92  0.2  7.13  0.8  4  0.27  0.1  0.56  0.1  0.28  0.1  0.83  0.1  5.48  0.6  2.14  0.2  8.73  0.9  5  0.35  0.1  0.66  0.1  0.34  0.1  1.01  0.2  6.29  0.7  2.42  0.3  10.07  1.1  6  0.29  0.1  0.56  0.1  0.29  0.1  0.86  0.1  5.48  0.6  2.02  0.2  8.65  0.9  (b) Series calcined at 650°C.  Methane  Ethene  Ethane  ±  ±  ±  Calcinations  C2  C3  C4  ±  ±  ±  ±  Total  1  0.04  0.1  0.05  0.1  0.04  0.1  0.08  0.1  0.23  0.1  0.12  0.1  0.48  0.1  2  0.11  0.1  0.22  0.1  0.15  0.1  0.37  0.1  2.53  0.3  0.81  0.1  3.83  0.4  3  0.12  0.1  0.26  0.1  0.15  0.1  0.42  0.1  3.11  0.3  0.97  0.2  4.62  0.5  4  0.13  0.1  0.28  0.1  0.16  0.1  0.44  0.1  3.25  0.4  1.03  0.2  4.84  0.5  5  0.11  0.1  0.25  0.1  0.15  0.1  0.39  0.1  2.94  0.3  0.92  0.1  4.36  0.5  6  0.08  0.1  0.16  0.1  0.11  0.1  0.27  0.1  1.85  0.2  0.57  0.1  2.78  0.3  Table 5.5 n-Hexane catalytic cracking product conversion percentages per unit mass of dry USY materials  calcined at (a) 550°C and (b) 650°C. All data have been normalized to a total conversion percentage of 10% for the standard sample LZY-84. Error estimates have been made using contributions from the uncertainty in the TGA data, distribution in catalytic cracking results from LZY-84 and a constant GC instrumental error of ± 0.02%.  163  5.4 Discussion  5.4.1  STRUCTURAL C H A N G E S INDICATED BY PHYSISORPTION AND P O W D E R X - R A Y DIFFRACTION D A T A  The Langmuir surface area decreases across both series of U S Y samples (Table 5.4). A n assessment of the changes in the surface area as a result of calcination is complicated by the absence of a starting H Y material representing zero calcinations because the conditions required to decompose ammonium ions to produce H Y also cause some framework dealumination. However, it is clear from the changes in the surface area across the 550-series that the surface area of this hypothetical starting material would be lower than that of NaY and approximately 12% greater than the surface area of the first member of the 550-series, i.e. a value of approximately 830 m g"\ 2  The decrease in surface area due to repeated calcination across the 650-series is considerably greater (ca. 45%) than across the 550-series (ca. 10%). There is also some indication that the surface area across the 550-series has levelled out after three or four calcinations. Additionally, B J H mesopore areas increase to a much greater extent at 650°C than at 550°C and across the 550-series the mesopore area passes through a maximum. The corresponding powder X R D data of the third through to the sixth members of the 550-series show no significant changes in the crystallinity of the samples (measured from the maximum peak intensities of the four lowest angle peaks) and there are no significant changes in the amplitude of the baseline to indicate an amorphous phase. A similar analysis of the X R P D patterns of the 650-series indicates that the corresponding changes are considerably greater. Peak intensities indicate that approximately 60-65% of the crystallinity has been lost and diffuse scattering intensity is apparent between the angles of 20 ~ 15-35°, with a maximum at 20 ~ 25°. The decrease in the sample crystallinity, as reflected in the X R P D data, could be caused by any changes that distort the framework, including partial or complete removal of aluminium leaving defect sites, or from more extensive framework decomposition and the formation of larger cavities within the framework. However, the diffuse scattering signal can only originate from 2  3  2  disordered regions large enough to scatter the X-rays (i.e. 10-10 A across) and can therefore  164  only originate from significant condensation of extra-framework aluminium containing species after extensive regions of framework collapse. The observed reduction in surface area (Table 5.4) could be a result of either pore blockage by the extra-framework material rendering regions of the crystallites inaccessible to the probe molecules, or by framework degradation. The faujasite structure has a three dimensional pore system and in order for regions of the structure to be cut off from probe molecules, the pore system in all three dimensions surrounding that region would have to be blocked by extra framework material. Since only 28% of the framework tetrahedral atoms in the parent NaY material are occupied by aluminium and at most 60% of this is removed from the lattice in the most extensively modified samples used for this study, it is highly unlikely that there is sufficient extra framework material present to cause the observed changes in surface area by pore blockage. Therefore, the major part of the observed changes in surface area must be the result of framework degradation. Mesopore formation and loss of crystallinity resulting from framework decomposition caused by steam calcination have been widely documented. ' ' ' Of particular interest are 3 4 31 32  those studies that also associate mesopore formation with an enhancement in surface aluminium concentration by comparison to the bulk ' ' since this infers some migration of aluminium to 5  31  33  the surface of mesopores and the exterior of crystallites. It is reasonable to suppose that this enhanced aluminium concentration is also associated with mesopore formation within the crystallites. The Saito-Foley modified Hovarth-Kawazoe pore diameter distribution analyses of the N2 isotherm data (Figure 5.7) show a pore size distribution with a framework-dominated maximum at ca. 9-10 A. The resolution of pore sizes by this method is limited and these data therefore indicate that there is no well defined size of mesopores present in these samples. However, a contribution from some wide distribution of larger sized pores within the structure, with similar adsorption characteristics to the intact framework, could account for the increase in median pore diameter that is observed to result from successive calcinations and with increasing calcination temperature (Table 5.4). Two extremes types of structural changes can be proposed to explain these data: In one extreme ("model-A") calcination causes framework decomposition associated with aluminium T-sites, such that overall framework dealumination occurs, leaving undamaged, mesopore-free  165  framework material (other than mesopores at the external surfaces of the crystallites) and particles of "extra-crystallite" amorphous material, some of which are large enough to produce discernable diffuse scattering of X-rays. If this complete phase separation of framework and extra-framework material occurs, the surface area per unit mass of completely intact framework would remain constant and close to that of the starting material. According to this model, since the surface area of any amorphous material is very small compared to the surface area of crystalline material, the change in the proportions of undisturbed framework and bulk amorphous material is given by the change in the total surface area. Since they are restricted to the exterior of the crystallites, the mesopore areas would be predicted by model-A to be small. In the other extreme ("model-B") there is no migration of extra-framework aluminium, dealumination would be associated with mesopore formation and extra-framework material would reside with the framework and the mesopores. Extra-crystallite amorphous material only results from mesopores that form on the exterior of crystallites and the change in surface area would be associated with a much greater increase in mesopore area than for model-A. Both models could account for the observed changes in the total surface area and median pore diameters. Changes in the Langmuir surface area and the BJH mesopore areas across the 550-series are much smaller than the changes across the 650-series. However, for the 650-series, these data suggest that the real system is most closely described by model-B, because mesopore areas constitute up to 34% of the total area of each sample or 18-20% of the area of the hypothetical "perfect" H Y material (Table 5.4(b)). However, the amorphous diffuse scattering signal in the powder X R D patterns indicates that some phase separation of bulk amorphous and crystalline material may also occur, i.e. model-A. A comparison of the physisorption data for samples in the 650-series with similar S i : A l 29  ratios to samples in the 550-series is instructive (Table 5.1), e.g. the comparison between the fourth and sixth members of the 550-series with the second and third members of the 650-series, respectively. In both cases the samples calcined at 650°C have lower Langmuir surface areas, higher mesopore areas, but very similar median pore diameters to the samples calcined at 550°C. These observations, and the concomitant increase in the diffuse scattering signal through the 650-series, can be explained by a mechanism of dealumination in which framework decomposition is more rapid at higher temperature and that it occurs throughout the crystallites,  166  including at the surface. This accounts for the large changes observed in the  Al MAS NMR  spectra of the 550-series that are not associated with dramatic changes in the X R P D patterns, since the random placement of regions of disturbed framework and decomposed framework and associated extra framework, intra-crystallite material would not produce significant diffuse scattering. More extensive dealumination and decomposition of the framework leads to a greater amount of amorphous material and the increase in temperature may also produce a greater degree of migration through the structure and condensation of that amorphous material. This would further enhance the observable diffuse scattering in the powder X R D patterns and could explain its dramatic increase in X R P D patterns of the 650-series as compared to the 550-series. The physisorption and powder X R D data of both series of samples are therefore indicative of structural changes most closely described by model-B, but with some "model-A character" to the changes across the 650-series in particular. This conclusion is consistent with studies that document extensive migration of extra-framework material leading to large quantities of extracrystallite amorphous material only after prolonged steam dealumination at higher temperatures than those used in the present work. '  5 33  It has long been known that bulk amorphous alumina and silica-alumina are catalytically inactive compared to U S Y and that the conditions used to synthesize and regenerate industrial U S Y catalysts generate significant amounts of extra-crystallite amorphous material. It would therefore be valuable to ascertain what proportion of the mass of each sample in the present study is composed of extra-crystallite amorphous material, since it is possible that changes in catalytic activity might reflect a change in the proportion of the amorphous phase rather than changes in the properties of the U S Y itself. The values for this "model-A correction" cannot be extracted from the physisorption data, since changes in the surface area per unit mass of the crystalline component and the contribution to the mesopore area from mesopores on external crystallite surfaces are unknown. However, it is clear that any such changes in the 550-series are small, such that uncorrected data closely reflects the properties of U S Y material. Clearly this is not the case for the 650-series. However, the trends in both the catalytic properties of and of the proportions of the aluminium species present in both series are similar (Table 5.3), which infers that changes in the proportions of extra-crystallite amorphous material are not primarily responsible these observations. Additionally it is possible from the physisorption data to determine the maximum allowable  167  correction, assuming that the remaining crystalline material has the same surface area per unit mass as that of the starting material. Catalytic cracking results and the abundance of aluminium species in each sample have been thus corrected and the results are discussed below.  5.4.2  RELATIONSHIP BETWEEN S i AND A l M A S N M R SPECTRA AND STRUCTURAL CHANGES 2 9  29  2 7  2 7  S i : A l values and Si:Al ratios calculated from an integration of the deconvolution of the  A l M A S N M R spectra are summarized in Table 5.1 and plotted against the number of  calcinations in Figure 5.9. At both 550°C and 650°C, the value of S i : A l  F w k T e t  follows the same trend as, but exceeds  the value of S i : A l . Thus, all of the aluminium atoms that contribute to A l 29  F w k T e t  are in the  framework, but all of the framework aluminium is not represented by this resonance. Si:Al  T o t a l T e t  values give a much closer approximation of S i : A l , but do not reproduce the trends  of S i : A l . S i : A l 29  29  T o t a l T e t  i s always less than S i : A l and whereas the values remain close for all 29  members of the 650-series (Figure 5.9(a)), the values diverge across the 550-series (Figure 5.9(b)). These data indicate that a variable but substantial proportion of the broad tetrahedral aluminium species are contained within the framework. At 550°C, the dealumination produced by the calcination produces a steady build up of the amount of exfra-framework A l  B r T e t  , but at  650°C the amount of this material remains small. Previous studies of U S Y systems (Chapters 14  3 and 4) have also shown that the proportion of extra-framework to framework A l  B r T e t  can vary  anywhere from 0-100%. These calculations and the associated changes in the physisorption data suggest that framework and extra-framework A l  B r T e t  represent the first two steps of framework  dealumination. First, some attack upon the framework leads to a distorted tetrahedral aluminium environment, but in such a way that does not noticeably broaden the S i M A S N M R 2 9  77  resonances. This condition is not difficult to envisage, given that A l is a quadrupolar nucleus and aluminium T-sites are always the focus of the chemical attack. Second, the reaction continues and the aluminium is removed from the framework. This model would predict a small reduction, or no change at all, in the surface area as a result of the first process. The change in 168  surface area as a result of the second step would depend upon the response of the remaining framework to the removal of an aluminium atom. However, the amount of extra-framework Al  B r T e t  remains small throughout the 650-series (unlike the 550-series) and the overall increase  in framework Si:Al is also greater across the 650-series. These two observations indicate that the conversion of extra-framework A l  B r T e t  to some other extra-framework species is favoured at  higher temperature.  Figure 5.9 Si:Al calculated for U S Y materials calcined at (a) 5 5 0 ° C and (b) 6 5 0 ° C . Data and notes on  calculations are presented in Table 5.1. 'o' represent Si:Al, '•' S i : A l and 'A' S i : A l . Trend lines result from least-squares-fitting of second order polynomials to the data. The mark V on the abscissa indicates the framework Si:Al value (2.6) of the starting material. 29  FwkTet  2  169  To,alTet  van Bokhoven et al.  have reported the appearance of a broad tetrahedral aluminium  resonance with an intensity proportional to the concentration of L a  3+  in the A l MAS NMR 27  spectra of (Na/. La ^)Y systems and have proposed that broad tetrahedral Al NMR intensity IH  27  x  in spectra of these samples and mixed Na^I-LY samples is due to framework aluminium adjacent to framework oxygen atoms co-ordinated to "La " or "Al ". Gola et al. draw similar 3+  3+  34  97  conclusions from Al MAS and 3QMAS NMR studies of USY materials, supported by XRPD data. It is therefore possible that some or all of the A l  Br T e t  intensity that is calculated to be from  aluminium in the framework is due to framework aluminium adjacent to isolated A l (such as, or similar to, those proposed to account for the sharp A l  ls0  framework A l  B r T e t  species  ° resonance) co-ordinated c t  to framework oxygen atoms. It does not seem likely that all of the framework A l results from coordination of Al  3 +  B r T e t  intensity  to the framework, since the proportion of framework to extra-  is greater in samples of similar framework Si:Al calcined at higher  temperature. Both the A l MAS and 3QMAS NMR data presented here and the work of Corma 27  et al.  indicate that the degree of "condensation" of the extra-framework aluminium increases  with the extent and severity of treatment.  Figure 5.10 (a) Percentage abundance of aluminium species in U S Y materials calcined at 5 5 0 ° C . (b) Relative percentages of aluminium species in U S Y materials calcined at 5 5 0 ° C normalized to a surface area of 1000  m g"'. '•' represent A l data, 'o' A l , '•' A l and '•' A l . Data and notes on calculations are presented in Table 5.3(a). Trend lines result from the least-squares-fitting of second order polynomials to the data. 2  F w k T e t  B r T e t  Pent  0ct  170  Figure 5.11 (a) Percentage abundance of aluminium species in U S Y materials calcined at 6 5 0 ° C . (b) Relative percentages of aluminium species in U S Y materials calcined at 6 5 0 ° C normalized to a surface area of 1000 m g"' and corrected for a trend in the percentage of extra-crystallite amorphous material such that the surface area per unit mass of remaining crystalline material is constant. ' • ' represent A l data, 'o' A l , 2  F w k T e t  B r T e t  '•' A l and '•' A l . (a) Data and notes on calculations are presented in Table 5.3(b). (b) This calculation supposes a composition of the amorphous material of 44% tetrahedral, 44% pentahedral and 12% octahedral aluminium. These values were obtained from the decovolution of a sample of U S Y subjected to prolonged steam calcination at 850°C to induce close to complete framework decomposition and are consistent with an extrapolation of data for the 650-series. Trend lines result from least-squares-fitting of second order polynomials to the data. Pent  0ct  There are obvious similarities across each series in the trends of the percentages of the other aluminium species (Figure 5.10 and Figure 5.11). At both 550°C and 650°C, the percentages of Al  P e n t  build up with successive calcinations and the percentages of A l  midway through each series. By contrast, the percentages of A l calcinations and the percentages of A l  B r T e t  F w k T e t  0 c t  pass through maxima  decline with successive  pass through a minimum. The build up of A l  Pent  isfar  greater across the 650-series than across the 550-series. As discussed above, Si:Al calculations (Table 5.1 and Figure 5.9) indicate that the conversion of extra-framework A l high temperature and therefore this build-up of A l formed from the extra framework A l  B r T e t  . Al  0 c t  P e n t  B r T e t  is favoured at  suggests that this is the species that is  and A l  B r T e t  follow approximately opposite  trends, which is consistent with the supposition that they represent the first and last stages of a  171  common reaction (Figure 5.12). The general nature of the trends across the 650-series are not changed by the correction that is made to account for the maximum possible proportion of extracrystallite amorphous material. However, this correction does seem to bring the trends closer to those seen across the 550-series, confirming that framework collapse and phase separation are indeed far more significant at the higher temperature.  ^|Fwk.Tet-^TT^^|Br.Tet  >  (framework)  A\  oa  ^jBr.Tet  (extra framework)  —  Al "" p  Figure 5.12 Proposed reaction scheme for the hydrothermal dealumination of  USY.  The reversibility of the reactions between the different types of extra-framework material may be inferred by the maxima and minima in their percentages. Although it may seem intuitively reasonable, there is no evidence to suggest or to refute that the formation of A l reversible or that A l  P e n t  is always formed intermediately between A l  B r T e t  P e n t  is  and Al° . These data ct  do not suggest that any species other than the tetrahedral aluminium constitutes part of the framework. ' ' Further insight into the nature of the competitive processes taking place is given by the mesopore surface areas. Across the 550-series, the mesopore area is small and passes through a maximum after five calcinations. This must indicate that "framework healing" takes place. ' '  31 38 3  Potentially this could be due to defect migration to the edge of crystallites, resulting in surface mesopores (i.e. model-A) or from a more localized healing process utilizing a nearby silica source (i.e. model-B). As discussed above, Langmuir surface area and porosity analyses of the physisorption data and the changes in the X R P D patterns suggest that the 550-series is best described by model-B. There is no indication of framework healing in the 650-series, where the mesopore area increases steadily across the whole series and to a much greater degree than the 550-series, indicating that model-B also best describes this system. 172  At 550°C the dealumination reaction occurs at a lower rate than at 650°C. When aluminium is removed from the framework, larger cavities are formed and the framework becomes damaged (indicated by an increase in median pore diameter and a decrease in surface area). The rate of dealumination must decrease along with the concentration of A l  F w k T e t  . The effects of the  framework healing reaction become apparent in the second half of the 550-series as a continued increase in median pore diameter and a decrease in mesopore area without the decrease in surface area seen earlier in the 550-series. At 550°C the formation of A l  P e n t  sufficiently slow that aluminium removed from the framework exists as A l  and A l ° is ct  B r T e t  in significant  quantities. At 650°C the rate of framework decomposition and the rate of consumption of the A l  B r T e t  removed from the framework remain sufficiently high that the trends in Langmuir and B J H mesopore areas indicative of the framework healing process are not observed. To a first approximation, every framework dealumination at 650°C is associated with framework destruction, the formation of cavities within the framework larger than those in the faujasite structure and with the formation of amorphous A l  P e n t  - and Al° -containing material within those ct  cavities.  5.4.3  RELATIONSHIP B E T W E E N STRUCTURAL C H A N G E S AND « - H E X A N E C R A C K I N G ACTIVITY  Conditions for the w-hexane cracking experiments were chosen to duplicate as closely as possible the conditions of the Mobil or-Test, typical of a non-diffusion controlled reaction with 40  19 7  first order kinetics. ' ' These conditions require a temperature of 540°C and a combination of catalyst time on stream and helium/hexane vapour flow rate such that total conversion remained below 15%  40,41  and there is no excessive coking. It is possible that the samples could be further  modified during the temperature ramp and the sample activation period and therefore the samples were not exposed to water at temperatures high enough to cause reactions similar to those occurring due to self steaming during calcination. The sample temperature was raised slowly (from room temperature to 540°C over the course of 2.5 hours) in a flow of dry helium to  173  allow the samples to dehydrate at considerably lower temperature than during the calcination treatments. Catalytic cracking activity results for both series of samples are show in Figure 5.13. Activity is generally higher for the 550-series than for the 650-series and passes through a maximum across both series. When calculated per unit area of dried, crystalline material, this maximum occurs after five calcinations for both the 550- and 650-series. Table 5.5 shows that the product distributions change between the first and second calcinations at both temperatures, but that thereafter they remain constant.  Number of Calcinations  Number of Calcinations  Figure 5.13 High temperature //-hexane catalytic cracking activity of U S Y materials calcined at (a) 5 5 0 ° C (b)  6 5 0 ° C . 'o' dry samples, raw data, '•' values calculated per 1000 m g"' of dried material, '•' values calculated per 1000 m g"' of dried, crystalline material (Table 5.5). 2  2  Some workers ' have suggested that the dominant factor governing increases of factors of 11 42  ca. 10 -10 in catalytic activity in these systems is the rate of diffusion to the catalytic sites and that the activity is enhanced by an increase in the external surface area of the crystallites and that the increases in activity are also accompanied by changes in the product distribution. These  174  studies consider the cracking reactions as being essentially instantaneous at the first reaction sites encountered at, or very close to, the crystallite surfaces. Such a model of the U S Y system does not appear to allow for the possibility of extensive /nfra-crystallite diffusion that is otherwise indicated by size and shape selectivity of zeolites in these types of reactions and nor does it address the effects on product distributions of variations in the concentrations different types of active sites that are inferred, for example, b y A l M A S N M R data. Therefore, the slight 2 7  change in product distributions observed at the start of both of the series of samples studied here does not necessarily indicate that a variability in the diffusion of n-hexane to catalytic sites is producing the observed trends in catalytic activity. Other workers, however, describe similar effects encountered with ultrastable ZSM-5, but 43  report that these effects vanish for small crystallites. ' It is therefore probable that the effects 43 44  of through space diffusion do not account for the trends observed in these data. Since long-range diffusion is held constant by the method of sample preparation (Section 5.2.2) the only diffusion effects that can vary between the samples used in this study are those associated with the microscopic structure of the individual crystallites, effects that have been shown by other workers not to be important. ' This conclusion is supported by the lack of correlation between 43 44  the greatest physical changes in these materials (which are seen for the 650-series) and the greatest changes in catalytic activity (which are seen for the 550-series). To eliminate the variable of surface area (which must also be proportional to external crystallite surface area) from the data, correlations between the percentage of each aluminium species per unit area versus the catalytic activity per unit area were examined (Figure 5.10(b) and Figure 5.11(b)) and the data for the 650-series were corrected for the presence of extracrystallite amorphous material. This latter correction effectively maximizes the proposed surface area of the crystallites (and therefore the external surface of the crystallites). Once corrected, the trends and correlations in these data were not substantially altered and therefore the mesoporerelated diffusion effects can be discounted. The data from this study can immediately disprove any special catalytic properties that have in the past been attributed to A l  P e n t  . In both the 550-series and the 650-series, the percentage of  this species continues to rise beyond the peak in catalytic activity and the overall activity of the 650-series is lower whereas the percentage of A l  P e n t  is higher than the 550-series and this  remains the case even after the corrections discussed above.  175  It has been suggested that catalytic activity is determined primarily by the framework Bransted acidity. Although the acidity of individual sites are reported to increase with framework dealumination, the number of acid sites on ultrastable zeolites accessible to reacting 1  molecules has been shown, from studies on dealuminated and acid washed H-mordenite materials, to be more important.  32  In the present study, the number of framework Bransted acid sites follows the same trend as the percentage of A l  F w k T e t  . For both the 550- and 650-series, the percentage of this species  declines gradually with successive calcinations (Figure 5.10 andFigure 5.11) and does not reflect the trends in catalytic activity (Figure 5.13). Therefore, although the overall decrease in activity with increasing calcination temperature could be attributed to the associated decrease in framework Bransted acidity, it is clear that this is not a dominant factor governing trends in the catalytic activity within each series. This conclusion is also supported by the lack of correlation 90  between trends in  Si:Al (Figure 5.9), an entirely independent measure of framework acidity,  and the catalytic activity. In neither series is there a clear similarity between the trends across the series of the percentages of A l  F w k T e t  or Al  P e n t  and the trend in catalytic activity. The single closest correlation  between catalytic activity and aluminium percentage is to A l  0 c t  .  In both cases the maximum catalytic activity occurs after the maximum in the percentage of Al  0 c t  . However, as discussed above, although the roles of Lewis and Bransted acidities remain a  subject of debate, the association of enhanced acidity of both types with the extra framework material and the positive effect on catalytic activity of this material is widely accepted. The simulations of the  A l M A S N M R spectra indicate a gradual increase in average quadrupolar  coupling across each series (Appendix IV), which reflects increased average distortion of local aluminium environments. A n increase in the activity of Bransted or Lewis acid centres as a consequence of this could temporarily offset the decline in percentage of A l ° , or the offset ct  could be due to the ongoing increase in the acidity of the framework Bransted acid sites across both series.  176  5.5  Conclusions  This work presents the results of a comprehensive characterization, using solid state 2 7  Si and  A l M A S N M R spectroscopy, nitrogen physisorption, powder X-ray diffraction and n-hexane  catalytic cracking activity, of a series of USY samples related to each other by successive ammonium exchange and calcination treatments. The changes associated with the self-steaming, framework dealumination reaction that results from these treatments are common to all hydrothermal dealuminations of U S Y catalysts. Physisorption data show gradual framework decomposition, mesopore formation (evaluated for pore sizes > 15 A ) and the formation of amorphous material associated with dealumination. The changes in the physisorption data are minimal in the 550-series, where there is in addition some evidence of framework healing, but quite substantial in the 650-series. Physisorption data are consistent with the changes in the proportions of the aluminium species present in the samples, as determined by A l M A S N M R spectroscopy. These data 2 7  imply a stepwise evolution of the nature of aluminium removed from the framework, via an intermediate framework aluminium species that appears as part of a broad A l M A S 2 7  NMR  resonance that has previously been assigned as an extra-framework aluminium resonance. Due to the interrelated nature of the changes taking place in these U S Y materials, it is an over simplification to attribute changes in catalytic activity solely to a particular aluminium species. However, there are clear indications of which factors are indeed important in determining the overall catalytic activity of U S Y materials (under the conditions of this study). Surface area, framework acidity and the amount of bulk amorphous material have all been shown to be factors determining catalytic activity. Corrections to the data were made for surface area and the proportion of bulk amorphous material (present in the samples of the 650-series) and although these corrections did make the data for the 650-series more similar to that of the 550-series, indicating that these factors do have some influence on catalytic activity, they did not qualitatively change the trends in either the proportions of the aluminium species or the catalytic activity. The trends in neither framework Si A l , measured by S i M A S NMR, nor the 2 9  percentages of A l  F w k T e t  , measured by A l M A S NMR, reflected the observed changes in 2 7  catalytic activity across each series of samples. Therefore, none of surface area, framework acidity, or the proportion of amorphous material, can account for the observed maximum in n-  177  hexane catalytic cracking activity as a function of successive calcinations, observed across each series of samples. The role of extra-framework aluminium was also examined. These results do not support the generally accepted proposal that pentahedral aluminium exhibits unusually high catalytic activity. Instead, the amount and average quadrupolar coupling parameters of octahedral, extraframework aluminium is shown to most accurately reflect the n-hexane cracking activity of USY.  178  5.6 References for Chapter 5 1.  Corma, A.; Orchilles, A . V . Micropor. Mesopor. Mater. 2000, 35-36, 21-30.  2.  Corma, A.; Fornes, V.; Mocholi, A.; Monton, J.B.; Rey, F. Influence of Superacid Sites in Ultrastable Y Zeolites on Gas Oil Cracking. In Fluid Catalytic Cracking II- Concepts in Catalytic Design; Occelli, M . L . , Ed.; A.C.S. Symposium Series 452: Washington, DC, 1991; pp 12-26.  3.  Parise, J.B.; Corbin, D.R.; Abrams, L. Acta Cryst. 1984, C40, 1493-1497.  4.  S-Aguiar, E.F.; Murta-Valle, M . L . ; Sobrinho, E.V.; Cardoso, D. Zeolites, 1995,15, AllAll.  5.  Wang, Q.L.; Giannetto, G.; Torrealba, M . ; Perot, G.; Kappenstein, C ; Guinset, M . J. Catal. 1991,130, 459-470.  6.  de Menezes, S.M.C.; Camorim, V . L . ; Lam, Y . L . ; San Gil, R.A.S.; Bailly, A.; Amoureux, J.P. Appl. Catal, B. 2001, 207, 361-311.  I.  Kuehne, M . A . ; Babitz, S.M.; Kung, H.H.; Miller, J.T. Appl. Catal, A. 1998,166, 293-299.  8.  Mirodatos, C ; Barthomeuf, D. J. Chem. Soc. Chem. Commun. 1981, 46, 39-40.  9.  Wang, Q.L.; Giannetto, G.; Guiset, G M . J. Catal, 1991,130,471-482.  10. Abbot, J. Appl. Catal, 1989, 47, 33-44. II. Kung, H.H.; Williams, B.A.; Babitz, S.M.; Miller, J.T.; Snurr, R.Q. Catalyis Today, 1999, 52, 91-98. 12. Williams, B.A.; Miller, J.T.; Snurr, R.Q.; Kung, H.H. Micropor. Mesopor. Mater. 2000, 3536, 61-1 A. 13. van Bokhoven, J.A.; Roest, A . L . ; Koningsberger, D.C.; Miller, J.T.; Nachtegaal, G.H.; Kentgens, A . P . M . J. Phys. Chem. B 2000,104, 6743-6754.  179  14. Fyfe, C.A.; Bremerton, J.L.; Lam, L . Y . J. Am. Chem. Soc. 2001, 123, 5285-5291. 15. Chen, T.-H.; Wouters, B.H.; Grobet, P.J. Eur. J. Inorg. Chem. 2000, 281-285. 16. Kokotailo, G.T.; Fyfe, C.A.; Feng, Y . ; Grondey, H . Characterization of Zeolite Cracking Catalysts. In Fluid Catalytic Cracking II- Concepts in Catalytic Design; Occelli, M . L . , Ed.; A.C.S. Symposium Series 452: Washington, DC, 1991; pp27-44. 17. Klinowski, J.; Thomas, J.M.; Fyfe, C.A.; Gobbi, G.C. Nature 1982, 296, 533-536. 18. Sherzer, J. The Preparation and Characterization of Aluminium-Deficient Zeolites. In Catalytic Materials: Relationship Between Structure and Reactivity; Whyte, T.E., Ed.; A.C.S. Symposium Series 248: Washington, DC, 1984; ppl57-200. 19. Klinowski, J.; Fyfe, C.A.; Gobbi, G.C. J. Chem. Soc, Faraday Trans. 71985, 81, 3003-  3019. 20. Freude, D.; Brunner, E.; Pfeifer, FL; Prager, D.; Jerschkewitz, H.-G.; Lohse, U.; Oehlmann, G. Chem. Phys. Lett. 1987,139, 325-330. 21. Samoson, A . ; Lippmaa, E.; Engelhardt, G.; Lohse, U.; Jerschkewitz, H.-G. Chem. Phys. Lett. 1987,134, 589-592. 22. Ray, G.J.; Samoson, A . Zeolites, 1993,13, 410-413. 23. Gilson, J.-P.; Edwards, G . C ; Peters, A.W.; Rajagopalan, K.; Wormsbecher, R.F.; Roberie, T.G.; Shatlock, M.P. J. Chem. Soc, Chem. Commun. 1987, 91-92. 24. Shertukde, P.V.; Hall, W.K.; Dereppe, J.-M.; Marcelin, G. J. Catal. 1993,139, 468-481. 25. Remy, M.J.; Stanica, D.; Poncelet, G.; Feijen, E.J.P.; Grobet, P.J.; Martens, A.J.; Jacobs, P.A. J. Phys. Chem. 1996,100, 12440-12447. 26. Grobet, P.J.; Geerts, H.; Tielen, M . ; Martens, J.A.; Jacobs, P.A. Stud. Surf. Sci. Catal. 46: Elsevier, Amsterdam, 1989; pp721-734.  180  27. Massiot, D.; Fayon, F.; Capron, M . ; King, I.; Le Calve, S.; Alonso, B., Durand, J.-O.; Bujoli, B.; Gan, Z.; Hoatson, G. Magnetic Resonance in Chemistry 2002, 40, 70-76. 28. Rakiewicz, E.F.; Mueller, K.T.; Jarvie, T.P.; Sutovich, K.J.; Roberie, T.G.; Peters, A.W. Microporous Mater. 1996, 7, 81-88. 29. de la Puente, G.; Sedran, U.A. Microporous Mater. 1997,12, 251-260. 30. Storck, S.; Bretinger, H.; Maier, W.F. Appl. Catal. A, 1998,174, 137-146. 31. Beyerlein, B.A.; Choi-Feng, C ; Hall, J.B.; Huggins, B.J.; Ray, G.J. Investigation of Mesopore Formation. In Fluid Catalytic Cracking III: Materials and Processes; Occelli, M . L . ; O'Connor, P. Eds.; A.C.S. Symposium Series 571: Washington, D C , 1994; pp81-97. 32. Lee, K . - H . ; Lee, Y.-W.; Ha, B.-H. J. Catal. 1998, 775, 328-337. 33. da Silva, J.G.; de Menezes, S.C.; Cardoso, M.J.-B. Zeolites, 1994,14, 533-540. 34. Gola, A . ; Rebours, B.; Milazzo, E.; Lynch, J.; Benazzi, E.; Lacombe, S.; Delevoye, L.; Fernandez, C. Micropor. Mesopor. Mater., 2000, 40, 72-83. 35. Corma, A.; Martinez, A.; Martinez, C. Appl. Catal. A, 1996,134, 169-182. 36. Wouters, B.H.; Chen, T.; Grobet, P.J. J. Am. Chem. Soc. 1998, 120, 11419-11425. 37. Wouters, B.H.; Chen. T.; Grobet, P.J. J. Phys. Chem. B 2001, 105, 1135-1139. 38. Stach, H.; Lohse, U . ; Thamm, H.; Schirmer, W. Zeolites, 1986, 6, 74-90. 39. Gore, K . U . ; Abraham, A . ; Hegde, S.G.; Kumar, R; Amoureux, J.P.; Ganapathy, S. J. Phys. Chem. B, 2002,106, 6115-6120. 40. Miale, J.N.; Chen, N . Y . ; Weisz, P.B. J. Catal, 1966, 6, 278-287. 41. Maile, J.N.; Weisz, P.B. J. Catal, 1965, 4, 527-529. 42. Williams, B.A.; Babitz, S.M.; Miller, J.T.; Snurr, R.Q.; Kung, H.H. Appl. Catal. A, 1999, 177, 161-175.  181  43. Babitz, S.M.; Williams, B.A.; Miller, J.T.; Snurr, R.Q.; Haag, W.O.; Kung, H.H. Appl. Catal. A, 1999,779,71-86. 44. Voogd, P.; Van Bekkum, H . Appl Catal, 1990, 59, 311-331. 45. Hopkins, P.D.; Miller, J.T.; Meyers, B.L.; Ray, G.L.; Roginski, R.T.; Keuhne, M . A . ; Kung, H.H. Appl. Catal. A, 1996,136, 29-48.  182  Chapter 6. Detection, Characterization and Quantitation of the Multiple Components in Complex Aluminophosphate Ceramic Mixtures.  6.1  Introduction  The ceramic mixtures formed from the reaction of phosphoric acid with various aluminium oxide substrates are of considerable current interest. " The products contain aluminium 1  3  metaphosphates (A1(P03)3) that are capable of further chemical bonding, giving an alternative to high temperature sintering and fusion, which can usually be carried out at relatively low temperatures. " In addition, these materials show excellent surface adhesion and hardness 4  6  properties, making them attractive candidates for use as coatings of, for example, metal surfaces. '  7 8  Previous characterization of such mixtures has been mainly by X R P D , using databases of the diffraction patterns of known phases. ' " Although the mixtures are often complex, this 5,6 9  11  approach can be successful in identifying the major crystalline phases present. However, quantitation is difficult and any amorphous materials present are not characterized at all and therefore it has not previously been possible to monitor the progress of reactions from amorphous reagents. Solid state M A S N M R spectroscopy is a complementary technique to X R D for the characterization of solid materials " since it can be used to study both crystalline 12  14  and non-crystalline phases. 77  The present study utilizes  11  A l and  P solid state M A S N M R spectroscopy. The measures  necessary to acquire quantitative M A S N M R spectra of spin I - Vz nuclei are well established  183  and in this work, the utility of the  A l M A S and 3QMAS N M R spectra as a quantitative tool for  a wide range of materials can be established by comparison to quantitatively acquired P M A S 3 1  N M R spectra. The purpose of the present investigation is to explore the use of solid state N M R at high field for the detection, characterization and particularly the quantitation of the components of these types of complex aluminophosphate mixtures, derived from three representative alumina sources reacted with phosphoric acid.  6.2  Materials and Methods  6.2.1  MATERIALS  Aluminophosphate ceramic mixtures were prepared by C. Moorlag at the Department of Metals and Materials Engineering, University of British Columbia.  6.2.1(a)  15  Alumina substrates  The a-alumina source was A16-SG calcined alumina from Alcoa Industrial Chemicals, used as received. The mean particle diameter was 0.41 urn and the material was shown by X R P D (not shown) to be mainly crystalline.  6.2.1(b)  Alumina sol-gel  Alumina sol was prepared as described by Yoldas. 1.5 moles (306 g) of aluminium 16  isopropoxide (98%) from Aldrich Chemicals, Al( OCH(CH ) )3, were added to 3 L of distilled 1  3  184  2  water at 75°C. The pH was adjusted to 4 by addition of 1.0 M HN03(aq) and the resulting bluewhite suspension stirred vigorously for 16 hours between 75-85°C until it became a clear sol solution. Excess solvent was slowly evaporated from the sol until a molarity of 1.0 M was reached. Alumina sol-gel was prepared by applying a 1-2 urn thick film to a glass plate and drying in air for two days, or by gel-casting in small moulds. The gel-cast samples were ground up before further study. The dried gel was determined by X R D to be composed primarily of the amorphous bohemite phase.  6.2.1(c)  Composite Sol-Gel (CSG)  CSG was formed by slurrying 100 mL of 1 M alumina-sol with 45 g of a-alumina and sonicating for 5 minutes. The pH of the mixture was adjusted to 4 by addition of approximately 20 mL of 1 M HN03(aq), producing a suspension sufficiently viscous so as not to separate during gel-casting yet not so viscous as to prevent uniform drying. The mole ratio of Al203:Al('OCH(CH3) ) in the mixture was 4.4:1, corresponding to 11% of Al coming from the 2  sol. CSG xerogel was produced from the suspension by gel-casting in 3 cm diameter plastic trays and drying in air for several weeks (resulting in ~ 40% shrinkage in volume), or by drying on glass plates as for the alumina sol-gel preparation (above). Air-dried samples were placed in a drying oven at 85°C for 3 hours before subsequent treatment. X R D and IR analysis showed the material to be composed primarily of crystalline a-alumina and a small amount of beohmite, as observed for the alumina sol-gel.  6.2.1(d)  Aluminophosphate ceramic mixtures  Ceramic mixtures were prepared from each of the alumina sources (Sections 6.2.1(a)6.2.1(c)) by reaction with phosphoric acid.  185  Small amounts of a-alumina (~ 0.3 g) were spread onto glass plates and  H3PO4  was added  to give an A1:P ratio of 1:1. The mixture was fired for 1 hour at 500°C. Ground alumina sol-gel, pre-fired at 500°C was reacted with  H3PO4  in an A1:P ratio of 1:1  (based on the fired sol weight where conversion from a boehmite phase to an A l 2 0 phase has 3  occurred). The mixture reacted to immediately form a plastic-like solid, which was fired at 500°C for 1 hour. Ground CSG, pre-fired to 500°C, was mixed with  H3PO4  in a 1:1 ratio of A1:P and the  mixture fired at 500°C for 1 hour to yield a white powder.  6.2.1(e)  Aluminium metaphosphates  Samples of aluminium metaphosphates-B and -C were prepared by Dr. H . Morell. Metaphosphate-B was prepared by mixing a-alumina and phosphorous pentoxide (P2O5), obtained from Fisher Scientific, in a 1:2 ratio by mass, placed in an open crucible in a furnace and heated from room temperature to 330°C over a 4 hour period, held at this temperature for 8 hours and then allowed to cool slowly. Metaphosphate-C was prepared by mixing phosphoric acid (85%) and aluminium hydroxide hydrate (Fisher Scientific) in a 12:1 molar ratio and heated at 90-95°C for 3 days. After cooling, the solution was placed in a furnace and heated from room temperature to 700°C at a rate of 150°C/h and held at this temperature for 10 hours. The temperature of the melt was then increased to 900°C at a rate of 150°C/h and the sample removed form the furnace and allowed to cool in air.  6.2.2  EXPERIMENTAL APPARATUS  Solid state M A S N M R spectra were obtained at 9.4 T using a Bruker DMX400 spectrometer and at 17.6 T using a Varian Inova spectrometer located at the Pacific Northwest  186  National Laboratory, Richland, W A , U S A (see Acknowledgements). Acquisition parameters are given in the figure captions. Powder X R D experiments were carried out by C. Moorlag using monochromatic C u K a radiation on a Siemens D5000 Diffractometer. Samples were placed on a glass plate in an ethanol slurry and allowed to dry in air prior to analysis.  6.3 Results and Discussion  6.3.1  P O W D E R X - R A Y DIFFRACTION  The three product mixtures formed from reacting 1:1 mixtures of the aluminium sources, alumina sol, a-alumina and CSG, with phosphoric acid were investigated as being typical of ceramic materials formed from reactions of these types. By comparison of X R P D patterns (Figure 6.1) to the ICSD powder diffraction database, these data indicate that the major phases present are metaphosphate-B (all samples), metaphosphate-A (alumina sol-gel), y-alumina (all samples) and corundum ( a - A l 0 3 and CSG). 2  However, small amounts of other phases and any amorphous materials present may not be detected and quantitatively reliable determinations of the relative amounts of the different phases cannot be obtained from the X R D patterns. These materials were investigated by solid state P M A S , A l M A S and A1 3QMAS 3 1  2 7  27  N M R at both 9.4 T and 17.6 T in order to fully describe the different phases present and to determine the appropriate conditions which are necessary for the N M R experiments in order to obtain quantitatively reliable values of their relative proportions in the mixtures.  187  A L ^ sol  a-A^Oj  CSG B\b «  Li  -r—  B  20  10  a  B  B B  bb  B  B a B  -T—  30  50  40  20 (degrees)  Figure 6.1 Powder X-ray diffractograms of alumina sol, a-alumina and composite sol-gel derived ceramic  mixtures. 'B' denotes Berlinite, 'a' metaphosphate-A, 'b' metaphosphate-B and 'a' a-alumina.  6.3.2  31  P M A S N M R SPECTROSCOPY  The isotropic chemical shift regions of the low field and high field P M A S N M R spectra 3 1  are shown in Figure 2, with data collection parameters as indicated in the figure caption and the samples identified by their alumina source. The number of P resonances for each phase must 3 1  be consistent with the number of crystallographic sites (corundum, Berlinite and metaphosphateA have also been studied by single crystal X R D ) . " Various components in the spectra can 17  21  therefore be identified using the approximate relative abundances indicated by the powder  188  diffraction data and the P M A S N M R resonances can be assigned on the basis of their number 3 1  and approximate relative intensities.  B  (b)  11  i  i  i  | i i i i | i i i 11 i i i i | i i i i | i i i i |  -10  -20  -30  -40  11111  iiii| iiii|ii -50  11 | 11 1 1 |  -60  (ppm)  Figure 6.2 Single pulse P M A S N M R spectra of alumina sol, a-alumina and composite sol-gel derived 3 1  ceramic mixtures at (a) 9.4 T and (b) 17.6 T . Spectra at 9.4 T were acquired at spinning rates between 8.6 kHz and 10.2 kHz, using a 15 pulse angle and a 900 s recycle delay. Spectra at 17.6 T were acquired at spinning rates between 12.0 kHz and 12.3 kHz, using an 11° pulse angle and a 30 s recycle delay. 'B' denotes Berlinite, 'a' metaphosphate-A, 'b' metaphosphate-B, Y an amorphous aluminophosphate impurity phase and '*' spinning sidebands.  190  Thus, the resonance at -26 ppm can be assigned to the dense aluminophosphate phase Berlinite, whose structure is analogous to that of quartz and which contains a single phosphorous site. Similarly, the single resonance at -52 ppm can be assigned to the cubic metaphosphate-A phase whose crystal structure was determined by Pauling and Sherman and 21  which also contains a single phosphorous environment, from PO2O2/2 units in a three phosphorous-site ring structure. The remaining three signals in the spectra at -37, -39 and -44 ppm are of approximately equal intensity in the P M A S N M R spectra of all three samples and 3 1  indicate that there are three N M R inequivalent phosphorous sites in this structure, identified from the X P R D data as metaphosphate-B. This material will subsequently be discussed in more detail. There is also evidence of the presence of amorphous aluminophosphate material in the reaction mixture derived from the alumina sol precursor as indicated by the broad resonance at -30 ppm. 31  The  P M A S N M R spectra do not show intensity from any non-phosphorous containing  materials present and therefore cannot be used to quantify the conversion of the various substrates. In addition, care must be taken to ensure that they are quantitatively reliable. Firstly, the different phosphorous sites have different symmetries and the resonances have different chemical shift anisotropics. These are scaled down at 9.4 T and are considerably reduced when fast spinning (13-15 kHz) is used. However, even with fast spinning, care must be taken to include the sideband intensities when calculating the intensities of the different resonances. A second problem is that the T values of the different P resonances are all quite large and there 3 1  {  is considerable variation among them. In this case the T values at 9.4 T vary between 26 s for {  the most downfield of the metaphosphate-B resonances and 174 s for the Berlinite resonance (Table 6.1), necessitating delays of 900 s between 90° pulses (5 x T,) in order to obtain quantitatively reliable data. Both high and low field P M A S N M R spectra (Figure 6.2) were simulated using the 3 1  program DmFit. A n example of a simulation, for the P M A S N M R spectrum of the sol-gel 22  3 1  derived ceramic at 17.6 T, is shown in Figure 6.3. Integration of the components of these simulated spectra yielded the relative percentages of the phosphorous containing components of the ceramic mixtures (Table 6.2). The P M A S N M R spectrum of the sol-gel derived ceramic 3 1  mixture, containing all phosphorous containing species, recorded at 18.8 T with a 40° pulse  191  angle and a 600 s recycle delay (not shown) gave relative peak intensities identical to those in the spectrum recorded at 17.6 T with an 11° pulse angle and a 30 s recycle delay. This suggests that the relative intensities of the resonances in the  P M A S N M R spectra of the ceramic  mixtures recorded at 17.6 T (Figure 6.2(b)) are not distorted due to saturation of the signal, within the error limits of the data. This is confirmed by the excellent agreement with data extracted from P M A S N M R spectra acquired at 9.4 T (Table 6.2) that are known from T 3 1  x  measurements at 9.4 T to be quantitatively reliable (Table 6.1).  T at x  Meta-A  Meta-B  Berlinite  9.4  T  d  (ppm)  iso  Acs  (sec) ± 0.5  ±0.2  27.0  -51.6  -118  ±4  0.50  29.7  -44.1  -111  ±3  0.65  26.6  -38.5  -108  ±2  0.70  25.5  -36.9  -108  ±2  0.65  173.8  -25.7  35  ±3  0.00  (ppm)  ±0.05  Summary of the spectral parameters for P M A S N M R spectra of ceramic mixtures at 9.4 3 1  17.6 T. Simulation parameters were obtained using the program DmFit. Errors were estimated from the distribution of optimized parameter values for spectra of all ceramic mixtures at 9.4 T and 17.6 T. 2  In general, the acquisition of quantitative P M A S N M R spectra of dense ceramic mixtures 3 1  requires a measurement of the T values, some of which may be extremely long. In addition, x  they are limited in that they provide no information at all on the conversions of the alumina substrates, a very important practical aspect of the reactions.  192  Alumina source AI2O3  sol  a-Al 0 2  CSG  3  9.4 T  17.6 T  9.4 T  17.6 T  9.4 T  17.6 T  Metaphosphate-A  22.9  22.0  0.3  0.4  1.7  1.5  Metaphosphate-B  23.4  20.9  10.0  10.2  16.6  16.8  Berlinite  43.2  47.4  87.6  87.2  80.2  81.0  Amorphous Aluminophosphate  10.5  9.7  2.1  2.2  1.5  0.7  Table 6.2 Percentages of the different phosphorous containing species present in the ceramic mixtures, from  integration of t h e P MAS NMR spectra. Relative integral percentages are obtained from the simulations using DmFit of the P MAS NMR spectra at 9.4 T and 17.6 T and corrected for the stoichiometrics of each species. Metaphosphate-A and -B have the composition A1(P0 ) , Berlinite A1P0 and the amorphous aluminophosphate is assumed to have an A1:P ratio of 1:3. Based on simulations of spectra recorded at different spinning rates, errors are estimated to be ± 4% of each value. 3  22  31  3  3  193  4  00  (b)  (c)  (d)  ii  JUL  (e)  /A A  A  A  (f) i  1  60  1  1  40  1  1  20  1  1  0  1  1  -20  r  f  -40  -60  i  i  i  i  i  1  -80 -100 -120 -140 -160  (ppm) Figure 6.3 Single pulse  P M A S N M R spectrum of alumina sol derived ceramic mixture at 9.4 T . Recorded at  a spinning rate of 5.62 kHz, using a 15° pulse angle and a 600 s recycle delay. The spectrum shows the spinning sideband manifolds of Berlinite, metaphosphate-A and metaphosphate-B. (b) Simulation of (a) using DmFit. (c)(e) Deconvolution of (a) separated into the signalsfromBerlinite, metaphosphate-B and metaphosphate-A respectively, (f) amorphous aluminophosphate material. The asymmetry of this resonance was simulated by two overlapping isotropic peaks in the simulated spectrum. 22  194  6.3.3  2 7  A l M A S A N D 3QMAS N M R S P E C T R O S C O P Y  b b  80  60  40  20 0 (ppm)  -20 -40 -60 '  80  60  40  20 0 (ppm)  a  i  -20 -40 -60  Figure 6.4 Single p u l s e A l M A S N M R spectra of alumina sol, a-alumina and composite sol-gel derived ceramic mixtures at (a) 9.4 T and (b) 17.6 T . Spectra at 9.4 T were acquired at a spinning rate of 13.5 kHz, using a 45 pulse angle and a 200 ms recycle delay for a-alumina and composite sol-gel derived ceramic mixtures and at a spinning rate of 11.0 kHz and a recycle delay of 1.0 s for the alumina sol derived ceramic mixture. Spectra at 17.6 T were acquired at spinning rates between 10.2 kHz and 13.4 kHz, using a 12° pulse angle and a 500 ms recycle delay. 'B' denotes Berlinite, 'a' metaphosphate-A, 'b' metaphosphate-B, 'a' and 'y' the corresponding alumina phases, 'x' amorphous aluminophosphate, '*' spinning sidebands and '**' a satellite transition. 27  Figure 6.4 shows the A l M A S spectra at 9.4 T and 17.6 T with the alumina sources identified as previously. The high-field spectra (Figure 6.4(b)) have much sharper and better  195  resolved resonances whose M A S shifts are closer to the isotropic chemical shift values and whose residual quadrupolar lineshapes are greatly reduced, compared to the those of the lowfield spectra (Figure 6.4(a)). Further, it has previously been shown (Chapter 3) that quantitatively reliable data can be obtained for a variety of aluminium environments at sufficiently high field strengths (> 14.1 T). The A l M A S N M R spectra at both 9.4 T and 17.6 2 7  T of the ceramic mixtures were also simulated using DmFit (Figure 6.5) and the parameters 22  from these fits are presented in Table 6.3.  * ^ ( p p » )  <v  c  X  Isotropic 3 Q M A S Shifts (ppm)  %  8„  ±  9.4 T  17.6 T  (kHz)  (Hz)  ±0.2  Calc.  if Expt  Metaphosphate-A  -22.4  0.1  -3.9  -1.1  174  250  0.7  -11.9  0.4  Metaphosphate-B  -15.0 0.0  -2.6  -0.7  163  211  0.0  -8.0  0.2  a  Calc.  ±*  Expt  1.1  0.6  -11.3 -7.2 26.5 9.7  Berlinite  41.9  0.2 -11.6  -3.3  602  15  0.4  24.0  0.6  a-Al 0  14.6  0.1  -2.9  -0.8  250  448  o.r  8.3  0.2  -11.9 -11.0 -8.3 -7.4 24.0 26.7 7.9 9.0  y - A l 0 : 4 co-ord^  69.0  0.4 -17.4  -5.0  724  100  0.0  39.5  1.1  42.6  43.5  2.7  42.0  : 6 co-ord^  10.6  0.7 -14.6  -4.2  659  110  0.0*  21.0  1.0  21.5  22.8  2.4  22.7  Amorphous Aluminophosphate  37.1  0.4  -2.2  325  190  0.3  7.2  1.3  8.4  10.5  2.7  9.3  2  2  3  3  -7.7  e  e  0.7 1.6  Table 6.3 Summary of the spectral simulation parameters for t h e A l M A S N M R spectra of ceramic 2 7  mixtures at 9.4 T and 17.6 T. Parameters were obtained using the program DmFit and used to calculate theoretical isotropic 3Q MAS shifts at both fields. "DmFit does not account for the contribution to the chemical shiftfromthe portion of E attributable to a distribution of quadrupolar couplings. Therefore these values include a correction that is inversely proportional to the magnetic field, calculated from differences in the isotropic chemical shift values used in DmFit. ^Equation 1.44. "These values are taken from simulations of spectra recorded at 17.6 T. d vQ - 3e2qQ/2I(2I -1) (Equation 1.38) Signals are simulated using an average value of vQ, with distributions 22  m  accounted for using an exponential broadening function (E ) with a chemical shift component proportional to the magnetic field strength and a quadrupolar coupling component inversely proportional to the magnetic field. Values are estimated to be accurate to ± 10%. The simulation is relatively insensitive to this variable and values are only assumed to be accurate to ± 0.5/Error values are obtained from estimates of the uncertainties in the simulation parameters for each resonance. ^Asymmetric signals from Y-A1 0 have been simulated using two broadened quadrupolar lines. Parameters used for the calculation of the 3QMAS isotropic (Fl) shifts are averages of the F l shifts of the two lines, weighted according to their integrals in the 17.6 T Al MAS spectral simulations. m  2  3  27  196  (a) 9.4 T  Spectrum Simulation Deconvolution  (b) 17.6 T  Simulation Deconvolution i  1  100 Figure 6.5 Single pulse  1  80  1  1  60  1  1  40  1  1 —  -i  20  1  1  1  -20  r-  -40  -60  (PPm)  A l M A S N M R spectra of alumina sol derived ceramic mixture at (a) 9.4 T and  (b) 17.6 T . Acquisition parameters are as for Figure 6.4. Each spectrum is displayed above its simulation and deconvolution, generated using DmFit and using the lineshape parameters given in Table 6.3. '*' denotes a spinning sideband and '**' a satellite transition.  197  In the 17.6 T A l M A S N M R spectra (Figure 6.4(b)), the aluminophosphate resonances are clearly resolved and can be assigned to Berlinite (41 ppm), metaphosphate-B (-17 ppm) and metaphosphate-A (-24 ppm). In addition, the resonances due to the alumina substrates are clearly observed; corundum  (CC-AI2O3)  at 14 ppm and Y-AI2O3 at 67 ppm (tetrahedral) and 12  ppm (octahedral). Table 6.3 summarizes the lineshape parameters used to simulate the A l 2 7  M A S N M R spectra and includes a comparison between the calculated 3QMAS isotropic shifts and those obtained experimentally at both 9.4 T and 17.6 T. This provides an important internal self consistency check of the simulation parameters.  F2 (ppm) Figure 6.6 Nutation AI 3 Q M A S N M R spectrum of alumina sol-derived ceramic mixture at 17.6 T . Acquired 27  at a spinning rate of 13.4 kHz, using a rf. power of 68 kHz and a recycle delay of 850 ms. Total experiment time was 7 hours. 'B' denotes Berlinite, 'a' metaphosphate-A, 'b' metaphosphate-B, 'x' amorphous aluminophosphate and '*' spinning sidebands. Skyline projections are shown.  198  There is an additional signal from tetrahedral aluminium, assigned to an amorphous aluminophosphate phase (discussed above) which is evident in the 3QMAS spectra of the alumina sol-derived ceramic mixture at both 9.4 T and 17.6 T (Figure 6.6) with a chemical shift range in F2 that is approximately the same as the shift range covered by the second order powder lineshape of the Berlinite resonance. This amorphous signal due to tetrahedral aluminium is therefore not clearly resolved in the A l MAS NMR spectra. However, the 27  lineshape parameters used to simulate this resonance (Table 6.3) are able to accurately simulate the A l MAS NMR spectra at both fields and predict the 3QMAS F l isotropic shifts at both 27  fields within experimental error. The average quadrupolar coupling of the amorphous aluminophosphate resonance is approximately half of the value for the Berlinite resonance, which explains the discrimination against the intensity of the Berlinite resonance with respect to the amorphous aluminophosphate resonance in the 3QMAS spectra.  Alumina source A 1 0 sol 2  9.4 T  (X-AI2O3  3  17.6 T  CSG  9.4 T  17.6 T  9.4 T  17.6 T  Metaphosphate-A  23.2  22.6  ±1  2.6  0.6 ± 0 . 5  7.4  2.2  ±2  Metaphosphate-B  23.6  22.2  ±2  17.7  11.6 ± 4  26.7  15.9  ±4  Berlinite  39.0  43.2  ±6  79.7  87.8 ± 6  65.9  81.9  ±6  Amorphous aluminophosphate  14.2  11.9  ±3  % of alumina source remaining  31.9  29.2  ±4  18.3  ±2  -  31.1  35.4 ± 3  26.7  Table 6.4 Percentages of the different aluminium containing species present in the ceramic mixtures, from  A1 M A S N M R . Single pulse A l MAS NMR spectra acquired at 9.4 T and 17.6 T were simulated using DmFit and relative mole percentages are obtained from the integration of these simulations. The percentages of the product species are summed to 100. The 17.6 T data are considered to be quantitatively reliable. 27  27  22  Table 6.4 presents the estimates of the relative amounts of all of the species present in the three ceramic mixtures, obtained from integration of the simulations of the Al MAS NMR 27  199  spectra at both 9.4 T and 17.6 T. The relative proportions of the phosphorous containing phases, including the amorphous aluminophosphate, are in good agreement with the values obtained from simulations of the P M A S spectra (Table 6.2) providing that an A1:P ratio of 1:3 is 31  proposed, i.e. if the amorphous phase is proposed to be a metaphosphate precursor. If an A1:P ratio of 1:1 is proposed (i.e. if the signal is assigned to be an A I P O 4 precursor) then the quantification of the components of the sol-gel derived ceramic mixture is very different 31  between  27  P and  A l data. Spectroscopy performed using the same parameters for the two other  ceramic mixtures, in addition to experience with other mixed crystalline/amorphous materials (Chapters 3, 4 and 5) strongly suggest that the discrepancy between P and A1 M A S N M R 3 1  27  data for quantifications based upon the A1:P ratio of 1:1 are well beyond the experimental errors associated with these techniques. In general there will be an underestimation of the intensities of A l M A S resonances at low 2 7  magnetic field strengths which increases with increasing quadrupolar coupling. This effect is also observed in present data where, in particular, the percentage of Berlinite is underestimated by approximately 25% in the 9.4 T experiments. Therefore only the 17.6 T A l spectra are 2 7  assumed to be quantitative.  6.3.4  M E T A P H O S P H A T E S - B AND - C  In his synthesis work, d'Yvoire prepared a series of different metaphosphates ' with the 25 26  general composition Al(POs)3 which he named metaphosphates-A (or -a) through -E (or E). Of these, a single crystal X R D structure of metaphosphate-A was determined by Pauling and Sherman  and more recently a single crystal structure of a material described only as 71  'metaphosphate' by van der Meer.  This latter structure consists of 'infinite' (0-P02)« chains,  bridge-bonded by octahedrally co-ordinated A l  3 +  cations. Because these are the only two  metaphosphate structures known, there is a tendency to refer to the "chain structure" of metaphosphate-B, based on the structure found by van der Meer.  200  However, a comparison o f the reference powder pattern o f metaphosphate-B (d'Yvoire ) 25  with that observed experimentally and with the theoretical powder diffraction pattern calculated from the single crystal structure o f van der Meer's "metaphosphate" demonstrates that these two materials are quite different (Figure 6.7).  Metaphosphate-B  _JULJJ  ILL.  Metaphosphate-C  Metaphosphate-C simulation  — I —  1^  20  10  20  —i— 30  i  40  50  Figure 6.7 Powder XRD patterns of metaphosphate-B and metaphosphate-C and a simulated pattern of "metaphosphate". The simulated metaphosphate powder XRD pattern was generated using the single crystal structure of van der Meer, assuming a pseudo Voigt function with 50% Lorentzian character, Cu K a radiation and a step size of 0.02°. 'a' denotes peaks due to a small metaphosphate-A impurity (see text). 21  27  201  The metaphosphate studied by van der Meer can be identified as metaphosphate-C by comparison with an authentic sample synthesized according to the method described by d'Yvoire.  The P M A S N M R spectra of this material, together with its simulation, is shown in  Figure 6.8. The structure is known to be monoclinic and the asymmetric unit contains nine inequivalent phosphorous atoms and three inequivalent aluminium atoms. '  25 26  (a) Deconvolution of isotropic region  g  (b) Simulation  (c) Spectrum  -1  40  0  1  1  1  -40  1  -80  1  1  -120  1  1  -160  (ppm)  Figure 6.8 (a) Deconvolution (of the isotropic region) of, (b) simulation of, and (c) experimental single pulse  P MAS spectrum of metaphosphate-C, recorded at 17.6 T. Acquired using a pulse angle of 30°, a recycle delay of 360 s and at a spinning rate of 10.0 kHz. The spectral simulation was generated using DmFit, using the lineshape parameters given in Table 5. The simulated lines comprising the isotropic region are shown above, labelled according to Table 5. V denotes a small contribution from residual reaction intermediate aluminium triphosphate hydrate and 'a' metaphosphate-A impurity. 3 1  22  15  202  8  /so  A  (ppm)  Relative Intensity  (ppm)  ±0.2  ±0.05  (%)  1  -53.2  -112  ±3  0.50  11.1  ±0.4  2  -52.1  -122  ±5  0.55  9.6  ±0.9  3 4  -50.9  -115  ±6  0.35  8.9  ±0.9  -50.3  -122  ±6  0.45  12.3  ± 1.5  5 6 7 8  -49.7  -102  ±1  0.40  11.1  ± 1.3  -47.9  -102  ±1  0.50  22.8  ± 1.8  -43.0  -112  ±4  0.50  9.8  ±0.1  -40.4  -115  ±2  0.45  9.6  ±0.8  Table 6.5 Summary of the spectral simulation parameters for the P M A S N M R spectra of metaphosphate3 1  C at 9.4 T and 17.6 T . Parameters were obtained using the program DmFit A signal from the metaphosphate-A impurity, totalling 5% of the overall intensity, was included in the simulation and subtracted to give the data in the table. 22  c«o(2)  "Integral  Isotropic 3QMAS Shifts (ppm)  (ppm)  17.6 T  K  b  Metaphosphate-C  (%)  9.4 T  17.6 T  (kHz) (Hz) ± 0 . 2  1  29.4  ± 1  -16.5  -1.3  -0.4  218  220  0.2  2  32.7  ±2  -16.8  -3.1  -0.9  344  35  0.7  3  32.4  ±2  -16.4  -13.8  -3.9  745  12  0.0  Metaphosphate-A  5.5  ± 1  -22.5  -3.9  -1.1  170  60  0.7  Calc.  ±  -8.9  0.1  -8.9  0.2  -7.7  0.4  -12.2  0.1  e  9.4 T  Expt  Calc.  ±  e  -8.7 -8.7 0.2 -8.8 -8.2 0.5 -7.8 -4.6 1.2 -11.9 -12.0 0.1  Expt.  -8.6 -8.0 -4.6 -11.3  Table 6.6 Summary of the spectral simulation parameters f o r A l M A S spectra of metaphosphate-C sample 2 7  at 9.4 T and 17.6 T . Parameters were obtained using the program DmFit and used to calculate theoretical isotropic 3Q MAS shifts at bothfields."Integrals at 9.4 T and 17.6 T were found to be consistent to within the estimated error limits. These values are taken from spectra recorded at 17.6 T. *DmFit does not account for the contribution to the chemical shift from the portion of E attributable to a distribution of quadrupolar couplings. Therefore these values include a correction that is inversely proportional to the magnetic field, calculated from 22  m  differences in the isotropic chemical shift values used in DmFit. v = 3e qQ/2I(2I -1) (Eqn 1.38) Signals are 2  0  Q  simulated using an average value of v , with distributions accounted for using an exponential broadening function (E ) with a chemical shift component proportional to the magnetic field strength and a quadrupolar coupling component inversely proportional to the magnetic field. Values are assumed to be accurate to ±10%. The value given is from the 17.6 T simulation. "Error values are from estimates of the uncertainties in the simulation parameters for each resonance. Q  m  rf  203  The P M A S spectra at 17.6 T (Figure 6.8) and at 9.4 T can both be simulated by eight 3 1  phosphorous signals in the approximate intensity ratios 1:1:1:1:1:2:1:1 and including a metaphosphate-A peak (of ~ 5% of the total intensity) using the P spectral parameters of 3 1  metaphosphate-A determined from the studies of the ceramic mixtures. This is consistent with the presence of nine inequivalent phosphorous environments, two of which are degenerate in 31  their  31  P M A S N M R chemical shifts. B y contrast, the  P spectrum of metaphosphate-B shows  only three inequivalent phosphorous sites. A summary of the P M A S N M R simulation 3 1  parameters and integrals is presented in Table 6.5.  Figure 6.9 Al MAS spectra of metaphosphate-C at (a) 17.6 T and (b) 9.4 T. The spectrum at 9.4 T was recorded using a pulse angle of 30°, a recycle delay of 200 ms and at a spinning rate of 14.0 kHz. The spectrum at 17.6 T was recorded using a pulse angle of 36°, a recycle delay of 500 ms and at a spinning rate of 12.2 kHz. Each spectrum is displayed above its simulation and deconvolution, generated using DmFit using the lineshape parameters given in Table 6. Not shown in the figure are two additional lines in each simulation to account for satellite transition intensity. 27  The  A l M A S spectra can be simulated (Figure 6.9) in terms of three inequivalent sites as  found in the crystal structure (confirmed b y A l 3QMAS at both 9.4 T and 17.6 T, Figure 6.10) 2 7  204  while the  A l M A S spectrum of metaphosphate-B shows only a single aluminium site (again  confirmed by 3QMAS, Figure 6.6). The sample of metaphosphate-C contains a small metaphosphate-A impurity, which can be detected in both the A1 M A S and 3QMAS N M R 27  spectra as well as the powder X R D pattern (Figure 6.7). A complete summary of the A1 M A S 27  and 3QMAS N M R simulation parameters and integrals is presented in Table 6.1. Although is has been possible to show that the structure of metaphosphate-B is different from the structure determined by van der Meer (now identified as metaphosphate-C), the structure is unknown. Using the synthetic methods described in Section 6.2.1(e), it has not been possible to synthesize large enough crystals suitable for a single crystal X R D study and attempts to apply direct methods refinement procedures to a powder diffraction data set have been unsuccessful to date.  6.4 Conclusions This work identifies the components of complex aluminophosphate and alumina containing ceramic mixtures using a combination of X R P D and multiple field solid state N M R spectroscopy. In particular, the viability of solid state N M R is demonstrated as a method of quantifying the compositions of such mixtures. The general method described for deriving N M R lineshape parameters from A1 M A S spectra at both high and low field, the internal 27  confirmation of these parameters with 3QMAS spectroscopy and the comparison between A l 27  and P M A S N M R spectra ensures the reliability of this approach as a means of determining the 3 1  conversion of both amorphous and crystalline substrates to their resulting products. These data have previously been unavailable. In addition, the study has confirmed the validity of using high field A l M A S N M R as a 2 7  quantitative tool for the analysis of such mixed samples, by comparison of the data to those •3 1  from quantitatively reliable P M A S N M R spectra. This will be of particular value for the study of materials for which the relaxation properties and/or sensitivity of P (or other spin-half 3 1  nuclei) make the acquisition of quantitative N M R spectra impracticable for these nuclei.  205  -1—I—I—I—I—I—I—I—I—I—I—I—I—I—I I 1 I—I  -10  I 1 I—I—I—I—I—I—I—I—J—  -20 F2 (ppm)  -30  -16  h- -12 Fl (ppm)  h  -8  \--A  F2 (ppm) Figure 6.10 " A l 3QMAS NMR spectra of metaphosphate-C at (a) 17.6 T and (b) 9.4 T. A' denotes a metaphosphate-A impurity and the labels '1', '2' and '3' relate to the peak numbering scheme used in Table 6.6. (a) Acquired using a 2-pulse rotor synchronized nutation pulse sequence at a spinning rate of 14.1 kHz, using a rf. pulse power of 45 kHz and a recycle delay of 0.1 ms. (b) Acquired using a z-filtered, rotor synchronized nutation phase cycle at a spinning rate of 11.8 kHz, using 53 kHz high power pulses and a soft 4.5 kHz z-filtering pulse.  206  References for Chapter 6  1. Chung, D.D.L. US Patent 5536686, 1996. 2.  Singh, D; Wagh, A.S.; Tlustochowicz, M.S.; Jeong Y . Waste Management 1998,18, 135143.  3. Craig, B.D.; Francis, L.F. J. Am. Ceram. Soc. 1998, 81, 3109-3116. 4. Kingery, W.D. J. Am. Ceram. Soc. 1950, 33, 239-241. 5. Gonzalez, F.J.; Halloran, J.W. Am. Ceram. Soc. Bull. 1980, 59,121-131. 6. Chiou, J.M.; Chung, D.D.L. J. Mat. Sci. 1993, 28, 1435-1446. 7. Rothon, R.N. Thin Solid Films 1981, 77, 149-153. 8. Wilson, S.; Hawthorne, H . M . ; Yang Q.; Troczynski., T. Surface and Coatings Technology 2000,133, 389-396. 9. Bothe Jr., J.V.; Brown, P.W. J. Am. Ceram. Soc. 1993, 76, 2353-2358. 10. Lukasiewicz, S.J.; Reed, J.S.Am. Ceram. Soc. Bull. 1987, 66,1134-1138. 11. Kominami, H.; Matsuo, K.; Kera, Y . J. Am. Ceram. Soc. 1996, 79, 2506-2508. 12. Livage, J.; Babonneau, F.; Chatry, M . ; Coury, L . Ceram. Int. 1997, 23, 13-18. 13. Quartararo, J.; Guelton M . ; Rigole, M ; Amoureux, J.-P.; Fernandez, C; Grimblot, J J. Mater. Chem. 1999, 9, 2637-2646. 14. Iwamoto, R.; Fernandez, C ; Amoureux, J.-P.; Grimbolt, J. J. Phys. Chem. B 1998,102, 4342-4349. 15. Moorlag, C. MSc. Thesis, University of British Columbia, 2000. 16. Yoldas, B.E. Am. Ceram. Soc. Bull. 1975, 54, 289-290.  207  17. Ishizawa, N . ; Miyata, T.; Minato, I.; Marumo, F.; Iwai, S. Acta. Cryst. 1980, B36, 228-230. 18. Schwarzenbach, D. Z. Kristallogr. 1966,123, 161-185. 19. Thong N . ; Schwarzenbach, D. Acta. Cryst. \919,A35, 658-664. 20. Pauling, L.; Sherman, J. Z. Kristallogr. 1937, 96, 481-487. 21. van der Meer, H. Acta. Cryst. 1976, B32, 2423-2426. 22. Massiot, D.; Fayon, F.; Capron, M . ; King, I.; Le Calve, S.; Alonso, B., Durand, J.-O.; Bujoli, B.; Gan, Z.; Hoatson, G. Magnetic Resonance in Chemistry 2002, 40, 1Q-16. 23. Massiot, D.; Touzo, B.; Trumeau, D.; Coutures, J.P.; Virlet, J.; Florian, P.; Grandinetti, P.J. Solid State NMR 1996, 6, 73-83. 24. Massiot, D. Quadrupolar Nuclei, notes from Bruker Symposium, Billercia, MA 1997,118. 25. d'Yviore, F. Bull. Soc. Chim. Fr. 1961, 1762-1776. 26. d'Yviore, F. Bull. Soc. Chim. Fr. 1962, 1237-1243. 27. Baerlocher, Ch.; McCusker, L.B. Database of Zeolite Structures: http://www.izastructure.org/  208  Chapter 7. A1 and P MAS NMR Structural 27  31  Investigation of Calcined, Hydrated AIPO4-I8.  7.1  Introduction  Aluminophosphate molecular sieves to zeolites, but composed of alternating  (AIPO4S)  are a class of synthetic materials analogous  and P 0  AIO4"  + 4  tetrahedra (Section 1.3.2). Although the  reasons are not well understood, the electrostatic neutrality of  AIPO4  frameworks leads to direct  coordination of adsorbed molecules to T-site atoms (most notably water coordinated to 1  aluminium T-atoms, resulting in 5 - and 6-coordinate framework aluminium atoms), in contrast to aluminium containing zeolites. Many A l P 0 s undergo well defined, reversible phase changes 4  at temperatures consistent with the removal of chemisorbed water. X R P D studies of several AIPO4S  (e.g.  AIPO4-H  and  AIPO4-I8)  2  show that this reversible hydration lowers the space  group symmetry. However, it has only been possible to obtain X-ray structure refinements on a small number of hydrated  AIPO4S, ' 3  4  predominantly due to a loss of resolution of the X R P D  patterns following hydration. IR and Raman spectroscopic and T G A measurements of the hydration processes of  AIPO4-5, AIPO4-8  and VPI-5 show that there are both chemi- and  physisorbed water molecules and the latter are likely to be the primary reason for the increase 5  in diffuse scattering of X-rays. As a result, it is often only possible to infer from diffraction studies that the general framework connectivities are unchanged following hydration. Solid state N M R spectra, although modified, indicate that the chemisorbed water molecules often form an ordered array. Solid state P and A l M A S N M R spectra must be consistent with 3 1  2 7  any structure proposed on the basis of X R P D (Section 1.2.2(e)) and it has been possible to determine the space group ' and in rare cases assign the T-sites of 6 7  8  77  of X R P D and N M R spectroscopy.  AIPO4S  using a combination  'X1  Al<-> P coherence transfer N M R spectroscopy potentially  yields considerably more information and it has been possible from such data to assign t h e A l 27  and P N M R resonances to crystallographic T-sites and to therefore learn the coordination of 3 1  AIPO4  T-sites, e.g. A1P0 -14 and A l P 0 - 3 4 . 9  4  3  4  209  Figure 7.1 Topological structure of calcined, dehydrated A I P O 4 - I 8 . Straight lines represent Al-O-P connections. Double 6-ring units are shaded for clarity and the unit cell axis system is shown. The figure was generated using the programme Diamond using crystal structure data taken from reference 18. 10  AIPO4-I8 represents a good candidate for such a study since solid state NMR studies indicate that it exists in an ordered hydrated phase in ambient conditions. This material is of 2  interest as an alkane oxidation catalyst when doped with transition metals. " Catalytic activity 12  15  is thought to derive from the redox properties of the transition metal T-sites (e.g. Co <-»Co ). n  in  12  In addition to their size and shape selectivities, variation of the amount and nature of the substituted T-sites of AIPO4-I8 materials offers far greater control over chemical selectivity than the substituted alumina catalysts that are conventionally used for these redoxtransformations. ' An understanding of the interactions between the frameworks of these 12 14  materials and adsorbed molecules is crucial to the eventual understanding of their activity. For example, there are indications for metal substituted analogues of AIPO4-I8 and other AIPO4 catalysts that the oxygen moieties of adsorbed molecules are preferentially coordinated to the frameworks and that adsorption does not take place at heteroatom T-sites. ' 16  15 17  210  The framework topology of A I P O 4 - I 8 and the structures of as-synthesised, dehydrated AIPO4-I8  and calcined, dehydrated A I P O 4 - I 8 were refined from XRPD patterns by Simmen et  al. (Figure 7.1). The calcined, dehydrated material adopts the monoclinic C2/c space group with unit cell parameters a = 13.711, b = 12.732 and c = 18.571 k,a=y= 90.00° andfi= 90.01°. The structure is composed of alternating layers of double 6-ring units joined by 4-rings related by a mirror plane. Three mutually orthogonal 8-ring channels with a diameter of 3.8 A, aligned along each of the unit cell axes (Figure 7.2), intersect at "pear shaped" cavities . Both 18  the as-synthesised and calcined, dehydrated structures have three inequivalent aluminium Tsites and three inequivalent phosphorous T-sites.  (a) View along c-axis.  a *<r-  (b) View along a-axis  Figure 7.2 Topological structure of calcined, dehydrated A1P0 -18 viewed along (a) the c-axis channel system and (b) the a-axis channel system. V marks the centre of a pear-shaped cavity, (b) The arrow indicates the direction of the final channel system. Figure was generated using the programme Diamond using crystal structure data taken from reference 18. 4  10  211  7.2  Materials and Methods  7.2.1  SYNTHESIS OF AIPO4-I8  A I P O 4 - I 8 was prepared by Dr. Y . Huang according to the method described by the  International Zeolite Association. ' The as-synthesized material crystallised from a gel of 19 20  composition Al O3:P2O :(TEA)2O:60H2O:6^C3H OH) prepared by the addition of 2  5  7  tetraethylammonium (TEA) hydroxide solution to a solution made from the phosphorous source, phosphoric acid, and the aluminium source, aluminium isopropoxide. The T E A organic template was removed from the as-synthesised material by calcination overnight at 550°C. The sample was allowed to hydrate in air to produce the calcined, hydrated AIPO4-I8 material. Solid state N M R experiments were performed on a sample that was later dehydrated at 100°C for 4 hours and rehydrated at room temperature in a saturated vapour pressure of D2O above a saturated solution of calcium chloride in D2O. This controlled hydration with deuterium oxide was designed to maximise P M A S N M R spectral resolution by the removal of any 3 1  dipolar coupling to the protons of adsorbed water molecules. However, P M A S N M R spectra 3 1  of A I P O 4 - I 8 hydrated in water showed the same degree of resolution as samples hydrated in deuterium hydroxide.  7.2.2  2 7  A 1 AND  3 1  P M A S N M R SPECTROSCOPY  Solid state P and A l M A S N M R experiments were performed at 9.4 T (161.9 MHz and 3 1  2 7  104.3 MHz, respectively) on a Bruker DMX400 spectrometer using a Bruker 4 mm triple tuned H X Y M A S probe. A l l experiments were performed at a spinning rate of 12.0 kHz. The single pulse P M A S N M R spectrum was obtained using a 35° pulse angle (2.0 u.s) at a rf. power of 3 I  56 kHz and a 10 s recycle delay. Single p u l s e A l M A S N M R experiments were obtained using 27  a solid pulse angle of 45° (1.0 us) at a rf. power of 42 kHz. The A l z-filtered 3QMAS N M R 2 7  experiment was performed using a rf. power of 50 kHz for the first two pulses and using a soft z-filtering pulse of 45 p,s at a rf. power of 2 kHz. The 2D refocused INEPT experiment was  212  performed using rf. powers of 56 kHz and 42 kHz for the P and A1 channels, respectively, 3 1  27  using a recycle delay of 50 ms. The rotor synchronised first and second r-delays were 2.2 ms and 1.9 ms, respectively. These values were empirically optimised in ID refocused INEPT experiments. Solid state P and A1 M A S N M R experiments were performed at 18.8 T (323.8 MHz and 3 1  27  208.4 MHz, respectively) on a Varian Inova spectrometer (see Acknowledgements) using a home built double channel probe incorporating a supersonic 5 mm Doty stator. The single pulse 31  P experiment was obtained using a pulse angle of 8° (0.5 us) at a rf. power of 42 kHz, a recycle delay of 15 s and at a spinning rate of 12.5 kHz. The single pulse A l M A S 2 7  NMR  experiment was performed using a pulse angle for the solid state of 50° (1.0 us) at a rf. power of 48 kHz, a 100 ms recycle delay and at spinning rate of 12.6 kHz. The 2D refocused INEPT experiment was performed using rf powers of 42 kHz and 48 kHz for the P and A l channels, 3 1  2 7  respectively, using a recycle delay of 50 ms at a spinning rate of 12.5 kHz. The rotor synchronised first and second r-delays were 2.2 ms and 2.6 ms, respectively. These values were empirically optimised using ID refocused INEPT experiments. The nutation A l 3QMAS N M R 2 7  spectrum at 17.6 T was acquired using a home-built single frequency probe incorporating a supersonic 5 mm Doty stator at a rf. pulse power of 20 kHz, using a 100 ms recycle delay and at a spinning rate of 12.4 kHz. At all fields the P N M R spectra were referenced to the P N M R signal of 85% H P 0 3 1  3 1  3  97  assigned a value of 0.0 ppm and  4  97  A l N M R spectra were referenced to the  A l N M R signal of 1  M A1(N03)3 solution assigned a value of 0.0 ppm. The X R P D pattern was acquired on Siemens D5000 diffractometer. Data were collected over the range 20 = 5° and 40° in steps of 0.0077°. Each scan took approximately 4 hours and 4 scans were taken for each sample. The instrument was located at the Institut fur Minerologie, Ruhr Universitat, Bochum, Germany and the data collected by the research group of Professor Hermann Gies.  213  7.3 Results and Discussion  7.3.1  2 7  A l M A S A N D 3QMAS N M R SPECTROSCOPIES  Single pulse A l M A S N M R spectra of calcined, hydrated AIPO4-I8 recorded at both 9.4 T 2 7  and 18.8 T (Figure 7.3) show more than three resonances, indicating a phase change from the dehydrated material. There is also a broad resonance at ca. 12 ppm from an amorphous impurity generated during synthesis.  (a)  (b)  80  60  40  20  0  -20  -40  -60  (ppm) Figure 7.3 Single pulse A I M A S N M R spectra of calcined, hydrated AIPO4-I8 recorded at (a) 9.4 T and (b) 18.8 T. Acquired using a rf. power of (a) 42 kHz (b) 48 kHz, a recycle delay of 0.1 s and at a spinning rate of (a) 12.0 kHz. An impurity phase V and satellite transitions V are shown, (b) 12.6 kHz. An impurity phase V and its spinning sideband '*' are shown. The intensity of the upfield sideband is extremely small. 27  214  Z/  A 1 3QMAS N M R spectra were recorded at 9.4 T and 17.6 T (Figure 7.4 and 7.5) in order  to facilitate an understanding of the single pulse A l M A S N M R spectra. These spectra clearly 2 7  77  show six A l resonances; one broad and two narrow resonances at chemical shifts characteristic of tetrahedral and octahedral aluminium atoms.  1  60  •  1  40  •  1  20  •  1  0  •  1  -20  •  1  -40  F2 (ppm) Figure 7.4 Z-filtered A13QMAS NMR spectrum of calcined, hydrated AIPO4-I8 recorded at 9.4 T. Acquired at a spinning rate of 12.0 kHz using rf. powers of 50 kHz and 2 kHz for the coherence selection and zfilter pulses, respectively. Exponential line broadenings of 75 Hz were applied to both dimensions. Skyline projections are shown. '*' denote spinning sidebands. 27  215  Alljw  •*  -5  A12W  0  1  0 +  \ A13  10-  Pi 15 *  A15  o  30"~i i i i I i i i i I i i i i  60  50  40  i  i i i i  30  i  i i n  20  r~i  10  ii ijii  0  i i | r~  -10 -20  F2 (ppm) Figure 7.5 Nutation  2 7  A 1 3 Q M A S N M R spectrum of calcined, hydrated AIPO4-I8 recorded at 17.6 T .  Acquired at a spinning rate of 12.5 kHz using a rf. power of 20 kHz. Exponential line broadenings of 90 Hz and 50 Hz were applied to the F2 and F l dimensions, respectively. Skyline projections are shown. '*' denote spinning sidebands.  7.3.2  SIMULATION OF SINGLE PULSE  2 7  A l AND P M A S N M R SPECTRA 3 1  The single pulse P M A S N M R spectra of calcined, hydrated A I P O 4 - I 8 are shown in 3 1  Figure 7.6. The spectrum recorded at 18.8 T (Figure 7.6(a)) shows six clearly resolved resonances in addition to a broad resonance from the same amorphous impurity that is evident in the A l M A S N M R spectra (Figure 7.3). The P M A S N M R spectrum recorded at 9.4 T is less 2 7  3 1  o 1  well resolved and shows only five distinct resonances. The single pulse  P M A S N M R spectra  of A I P O 4 - I 8 acquired at both 9.4 T and 18.8 T were simulated using peaks of mixed Gaussian 216  and Lorentzian character (Figure 7.6). At the spinning rates employed, spinning sideband intensities were small and were ignored in these simulations. The 9.4 T P M A S N M R 3 1  spectrum was simulated using the same six peaks as those that were used to simulated the 18.8 T spectrum, only allowing for modifications to peak intensities and minor changes to the peaks' Lorentzian/Gaussian characters. Integrated intensities of the peaks used to simulate both the high and low field spectra (Table 7.1) show six phosphorous environments of equal occupancy.  9.4 T  18.8 T  Average  PI  16.6  15.5  16.1  P2  15.8  17.7  16.8  P3  15.3  13.9  14.6  P4  17.3  16.6  17.0  P5  20.3  18.3  19.3  P6  14.7  18.0  16.4  Table 7.1 Integrals of peaks used to simulate the P MAS NMR resonances, recorded at 9.4 T and 18.8 T , from the crystalline component of calcined, hydrated AIPO4-I8. All values are estimated to be accurate to ±2%. Peaks are numbered according to the resonances labelled in Figure 7.6. 3 1  The single pulse  A l M A S N M R spectra (Figure 7.3) were simulated using the methods  described in Chapters 3-6. The lineshape parameters used to simulate the spectra at both 9.4 T and 18.8 T (Figure 7.7) were chosen such that calculated 3QMAS isotropic F l shifts were in good agreement with the experimental values. A summary of these data is given in Table 7.2. 97  The  A l M A S N M R spectra of calcined, hydrated  AIPO4-I8  have contributions from a larger  number of overlapping resonances than are present in the spectra of other systems that are presented in this thesis. Therefore, although integrated intensities of the peaks used to simulate these data (Figure 7.7) are consistent with the presence of six aluminium T-sites of equal occupancy, errors introduced by overlapping intensity from satellite transitions, the amorphous material and spinning sidebands requires that this conclusion be supported by the  P MAS  N M R spectra. It is for the same reason that any discrimination against resonances from  217  aluminium atoms with greater quadrupolar couplings is not reflected i n the integrations o f the peaks used to simulate the 9.4 T spectrum.  (a) 18.8 T  Spectrum Simulation Deconvolution  (a) 9.4 T  Spectrum  Simulation Deconvolution -5  -10  -15  -20  -25  -30  -35  -40  -45 (ppm)  Figure 7.6 Single pulse P MAS NMR spectra recorded at (a) 18.8 T, (b) 9.4 T. Simulations and deconvolutions are shown below each spectrum. Simulations generated using DmFit. Peak numbering is consistent with Table 7.1. Three peaks were used to simulate the signalfroman amorphous aluminophosphate impurity phase. Spectra were recorded using (a) a pulse angle of 8°and a recycle delay of 15 s at a spinning rate of 12.5 kHz, (b) a pulse angle of 35° and a recycle delay of 10 s at a spinning rate of 12.0 kHz. 3 ,  21  218  The crystal structure of calcined, dried A I P O 4 - I 8 has three inequivalent aluminium and phosphorous T-sites. The 18  3 1  P and A l N M R spectra of the calcined, hydrated material show 2 7  that hydration exactly doubles the number of inequivalent T-sites. This suggests the removal of a single symmetry element from the crystal structure, i.e. that the space group of calcined, hydrated A I P O 4 - I 8 is a subgroup of C2/c. The connectivity patterns of the atoms in any proposed structure of calcined, hydrated AIPO4-I8  must be consistent with the connectivities observed by A 1 - P refocused INEPT 27  31  M A S N M R spectroscopy.  "Integral (± 3%)  So  Isotropic 3QMAS Shifts (ppm)  c eiso(2) (ppm)  VQ  9.4 T  18.8 T  (ppm) 9.4 T  18.8 T  9.4 T  All  12.3  14.5  -10.1  -1.7  -0.4  273  30  A12  10.8  15.8  -8.1  -2.5  -0.6  425  A13  12.1  18.1  -9.9  -17.0  -4.3  A14  13.1  17.5  46.5  -2.5  A15  14.6  17.2  40.0  A16  12.9  17.0  48.1  18.8 T (kHz) (Hz) ± 0 . 2  Calc.  if  0.4  -5.0  0.5  -5.4  -5.4  0.3 -5.3  22  0.4  -3.7  1.4  -3.9  -4.3  0.5 -4.2  800  22  0.2  0.1  1.7  0.4  -4.0  0.6 -3.8  -0.6  330  36  0.4  26.3  0.4  25.9  25.7  0.1 25.8  -2.7  -0.7  284  60  0.9  22.8  0.4  22.1  22.2  0.1 22.0  -12.5  -3.1  651  34  0.3  30.4  1.4  30.4  27.4  0.4 27.8  ExpU  c  Calc.  if  Expt.  c  Table 7.2 Summary of the spectral simulation parameters for A1 MAS spectra of calcined, hydrated AIPO418 at 9.4 T and 18.8 T. Parameters were obtained using the program DmFit and used to calculate theoretical isotropic 3Q MAS shifts at both fields. "Integrals at 9.4 T and 18.8 T were found to be consistent to within the estimated error limits. These values are taken from spectra recorded at 18.8 T. *DmFit does not account for the contribution to the chemical shift from the portion of E attributable to a distribution of quadrupolar couplings. Therefore these values include a correction that is inversely proportional to the magnetic field, calculated from differences in the isotropic chemical shift values used in DmFit. ±0.2 ppm. v - 3e qQ/2I(2I-1) (Eqn 1.38) 27  21  m  c  d  2  Q  Signals are simulated using an average value of v , with distributions accounted for using an exponential broadening function (E ) with a chemical shift component proportional to the magnetic field strength and a quadrupolar coupling component inversely proportional to the magnetic field. Values are assumed to be accurate to ±10%. The value given is from the 18.8 T simulation. •'Error values are from estimates of the uncertainties in the simulation parameters for each resonance. Q  m  219  (a) 18.8 T  Spectrum  Simulation  Deconvolution i  80  60  1  40  1  1  1  1  r  1  1  -20  0  20  1  1  -40  -60  (ppm)  (b) 9.4 T  Spectrum  Simulation  Deconvolution i  1  80  1  ^/3vVS-^ 1  60  1  1  40  1  1  1  1  0  20  1  1  -20  1  1  -40  1  1  1  -60  (ppm) Figure 7.7 Simulations and deconvolutions of single p u l s e A l M A S N M R spectra of AIPO4-I8 recorded at 27  (a) 18.8 T and (b) 9.4 T . Spectra were simulated using the programme DmFit using the methods described in earlier chapters. A summary of the simulation parameters is given in Table 7.2. Peaks used to simulate spinning sidebands and satellite transitions are omitted for clarity. The amorphous impurity resonance is simulated using three peaks of mixed Gaussian/Lorentzian character. 21  220  7.3.3 27  27  A1- P REFOCUSED I N E P T M A S N M R 31  SPECTROSCOPY  A 1 - P refocused INEPT M A S N M R spectra of calcined, hydrated A I P O 4 - I 8 were 31  recorded at both 9.4 T and 18.8 T (Figure 7.8). The two magnetic field strengths provide spectra with complimentary resolution in the indirectly detected Al M A S dimension. At 9.4 T, S:N in 27  the octahedral aluminium region is restricted, but is acceptable in the tetrahedral aluminium region, where resolution is better than at 18.8 T. The S:N of the 18.8 T refocused INEPT spectrum is sufficient to observe most or all of the Al-P connectivities. These data are only semi-quantitative since a single value must be chosen for the r delay of the INEPT experiment (refer to Section 1.1.10) that necessarily transfers magnetisation with efficiencies that vary with the scalar couplings. It is immediately obvious, however, that there are insufficient cross peaks for each of the T-sites to be connected to four inequivalent nearest neighbour T-sites and that some cross peaks are of sufficiently greater intensity than the others along the same F l or F2 chemical shift to propose the presence of multiple connections between T-sites. It is reasonable to suppose, based on numerous studies of the reversible hydration of  AIPO4  molecular sieve materials (refer to Reference 22 and literature cited therein), that the framework connectivities of  AIPO4-I8  are not changed during hydration and that, therefore, the unit cell  parameters and atomic coordinates of the framework atoms of the dehydrated and hydrated materials will be similar. Using the fractional coordinates and the unit cell parameters of the calcined, dehydrated material as a starting point, model structures were generated for all seven of the possible subgroups of C2/c;  C121  23  PT  Clcl  P2/c  P2/n  From the atomic coordinates of the calcined, dehydrated  P2i/n V2Jc  AIPO4-I8  crystal structure  and  using the programme CrysCon, aluminium and phosphorous T-site fractional coordinates were 24  calculated for each of the subgroups. These fractional coordinates and the unit cell parameters of the dehydrated material were used to generate model structures for each of the subgroups using the programme Diamond. Only four of the subgroups produced model structures with the 10  correct three dimensional A I P O 4 - I 8 framework; PT  Clcl  221  P2/c  P2i/n  60  1  11  * * i  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  1  '  1  -5 -10 -15 -20 -25 -30 -35 -40 -45  -5 -10 -15 -20 -25 -30 -35 -40 -45 F2 ( P ppm) 31  Figure 7.8 A 1 - P refocused I N E P T M A S N M R spectra of calcined, hydrated AIPO4-I8, recorded at 27  31  (a) 9.4 T and (b) 18.8 T. (a) Recorded at a spinning rate of 12.0 kHz with rf. powers of 53 kHz and 42 kHz at the P and A1frequencies,respectively and a recycle delay of 50 ms. Exponential line broadenings of 50 Hz and 30 Hz were applied in F2 and F l , respectively. Total experiment time was 72 hours, (b) Recorded at a spinning rate of 12.5 kHz with rf. powers of 42 kHz and 48 kHz at the P and A1frequencies,respectively and a recycle delay of 50 ms. Exponential line broadenings of 30 Hz and 80 Hz were applied in F2 and F l , respectively. Total experiment time was 43 hours. Skyline projections are shown. 31  27  3  222  27  Of these, the structures of the Clcl and P2i/c space groups have connectionsfromeach Tsite to four different nearest neighbour T-sites. Several of the aluminium T-sites in the P1 and P2/c space groups have connections to less than four different neighbouring T-sites. Therefore, the space group of calcined, hydrated A I P O 4 - I 8 must be either P1 or P2/c. (a) P2/c model structure connectivities. All  A12  A13  PI  X  X  X  P2  X  X  X  P3  XX  X  X  X  P4  X  P5 P6  (b)Pl  A14  A15  A16  X X  X  X  X  X  X  X  XX  X  X  A14  A15  A16  lei structure connectivities. All  A12  A13  PI  X  XX  X  P2  X  X  XX  P3  X  X  X  X  P4  X  XX  X  P5  X  X  XX  X  X  X  P6  X  Table 7.3 Nearest neighbour A1<->P T-site connectivities for the AIPO4-I8 framework in the in the (a) P2/c and (b) P 1 space group. 'X' represents a single connection, such that each T-atom has a total of four connections.  To determine which of these two space groups is adopted by calcined, hydrated A I P O 4 - I 8 requires that the known connectivities of the model structure be consistent with the observed connectivitiesfromthe refocused INEPT MAS NMR spectra. However, it is not possible by simple inspection to relate the aluminium and phosphorous resonances (A11-A16, Figure 7.5 and  223  P1-P6, Figure 7.6) to the aluminium and phosphorous T-sites in each of the two model structures (A11-A16 and P1-P6). Therefore, all possible permutations of the assignment of Aln and Pn N M R resonances to Aln and Pn (n = 1-6) crystallographic sites must be considered. Without any constraints, the number of such permutations numbers (6!) , i.e. greater than 2  520,000 possible assignments. Using the Mathematica programme "INADEQUATE.nb" written by D.H. Brouwer, a table 25  of known connectivities for each model structure (Table 7.3) was compared to a table of experimental connectivities obtained from the A 1 - P refocused INEPT M A S N M R spectra 27  31  and a list of possible peak assignments generated. Using INADEQUATE.nb,  it was possible to input probable single connections, definite  single connections, possible double connections, and definite double connections into the experimental connectivity table. A preliminary analysis of the A 1 - P refocused INEPT M A S 27  31  N M R spectra of calcined, hydrated A I P O 4 - I 8 gave the connectivities listed in Table 7.4.  PI  All  A12  Xx  Xx  P2  A13  A14  A15  A16  X  Xx  Xx  X  X  P3  X  Xx  X  Xx  P4  Xx  Y  X  Xx  P5 P6  X  X  X  X  X  X  Xx  Table 7.4 Observed nearest neighbour A1<->P T-atom connectivities from A 1 - P refocused I N E P T M A S 27  31  N M R spectra of calcined, hydrated AIPO4-I8, recorded at 9.4 T and 18.8 T . 'X' represents a definite cross-peak observed in the 18.8 T spectrum, indicating a certain connection. 'Y' represents a definite cross-peak observed in the 9.4 T spectrum that is ambiguous in the 18.8 T spectrum, 'x' represents a possible connection.  The initial connectivity table ('X's, Table 7.4) was composed by examining cross-peak intensity in the A 1 - P refocused INEPT M A S N M R spectrum recorded at 18.8 T at chemical 27  31  shifts characteristic of a single aluminium resonance. A l l connections were found to be consistent with the spectrum acquired at 9.4 T, although several of these (e.g. A15-P2) would be considered ambiguous i f based on the 9.4 T data alone. The converse is also true for the A12-P5 cross-peak ('Y', Table 7.4): Only the cross-peaks for A12 are observed in the octahedral 224  aluminium region of the Al- P refocused INEPT MAS NMR spectrum recorded at 9.4 T. Here, four resonances are clearly seen, including the A12-P4 resonance that would not have been proposed on the basis of the 18.8 T data alone. This connection is most probably obscured in the 77  ^1  Al- P refocused INEPT MAS NMR spectrum recorded at 18.8 T by sine distortions in the Fl dimension. Similar distortions can be seen between the A14-A15/P5-P6 region of the spectrum. Several "possible connections" were added (Ys, Table 7.4) and the connections as listed in Table 7.4 were compared for possible peak assignment to the connectivity tables derived from the two model structures (Table 7.3) using INADEQUATE.nb. Table 7.4 contains complete connectivity information for just one of the 12 T-atoms (P5), yet this pattern of connectivities was found to be inconsistent with any assignment of NMR resonances for the P2/ space group c  and consistent with 128 possible peak assignments for the P1 space group. Therefore, P and 31  A1 MAS, A1 3QMAS and A1- P refocused INEPT MAS NMR spectroscopy performed on  27  27  27  31  calcined, hydrated A I P O 4 - I 8 is consistent only with the space group P1 .  7.3.4  ASSIGNMENT OF A l AND 2 7  3 1  P M A S N M R RESONANCES  The constraint that each T-atom must have four connections allows the remaining connectivities to be deduced (Table 7.5). The only remaining assumption required to complete the connectivity table is the choice between PI and P4 of the double connection to All. All cross-peaks are seen in the 18.8 T A1- P refocused INEPT MAS NMR spectrum only and at 27  31  this field there is overlap between the All and A12 resonances. For this reason, although intensity is only seenfromconnections to PI, P3 and P4, it is not clear that the relative intensity of the three cross-peaks represents the relative number of connections. However, the 9.4 T INEPT spectrum indicates that the three resonances in question have an equal degree of connectivity to A12 and it is therefore probable that the most intense of these resonances represents a double connection to All, i.e. A11-P4. This extra intensity from the narrow All resonance can also explain the presence of the sine distortions that obscure the A12-P4 connection at 18.8 T.  225  PI  All  A12  X  X  P2  A13  A14  A15  A16  Z XZ  Z  X  X  P3  X  X  XZ  P4  xz  Y  X  P5  X  P6  X  X  X  X  X  xz  Table 7.5 Nearest neighbour A1<-»P T-atom connectivities, derived from A 1 - P refocused I N E P T M A S 27  3I  N M R spectra of calcined, hydrated AIPO4-I8, recorded at 9.4 T and 18.8 T. 'X' represents a definite cross-peak observed in the 18.8 T spectrum, indicating a certain connection. 'Y' represents a definite cross-peak observed in the 9.4 T spectrum that is ambiguous in the 18.8 T spectrum. 'Z' represents a connection deduced from the constraint that each T-atom must have 4 connections.  Following this final assumption, schematic A 1 - P refocused INEPT M A S N M R spectra 27  31  (Figure 7.9 and Figure 7.10) based on Table 7.5 are still consistent with the experimental spectra acquired at both fields. Further, these connectivities (Table 7.5) are consistent with eight possible peak assignments (Table 7.6) in the P1 space group. The space group symmetry is such that this is the minimum possible number of peak assignments that can be made by this method. To confirm this, the known connectivities of P1 (Table 7.3(b)) were "assigned" in the same manner as the experimentally derived connectivities and gave the same eight possible assignments.  1 2 3 4 5 6 7 8  All  A12  A13  A14  A15  A16  PI  P2  P3  P4  P5  P6  A12 A12 A13 A13 A15 A15 A16 A16  All All All All AU AU AU AU  A15 A16 A15 A16 A12 A13 A12 A13  AU AU AU AU All All All All  A16 A15 A16 A15 A13 A12 A13 A12  A13 A13 A12 A12 A16 A16 A15 A15  P3 P3 P3 P3 P6 P6 P6 P6  P4 P5 P4 P5 PI P2 PI P2  P2 P2 PI PI P5 P5 P4 P4  PI PI P2 P2 P4 P4 P5 P5  P6 P6 P6 P6 P3 P3 P3 P3  P5 P4 P5 P4 P2 PI P2 PI  Table 7.6 Possible peak assignments of the resonances of A1 and P M A S N M R spectra of calcined, 27  3 1  hydrated AIPO4-I8. Peak labelling Aln and Pn (n = 1-6) is shown in Figures 7.5 and 7.6. Site labelling Aln and Pn (n = 1-6) is according to Table 7.3 and Appendix V.  226  227  228  7.3.5  P O W D E R X - R A Y DIFFRACTION  P I unit cell parameters were extracted from the positions of the first 20 peaks of the high resolution X R P D pattern of calcined, hydrated A I P O 4 - I 8 (Figure 7.11) using the indexing programme Crysfire . Based on these data, the most probably unit cell parameters, of calcined, 26  hydrated A I P O 4 - I 8 in the P1 space group are:  a  15.793 A  a  90.00°  b  11.594 A  B  91.96°  c  18.446 A  y  90.00°  Other unit cell parameters are consistent with these data but are less probable and show far greater deviations of unit cell dimension and/or angles from those of the dehydrated structure of calcined A I P O 4 - I 8 and are therefore far less consistent with the relatively minor changes in unit cell parameters reported to accompany the hydration of A I P O 4 molecular sieve materials (refer to Reference 22 and literature cited therein). Without further structural information with which to differentiate between the properties of the atoms of the model structure, Aln and Pn, no further structural assignments can be made. Empirical relationships have been documented that correlate properties such as tetrahedral 77 7R  distortion, average bond angles and average bond lengths '  and have been used to assign  N M R resonances to structural models. ' However, the structural model derived from fractional 3 8  coordinates for calcined, dried A I P O 4 - I 8 does not necessarily provide representative bond angles or distances, even when generated using the unit cell parameters of the hydrated material. Further, the relative patterns of bond angles and distances using any unit cell parameters reflect only the dried material crystal structure and do not provide a means of further differentiating the eight possible assignments (Table 7.6). X R P D data acquired to date are not of sufficient quality to undergo Rietveld refinement in order to obtain framework atomic coordinates of calcined, hydrated A I P O 4 - I 8 .  229  4  8  12  16  20  24  28  32  36  40  20 (degrees)  Figure 7.11 XRPD pattern of calcined, hydrated A I P O 4 - I 8 . Acquired on Siemens D5000 diffractometer. Data were collected over the range 20 = 5° and 40° in steps of 0.0077°.  7.4 Conclusions 77  Calcined, hydrated AIPO4-18 was studied using A l and  ^1  P M A S N M R spectroscopy at  9.4 T and 18.8 T and the crystallographic space group of this material was determined to be PI . Indexing of the X R P D pattern of this material showed it to have unit cell parameters of a = 15.793 A, b = 11.594 A and c = 18.446 A, a = y=90.00°  and p =91.96°, that are close to  the unit cell parameters of calcined, dehydrated A I P O 4 - I 8 . Single pulse  77  A l M A S spectra at these fields were simulated using reduced second order 97  quadrupolar M A S lineshapes with parameters that gave calculated A l 3QMAS N M R isotropic F l shifts that were consistent with A l 3QMAS N M R spectra acquired at both 9.4 T and 17.6 T. 2 7  Integrations of the peaks comprising these simulations were in good agreement with simulations 230  of single pulse P MAS NMR spectra acquired at 9.4 T and 18.8 T and show that this material contains six NMR inequivalent aluminium and phosphorous T-sites in addition to an amorphous impurity phase. Due to the agreement between the number of observed resonances, multiple field Al and P MAS NMR spectroscopy strongly suggests that the space group of calcined, 27  31  hydrated A I P O 4 - I 8 is a subgroup of C2/c, the space group of the calcined, dehydrated material. Using crystal structure data for calcined, dehydrated A I P O 4 - I 8 and nearest neighbour Al1 8  P T-atom connectivities derived from A1- P refocused INEPT MAS NMR spectra acquired at 27  31  9.4 T and 18.8 T, the crystallographic space group of calcined, hydrated A I P O 4 - I 8 was determined to be PI and the assignment of the twelve Al and P MAS NMR resonances to 27  31  crystallographic sites was reduced to eight possibilities. An analysis of powder XRD data strongly suggests unit cell parameters of a = 15.793 A, b = 11.594 A and c = 18.446 A, a=90.00°, p =91.96° and ^=90.00°. These XRPD data were not, however, suitable for Rietveld refinement and therefore no further structural information was available with which to reduce the number of possible assignments of the NMR resonances.  7.5 Acknowledgement I would like to thank Dr. A.R. Lewis for his assistance with the crystallographic aspects of this project.  231  7.6 References for Chapter 7 1. Kustanovich, I.; Goldfarb, D. J. Chem. Phys., 1991, 95, 8818-23. 2.  Satyanarayana, C.V.; Gupta, R.; Damodaran, K.; Sivasanker, S.; Ganapathy, S. Effects of hydration on A1P0 -14 and A I P O 4 - I 8 structures: P M A S and A1 3QMAS N M R study. In 3 1  27  4  Studies in Surface Science and Catalysis, 135; A C S , 2001; pp2098-2104. 3. Tuel, A.; Caldarelli, S.; Meden, A.; McLusker, L.B.; Baerlocher, C ; Ristic, A . ; Rajic, N . ; Mali, G.; Kaucic, V . J. Phys. Chem. B., 2000,104, 5697-5705. 4. McLusker, L.B.; Baerlochler, Ch.; Jahn, E.; Biilow, M . Zeolites, 1991,11, 308-313. 5. Knops-Gerrits, P.-P.; Toufar, H.; L i , X . - Y . Grobet, P.; Schoonheydt, R.A.; Jacobs, P.A.; Goddard, W.A. J. Phys. Chem. A., 2000,104, 2410-2422. 6. Kennedy, G. J.; Higgins, J.B.; Ridenour, C.F.; L i , H.-X.; Davis, M . E . Solid State Nucl. Magn. Reson., 1995, 4, 173-178. 7. Peeters, M.P.; de Hahn, J.W.; van de Ven, L.J.M.; van Hooff, J.H.C. J. Phys. Chem., 1993, 97, 5363-5369. 8. Khouzamo, R.; Coudurier, G.; Lefebvre, F.; Vedrine, J . C ; Mentzen, B.F. Zeolites, 1990,10, 183-188. 9. Fyfe, C.A., Meyer, H.; Wong-Moon, K . C . ; Grondey, H.; Chezeau, J.M. Solid State Nucl. Magn. Reson., 1997, 9, 97-106. 10. Diamond Visual Crystal Structure Information System version 2.1c, Crystal Impact GmbH Bonn, Germany, 1999. 11. von Ballmoos, R.; Higgins, J.B. Collection of Simulated XRD Powder Patterns for Zeolites2nded.; Butterworth-Heinmann: M A , 1990. 12. Concepcion, P.; Nieto, J.M.L. Catal. Commun., 2001,2, 363-367. 13. Thomas, J.M. Topics in Catalysis, 2001,15, 85-91. 14. Djieugoue, M . A . ; Prakash, A . M . ; Kevan, L. J. Phys. Chem. B., 2000,104, 6452-6461. 15. Chen, J.; Thomas, J.M.; Sankar, G. J. Chem. Soc. Faraday Trans., 1994, 90, 3455-3459. 16. Izamailova, S.G.; Vasiljeva, E.A.; Karetina, I.V.; Feoktistova, N . N . ; Khvoshchev, S.S. J. Colloid. Int. Sci., 1996,179, 374-379.  232  17. Canesson, L.; Arcon, I.; Calderelli, S.; Tuel, A . Micropor. Mesopor. Mater., 1998, 26, 117131. 18. Simmen, A.; McLusker, L.B.; Baerlocher, Ch.; Meier, W . M . Zeolites, 1991,11, 654-661. 19. Robson, H . Ed. Verified Syntheses ofZeolitic Materials - Second Revised Edition; Elsevier: Amsterdam, 2001; pp78-79. 20. Wilson, S.T.; Flanigen, E . M . US Patent 4310440,1982. 21. Massiot, D.; Fayon, F.; Capron, M . ; King, I.; Le Calve, S.; Alonso, B., Durand, J.-O.; Bujoli, B.; Gan, Z.; Hoatson, G. Magnetic Resonance in Chemistry 2002, 40, 70-76. 22. Ch. Baerlocher and L . B . McCusker, Database of Zeolite Structures: http://www.izastructure.org/ 23. International Tables for X-Ray Crystallography - Volume A.; Hahn, T. Ed.; Kynoch Press: Dordrecht, 1993. 24. Dowty, E. Cryscon for Windows - v. 0.91; Shape Software, 1999. 25. Mathematica; Wolfrram Research Inc., 1996. 26. Shirley, R. Crysfire - An Interactive Powder Indexing Support Program - v.9.34g; University of Surrey: Guildford, Surrey, 2000. 27. Miiller, D.; Jahn, E.; Ladwig, G.; Haubenreisser, U . Chem. Phys. Lett., 1984,109, 332-336. 28. Matijasic, A.; Lewis, A.R.; Marichal, C ; Delmotte, L.; Chezeau, J.M.; Patarin, J. Phys. Chem. Chem. Phys., 2000, 2, 2807-2813.  233  Proposals for Further Work  PI.  Ultrastable Zeolites A large number of studies would benefit from the availability of quantitative  Al MAS  N M R spectra and the evaluation and characterisation of the resonances that comprise them. With particular regard to the work presented in this thesis, the presence and nature of broad, tetrahedral framework aluminium species requires further investigation. Preliminary Al-Si 77  70  connectivity experiments have been performed at both 9.4 T and 18.8 T. A 2D A l - Si Cross Polarisation (CP) M A S N M R experiment (Appendix III) or a 2D A l - S i Transferred Echo 2 7  2 9  Double Resonance (TEDOR) M A S N M R experiment has the potential to correlate the A1 and 27  70  Si M A S N M R dimensions with cross peaks originating only from nuclei sharing strong dipolar coupling, i.e. nuclei that are in close proximity to one another, but that are not necessarily chemically bonded. Therefore these experiments have the potential to provide a direct observation of the connectivity between broad, tetrahedral aluminium (species generally regarded to be extra-framework in nature) and framework tetrahedral silicon atoms. However, the inefficiency of spin-locking of quadrupolar nuclei (refer to Section 1.1.10(e)) 7Q  77  and the low natural abundance of Si makes the  70  A l - Si C P M A S N M R experiment an  inherently low sensitivity experiment. This is also the case for the A l - S i TEDOR M A S N M R 2 7  2 9  77  experiment, due to the rapid T relaxation of A l . 2  The  2 7  A l - S i C P M A S N M R experiment has successfully been performed at both high and 2 9  low field on zeolite NaA using home-built probes equipped with a 10 mm supersonic Doty spinning system. Although larger spinning systems are available, the 10 mm system is the largest that is capable of spinning rates high enough (ca. 5 kHz) to move the spinning sidebands clear of the A l  B r T e t  resonance. The 10 mm stator-equipped probes are more efficient than those  used previously for this experiment. The use of a cage coil around the exterior of the stator for the indirectly detected frequency has enabled the complete separation of the channels, allowing the probes to deliver higher rf. powers to the sample than are possible using two channels 234  connected to a single rf. coil. As a result, the S:N achieved in a given time using this new equipment was approximately the same as was previously available using a 14 mm Doty spinning system and a conventional probe design. However, it has not been possible to observe a Al- Si CP MAS NMR signal even in a one 27  29  dimensional experiment from samples of USY. Both the Si and A1 single pulse MAS NMR 29  27  spectra of NaA have a single resonance. Therefore, by comparison to NaA, not only is there approximately 50% of the aluminium present in zeolite-Y materials, but only approximately 50% of this aluminium is likely to be directly connected to silicon. In addition, intensity is divided between a greater number of resonances at both frequencies, each of which is broader than the resonances of NaA. Therefore, approximately 1/16 of the maximum intensity th  available from the experiment with NaA is available for the experiment with USY, representing an increase in experiment time of a factor ~ 250 for a given S:N. Clearly, since approximately 10-15 minutes is required to achieve a S:N = 3 for NaA at 18.8 T, approximately two days would be required for each increment of a 2D CPMAS NMR experiment on USY, which is greatly in excess of the amount of instrument time available. However, sample design could greatly reduce the required experiment time. Zeolite-X is isostructural with zeolite-Y, but can be synthesised with lower framework Si: Al. 2D Al- Si 27  29  CPMAS NMR experiments might therefore be performed on mildly dealuminated "USX" materials within the available instrument time. In addition to further studies of ultrastable faujasite materials, the role of highly charged, extra framework metal ions should be further investigated. The presence and nature of at least some of the broad, framework tetrahedral aluminium that has been proposed in this thesis to be present in USY systems, might be explained by the study of highly crystalline, ion-exchanged zeolite-X and -Y materials without the need for framework dealumination and the broadening of all NMR resonances that results. The acquisition 2D Al- Si CPMAS NMR spectra of such 27  29  materials is, in principle, considerably easier than for USY materials. All such connectivity experiments will be made more practicable with the use of even higher magnetic field strengths. 21.1 T (900 MHz for protons) NMR magnets are now available and even the increase from 18.8 T represents a 20% reduction in experiment time.  235  P2.  Aluminophosphate Molecular Sieves Additional structural information is required for a full assignment of the P and A l M A S 3 1  2 7  N M R resonances of calcined, hydrated AIPO4-I8. This information might be available from a Rietveld refinement of X R P D data that is of higher quality than that the data obtained thus far. However, it is unlikely that the location of water molecules coordinated to framework aluminium atoms could be found using X-ray diffraction. A neutron diffraction study of calcined AIPO4-I8 hydrated in D2O is far more likely to provide the location of these water molecules, however. The neutron scattering cross section of protons is small. Therefore, the difference between the neutron diffraction pattern from calcined A I P O 4 - I 8 hydrated in H2O and from calcined AIPO4-I8 hydrated in D2O will be due only to scattering from deuterium and it should therefore be possible to locate the positions of the water molecules with such data. Even knowing only the fractional coordinates of the framework atoms of the dehydrated material and the space group and unit cell parameters of the calcined, hydrated material, the relationship of the positions of the deuterium atoms of coordinated water D2O molecules to the atoms in the model framework structure is very likely to reveal which of the T-sites of the model structure relate to hydrated framework aluminium atoms, thereby facilitating a complete structural assignment of calcined, hydrated A I P O 4 - I 8 . The experimental techniques presented in Chapters 6 and 7 can be used to solve the structures of other 97  AIPO4  molecular sieve materials. For example, a A l M A S and 3 Q M A S , 2 7  3 1  P  ^1  M A S and A l - P refocused INEPT M A S N M R study has been carried out on calcined, hydrated AIPO4-36. However, it has not yet been possible to deconvolve these spectra, due to a 97  lack of resolution even in A l 3 Q M A S N M R spectra recorded at 18.8 T. This material cocrystallizes with small quantities of A I P O 4 - H 3 , which becomes AIPO4-D upon calcination. The chemical shifts of the P and A 1 M A S N M R resonances of this impurity phase lie close to 3 1  2 7  and/or overlapping with resonances of AIPO4-36, further complicating the deconvolution of the spectra and preventing an unambiguous assigment of the number and relative intensities of NMR-inequivalent T-sites in calcined, hydrated AIPO4-36. As a result, it has also not been 97  ^1  possible to obtain the connectivities of the aluminium and phosphorous T-sites from A l - P refocused INEPT M A S N M R spectroscopy (Figure PI).  236  In principle, the refocused INEPT experiment should distinguish between A1PCV36 and crystalline impurity phases, since cross peaks from A I P O 4 - 3 6 and impurity phases should occur at mutually exclusive chemical shifts in both F2 and Fl. However, the Al MAS NMR 27  dimension is not well enough resolved for this to be possible in the case of calcined, hydrated A1PCV36.  Figure P I . A 1 - P refocused I N E P T M A S N M R spectrum of calcined, hydrated A l P 0 - 3 6 , recorded at 18.8 T at a spinning rate of 12.1 kHz. Acquisition parameters are identical to those for Figure 7.8. Skyline projections are shown. 27  31  4  The correlation of the P MAS dimension and the Al 3QMAS dimension would provide extra resolution. A preliminary Al 3QMAS- P refocused INEPT MAS NMR experiment 27  31  (Appendix HI) was successfully performed on calcined, hydrated A1PCV14. This experiment transfers triple quantum modulated magnetisation to the P spin system via scalar coupling. The 31  237  simple 2-pulse nutation 3QMAS pulse sequence was combined with the refocused INEPT pulse sequence (Appendix III). Even using the simples version of the 3QMAS pulse sequence, extensive 192-step phase cycling was required to combine the pulse sequences. This structure of calcined, dehydrated A1PCV14 is known and there are four inequivalent phosphorous and aluminium T-sites. A1 M A S N M R spectra of the calcined, hydrated material also show the 27  presence of four aluminium environments, with all of 4-, 5- and 6-coordinate aluminium environments represented. Further, the 2D A 1 - P refocused INEPT M A S N M R experiment 27  performed on calcined, hydrated hydrated  AIPO4-I8  AIPO4-I4  31  gave considerably better S:N that either calcined,  or AlP0 -36. Calcined, hydrated 4  AIPO4-I4  was therefore chosen as a test  material for this experiment.  FigureP2. A1- P 3QMAS-refocused I N E P T M A S N M R spectrum of calcined, hydrated A1P0 -14 recorded at 18.8 T . F2 ( P ppm) axis is correlated against F l ( A1 3Q, Hz) axis. 48 hours' total experiment time. 27  31  4  31  27  238  Although the experiment was successfully performed for the first time at 18.8 T, S:N was very low after approximately 48 hours' acquisition time (Figure P2). This experiment does, however have the potential to provide enough extra resolution in the indirectly detected F l  2 7  Al  M A S dimension to facilitate a structural assignment of calcined, hydrated A1PCV36 (and other AIPO4  molecular sieve materials) given sufficient S:N. Additional S:N may be available from a  simple optimisation of the experimental parameters. For example, the r delay chosen for the preliminary experiment was based on the value for the normal refocused INEPT experiment and the optimal value for the 3QMAS-refocused INEPT experiment may be different if the T  2  relaxation time of the 3Q-modulated magnetisation is significantly different from the single quantum magnetisation. This has not yet been investigated. Further, the nutation 3QMAS sequence used in this case may not be the version of the 3QMAS experiment that gives the greatest S:N when combined with the refocused INEPT experiment. The 2-pulse nutation sequence is the simplest of the 3QMAS experiments and was the version chosen as most likely to be successful for this new technique but other variations of the 3QMAS N M R experiment may give better results (refer to Section 1.1.10(e) and associated references). Further optimisation of the A1- P 3QMAS-refocused INEPT M A S N M R experiment and 27  31  attempts to duplicate analogous 3QMAS-TEDOR M A S and 3 Q M A S - C P M A S N M R experiments, each with various 3QMAS sequences, may lead to the solution of the structures many more  AIPO4  materials.  239  Appendices: Table of Contents Appendix I. Laboratory to Rotating Frame Transformation Matrix  A3  Appendix II. Normalised Spherical Harmonic Functions  A4  Appendix III. Pulse Sequences and Macros  A5  III. A  Pulse sequences and macros for Bruker DMX400  III.A(i)  SINGLE PULSE EXPERIMENT: zg.nodec.arl  A5 A5  III.A(ii)  90°-180° ECHO EXPERIMENT: 90.180.echo.jlb  A6  III. A(iii)  1D INEPT EXPERIMENT: e 1 qinept.jlb  A6  III.A(iv)  ID REFOCUSSED INEPT EXPERIMENT: elqldrin.jlb  A7  III.A(v)  2D INEPT EXPERIMENT: e 1 q2din.jlb  A9  III.A(vi)  2D RECOCUSSED INEPT EXPERIMENT: elq2drin.jlb  A10  IILA(vii)  Alternate 2D REFOCUSSED INEPT: elq2drin.tppi.jlb  A12  III.A(viii)  2D N U T A T I O N 3QMAS EXPERIMENT: z3q2pseq.jlb  A14  III.A(ix)  2D RIACTII 3QMAS EXPERIMENT: e3q3pseq.jlb  A15  III.A(x)  Z - Q U A N T U M FILTERED 3QMAS EXPERIMENT: mqzqf.cf  Al7  III.A(xi)  M Q M A S PROCESSING M A C R O : mqxfb  A18  III.A(xii)  TEST P R O G R A M : zg.Al-Si.test.jlb  A37  IILA(xiii)  ID CP EXPERIMENT: cp.quad.jlb  A38  III.B  Pulse sequences for Varian Inova  III.B(i)  A3 9  SINGLE PULSE E X P E R I M E N T (no phase cycling): s2pul.c  A39  IILB(ii)  G E N E R A L SPIN-ECHO EXPERIEMENT: echo.c  A39  III.B(iii)  Tj INV-REC EXPERIMENT: echotl .c  A40  III.B(iv)  ID INEPT EXPERIMENT: elqinept.c  A41  III.B(v)  ID REFOCUSSED INEPT EXPERIMENT: elqrinept.c  A43  III.B(vi)  2D REFOCUSSED INEPT EXPERIMENT: elq2drin.c  A44  IILB(vii)  2D N U T A T I O N 3QMAS EXPERIMENT: 3qmas 1 .c  III.B(vm)  2D RIACTII 3QMAS EXPERIMENT: 3qRIACT2p.c  A50  III.B(ix)  1D/2D CP EXPERIMENT: 2dcpmas.c  A52  Al  A47  III.B(x)  3QMAS-refocused INEPT M A S experiment: e3q2drinc.c  A53  Appendix IV. Simulation Parameters for 17.6 T Single Pulse and Echo A l M A S N M R Spectra 2 7  of U S Y Materials  A58  IV.A  Series of Samples Calcined at 550°C  A58  JV.B  Series of Samples Calcined at 650°C  A61  Appendix V . Fractional Coordinates of the Framework Atoms of Calcined, Hydrated ALPO4-I8. A65  A2  Appendix I. Laboratory to rotating frame transformation matrix Taken from Chapter 1 reference 13.  3sin 0 - 1 + 7/cos 0cos2^  ?7COS(9sin2^  sin6cos0(3-rjcos2<f)  ;7cos«9sin2^  -\-ncos2<j>  TJ sin 9 cos 6 sin 2^  sin6cos#(3-7cos2<f)  nsin6?cos0sin2^  3sin ^ - 1 + ^cos 6cos2<f>  2  2  Where 7 is defined by Equation 1.12.  A3  2  2  Appendix II. Normalised spherical harmonic functions Taken from Chapter 1 reference 14, p337.  m„  ( ] 1  ±1  ±2  2  f l i cos 6 3  f  (3cos 0-l) 2  V \6TT  '15^  cosflsinfle*"  '_15_^ \?>27T j  sin 6e 2  ±2i4  —V  (5 cos 6? - 3 cos e) 3  + —  r(5cos 6?-l)sin6b 2  105 32;r  sin (9cosc% ' 2  ±2 v  ( 35 ^ /r,±i(J sin 3 $2 l,64;rj  ±3  A4  ±,v  Appendix III. Pulse sequences  III.A Pulse sequences for Bruker DMX400  IILA(i)  SINGLE PULSE EXPERIMENT: zg.nodec.arl  ; zg.nodec.arl ; Avance-version ; ID one-pulse sequence ;One pulse experiment ;no decoupling!  #include <Avance.incl>  1 ze 2dl p i phi d4 go=2 ph31 wr#0 exit phl=0 2 3 1 ph31=0 2 3 1  ;pll : f l channel - power level for pulse (default) ;pl : f l channel - high power pulse ;dl : relaxation delay;  A5  IILA(ii)  90°-180° ECHO EXPERIMENT: 90.180.echo.jlb  ; Phase cycle copied from a programme by Massiot. lOu pllrfl 1 ze 2dl 3 3u:fl phi (pi phl):fl d6 (p2 ph2):fl d7 go=2 ph31 wr#0 exit phl= 0 12 3 ph2= 0 0 0 0 1 1 1 1 2 2 2 2 3 3 3 3 ph31=03 2 1 2 1 03 03 2 1 2 1 03  III.A(iii)  ID INEPT EXPERIMENT: elqineptjlb  ;elqinept.pc ;1D sequence ;For 27-A1/31-P experiments #include <Avance.incl> #include <Solids.incl>  1 ze 2 dl 10upll:fl 10upl2:f2 3 (p21 ph3):f2 dlO 4 (p22 ph4):f2  ; 90 degree 27-A1 pulse ; Taul : Scalar coupling evolution period ; 180 deg 27-A1 pulse  A6  (p2 ph6):fl dlO 5 (p21 ph5):f2 (pi ph7):fl go=2 ph31 6wr#0 exit  ; 180deg31-Ppulse ; Taul ; 90 deg 27-A1 transfer pulse ; 90 deg 31-P transfer pulse  ph3= {0}*4 {3}*4 {2}*4 {1}*4 ph4= (0}*4 {3}*4 {2}*4 {1}*4 ph5= 1 1 3 3 0 0 2 2 3 3 1 1 2 2 0 0 ph6= {0}*4 {3}*4 {2}*4 {1}*4 ph7= {0}*4 {3}*4 {2}*4 {1}*4 ph31=00223 3 1 1 2 2 0 0 1 1 3 3  p l l : f l channel - power level for pulse (default) ;p21 : Y nucleus - 90 degree pulse ;dl : relaxation delay; 1-5 * TI ;dl0 : Taul - scalar coupling evolution time ;p22 - Y nucleus - 180 degree pulse p i - X nucleus - 90 degree tansfer pulse ;p2 - X nucleus -180 degree pulse  IILA(iv)  ID REFOCUSSED INEPT EXPERIMENT: elqldrin.jlb  ;elqldrin.pc ;1D refocussed INEPT sequence ;phase cycling from 2D programme. #include <Avance.incl> #include <Solids.incl>  "dl2=dl 1-0.000020" 1 ze 2 dl lOu p l l : f l  A7  lOu pl2:f2 3 (p21 phl):f2 dlO (p22 ph3):f2 2u (p2 ph4):fl dlO 4(p21 ph5):f2 2u (pi ph6):fl dll (p22 ph7):f2 2u (p2 ph8):fl dl2 go=2 ph31  ; 90 degree 27-A1 preparation pulse Taul : Scalar coupling evolution period ; 180 deg 27-A1 pulse ; 180 deg 31-P pulse Taul ; 90 deg 27-A1 transfer pulse ; 90 deg 31-P transfer pulse Refocussing delay - tau2 ; 180 degree refocussing pulse  Acquire at echo top  10mwr#0 exit phl= {0}*16 {3}*16 {2}*16 {1}*16 ph3= {0}*8 {2}*8 {3}*8 {1}*8 {2}*8 {0}*8 {1}*8 {3}*8 ph4= {0}*16 {3}*16 {2}*16 {1}*16 ph5=1313131313131313020202020202020231313131313131312 020202020202020 ph6= {1}*16 {0}*16 {3}*16 {2}*16 ph7= {1}*16 {0}*16 {3}*16 {2}*16 ph8=1133113311331133002200220022002233113311331133112 200220022002200 ph31=l 3 1 3 1 3 1 3 1 3 1 3 1 3 1 3 0 2 0 2 0 2 0 2 0 2 0 2 0 2 0 2 3 1 3 1 3 1 3 1 3 1 3 1 3 1 3 1 2020202020202020  ;pll : X nucleus - power level ;pl : X nucleus - 90 deg. observe pulse ;p2 : X nucleus -180 deg. observe pulse ;p21 : Y nucleus - 90 degree pulse ;p22 : Y nucleus - 189 degree pulse ;d0 : first delay 3u incremented by IN0 ;dl0 : taul ;dl 1 : tau2 ;dl : relaxation delay ;ns: multiple of 64  A8  III.A(v)  2D INEPT EXPERIMENT: elq2din.jlb  ;elq2din.pc ;2D INEPT sequence #include <Avance.incl> #include <Solids.incl>  define loopcounter nsc "nsc=tdl/2" 1 ze 2 dl lOu p l l : f l 10upl2:f2 3 (p21 phl):f2 ; 90 degree preparation pulse dO (p21 ph3):f2 ; 90 degree 27-Al pulse: select component dlO ; Taul : Scalar coupling evolution period (p22 ph4):f2 ; 180 deg 27-Al pulse (p2 ph6):fl ; 180 deg 31-P pulse dlO ; Taul 4 (p21 ph5):f2 ; 90 deg 27-A1 transfer pulse (pi ph7):fl ; 90 deg 31-P transfer pulse go=2ph31 10mwr#0 if #0 i p l zd 5 ze 6 dl 10upll:fl 10upl2:f2 7 (p21 phl):f2 dO (p21 ph3):f2 dlO (p22 ph4):f2 (p2 ph6):fl dlO 8 (p21 ph5):f2  A9  (pi ph7):fl go=6 ph31 10mwr#0 if #0 idOdpl ze lo to 2 times nsc exit phl= {3}*4 {2}*4 {1}*4 {0}*4 ph3= {0}*4 {3}*4 {2}*4 {1}*4 ph4= {0}*4 {3}*4 {2}*4 {1}*4 ph5= 1 1 3 3 0 0 2 2 3 3 1 1 2 2 0 0 ph6= {0}*4 {3}*4 {2}*4 {1}*4 ph7= {0}*4 {3}*4 {2}*4 {1}*4 ph31=0 0 2 2 3 3 1 1 2 2 0 0 1 1 3 3  ;pll : f l channel - power level ;pl : f l channel - 90 deg. observe pulse ;p2 : f l channel - 180 deg. observe pulse ;d0 : first delay 3u incremented by IN0 dlO : taul TN0 : incrementable delay, l/swh{Fl} ;dl : relaxation delay ns: multiple of 16  IILA(vi)  2D RECOCUSSED INEPT EXPERIMENT: elq2drin.jlb  ;elq2drin.pc ;2D INEPT sequence ;Process using States #include <Avance.incl> #include <Solids.incl>  define loopcounter nsc "nsc=tdl/2" "dl2=dl 1-0.000020"  AlO  1 ze 2 dl 10upll:fl lOu pl2:f2 3 (p21 phl):f2 dO (p21 ph2):f2 dlO (p22 ph3):f2 2u (p2 ph4):fl dlO 4(p21ph5):f2 2u (pi ph6):fl dll (p22 ph7):f2 2u (p2 ph8):fl dl2 go=2ph31  ; 90 degree preparation pulse ; 90 degree 27-Al pulse: select component ; Taul : Scalar coupling evolution period ; 180 deg 27-Al pulse ; 180 deg 31-P pulse ; Taul ; 90 deg 27-Al transfer pulse ; 90 deg 31-P transfer pulse ; Refocussing delay - tau2 ; 180 degree refocussing pulse  ; Acquire at echo top  10mwr#0 if #0 i p l ze 5 ze 6 dl 10upll:fl 10upl2:f2 7 (p21 phl):£2 dO (p21 ph2):f2 dlO (p22 ph3):f2 2u (p2 ph4):fl dlO 8 (p21 ph5):f2 2u (pi ph6):fl dll (p22 ph7):f2 2u (p2 ph8):fl  All  dl2 go=6 ph31 10mwr#0 i f #0 idOdpl ze lo to 2 times nsc exit phl= {0}*16 {3}*16 {2}*16 {1}*16 ph2= {1}*4 {3}*4 {1}*4 {3}*4 {0}*4 {2}*4 {0}*4 {2}*4 {3}*4 {1}*4 {3}*4 {1}*4 {2}*4 {0}*4 {2}*4 {0}*4 ph3= {0}*8 {2}*8 {3}*8 {1}*8 {2}*8 {0}*8 {1}*8 {3}*8 ph4= {0}*16 {3}*16 {2}*16 {1}*16 ph5=1313131313131313020202020202020231313131313131312 020 2 0 2 0 2 0 2 0 2 0 2 0 ph6= {1}*16 {0}*16 {3}*16 {2}*16 ph7= {1}*16 {0}*16 {3}*16 {2}*16 ph8=1133113311331133002200220022002233113311331133112 200220022002200 ph31=l 31313131313131302020202020202023131313131313131 2020202020202020  ;pll : f l channel - power level ;pl : f l channel - 90 deg. observe pulse ;p2 : f l channel - 180 deg. observe pulse ;p21 : f2 channel - 90 deg. ;p22 : f l channel -180 deg. ;d0 : first delay 3u incremented by PN0 ;dl0 : taul ;dl 1 : tau2 ;PN0 : incrementable delay, l/swh{Fl} ;dl : relaxation delay ;ns: multiple of 64  III.A(vii)  Alternate 2D REFOCUSSED INEPT: elq2drin.tppi.jlb  ;elq2drin.pc ;2D INEPT sequence ;Process using TPPI  A12  ;Set nd0=2 and set twice desired SW{F1} #include <Avance.incl> #include <Solids.incl>  define loopcounter nsc "nsc=tdl" "dl2=dl 1-0.000020 1 ze 2 dl lOu p l l : f l lOu pl2:f2 3 (p21 phl):f2 dO (p21 ph2):f2 dlO (p22 ph3):f2 (p2 ph4):fl dlO 4 (p21 ph5):f2 (pi ph6):fl dll (p22 ph7):f2 (p2 ph8):fl dl2 go=2 ph31  ; 90 degree preparation pulse ; 90 degree 27-Al pulse:select component ; Taul : Scalar coupling evolution period ; 180 deg 27-A1 pulse ; 180 deg 31-P pulse ; Taul ; 90 deg 27-A1 transfer pulse ; 90 deg 31-P transfer pulse ; Refocussing delay - tau2 ; 180 degree refocussing pulses ; Acquire at echo top  10mwr#0 if #0 idO i p l ze lo to 2 times nsc exit phl= {0}*16 {3}*16 {2}*16 {1}* 16 ph2= {1}*4 {3}*4 {1}*4 {3}*4 {0}*4 {2}*4 {0}*4 {2}*4 {3}*4 {1}*4 {3}*4 {1}*4 {2}*4 {0}*4 {2}*4 {0}*4 ph3= {0}*8 {2}*8 {3}*8 {1}*8 {2}*8 {0}*8 {1}*8 {3}*8 ph4= {0}*16 {3}*16 {2}*16 {1}*16 ph5=1313131313131313020202020202020231313131313131312 020202020202020 ph6= {1}*16 {0}*16 {3}*16 {2}*16  A13  ph7= {1}*16 {0}*16 {3}*16 {2}*16 ph8=1133113311331133002200220022002233113311331133112 200220022002200 ph31=l 3 1 3 1 3 1 3 1 3 1 3 1 3 1 3 0 2 0 2 0 2 0 2 0 2 0 2 0 2 0 2 3 1 3 1 3 1 3 1 3 1 3 1 3 1 3 1 2020202020202020  ;pll : f l channel - power level ;pl : f 1 channel - 90 deg. observe pulse ;p2 : f l channel - 180 deg. observe pulse ;p21 : f2 channel - 90 deg. ;p22 : f l channel - 180 deg. ;d0 : first delay 3u incremented by INO ;dl0 : taul ;dl 1 : tau2 ;IN0 : incrementable delay, l/swh{Fl} ;dl : relaxation delay ;ns: multiple of 64  III.A(viii)  2D NUTATION 3QMAS EXPERIMENT: z3q2pseq.jlb  ;z3q2pseq.pc ;2D sequence ;phase cycling from MSL400 programme of same name. ;F1 axis may require reversal #include <Avance.incl> #include <Solids.incl>  define loopcounter nsc "nsc=tdl/2" 1 ze lOu p l l : f l 2 dl 3 (pi phl):fl dO (p2 ph2):fl  A14  go=2 ph31 10mwr#0 if #0 i p l zd 4 dl 5 (pi phl):fl dO (p2 ph2):fl go=4 ph31 10mwr#0 if #0 idOdpl ze lo to 2 times nsc exit phl= (12)00 1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9 10 10 11 11 ph2= 0 2 13 ph31=0 0 3 3 2 2 1 1  p l l : f l channel - power level p i : f l channel - 1st high power pulse ;p2 : f l channel - 2nd high power pulse ;d0 : first delay 3u incremented by JNO ;IN0 : increment (generally 1/vRot) ;dl : relaxation delay; 1-5 * TI ;ns: multiple of 24  IILA(ix)  2D RIACTII 3QMAS EXPERIMENT: e3q3pseq.jlb  ;e3q3pseq.pc ;2D sequence ;phase cycling from 2 pulse programe - first pulse ;90degrees behind first adiabatic pulse. ;F1 axis may require reversal #include <Avance.incl> #include <Solids.incl>  A15  define loopcounter nsc "nsc=tdl/2" 1 ze 10upll:fi 2 dl 3 (pi phl):fl 4 (p2 ph2):fl dO (p3 ph3):fl go=2ph31 10mwr#0 if #0 i p l zd 5 dl 6 (pi phl):fl 7 (p2 ph2):fl dO (p3 ph3):fl go=5 ph31 10mwr#0 if #0 idOdpl ze lo to 2 times nsc exit phl= (12) 9 9 10 10 11 11 0 0 1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 ph2=(12)0 0 1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9 10 10 11 11 ph3= 0 2 13 ph31=003 3 2 2 1 1  ;pll : f l channel - power level ;pl : 90 degree solid pulse ;p2 : First Adiabatic pulse (l/4)vRot ;p3 : Second adiabatic pulse ;d0 : first delay 3u incremented by IN0 ;IN0 : increment (generally 1/vRot) ;dl : relaxation delay; 1-5 * T l ;ns: multiple of 24  A16  IILA(x)  Z-QUANTUM FILTERED 3QMAS EXPERIMENT: mqzqf.cf  MQZQF 2D sequence cf-000517 #include <Avance.incl>  define loopcounter nsc "nsc=tdl/2" 1 ze 2 dl lOu p l l : f l 3 (pi phl):fl dO (p2 ph2):fl d4 pl2:fl (p3 ph3):fl go=2 ph31 10mwr#0 if #0 i p l zd 4  dl lOu pllrfl 5 (pi phl):fl dO (p2 ph2):fl d4 pl2:fl (p3 ph3):fl go=4 ph31 10mwr#0 if #0 idO dpi ze lo to 2 times nsc  exit phl=(12){0}  A17  ph2= (6) {0000 1111 2222 3333 4444 5555} ph3= 0 2 13 ph31=0 2 1 3 203 1 ;pll : fl channel - power level for pulse pi and p2 ;pl2 : fl channel - power level for the soft pulse p3 ;pl : fl channel - 1st high power pulse ;p2 : fl channel - 2nd high power pulse ;p3 : fl channel - soft pulse Zfilter dO : first delay 3u incremented by IN0 ;IN0 : increment (generally 1/vRot) ;d4 : duration of the z-filter ;dl : relaxation delay; 1-5 * Tl ;ns: multiple of 24  IILA(xi)  MQMAS PROCESSING MACRO: mqxfb  /*  xfshear  08.11.1996  */  /* /* /* /*  Short Description: */ Program for shearing of 2D MQMAS spectra of odd half */ integer quadrupolar nuclei. Data need to be aquired in */ States Mode */  /* /*  Keywords: */ shear,half integer spin, multi quantum, MQ, headache */  /* /* /* /* /* /* /* /* /* /* /*  Description/Usage: */ Program is used for 2D MQ experiments on nuclei with */ odd half integer spin for shearing-FT of 2D spectrum */ including 2DFT and referencing in Fl dimension */ according to delta(MQ)=delta(iso)+(p-R)delta(qis) */ It checks for the parameter "nucleus" (AMX) or "NUC1" */ (Avance) and sets the spin value accordingly. */ If the nucleus is not in the list it asks for the spin, */ too. Only most common nuclei are in the list. */ If the pulse program name matches mpxq, where */ x corresponds to the multiple quantum order of */  A18  /* /* /* /* /* /* /* /* /* /* /* /* /* /* /*  the experiment, then this is assumed, otherwise it is asked for by the program. Program asks i f abs2 in F2 dimension is desired. If answered with yes, then a Hilbert transform is automatically done, which requires xwinnmrl .3 */ or later. In older versions, this command has to */ be eliminated from the program code. A F l frequency shift in ppm is asked for to compensate for malchosen o l setting. This value is stored and */ suggested for any repeated processing. */ JLB: This will totally ruin any prospect of correctly */ referencing the F l axis. Use only in emergencies, e.g. to learn where to put the offset the next time. This */ is only a big consideration in the confined of rotor- */ synchronous acquisition. */  */ */ */ */  */ */  */  /****************************************************************^ /* /* /* /*  Author(s): Name : Christian Fernandez Organisation : Universite de Lille, CNRS-801 Email : christian.fernandez@univ.lille.fr  /*  */ */ */ */  */  /* Name : Stefan Steuernagel */ /* Organisation : Keystone Cops */ /* Email : stefan.steuernagel@bruker.de */ /****************************************************************/ /* /* /* /* /* /* /* /* /* /* /* /* /* /* /* /* /* /*  Name cf ste ste  ste  ste  jib  Date Modification: */ 960201 created 980813 check xdim's and status sizes at right time */ 980813 XHT2 included after XF1 to allow £2 phase correction after */ processing */ 981101 close of input data files and */ "mv temp-file to 2rr" etc */ included */ 000105 XIF2 and X F B included after shearing to allow apodisation */ in F2, set-up to be done prior */ to execution of xfshear, all */ handling done by this program 001110 Referencing in F l modified. */ Literature and method described */ at line 357. */  A19  */ */ */  */  */  /* $Id: xfshear,v 1.3.2.2 1999/03/01 10:55:04 gsc Exp $ */ #include <math.h> int parmode, si2, s i l , outexpno, tempfile2rr, infile2rr, tempfile2ir, infile2ir, in2rr[4096], in2ir[4096], temp2rr[4096], temp2ir[4096], temp2rr2[4096], temp2ir2[4096], temp2rrB[4096], temp2irB[4096], temp2rrB2[4096], temp2irB2[4096], nbytes, sizeofmt, nbytesread, xdim2, xdiml, loopcount4, sig, mq, nspin, wdw2, bcmod, phmod; double s w l , sw2, ratio, ph, p h i , ph2, phcl, phc2, inO, sfol, sfol2, sfol 1, bfl 1, bfl2; float Noise, spin, offset 1, offset2, off3, off2, offl, sr2, srl; char inname2rr [P A T H L E N G T H ] , tempname2rr[PATH_LENGTH], inname2ir[P A T H L E N G T H ] , tempname2ir[P A T H L E N G T H ] , outname [P A T H L E N G T H ] , nucl[80], yes[20], pulprog[80], ti[80]; /* select current dataset  ====================== */ GETCURDATA; (void) strcpy (yes, "no"); /* is-it a 2D spectrum ?  =============== *  ;  FETCHPAR("PARMODE",&parmode); if (parmode != 1) { STOPMSG(" Not a 2D spectrum!");} /* init some variables  =============== */ FETCHPARS("PULPROG",pulprog); FETCHPARS("SW",&sw2); FETCHPARS("SF01",&sfol); FETCHPAR1 S("IN 0",&in0); FETCHPAR("SI",&si2); FETCHPAR("WDW",&wdw2) STOREPAR("WDW",0)  A20  /* shearing ratio calculation */ /  *  =  =  =  =  =  =  =  =  =  =  =  =  =  =  =  =  =  =  =  =  =  =  ^  =  =  *  /  spin=5./2.; mq=3; FETCHP AR("NUC 1 ,nucl) if(!strcmp(nucl,"off *)) F E T C H P A R ("NUCLEUS",nucl) if(!strcmp(nucl,"7Li")) spin=l.5; else if(!strcmp(nucl,"l IB")) spinal.5; else if(!strcmp(nucl,"170")) spin=2.5; else if(!strcmp(nucl,"23Na")) spin=1.5; else if(!strcmp(nucl,"27Al")) spin=2.5; else if(!strcmp(nucl,"35Cl")) spin=1.5; else if(!strcmp(nucl,"39K")) spin=1.5; else if(!strcmp(nucl,"45Sc")) spin=3.5; else if(!strcmp(nucl,"45Sc")) spin=3.5; else if(!strcmp(nucl,"51V")) spin=3.5; else if(!strcmp(nucl,"55Mn")) spin=2.5; else if(!strcmp(nucl,"63Cu")) spin=1.5; else if(!strcmp(nucl,"65Cu")) spin=1.5; else if(!strcmp(nucl,"69Ga")) spin=1.5; else if(!strcmp(nucl,"71Ga")) spin=1.5; else if(!strcmp(nucl,"85Rb")) spin=2.5; else if(!strcmp(nucl,"87Rb")) spin=1.5; else if(!strcmp(nucl,"93Nb")) spin=4.5; else if(!strcmp(nucl,"121Sb")) spin=2.5; else GETFLOAT(" Enter spin number : ",spin); nspin = (int) ceil ( (double)(2*spin)); if (nspin == 3 ) {mq = -3; ratio = 7./9.;} else { i f (!strncmp(pulprog,"mp3q",4)) mq=3; else if(!strncmp(pulprog,"mp5q",4)) mq=5; else if(!strncmp(pulprog,"mp7q",4)) mq=7; else if(!strncmp(pulprog,"mp9q",4)) mq=9; else GETINT( " Enter the pQ order : ",mq); mq = (int) ceil ( (double)mq); if ( mq == nspin) mq = -mq; if ( nspin == 5 & & mq == -5) ratio = 25./12.; if (nspin == 5 & & mq == 3) ratio = 19./12.; if (nspin == 7 & & mq == -7) ratio =161745.; if (nspin == 7 & & mq = 5) ratio =11.19.; if (nspin = 7 & & mq = 3) ratio = 101./45.; if (nspin == 9 & & mq == -9) ratio = 31.16.; M  A21  if (nspin == 9 & & mq = 7) ratio = 7./18.; if (nspin == 9 & & mq = 5) ratio = 95736.; if (nspin == 9 & & mq == 3) ratio = 91./36.; } ; /* phase for shearing phl= -2.*3.14159*ratio*sw2*sfol*in0/(double)(si2-l); */ /* corrected formula */ phl=-2.*3.14159*ratio*sw2*sfol*in0/(double)(si2); if(mq<0)phl=-phl; dl= fabs(ratio - (double)mq); sfol 1= sfol *(ratio+ 1); /* additional F l shift?  ===================== */ fl=0.; FETCHPARlS("NOISFl",&Noise); FETCHP ARS("TI" ,ti); if (!strcmp(ti,"shearing done")) fl=Noise; GETFLOAT(" F l shift in ppm ? : ",fl); ph2= (double)fl; phc2= -2.*3.14159*ph2*sfoll*in0; /* 1st FFT : acquisition must be done in STATES mode! /* and automatic baseline correction ?  ============================= */ GETSTRING("Apply ABS2 ? : ",yes); ST0REPAR1("MC2",3); STOREPAR1 S("MC2",3);  (void)sprintf(inname2rr; %s/data/%s/nmr/%s/%d/pdata/%d/2rr", disk,user,name,expno,procno); (void)unlink(inname2rr); ,  (void)sprintf(inname2ir,"%s/data/ /os/nmr/%s/%d/pdata/%d/2ir", disk,user,name,expno,procno); (void)unlink(inname2ir); 0  XF2 if ( y e s [ 0 ] = y ) {ABS2 XHT2} /* init status processing variables  ===============================*/  A22  FETCHP ARlS("STSI",&sil); FETCHP A R l S ( " X D I M " , & x d i m l ) ; FETCHPARS("XDIM",&xdim2); sizeofint=sizeof(int); nbytes=sizeofint*xdim2*2; /* Open source file R R and IR  ========================== */ if ((infile2rr=open(inname2rr,0))==-1) {(void)sprintf(text," I/O Error (Open) \n%s ",inname2rr); STOPMSG(text);} if ((infile2ir=open(inname2ir,0))==-1) {(void)sprintf(text," I/O Error (Open) \n%s ",inname2ir); STOPMSG(text);} /* Create temporary files : 2rrtemp et 2irtemp  ========================================= */ (void)sprintf( tempname2rr,"%s/data/%s/nmr/%s/%d/2rrtemp", disk,user,name,expno); if ((tempfile2rr=creat(tempname2rr,0664))==-1) {(void)sprintf(text," I/O Error (Create) \n%s ",tempname2rr); STOPMSG(text);} (void)sprintf(tempname2ir,"%s/data/%s/nmr/%s/%d/2irtemp", disk,user,name,expno); if ((tempfile2ir=creat(tempname2ir,0664))==-1) { (void)sprintf(text," I/O Error (Create) \n%s ",tempname2ir); STOPMSG(text);} (void)sprintf(text,"shear: calculating"); Show_status(text); /* Read data in submatrix ====================== */ TEVIESfsil/xdiml) TIMES2 (si2/xdim2) TIMES3 (xdiml/2) if ((nbytesread=read(infile2rr,in2rr,nbytes))<=0) /* Read 2 rows ! */ {(void)unlink(tempname2rr); S T O P M S G f 2rr file corrupted! ");}  A23  if ((nbytesread=read(infile2ir,in2ir,nbytes))<=0) {(void)unlink(tempname2ir); STOPMSG(" 2ir file corrupted! ");} /* Store Sx */ for (loopcount4=0; loopcount4<xdim2; loopcount4++) {temp2rr[loopcount4]= in2rr[loopcount4]; temp2ir[loopcount4]= in2ir[loopcount4];} /* Store iSy */ for (Ioopcount4=0;loopcount4<xdim2;loopcount4++) {temp2rr2[loopcount4]= in2rr[loopcount4+xdim2]; temp2ir2[loopcount4]= in2ir[loopcount4+xdim2];} /* combine Sx and iSy to create ...  ================================*/ for (Ioopcount4=0;loopcount4<xdim2;loopcount4++) { /* echo */ temp2rrB[loopcount4]=(temp2rr[loopcount4]-temp2ir2[loopcount4])/2; temp2irB[loopcount4]=(temp2ir[loopcount4]+temp2rr2[loopcount4])/2; /* and antiecho */ temp2rrB2 [loopcount4]=(temp2rr [loopcount4] +temp2ir2 [loopcount4] )/2; temp2irB2 [loopcount4]=(temp2ir [loopcount4] -temp2rr2 [loopcount4] )/2;} /* perform shearing transformation = = = = = = = = = = = = = = = = = = = = = = = = = = = =  _ */  i l = loopcount3 +(loopcountl)*xdiml/2; for (loopcount4=0; loopcount4<xdim2; loopcount4++) {i2= loopcount4+(loopcount2)*xdim2; phcl=phl*(double)((i2-si2/2)*il); phc 1 = phc 1 +phc2 * (double)i 1; in2rr[loopcount4]= (int)(((double)temp2rrB[loopcount4])*cos(phc 1) +((double)temp2irB[loopcount4])*sin(phcl))/2; in2ir[loopcount4]= (int)(((double)temp2irB[loopcount4])*cos(phcl) -((double)temp2rrB[loopcount4])*sin(phcl))/2; in2rr[loopcount4+xdim2]= (int)(((double)temp2rrB2 [loopcount4])*cos(phc 1) -((double)temp2irB2[loopcount4])*sin(phcl))/2; in2ir[loopcount4+xdim2]= (int)(((double)temp2irB2[loopcount4])*cos(phcl)  A24  +((double)temp2rrB2[loopcount4])*sin(phcl))/2;} for (Ioopcount4=0;loopcount4<xdim2;loopcount4++) {temp2rr [loopcount4]=0.5 * ( in2rr[loopcount4]+in2rr[loopcount4+xdim2]); temp2ir[loopcount4]=0.5*( in2ir[loopcount4]+in2ir[loopcount4+xdim2]); temp2irB[loopcount4]=0.5*( in2rr [loopcount4] -in2rr[loopcount4+xdim2]); temp2rrB[loopcount4]=0.5*( in2ir[loopcount4]-in2ir[loopcount4+xdim2]);} for (Ioopcount4=0;loopcount4<xdim2;loopcount4++) {in2rr[loopcount4]= temp2rr[loopcount4]; in2ir[loopcount4]= temp2ir[loopcount4]; in2rr [loopcount4+xdim2] = temp2rrB [loopcount4]; in2ir[loopcount4+xdim2]= temp2irB2[loopcount4];} /* Save write(tempfile2rr,in2rr,nbytesread); write(tempfile2ir,in2ir,nbytesread); E N D ; /* loop3 */ END; /* loop2 */ END; /*loopl */ close(tempfile2rr); close(tempfile2ir); close(infile2rr); close(infile2ir); /* Copy  (void)unlink(inname2rr); if (rename(tempname2rr,inname2rr)!=0) Proc_err(l," error %d",errno) (void)unlink(inname2ir); if (rename(tempname2ir,inname2ir)!=0) Proc_err(l," error %d",errno) (void)sprintf(text,"shear: finished"); Show_status(text); /* 2nd FFT and Scaling */ VIEWDATA FETCHPARS("BF1 ",&bfl2); FETCHPARS("SR",&sr2);  A25  bfll=bfl2*(ratio+ 1); STOREPARlS("BFl",bfl 1); S T O R E P A R l ( " B F l " , b f l 1); STOREPARlS("SF01",sfoll); ST0REPAR1("SF01 ",sfol 1); /* store T l to recall shearing and shifting */ STOREPARS("Tr\"shearing done"); STOREPAR1 S("NOISF 1 ",fl); /* F l transform */ XHT2 SETCURDATA AUERR=CPR_exec("xif2",WAIT_TERM); FETCHPAR("PH_mod",&phmod) STOREPAR("PH_mod",0) STOREPAR("WDW",wdw2) FETCHPAR("BC_mod",&bcmod) STOREPAR("BC_mod",0) XFB STOREPAR("PH_mod",phmod) STOREPAR("BC_mod",bcmod)  /* **************************************************** /•Referencing in F l  */  y * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *^  /*For shearing transformations performed after the first fourier */ /•transform, additional scaling in F l is required. (1) F l frequency */ /*is multiplied by (ratio + 1), where ratio=T9/12 for 5/2 nuclei, etc. */ /*This is to account for the true shift of the MQ-echo in t l . For */ /*echo experiments, setting ndO = 2 would suffice. (2) M Q M A S F l axes are*/ /•referenced with respect to the transmitter. The transmitter is in the*/ /*middle of the F l axis. (3) Chemical shifts are scaled by a factor of */ /*(ratio - mq order), e.g. for 3qMAS on a spin=5/2 nucleus, chemical */ /*shifts are 17/12 their value in F2. Therefore the frequency difference*/ /*between 0 ppm and the transmitter (in ppm units) in F l should be */ /*(ratio - mq order) * the value in F2. */ /*The sign of the F l frequencies depends on the pulse sequence, */ /* i.e. does the phase cycling collect the echo or anti-echo pathway? */ /*Since the majority of M Q M A S pulse sequences where originally designed */ /*for Na-23, even if the comments to that effect in the code have long */  A26  /* since fallen by the wayside during numerous updates to keep up with */ /*new software, I have found it probably that the phase cycling selects */ /*the anti-echo pathway for 5/2 nuclei, making the chemical shift */ /*contribution to F l positive and the quadrupolar shift contribution */ /*negative (i.e. the same as in F2). */ /*This programme (i) Scales S F O l ( F l ) by (ratio + 1), (ii) zeroes the F l */ /*axis at its mid-point, (iii) calculates the ppm difference between */ /*SF01(F2) and 0 ppm (F2), (iv) scales this value by a factor of */ /*(ratio - mq order)/(ratio + 1), (y) corrects 0 ppm (Fl) by this value.*/ /*Literature references: */ /*Youngman et. al., Z.Naturforschung, 51a, 321-329, 1996. */ /*Wu, Rovnyank, Sun, Griffin, Chem. Phys. Lett., 249, 210-217, 1995. */ /*Medek, Harwood, Frydman, J .Am. Chem. Soc, 117, 12779, 1995. /* and most usefully */ /*Massiot, Touzo, Trumeau, Coutures, Virlet, Florian, Grandinetti, Solid*/ /*State N M R , 6, 73-83, 1996. */  */  FETCHP AR1 S("SWH",&offsetl); offl=offsetl/(2*sfoll); STOREPAR1 ("OFFSET",offl); FETCHP ARS("OFFSET",&offset2); off2 = offset2 - sw2/2; srl =dl*off2/(ratio+ 1); ofO = offl + s r l ; STOREPARl("OFFSET",off3); STOREPAR1 S("OFFSET",off3); VIEWDATA QUITMSGC Shearing-FT done!")  /*  xfshear  /* /* /* /*  Short Description: */ Program for shearing of 2D M Q M A S spectra of odd half integer quadrupolar nuclei. Data need to be aquired in States Mode */  j^fi s(s  /*  08.11.1996  */  */ */  s|s s|c s)c sfc s)c s)c s)c s)c s|c s|c s|c s|c sfc sfc s|c sfc s{c s{c s(c sfc sfc 3(c s|c s(c s{c 3}c s|c s|c s|c s|c s)c s)c s)c sjc sjc sfc s{c s{c sfc sfc 3|c s(c s)c sfc sfc sfc s(c sfc s)c s)c s|c s|c sfc s{c s)c s|c  Keywords:  */  A27  s|c s|c sjc sjc ^  /*  shear,half integer spin, multi quantum, M Q , headache  */  ^/* ***************************************************************/ /* /* /* /* /* /* /* /* /* /* /* /* /* /* /* /* /* /* /* /* /* /* /* /* /* /*  Description/Usage: */ Program is used for 2D M Q experiments on nuclei with */ odd half integer spin for shearing-FT of 2D spectrum */ including 2DFT and referencing in F l dimension */ according to delta(MQ)=delta(iso)+(p-R)delta(qis) */ It checks for the parameter "nucleus" ( A M X ) or "NUC1" */ (Avance) and sets the spin value accordingly. */ If the nucleus is not in the list it asks for the spin, */ too. Only most common nuclei are in the list. */ If the pulse program name matches mpxq, where */ x corresponds to the multiple quantum order of */ the experiment, then this is assumed, otherwise */ it is asked for by the program. */ Program asks i f abs2 in F2 dimension is desired. */ If answered with yes, then a Hilbert transform */ is automatically done, which requires xwinnmrl .3 */ or later. In older versions, this command has to */ be eliminated from the program code. */ A F l frequency shift in ppm is asked for to compensate */ for malchosen o l setting. This value is stored and */ suggested for any repeated processing. */ JLB: This will totally ruin any prospect of correctly */ referencing the F l axis. Use only in emergencies, e.g. */ to learn where to put the offset the next time. This */ is only a big consideration in the confined of rotor- */ synchronous acquisition. */  /****************************************************************/ /* /* /* /*  Author(s): Name Organisation Email  /* /* /*  Name : Stefan Steuernagel Organisation : Keystone Cops Email : stefan.steuernagel@bruker.de  /*  : Christian Fernandez : Universite de Lille, CNRS-801 : criristian.fernandez@univ.lille.fr  */ */ */ */  */  */ */ */  ^****************************************************************/ /* /* /* /* /* /*  Name cf ste ste  Date Modification: */ 960201 created */ 980813 check xdim's and status sizes */ at right time */ 980813 X H T 2 included after XF1 to */ allow f2 phase correction after */  A28  /* /* /* /* /* /* /* /* /* /* /* /*  ste  ste  jib  processing */ 981101 close of input data files and */ "mv temp-file to 2rr" etc */ included */ 000105 XIF2 and X F B included after shearing to allow apodisation */ in F2, set-up to be done prior */ to execution of xfshear, all */ handling done by this program */ 001110 Referencing in F l modified. */ Literature and method described */ at line 357. */  /* $Id: xfshear,v 1.3.2.2 1999/03/01 10:55:04 gsc Exp $ */ #include <math.h> int parmode, si2, s i l , outexpno, tempfile2rr, infile2rr, tempfile2ir, infile2ir, in2rr[4096], in2ir[4096], temp2rr[4096], temp2ir[4096], temp2rr2[4096], temp2ir2[4096], temp2rrB[4096], temp2irB[4096], temp2rrB2[4096], temp2irB2[4096], nbytes, sizeofint, nbytesread, xdim2, xdiml, loopcount4, sig, mq, nspin, wdw2, bcmod, phmod; double s w l , sw2, ratio, ph, p h i , ph2, phcl, phc2, inO, sfol, sfol2, s f o l l , b f l l , bfl2; float Noise, spin, offset 1, offset2, off3, off2, offl, sr2, srl; char inname2rr[PATH_LENGTH], tempname2rr[PATH_LENGTH], inname2ir[P A T H L E N G T H ] , tempname2ir[P A T H _ L E N G T H ] , outname[P A T H L E N G T H ] , nucl[80], yes[20], pulprog[80], ti[80]; /* select current dataset  ==================== */ GETCURDATA; (void) strcpy (yes,"no ); M  /* is-it a 2D spectrum ?  ==================== */ FETCHPAR("PARMODE",&parmode);  A29  */  if (parmode != 1) { STOPMSG(" Not a 2D spectrum!");} /* init some variables  _  ===  _  =======  = */  FETCHPARS("PULPROG",pulprog); FETCHPARS("SW",&sw2); FETCHPARS("SF01",&sfol); FETCHPAR1S("IN 0",&inO); FETCHPAR("SI",&si2); FETCHPAR("WDW",&wdw2) STOREPAR("WDW",0) /* shearing ratio calculation */  I*  */  spin=5./2.; mq=3; FETCHP A R ( " N U C 1 ",nucl) if(!strcmp(nucl,"off')) F E T C H P A R ("NUCLEUS",nucl) if(!strcmp(nucl,"7Li")) spin=l.5; else if(!strcmp(nucl,"HB")) spin=1.5; else if(!strcmp(nucl,"170")) spin=2.5; else if(!strcmp(nucl,"23Na")) spin=1.5; else if(!strcmp(nucl,"27Al")) spin=2.5; else if(!strcmp(nucl,"35Cl")) spin=1.5; else if(!strcmp(nucl,"39K")) spin=1.5; else if(!strcmp(nucl,"45Sc")) spin=3.5; else if(!strcmp(nucl,"45Sc")) spin=3.5; else if(!strcmp(nucl,"51V")) spin=3.5; else if(!strcmp(nucl,"55Mn")) spin=2.5; else if(!strcmp(nucl,"63Cu")) spin=1.5; else if(!strcmp(nucl,"65Cu")) spin=1.5; else if(!strcmp(nucl,"69Ga")) spin=1.5; else if(!strcmp(nucl,"71Ga")) spin=1.5; else if(!strcmp(nucl,"85Rb")) spin=2.5; else if(!strcmp(nucl,"87Rb")) spin=1.5; else if(!strcmp(nucl,"93Nb")) spin=4.5; else if(!strcmp(nucl,"121Sb")) spin=2.5; else GETFLOAT(" Enter spin number : ",spin); nspin = (int) ceil ((double)(2*spin)); if (nspin == 3 ) {mq = -3; ratio = 7./9.;} else  A30  { i f (!strncmp(pulprog,"mp3q",4)) mq=3; else if(!stmcmp(pulprog,"mp5q",4)) mq=5; else if(!strncmp(pulprog,"mp7q",4)) mq=7; else if(!stmcmp(pulprog,"mp9q",4)) m q ^ ; else GETINT( " Enter the pQ order : ",mq); mq = (int) ceil ((double)mq); if ( mq - nspin) mq = -mq; if (nspin == 5 & & mq == -5) ratio = 25./12.; if ( nspin == 5 & & mq == 3) ratio = 19./12.; if ( nspin == 7 & & mq - - -7) ratio = 161./45.; if ( nspin == 7 & & mq == 5) ratio = 11./9.; if (nspin = 7 & & mq == 3) ratio = 101./45.; if (nspin == 9 & & mq == -9) ratio = 31.16.; if (nspin == 9 & & mq == 7) ratio = 7./18.; if (nspin == 9 & & mq == 5) ratio = 95736.; if ( nspin == 9 & & mq == 3) ratio = 91736.; } ; /* phase for shearing phl=-2.*3.14159*ratio*sw2*sfol*in0/(double)(si2-l); /* corrected formula */ phl=-2.*3.14159*ratio*sw2*sfol*in0/(double)(si2); if(mq<0)phl=-phl; dl= fabs(ratio - (double)mq); sfoll=sfol*(ratio+ 1); /* additional F l shift ?  =============== */ fl=0.; FETCHPARlS("NOISFl",&Noise); FETCHP ARS(''TT',ti); if (!strcmp(ti,"shearing done")) fl=Noise; GETFLOAT(" F l shift in ppm ? : ",fl); ph2= (double)fl; phc2= -2.*3.14159*ph2*sfoll*in0; /* 1st FFT : acquisition must be done in STATES mode! /* and automatic baseline correction ?  = = = = = = = = = = = = = = = = _ = „ = = = = = */ GETSTPJNG("Apply A B S 2 ? : ",yes); ST0REPAR1("MC2",3); ST0REPAR1 S("MC2",3);  A31  (void)sprintf( irmame2r^ disk,user,name,expno,procno); (void)unlink(inname2rr); (void)sprintf(irmame2ir;'%s/data/%s/nmr/%s/%d/pdata/%^ disk,user,name,expno,procno); (void)unlink(inname2ir); XF2 if (yes[0]=='y') {ABS2 XHT2}  _ _ __  /*• init status processing variables  ===  ==  ==  ========:====#/  FETCHPARlS("STSI ,&sil); FETCHPAR1 S("XDIM",&xdiml); FETCHP ARS("XDIM",&xdim2); sizeofint=sizeof(int); nbytes=sizeofint*xdim2*2; M  /* Open source file R R and IR  ===================== */  if ((infile2rr=open(inname2rr,0))==-1) {(void)sprintf(text," I/O Error (Open) \n%s ,inname2rr); STOPMSG(text);} M  if ((infile2ir=open(inname2ir,0))==-1) {(void)sprintf(text," I/O Error (Open) \n%s ",inname2ir); STOPMSG(text);} /* Create temporary files : 2rrtemp et 2irtemp  =================================== */ (void)sprintf(  tempname2rr,"%s/data/%s/nmr/%s/%d/2rrtemp", disk,user,name,expno); if ((tempfile2rr=creat(tempname2rr,0664))==-1) {(void)sprintf(text," I/O Error (Create) \n%s ",tempname2rr); STOPMSG(text);} (void)sprintf( tempname2ir,"%s/data/%s/nmr/%s/%d/2irtemp", disk,user,name,expno); if ((tempfile2ir=creat(tempname2ir,0664))==-l) {(void)sprintf(text," I/O Error (Create) \n%s ",tempname2ir); STOPMSG(text);}  A32  (void)sprintf(text,"shear: calculating"); Show_status(text); /* Read data in submatrix  =================== */ TIMES(sil/xdiml) TIMES2 (si2/xdim2) TEVIES3 (xdiml/2) if ((nbytesread=read(infile2rr,in2rr,nbytes))<=0) /* Read 2 rows ! */ {(void)unlink(tempname2rr); STOPMSG(" 2rr file corrupted! ");} if ((nbytesread=read(infile2ir,in2ir,nbytes))<=0) {(void)unlink(tempname2ir); STOPMSG(" 2ir file corrupted! ");} /* Store Sx */ for (loopcount4=0;loopcount4<xdim2;loopcount4++) {temp2rr[loopcount4]= in2rr[loopcount4]; temp2ir[loopcount4]= in2ir[loopcount4];} /* Store iSy */ for (Ioopcount4=0;loopcount4<xdim2;loopcount4++) {temp2rr2[loopcount4]= in2rr[loopcount4+xdim2]; temp2ir2[loopcount4]= in2ir[loopcount4+xdim2];} /* combine Sx and iSy to create ...  ==============================*/ for (loopcount4=0; loopcount4<xdim2; loopcount4++) { /* echo */ temp2rrB[loopcount4]=(temp2rr[loopcount4]-temp2ir2[loopcount4])/2; temp2irB[loopcount4]=(temp2ir[loopcount4]+temp2rr2[loopcount4])/2; /* and antiecho */ temp2rrB2[loopcount4]=(temp2rr[loopcount4]4-temp2ir2[loopcount4])/2; temp2irB2[loopcount4]=(temp2ir[loopcount4]-temp2rr2[loopcount4])/2;} /* perform shearing transformation  =============================*/ i l = loopcount3 +(loopcountl)*xdiml/2; for (Ioopcount4=0;loopcount4<xdim2;loopcount4++)  A33  {i2= loopcount4+(loopcount2)*xdim2; phcl= phi *(double)((i2-si2/2)*il); phc 1 = phc 1 +phc2 * (double)i 1; in2rr [loopcount4] = (int)(((double)temp2rrB [loopcount4]) * cos(phc 1) +((double)temp2irB[loopcount4])*sin(phcl))/2; in2ir [loopcount4] = (int)(((double)temp2irB [loopcount4]) * cos(phc 1) -((double)temp2rrB[loopcount4])*sin(phcl))/2; in2rr [loopcount4+xdim2]= (int)(((double)temp2rrB2 [loopcount4]) * cos(phc 1) -((double)temp2irB2[loopcount4])*sin(phc 1 ))/2; in2ir[loopcount4+xdim2]= (int)(((double)temp2irB2[loopcount4])*cos(phcl) +((double)temp2rrB2 [loopcount4]) * sin(phc 1 ))/2;} for (Ioopcount4=0;loopcount4<xdim2;loopcount4-H-) {temp2rr [loopcount4]=0.5 * ( in2rr[loopcount4]+in2rr[loopcount4+xdim2]); temp2ir [loopcount4]=0.5 * ( in2ir[loopcount4]+in2ir[loopcount4+xdim2]); temp2irB[loopcount4]=0.5*( in2rr [loopcount4] -in2rr [loopcount4+xdim2]); temp2rrB[loopcount4]=0.5*( in2ir[loopcount4]-in2ir[loopcount4+xdim2]);} for (loopcount4=0; loopcount4<xdim2 ;loopcount4+4-) {in2ir[loopcount4]= temp2rr[loopcount4]; in2ir[loopcount4]= temp2ir[loopcount4]; in2rr[loopcount4+xdim2]= temp2rrB [loopcount4]; in2ir[loopcount4+xdim2]= temp2irB2 [loopcount4];} /* Save _ */ write(tempfile2rr,in2rr,nbytesread); write(tempfile2ir,in2ir,nbytesread); E N D ; /* loop3 */ E N D ; /* loop2 */ E N D ; /*loopl */ close(tempfile2rr); close(tempfile2ir); close(infile2rr); close(infile2ir); /* Copy  A34  (void)unlink(inname2rr); if (rename(tempname2rr,inname2rr)!=0) Proc_err(l,"error %d",errno); (void)unlink(inname2ir); if (rename(tempname2ir,inname2ir)!=0) Proc_err(l,"error %d",errno); (void)sprintf(text,"shear: finished"); Showstatus(texf);  /* 2nd FFT and Scaling */ VIEWDATA FETCHPARS("BF1 ",&bfl2); FETCHP ARS("SR",&sr2); bfll=bfl2*(ratio+ 1); STOREPAR1 S("BF 1 ",bf 11); S T O R E P A R l ( " B F l " , b f l 1); STOREPARlS("SF01",sfoll); STOREPARl("SF01'\sfoll); /* store T l to recall shearing and shifting */ STOREPARS("TI","shearing done"); STOREP AR1 S("NOISF 1 ",fl); /* F l transform */ XHT2 SETCURDATA AUERR=CPR_exec("xif2",WAIT_TERM); FETCHP AR("PH_mod",&phmod) STOREP AR("PH_mod" ,0) STOREP AR("WDW",wdw2) FETCHP AR("BC_mod",&bcmod) STOREP AR("BC_mod",0) XFB STOREP AR("PH_mod",phmod) STOREP AR("BC_mod",bcmod)  I***************************** /* Referencing in F1  */  I***************************** /*For shearing transformations performed after the first fourier */ /•transform, additional scaling in F l is required. (1) F l frequency */  A35  /*is multiplied by (ratio + 1), where ratio=19/12 for 5/2 nuclei, etc. */ / T h i s is to account for the true shift of the MQ-echo in t l . For */ /*echo experiments, setting ndO = 2 would suffice. (2) M Q M A S F l axes are*/ /•referenced with respect to the transmitter. The transmitter is in the*/ /*middle of the F l axis. (3) Chemical shifts are scaled by a factor of */ /•(ratio - mq order), e.g. for 3qMAS on a spin=5/2 nucleus, chemical */ /*shifts are 17/12 their value in F2. Therefore the frequency difference*/ /*between 0 ppm and the transmitter (in ppm units) in F l should be */ /*(ratio - mq order) * the value in F2. */ /*The sign of the F l frequencies depends on the pulse sequence, */ /* i.e. does the phase cycling collect the echo or anti-echo pathway? */ /*Since the majority of M Q M A S pulse sequences where originally designed */ /*for Na-23, even if the comments to that effect in the code have long */ /*since fallen by the wayside during numerous updates to keep up with */ /•new software, I have found it probably that the phase cycling selects */ /•the anti-echo pathway for 5/2 nuclei, making the chemical shift */ /•contribution to F l positive and the quadrupolar shift contribution */ /•negative (i.e. the same as in F2). */  /************************************************************************y  /•This programme (i) Scales S F O l ( F l ) by (ratio + 1), (ii) zeroes the F l •/ /*axis at its mid-point, (iii) calculates the ppm difference between •/ /•SF01(F2) and 0 ppm (F2), (iv) scales this value by a factor of •/ /•(ratio - mq order)/(ratio + 1), (v) corrects 0 ppm (Fl) by this valued/  /************************************************************************y  /•Literature references: •/ /•Youngman et. al., Z.Naturforschung, 51a, 321-329, 1996. •/ /*Wu, Rovnyank, Sun, Griffin, Chem. Phys. Lett., 249, 210-217, 1995. */ /*Medek, Harwood, Frydman, J A m . Chem. Soc, 117, 12779, 1995. */ /* and most usefully */ /•Massiot, Touzo, Trumeau, Coutures, Virlet, Florian, Grandinetti, Solid*/ /•State N M R , 6, 73-83, 1996. */ /************************************************************************/ FETCHP ARlS("SWH",&offsetl); offl=offsetl/(2 sfoll); STOREP ARl("OFFSET",offl); #  FETCHP ARS("OFFSET",&offset2); off2 = offset2 - sw2/2; srl =dPoff2/(ratio+ 1); off3 = offl + s r l ; STOREP AR1 ("OFFSET",off3); STOREP AR1 S("OFFSET",off3);  A36  VIEWDATA QUITMSG(" Shearing-FT done!")  III.A(xii)  TEST PROGRAM: zg.Al-Si.test.jlb  ;zg.Si-Al.test.jlb ;Test programme for P N N L Si-Al probe. Pulse simultaneously on both channels, ;no decoupling during acquisition. ;avance-version ;single pulse excitation with high power decoupling ; !!! recable 1H channel £2 direct to probehead !!! ; !!! do not connect via the proton preamplifier !!!  #include <Avance.incl>  1 ze 2 d l ;do:f2 10upll:f3 ;2u cw:f2 (pi phl):f3 ;2u do:£2 go=2ph31 wr#0 exit  phl=0 2 1 3 ph31=0 2 1 3  ;pll : f l channel - power level for pulse (default) ;pl2: f2 channel - power level for C W decoupling ;pl : f l channel - 90 degree high power pulse ;dl : relaxation delay: 5 sec ;NS: 4 * n (not exactly as D M X text file - irrelevant comments deleted)  A37  IILA(xiii)  ID CP EXPERIMENT: cp.quad.jlb  ;cp.quad.jlb ;CP experiment, no decoupling ; independent power level for 90 degree Y-pulse. ; f l is the X channel (observed) ; £2 is the nucleus that is the source for CP (usually H) ;Andrew Lewis, Jerrikins, U B C ;2001 ;avance-version #include <Avance.incl>  1 ze 2dl 3u pl21:f2 (p3 pl21):f2 3u p l l i f l 12:f2 (pl5ph2):fl (pl5ph3):f2 go=2 ph31 wr#0 exit P  phl=l3 ph2=0 0 1 1 2 2 3 3 ph3=0 ph31=0 2 1 3 2 0 3 1 ;pl21 £2 - 90 degree pulse power level ;pll f l - CP power level ;pl2 £2 - CP power level ;p3 £ 2 - 9 0 degree high power pulse ;pl5 contact pulse on f l and £2 ;dl relaxation delay ;NS: 8 * n ;DS: 0  A38  III.B Pulse sequences for Varian Inova  III.B(i)  SINGLE PULSE EXPERIMENT (no phase cycling): s2pul.c  #imdefLINT static char SCCSid[] = "@(#)s2pul.c 14.1 10/10/97 Copyright (c) 1991-1996 Varian Assoc.,Inc. A l l Rights Reserved"; #endif /* s2pul - standard two-pulse sequence */ #include <standard.h> pulsesequence() { /* equilibrium period */ status(A); hsdelay(dl); /* — tau delay — */ status(B); pulse(pl, zero); hsdelay(d2); /* — observe period — */ status(C); pulse(pw,oph);  }  m.B(ii)  GENERAL SPIN-ECHO EXPERIEMENT: echo.c  #imdefLTNT  A39  static char SCCSid[] = "@(#)s2pul.c 14.1 10/10/97 Copyright (c) 1991-1996 Varian Assoc.,Inc. A l l Rights Reserved"; #endif /* Jerry - modified from s2pul - standard two-pulse sequence */ #include <standard.h> static inttablel[8] = {0,0,1,1,2,2,3,3}; static inttable2[8] = {0,2,1,3,2,0,3,1}; static int table3[8] = {0,0,1,1,2,2,3,3}; pulsesequence()  settable(tl,8,tablel); settable(t2,8,table2); settable(t3,8,table3); setreceiver(t3); /* equilibrium period */ status(A); hsdelay(dl); /* — tau delay — */ status(B); pulse(pl,tl); delay(d2); /* — observe period — */ status(C); pulse(pw,t2); delay(d2-50e-6-alfa); acquire(np,l/sw); }  III.B(iii)  Ti INV-REC EXPERIMENT: echotl.c  #ifhdefLINT static char SCCSid[] = "@(#)s2pul.c 14.1 10/10/97 Copyright (c) 1991-1996 Varian Assoc.Inc. A l l Rights Reserved"; #endif  A40  /* Jerry - modified from s2pul - standard two-pulse sequence */ #include <standard.h> static int tablel[4] = {0,1,2,3}; static inttable2[12] = {0,0,0,1,1,1,2,2,2,3,3,3}; static inttable3[12] = {0,0,0,1,1,1,2,2,2,3,3,3}; pulsesequence()  { settable(tl,4,tablel); settable(t2,12,table2); settable(t3,12,table3); setreceiver(t3); /* equilibrium period */ status(A); hsdelay(dl); /* — tau delay — */ status(B); pulse(pl,tl); delay(d2); /* — observe period — */ status(C); pulse(pw,t2); acquire(np,l/sw); }  III.B(iv)  ID INEPT EXPERIMENT: elqineptc  #ifhdefLINT static char SCCSid[] = "@(#)1-D INEPT Experiment"; #endif /*einept.c Number of transients (nt) must be a multiple of 16 to allow for phase cycling. Phase cycling copied from pulse programme used on DSX400  A41  pw90 - 90 degree pulse width (observe channel) p90dec - 90 degree pulse width (decouple channel) taul = delay for evolution of J-coupling (i.e. 1/4J delay) np - # of acquisition points (t2-dimension) */ #include <standard.h> static static static static static static  int table3[16]={0,0,0,0,3,3,3,3,2,2,2,2,1,1,1,1}; int table4[ 16]={0,0,0,0,3,3,3,3,2,2,2,2,1,1,1,1}; int table5[16]={l, 1,3,3,0,0,2,2,3,3,1,1,2,2,0,0}; int table6[ 16]= {0,0,0,0,3,3,3,3,2,2,2,2,1,1,1,1}; int table7[16]={0,0,0,0,3,3,3,3,2,2,2,2,l,1,1,1}; int table9[16]={0,0,2,2,3,3,l,1,2,2,0,0,1,1,3,3};  pulsesequence() { /* declare new variables */ double pw90,p90dec,p90decB,taul; pw90=getval("pw90"); p90dec=getval("p90dec"); p90decB=getval("p90decB"); taul=getval("taul"); settable(t3,16,table3); settable(t4,16,table4); settable(t5,16,table5); settable(t6,16,table6); settable(t7,16,table7); settable(t9,16,table9); setreceiver(t9); status(A); txphase(zero); delay(dl);  /* recycle delay */  status(B); decrgpulse(p90decB,t3,rofl,rof2); /* select one orthogonal component */ delay(taul); /* 1/4J delay */ rgpulse(2*pw90,t6,rofl,rof2); /*"Simultaneous" 180 pulses*/ decrgpulse(2*p90dec,t4,rofl,rof2);  A42  delay(taul); /* 1/4J delay */ rgpulse(pw90,t7,rofl,rof2); /*"Simultaneous" 90 transfer pulses*/ decrgpulse(p90dec,t5,rof 1 ,ro£2); acquire(np, 1.0/sw); /* begin acquisition */ }  III.B(v)  ID REFOCUSSED INEPT EXPERIMENT: elqrinept.c  #imdefLINT  static char SCCSid[] = "@(#)1-D INEPT Experiment"; #endif /*einept.c Number of transients (nt) must be a multiple of 16 to allow for phase cycling. Phase cycling copied from pulse programme used on DMX400 pw90 - 90 degree pulse width (observe channel) p90dec - 90 degree pulse width (decouple channel) taul = delay for evolution of J-coupling (i.e. 1/4J delay) np - # of acquisition points (t2-dimension) */ #include <standard.h> static int table3[ 16]={0,0,0,0,3,3,3,3,2,2,2,2,1,1,1,1}; static int table4[16]={0,0,0,0,3,3,3,3,2,2,2,2,l, 1,1,1}; static int table5[16]={l,1,3,3,0,0,2,2,3,3,1,1,2,2,0,0}; static int table6[ 16]={0,0,0,0,3,3,3,3,2,2,2,2,1,1,1,1}; static int table7[16]={0,0,0,0,3,3,3,3,2,2,2,2,l,1,1,1}; static int table9[16]={0,0,2,2,3,3,l,1,2,2,0,0,1,1,3,3}; static int tablel0[16]={ 1,3,1,3,0,2,0,2,3,1,3,1,2,0,2,0}; static int tablel 1 [16]={ 1,1,1,1,0,0,0,0,3,3,3,3,2,2,2,2}; pulsesequence() { /* declare new variables */ double pw90,p90dec,p90decB,tau 1 ,tau2;  A43  pw90=getval("pw90"); p90dec=getval("p90dec"); p90decB=getval("p90decB"); taul=getval("taul"); tau2=getval("tau2"); settable(t3,16,table3); settable(t4,16,table4); settable(t5,16,table5); settable(t6,16,table6); settable(t7,16,table7); settable(t9,16,table9); settable(tl0,16,tablel0); settable(tll,16,tablell); setreceiver(t9); status(A); txphase(zero); delay(dl);  /* recycle delay */  status(B); decrgpulse(p90decB,t3,rofl,rof2); /* select one orthogonal component */ delay(taul); /* 1/4J delay */ rgpulse(2*pw90,t6,rofl,rof2); /*" Simultaneous" 180 pulses*/ decrgpulse(2*p90dec,t4,rofl ,rof2); delay(taul); /* 1/4J delay */ rgpulse(pw90,t7,rofl,rof2); /*"Simultaneous" 90 transfer pulses*/ decrgpulse(p90dec,t5,rof 1 ,rof2); delay(tau2); /* Refocussing delay */ rgpulse(2 *pw90,t 10,rof1 ,rof2); decrgpulse(2*p90dec,tl 1 ,rof1 ,rof2); /*Refocussing 180degree pulses*/ delay(tau2-0.000100); /* Refocussing delay. Begin acq. at top of echo */ acquirefnp, 1.0/sw); /* begin acquisition */  }  IILB(vi)  2D REFOCUSSED INEPT EXPERIMENT: elq2drin.c  #ifhdefLINT static char SCCSid[] = "@(#)2D INEPT Experiment";  A44  #endif /*e2drinept.c Number of transients (nt) must be a multiple of 32 to allow for phase cycling. Phase cycling developed from ID pulse programme used on DMX400 pw90 - 90 degree pulse width (observe channel) p90dec - 90 degree pulse width (decouple channel) p90decB - 90 degree pulse (decoupler) during preparation phase taul - delay for evolution of J-coupling (i.e. 1/4J delay - with TI relaxation trade-off) tau2 - J-coupling refocussing delay. phase - if phase=0, all phase tables read as written. If phase=3, id2 (increment of d2) is added to phase of first decoupler pulse, t2.  */ #include <standard.h> static int table2[64]={0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,2,2,2,2,2,2,2,2 2,2,2,2,2, 2,2,2,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1}; static int table3[64]={l^aJ3,3,3,3,14,1433,33A0A0,2,2^,2A0A0,2,2,2,2,3,3,3,3,l,l,l,l,3,3,3,3,l, 1,1,1,2,2,2,2,0,0,0,0,2,2,2,2,0,0,0,0}; static int table4[64]={0,0,0,0,0,0,0,0,2,2,2,2,2,2,2,2,3,3,3,3,3,3,3,3,1,1,1,1,1,1,1,1,2,2,2,2,2,2,2,2,0,0,0,0,0, 0,0,0,1,1,1,1,1,1,1,1,3,3,3,3,3,3,3,3}; static int table5[64]={ 1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,3,1,3,1,3,1,3,1,3,1,3,1,3, 1,3,1,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0}; static int table6[64]={0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,3,3,3,3,3,3,3,3,3,3,3,3,3 3,3,3,2,2,2,2,2,2,2,2,2,2,2,2,2, 2,2,2,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1}; static int table7[64]={14,14,l,l,lJ4a,lJ4,l,l,lA0A0A0A0A0A0A0A0,3,3,3,3,3,3,3,3,3,3,3,3,3, 3,3,3,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2}; static int table9[64]={13,13,13,1343,13a,3J3A2A2A2A2A2A2A2A2 343,l,3,134,3a3,l,3, 1,3,1,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0}; static int tablel0[64]={l,l,3,3,l,l,3,3,l,l,3,3,l,l,3,3,0A2,2,0,0,2,2,0,0,2,2,0,0,2,2,3,3,l,l^ 3,3,1,1,2,2,0,0,2,2,0,0,2,2,0,0,2,2,0,0}; 5  5  J  A45  static int tablet l[64]={iaj,l,l,l,l,l,l,l,l,l,l,iaj,0A0A0A0,0,0,0,0,0,0,0,0,0,3,3,3,3,3,3,3,3,3,3,3,3, 3,3,3,3,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2}; pulsesequence() { /* declare new variables */ double pw90,p90dec,p90decB,taul,tau2,phasel; pw9Ogetval("pw90"); p90dec=getval("p90dec"); p90decB=getval("p90decB"); taul=getval("taul"); tau2=getval("tau2"); phase 1 =getval("phase"); settable(t2,64,table2); settable(t3,64,table3); settable(t4,64,table4); settable(t5,64,table5); settable(t6,64,table6); settable(t7,64,table7); settable(t9,64,table9); settable(tl0,64,tablel0); settable(tll,64,tablell); setreceiver(t9); if (phasel==3); add(t2,id2,t2);  /* TPPI phase incrementation. If used, set swl to 2*desired value. */  status(A); txphase(zero); delay(dl);  /* recycle delay */  status(B); decrgpulse(p90decB,t2,rofl,rof2); /* 90degree preparation pulse */ delay(d2); decrgpulse(p90decB,t3,rofl,rof2); /* select one orthogonal component */ delay(taul); /* 1/4J delay */ rgpulse(2*pw90,t6,rofl,rof2); /*" Simultaneous" 180 pulses*/ decrgpulse(2*p90dec,t4,rofl ,rof2); delay(taul); /* 1/4J delay */  A46  rgpulse(pw90,f7,rofl,rof2); /*"Simultaneous" 90 transfer pulses*/ decrgpulse(p90dec,t5,rof 1 ,rof2); delay(tau2); /* Refocussing delay */ rgpulse(2*pw90,tl 0,rofl ,ro£2); decrgpulse(2*p90dec,tl l,rofl,rof2); /*Refocussing 180degree pulses*/ delay(tau2-0.000150); /* Refocussing delay. Begin acq. at top of echo */ acquire(np, 1.0/sw); /* begin acquisition */ }  IILB(vii)  2D NUTATION 3QMAS EXPERIMENT: 3qmasl.c  /* 3qmasl.c - Sequence for triple-quantum 2D M A S for quadrupole nuclei with optional selective refocussing pulse. D.Rice 6/19/96 D. Massiot, B Touzo, D. Truman, J.P, Coutures, J. Virlet, P. Florian, P.J. Grandinetti; Solid State N M R 6, 73-83 (1996). If pws = 0.0 this sequence uses the hypercomplex phase cycle of Massiot et. al. (equations 3 and 4 - p75). If pws > 0.0 this sequence uses the shifted echo, hypercomplex phase cycle (equations 18 and 19 - p78). Note that equations 3 and 19 (phase = 2) must be left shifted by 1 column to match the phase cycle in this pulse sequence.  */ #include <standard.h> static int tablel[6] = {0,0,1,2,2,3}; static int table2[3] = {0,60,30}; static inttable3[48]= {0,0,0,0,0,0,0,0,0,0,0,0, 1,1,1,1,1,1,1,1,1,1,1,1, 2,2,2,2,2,2,2,2,2,2,2,2, 3,3,3,3,3,3,3,3,3,3,3,3}; static inttable4[12]= {0,0,0,0,0,0,45,45,45,45,45,45}; static inttable5[24] = {0,2,0,2,0,2,0,2,0,2,0,2, 0,2,0,2,0,2,0,2,0,2,0,2}; static int table6[24]= {0,0,0,0,0,0,1,1,1,1,1,1, 2,2,2,2,2,2,3,3,3,3,3,3}; pulsesequence() {  A47  /*define new variables*/ double  lvlshft, phase, pw3q, pw3qlq, pws, pshft, periods, srate, tau, tpwrm, tpwrms, tpwrs;  /*initialize parameters*/ phase = getval("phase"); pw3q = getval("pw3q"); pw3qlq = getval("pw3qlq"); pws = getval("pws"); periods = getval("periods ); srate = getval("srate"); tpwrm = getval("tpwrm"); tpwrms = getval("tpwrms"); tpwrs = getval("tpwrs"); M  tau = periods/srate - pws/2.0 - pw3q/2.0; lvlshft = 0.5e-6; if (tpwr != tpwrs) { lvlshft = lvlshft + 0.5e-6; } pshft = 0.5e-6; /*set phase tables and correct for A P delays*/ d l = d l - pshft; d2 = d2 - pshft; settable(tl ,6,tablel);  A48  settable(t2,3,table2); settable(t5,24,table5); settable(t6,24,table6); if(pws>0.0) { settable(t3,48,table3); settable(t4,12,table4); settable(t6,24,table6); ttadd(t6,t5,t5); d l = d l - lvlshft; tau = tau - lvlshft - pshft; } if (phase == 2) { tsadd(tl,l,4); tsadd(t5,2,4); } i f ( d K O . O ) d l =0.0; if(d2<0.0)d2 = 0.0; if(tau<0.0)tau = 0.0; setreceiver(t5); stepsize(1.0,TODEV); /*begin pulsesequence*/ status(A); if(pws>0.0) { rlpwrm(tpwrrn,TODEV); if (tpwrs != tpwr) rlpower(tpwr,TODE V);  } xmtrphase(t2); delay(dl); rcvroff(); splon(); delay(rofl); rgpulse(pw3q, t l , 0.0, 0.0); xmtrphase(zero);  delay(d2); rgpulse(pw3qlq, zero, 0.0, 0.0); sploff(); if(pws>0.0) { rlpwrm(tpwrms,TODEV); if (tpwrs != tpwr) rlpower(tpwrs,TODEV); xmtrphase(t4); delay(tau); rgpulse(pws, t3, 0.0, 0.0); } /*begin acquisition*/ status(C); delay(rof2); rcvron(); }  IILBfviii)  2D RIACTII 3QMAS EXPERIMENT: 3qRIACT2p.c  /* 3qPJACT2 - Sequence for triple-quantum 2D M A S for quadrupole nuclei (I > 3/2) modified October 15, 1997 to include acquisition of phase sensitive data using adiabatic passage excitation - change phase cycle for I = 3/2 nuclei eg. table5[8] = {0,2,1,3,2,0,3,1}  */ #include <standard.h> static inttablel[24] = {0,0,30,30,60,60,90,90,120,120,150,150,180,180, 210,210,240,240,270,270,300,300,330,330}; static int table2[2] = {0,0}; static int table3[2] = {1,3}; static inttable5[8] = {0,2,3,1,2,0,1,3}; pulsesequence() { /*define new variables*/  A50  double  pw3q, pw3qlq, pw90, phase, pshft; /*initialize parameters*/ pw90 = getval("pw90"); pw3q = getval("pw3q"); pw3qlq = getval("pw3qlq"); phase = getval("phase"); /*set phase tables and correct for A P delays*/ pshft=0.5e-6; d l = d l - pshft; d2 = d2 - pshft; settable(tl,24,tablel); settable(t2,2,table2); settable(t3,2,table3); settable(t5,8,table5); if (dl < 0.0) dl =0.0; if(d2<0.0)d2 = 0.0; if (phase==2) { tsadd(t2,l,4); tsadd(t3,l,4); tsadd(t5,2,4); } setreceiver(t5); stepsize(l.OJODEV); /*begin pulsesequence*/ status(A); xmtrphase(tl); delay(dl); rcvroff(); splonQ;  A51  delay(rofl); rgpulse(pw90,t2,0.0,0.0); rgpulse(pw3q,t3,0.0,0.0); xmtrphase(zero); delay(d2); rgpulse(pw3qlq,zero,0.0,0.0); sploff(); /*begin acquisition*/ status(C); delay(rof2); rcvron(); }  III.B(ix)  1D/2D CP EXPERIMENT: 2dcpmas.c  #ifndefLINT static char SCCSid[] = "@(#)2D CP M A S Experiment"; #endif /*2dcpmas.c 2D C P M A S sequence written for 27-A1/29-SL No decoupling during acquisition p90 - 90 degree preparation pulse on aluminium (decouple channel) lock - spin locking pulse duration phase - if phase=0, all phase tables read as written. If phase=3, id2 (increment of d2) is added to phase of first decoupler pulse (tl) and sw must be doubled. If phase=l,2 then phase 1 counter alternates between T with phase t l and '2' with phase tl+1 (remainder of, after division by 4) */ #include <standard.h> static int tablel[4]={0,l,2,3}; static int table2[16]={l,2,l,2,l,2,l,2,3,0,3,0,3,0,3,0}; static int table3[8]={l,2,l,2,3,0,3,0}; static int table4[4]={0,l,2,3};  pulsesequence() { /* declare new variables */ double p90,lock,phasel;  A52  p90=getval("p90"); /*Use create('parameterVpulse') for units of microseconds.*/ lock=getval("lock"); phasel=getval("phase"); settable(tl,4,tablel); settable(t2,16,table2); settable(t3,8,table3); settable(t4,4,table4); setreceiver(t4); if (phase 1 ==3); add(tl,id2,tl);  /*TPPI phase incrementation. If used, set swl to 2*desired value*/  if (phase 1 ==2); tsadd(t 1,1,4); /*Hypercomplex phase incrementation*/ status(A); txphase(zero); delay(dl);  /* recycle delay */  status(B); decrgpulse(p90,tl,rofl,rof2); /* 90degree decoupler preparation pulse */ delay(d2); /*incrementable delay*/ simpulse(lock,lock,t2,t3,rofl,rof2); acquire(np, 1.0/sw); /* begin acquisition */ }  III.B(x)  3QMAS-refocused INEPT MAS experiment: e3q2drinc.c  /* e3q2drinc.c D. Massiot, B Touzo, D. Truman, J.P, Coutures, J. Virlet, P. Florian, P.J. Grandinetti; Solid State N M R 6, 73-83 (1996). Adapted from elq2drin.c and 3qmastop.c. elq2drin.c phase lists of 4*16 were reduced to 4*4=16. Each of these was then mated to phase lists from the latter, where 180 alternations of phase arise from similar pattern in receiver phase of 3qmastop.c  A53  This results in phase lists of 4*12=48, which have then been quadrature cycled to produce 4*48=192 Phase sensitivity of elq2drin.c included in preparation pulse. Two alternatives are presented here. Either obtain phase sensitivity from selection pulse on top of 3q echo, or from the original 3q-generation pulse. */ #include <standard.h> static inttable31[192] ={0A1,2 2,3A0J,2,23A04,2,23A04,2,2,3A0,1A23A0,1,2,23A0J,2,23,0,0,1,2,2,3 1,1 2,3,3A1,1,2,3 3A1,1,2,3,3A1,1,2 3,3A1,1 2,3,3A1,1,2,3,3A1,1 2 3 3A1,1,2,3,3A2 2,3A0,1, 2,2,3,0,0,1,2,2,3,0,0,1,2,2,3,0,0,1,2,2,3,0,0,1,2,2,3,0,0,1,2,2,3,0,0,1,2,2,3,0,0,1,3,3,0,1,1,2,3,3,0,1, 1,2,3,3,0,1,1,2,3,3,0,1,1,2,3,3,0,1,1,2,3,3,0,1,1,2,3,3,0,1,1,2,3,3,0,1,1,2}; >  >  J  J  >  >  >  >  >  J  static int table32[3] ={0,60,30}; static inttable33[192] ={0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,1, 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,2,2,2,2,2,2, 2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,3,3,3,3,3,3,3,3,3,3,  static inttable3[192] ={1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,2,0, 2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,3,1,3,1,3,1, 3,1A1,3 1 3,1,3,1,3,1,3,1,3,1,3 1,1,3,1,3,1,3,1,3,1 3,1 3,1,3,1,3,1 3,1,3 1,3,1,3A2A2A2A2A2, 0,2,0,2,0,2,0,2,0,2,0,2,0,2,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0}; J  >  >  J  J  >  >  static inttable4[192] ={0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,1,3, 1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,2,0,2,0,2,0, 2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,3,1,3,1,3,1,3,1,3,1, 3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1}; static inttable5[192] ={1,3,1,3,1,3,1,3,1,3,1,3,3,1,3,1,3,1,3,1,3,1,3,1,1,3,1,3,1,3,1,3,1,3,1,3,3,1,3,1,3,1,3,1,3,1,3,1,2,0, 2,0,2,0,2,0,2,0,2,0,0,2,0,2,0,2,0,2,0,2,0,2,2,0,2,0,2,0,2,0,2,0,2,0,0,2,0,2,0,2,0,2,0,2,0,2,3,1,3,1,3,1, 3 13,l,3,ia3a3J,3,l,3,l,3,l,3,3,l,3,l,3,l,3,l,3,l,3,l,l,3,l,3,l,3,l,3,l,3,l,3A2A2A2A2A2, 0,2,2,0,2,0,2,0,2,0,2,0,2,0,0,2,0,2,0,2,0,2,0,2,0,2,2,0,2,0,2,0,2,0,2,0,2,0}; >  A54  static inttable6[192] ={0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,1,3, 13^3,1,3^,3,13^3^,3^3,1,3^,3^,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,1,3,2,0,2,0,2,0, 2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,3,1,3,13,13,13,1, 3,13,13,13,13,13,13,13,13,13,13,13,13,13,13,13,13,13,13,1}; static inttable7[192] ={13,13,13,13,13,13,13,13,13,13,13,13,13,13,13,13,13,13,13,13,13,13,13,13,2,0, 2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,3,1,3,1,3,1, 3^3^3^,3,13,13,13,13,1,3,1,3,13,13,134,3,13^3^3^3,1,3,13^3,1,0,2,0,2,0,2,0,2,0,2, 0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2}; static int table9[192] ={1,3,1,3,1,3,1,3,1,3,1,33,1,3,1,3,1,3,1,3,1,3,1,1,3,1,3,1,3,1,3,1,3,1,3,3,1,3,1,3,1,3,1,3,1,3,1,2,0, 2,0,2,0,2,0,2,0,2,0,0,2,0,2,0,2,0,2,0,2,0,2,2,0,2,0,2,0,2,0,2,0,2,0,0,2,0,2,0,2,0,2,0,2,0,2,3,1,3,1,3,1, 3434,3443434343434334343434343^^ 0,2,2,0,2,0,2,0,2,0,2,0,2,0,0,2,0,2,0,2,0,2,0,2,0,2,2,0,2,0,2,0,2,0,2,0,2,0}; static int table 10[ 192] ={134343434343343434343434343434343434434343434343,2,0, 2,0,2,0,2,0,2,0,2,0,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,2,0,2,0,2,0,2,0,2,0,2,0,3,1,3,13,1, 34343443434343434343434343434334 0,2,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,0,2,0,2,0,2,0,2,0,2,0,2}; static inttablell[192] ={134343434343434343434343434343434343434343434343,2,0, 2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,343434, 3434343434343434343434343434 0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2,0,2};  pulsesequence() { /*define new variables*/ double pw3q,pw3q 1 q,pw90,p90dec,p90decB,tau 1 ,tau2,phase 1;  /*initialize parameters*/ phasel = getval("phase"); pw3q = getval("pw3q"); pw3qlq = getval("pw3qlq");  A55  pshft = 0.5e-6; pw90=getval("pw90"); p90dec=getval("p90dec ); p90decB=getval("p90decB"); taul=getval("taul"); tau2=getval("tau2"); M  /*set phase tables and correct for A P delays*/ d l = d l - pshft; d2 = d2 - pshft; settable(f31,192,table31); settable(f32,3,table32); settable(t33,192,table3 3); settable(t3,192,table3); settable(t4,192,table4); settable(t5,192,table5); settable(t6,192,table6); settable(f7,192,table7); settable(t9,192,table9); settable(t 10,192,table 10); settable(tll,192,tablell); setreceiver(t9); if (phase 1 ==3); add(t3,id2,t3);  /*TPPI phase incrementation. If used, set swl to 2*desired value*/  if(phasel==2); tsadd(t3,l,4); /*Hypercomplex phase incrementation*/ / * i f (phase == 2) { tsadd(t31,1,4); /*Hypercomplex phase sensitivity. Add integer, div. By 4, phase is the integer number left over. Adapted from M Q M A S phase sensitive cycle. Moves magnetization back by 90degrees, and receiver back by 180degrees. Complimentary approach to moving phase of component selection pulse.*/ tsadd(t9,2,4);  A56  }  */ if (dl < 0.0) dl = 0.0; if(d2<0.0)d2 = 0.0; stepsize(1.0,DODEV); /*Phase increments not multiples of 90 degrees. Base is l.Odegrees. Device is Transmitter. Command only associated with dcplrphase command. Anything product of Base and Multiplier that is more than 90degrees is divided by 90 and the leftover is the value of dcplrphase. The next decrgpulse command adds some multiple of 90 to dcplrphase, hence the need for dcplrphase(zero) statements to get back to normal phase cycling.*/ /*begin pulsesequence*/ status(A); dcplrphase(t32); /*Small angle portion of transmitter phase*/ delay(dl); delay(rofl); decrgpulse(pw3q, t31, 0.0, 0.0); /*Adds t l to t2, as: 0+0=0, 0+60=60, 1+30=120, 2+0=180, 2+60=240, 3+30=300*/ dcplrphase(zero); /*Zeroes the non-90 part of the transmitter phase*/ delay(d2); decrgpulse(pw3qlq, t33, 0.0, 0.0); delay((19./12.)*d2); /*Shifted echo for 1=5/2*/ /*begin INEPT pulse sequence*/ status(B); decrgpulse(p90decB,t3,rofl,rof2); /* select one orthogonal component */ delay(taul); /* 1/4J delay */ rgpulse(2*pw90,t6,rofl ,rof2); /*" Simultaneous" 180 pulses*/ decrgpulse(2*p90dec,t4,rofl ,ro£2); delay(taul); /* 1/4J delay */ rgpulse(pw90,t7,rofl,ro£2); /*"Simultaneous" 90 transfer pulses*/ decrgpulse(p90dec,t5,rof 1 ,rof2); delay(tau2); /* Refocussing delay */ rgpulse(2*pw90,tl0,rofl,rof2); decrgpulse(2*p90dec,tl 1 ,rofl ,rof2); /*Refocussing 180degree pulses*/ delay(tau2-0.000150); /* Refocussing delay. Begin acq. at top of echo */ /* Acquisition*/ status(C) acquire(np, 1.0/sw); /* begin acquisition */ }  A57  Appendix IV. Simulation Parameters for 17.6 T Single Pulse and Echo A1 MAS NMR Spectra of USY Materials 27  IV.A Series of Samples Calcined at 550°C  \  Position  Width  brp  (ppm)  (ppm)  (Hz)  -0.7  1.8  -  0.2  -  3.8  -  80  -  0.9  -  86  33.0  -  ^jFwk.Tet  61.2  Br.Tet  59.0  a  (l)  ^jlso.Oct  Al°  A 1  A1  ct  Pe„t  /(l-x)L  Integral Percentage  (kHz)  Single Pulse  Echo  -  3.9  2.0  457  0.3  3.9  3.9  -  729  0.1  8.5  6.8  100  -  758  0.3  1-7  2.6  -  280  -  309  0.0  34.9  28.4  -  95  -  829  0.0  47.1  56.3  "Number of calcinations. * v - 3e qQ/2I(2I -1) Signals are simulated using an average value of v , with 2  Q  Q  distributions accounted for using an exponential broadening function (E ) with a chemical shift component proportional to the magnetic field strength and a quadrupolar coupling component inversely proportional to the magnetic field. Values are estimated to be accurate to ±10%. 'Proportion of Gaussian (G) and Lorentzian (L) character of the peak. The simulation is relatively insensitive to this variable and values are assumed to accurate to ±0.3. m  d  A58  (2)  Integral Percentage  Em  (ppm)  (Hz)  -1.2  1.7  -  0.1  -  3.2  -  80  -  0.7  -  100  Pe„t  31.6  -  jFwk.Tet  59.7  ,Br.Tet  |Iso.Oct  A 1  A 1  A  Oct  (3)  (kHz)  Echo  -  6.8  3.7  451  0.3  5.5  6.1  -  754  0.1  13.3  12.5  100  -  758  0.3  1.8  3.4  -  315  -  319  0.0  31.6  27.9  58.0  _  80  _  818  0.0  40.9  46.4  Position  Width  E  (Hz)  -1.5  1.7  -  0.3  -  3.0  -  90  -  1.7  -  101  Pe„t  32.0  -  jFwk.Tet  59.1  Br.Tet  56.7  A 1  A 1  A1  Oct  Integral Percentage  VQ  m  (ppm)  jIso.Oct  MQ  Single Pulse  (ppm)  A  A  Width  (ppm)  A  A  Position  (kHz)  MQ  Single Pulse  Echo  -  4.0  2.3  470  0.3  8.5  7.4  -  781  0.1  14.1  13.3  90  -  758  0.3  5.1  6.2  -  320  -  346  0.0  27.8  26.7  -  90  -  839  0.0  40.4  44.1  A59  (4)  Position  Width  E  Integral Percentage  m  % x ) L  (ppm)  (ppm)  (Hz)  -1.8  1.7  -  0.0  -  2.6  -  70.0  -  0.4  -  101.0  Pe»t  32.1  -  ^jFwk-Tet  59.1  B,Tet  A  jIso.Oct  Al°  A 1  A 1  £t  (5)  Single Pulse  Echo  -  4.8  3.4  485  0.2  6.8  6.6  -  791  0.1  10.7  9.8  90.0  -  758  0.3  5.3  6.3  -  320.0  -  346  0.0  26.0  23.1  57.2  _  90.0  _  902  0.0  46.4  50.8  Position  Width  E  Integral Percentage  m  (ppm)  (ppm)  (Hz)  % x ) L  (kHz)  -2.0  1.8  -  0.0  -  2.7  -  60.0  -  -0.6  -  90.0  Pe„t  32.7  -  jFwk.Tet  59.2  B,Tet  57.1  A  |Iso.Oct  Al°  A 1  A  (kHz)  A 1  ct  MQ  MQ  Single Pulse  Echo  -  6.6  4.4  490  0.2  6.2  6.9  -  789  0.1  10.0  8.4  90.0  -  781  0.3  5.4  7.6  -  350.0  -  336  0.0  26.0  20.9  -  80.0  -  895  0.0  45.9  51.7  A60  Position  Width  (ppm)  (ppm)  -1.7  1.3  2.5  -  -0.2 Pe„t  (6)  m  Integral Percentage  xG/  Al-x)L  (kHz)  0.0  -  55.0  -  32.3  jFwk.Tet Br.Tet  A  jIso.Oct  A 1  A 1  A  E  A1  Oct  (Hz)  Single Pulse  Echo  -  5.3  3.5  498  0.2  4.7  6.1  85.0  111  0.1  10.8  8.2  -  100.0  851  0.3  6.9  7.9  59.3  -  340.0  346  0.0  23.6  21.1  58.5  -  80.0  950  0.0  48.8  53.2  IV.B Series of Samples Calcined at 650°C  (1)  Width  E  Q  % x ) L  (ppm)  (Hz)  -1.3  3.7  -  0.0  -  0.2  -  76  -  4.0  -  100  Pe„t  32.1  -  jFwk.Tet  60.9  Br.Tet  59.1  jIso.Oct  Al°  A 1  A1  £t  Integral Percentage  V  m  (ppm)  A  A  Position  (kHz)  MQ  Single Pulse  Echo  -  2.2  0.2  536  0.1  6.6  6.4  -  773  0.0  6.0  8.5  110  -  674  0.2  3.1  5.9  -  300  -  346  0.0  28.9  23.8  -  90  -  854  0.0  49.6  50.6  A61  (2)  (Hz)  -1.6  3.3  -  0.2  -  3.2  -  80.2  -  0.5  -  100.2  Pent  33.7  -  |Fwk.Tet  58.3  Br.Tet  jIso.Oct  A1  A1  (kHz)  Single Pulse  Echo  -  1.5  1.9  552  0.4  10.5  7.9  -  825  0.0  10.1  10.3  110.0  -  833  0.6  17.3  19.0  -  420.0  -  363  0.0  22.0  21.5  55.8  _  80.6  _  947  0.0  34.9  39.1  Position  Width  E  Integral Percentage  m  (ppm)  (ppm)  (Hz)  -1.6  3.3  -  0.0  -  3.7  -  95.2  -  -0.6  -  100.2  Pe„t  34.3  -  jFwk.Tet  59.9  B,Tet  57.2  |Iso.Oct  Al°  A 1  A 1  MQ  ct  (3)  A  Integral Percentage  VQ  (ppm)  Al°  A  Width  (ppm)  A  A  Position  (kHz)  MQ  Single Pulse  Echo  -  1.2  1.5  578  0.4  15.2  11.5  -  888  0.0  9.3  9.5  110.0  -  854  0.7  23.7  26.0  -  420.0  -  363  0.0  21.9  21.4  -  65.6  -  953  0.0  26.9  30.1  ct  A62  Position (4)  % x ) L  (Hz)  -1.8  3.3  -  0.0  -  3.4  -  100.0  -  0.1  -  105.0  Pe„t  34.5  - •  |Fwk.Tet  59.2  Br.Tet  jIso.Oct  A 1  A1  ct  (5)  (kHz)  Echo  -  1.4  0.0  570  0.2  13.8  13.6  -  915  0.0  9.7  10.5  110.0  -  888  0.5  29.9  30.9  -  430.0  -  346  0.0  16.1  14.8  58.7  _  60.0  _  1014  0.0  29.0  30.3  Position  Width  E  (Hz)  -1.8  4.3  -  1.0  -  2.7  -  110.0  -  -1.0  -  105.0  Pent  34.9  -  jFwk.Tet  60.0  B,Tet  59.8  A 1  A1  A 1  Oct  Integral Percentage  VQ  m  (ppm)  jIso.Oct  MQ  Single Pulse  (ppm)  A  Integral Percentage  Q  V  m  (ppm)  Al°  A  E  (ppm)  A  A  Width  (kHz)  MQ  Single Pulse  Echo  -  0.3  0.0  589  0.3  9.9  10.3  -  953  0.0  7.2  11.4  95.0  -  972  0.2  28.5  35.7  -  430.0  -  360  0.0  8.6  5.5  -  70.0  -  1048  0.0  29.6  37.0  A63  Position  (6)  (ppm)  (Hz)  % x ) L  (kHz)  -1.6  4.3  -  1.0  -  3.4  -  110.0  -  3.6  -  105.0  Pe„t  35.0  -  jFwk.Tet  58.4  Br.Te,  59.9  jIso.Oct  Al°  A 1  A1  ct  Integral Percentage  VQ  (ppm)  A  A  Width  Me  Single Pulse  Echo  -  0.5  0.0  589  0.3  5.7  5.0  -  953  0.0  5.7  9.8  120.0  -  990  0.2  41.7  47.9  -  430.0  -  394  0.0  6.3  6.5  -  65.0  -  1120  0.0  26.2  30.9  A64  Appendix V. Fractional Coordinates of the Framework Atoms of Calcined, Hydrated ALP0 -i8. 4  Coordinates are derived from the fractional atomic coordinates in crystal structure of calcined, dehydrated ALPO4-I8 transformed by the symmetry elements of the P1 ( C 1 ) space group. The coordinates of All, A12, A13, PI, P2, P3 and 01a-012a are taken from Chapter 7, reference 18. Remaining atomic coordinates are generated.  Fractional coordinates Label  a  b  c  All  0.11000  0.04100  0.16600  A12  0.11800  0.77300  0.93800  A13  0.77000  0.90300  0.04900  AU  0.89000  0.04100  0.33400  A15  0.88200  0.77300  0.56200  A16  0.23000  0.90300  0.45100  PI  0.22700  0.90500  0.05400  P2  0.89100  0.76900  0.93800  P3  0.88300  0.03000  0.16400  P4  0.77300  0.90500  0.44600  P5  0.10900  0.76900  0.56200  P6  0.11700  0.03000  0.33600  Ola  0.82200  -0.05300  0.12900  02a  0.85800  0.83300  0.00200  A65  Fractional coordinates Label  a  b  c  03a  0.00300  0.74100  0.94600  04a  0.15300  0.83200  0.02000  05a  0.18200  -0.04600  0.12000  06a  -0.01000  0.00200  0.15600  07a  0.86400  0.13700  0.13300  08a  0.66800  0.83000  0.07000  09a  0.14100  0.03400  0.25700  OlOa  0.18400  0.65900  0.92500  Olla  0.87200  0.83800  0.87300  012a  0.73800  0.01200  0.99800  Olb  -0.82200  -0.05300  0.37100  02b  -0.85800  0.83300  0.49800  0.00300  0.74100  -0.44600  04b  0.15300  0.83200  0.48000  05b  -0.18200  -0.04600  0.38000  06b  0.01000  0.00200  0.34400  07b  -0.86400  0.13700  0.36700  08b  -0.66800  0.83000  0.43000  09b  -0.14100  0.03400  0.24300  01 Ob  0.18400  0.65900  -0.42500  Ollb  -0.87200  0.83800  -0.37300  012b  -0.73800  0.01200  -0.49800  03b  ......  A66  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share