1997 Ageev

APPLICATION OF DENSITY MATRIX METHODS TO
QUADRUPOLAR SPINS IN SOLID STATE NMR AND NQR
by
Sergei Z. Ageev
Department of Chemistry
McGill University
Montreal, Canada
March 1997
A thesis submitted to the Faculty of Graduate Studies and Research in partial
fulfillment of the requirements for the degree ofDoctor ofPhilosophy.
(:) 1997 by Sergei Z. Ageev

+1 National Library
of Canada
Bibliotheque nationale
du Canada
Acquisitions and Acquisitions et

Bibliographic Services services bibliographiques
395 Wellington Street 395. rue Wellington
Ottawa ON K1A ON4 Ottawa ON K1A ON4
canada Canada
Your file Vo/re refe_
Our file NoIre nffllrence
The author has granted a non L' auteur a accorde une licence non
exclusive licence allowing the exclusive permettant a la
National Library of Canada to Bibliotheque nationale du Canada de
reproduce, loan, distribute or sell reproduire, preter, distribuer ou
copies of this thesis in microform, vendre des copies de cette these sous
paper or electronic formats. la forme de microfiche/film, de
reproduction sur papier ou sur format
electronique.
The author retains ownership ofthe L' auteur conserve la propriete du

copyright in this thesis. Neither the droit d'auteur qui protege cette these.
thesis nor substantial· extracts from it Ni la these ni des extraits substantiels
may be printed or otherwise de celle-ci ne doivent etre imprimes
reproduced without the author's ou autrement reproduits sans son
permISSion. autorisation.
0-612-29870-1
Canada
To my father
Abstract
Spin dynamics in solid state NMR and NQR are studied using spin density matrix
theory. First, the response of spin 7/2 subject to the first order quadrupolar interaction,
excited by one and two pulse sequences is examined. Specific pulse sequences with
appropriate phase cycling designed for detection of MQ coherences developed during
the first pulse are calculated analytically. The results are applied to the determination
of quadrupolar parameters and true chemical shifts utilizing a ID nutation experiment.
Solomon echoes under soft pulse excitation are also considered for spin 7/2. Second,
analytical solutions of off-resonance nutation line intensities for spin 3/2 are presented.
The first order quadrupolar interaction is retained during the pulse. The third case
puts forward a new theory of composite pulses in NQR. Shaped pulses are also
considered. The calculation is valid for a non-zero asymmetry parameter and arbitrary
orientation of the rf field. The results are generalized for half integer spins of
magnitude greater then 3/2.

Resume
La dynamique du spin dans Ie RMN et RQN de l' etat solide est erudiee en utilisant la
theorie de matrice de densite du spin. Le premier cas examine la reponse du spin 7/2
au premier ordre d' interaction quadrupolaire en l' excitant par un ou deux sequences
de pulsation. Des sequences de pulsation possMant des phases cyclees appropriees

pour la detection des coherences de QM developees pendant la premiere pulsation ont
ete etudiees analitiquement. Les resultats sont appliques pour la determination des
parametres quadrupolaires et les relais chimiques reels pendant une experience de ID
nutation. Les echos de Solomon pendant une excitation douce de pulsation ont ete
consideres pour Ie spin 7/2. Le deuxieme cas presente des solutions analytiques des
intensites de ligne de nutation hors-resonance pour Ie spin 3/2. L' interraction
quadrupolaire du premier ordre est retenue pendant la pulsation. La troisieme cas

presente une nouvelle theorie pous les pulsations composes dans la RMN. Les
pulsations deformees ont ete consideres. Le calcul est valide pour un parametre d'
assymetrie non-nul et orientation arbitraire du champs rf. Les resuitats sont generalises
pour les spins demi-entiers de magnitude plus grande que 3/2.
ii
Acknowledgment
It is my pleasure to thank to Professor B. C. Sanctuary for professional and
personal support and guidance throughout my graduate studies in Chemistry.
Many thanks are due to Dr. P. P. Man for great collaborative efforts put forth
in studies of spin 712 dynamics in solid state NMR.
Valuable discussions with Drs. M. S. Krishnan, D. J. Isbister and F. P. Temme
are gratefully acknowledged.
I wish to offer my gratitude to James, Alvin, Mazen and Hikmat for providing
a pleasant research environment.
The financial support of McGill University is also recognized.
iii
Official Statement
The material presented in this thesis has appeared in the literature or submitted
for publication as follows
1. Ageev, S. Z., and Sanctuary, B. C., Analytical solutions for spin 7/2 line intensities
in solid state NMR, Molec. Phys., 84, 835 ( 1995 ).
2. Ageev, S. Z., Man, P. P., Fraissard, J., and Sanctuary, B. C., Determination of
quadrupolar spin coupling for spin 7/2 using two pulse sequences, 1997, Molec. Phys.
(in press).
3. Ageev, S. Z., Man, P. P., and Sanctuary, B. C., Solomon echoes for spin 7/2 by soft
pulse excitation, Molec. Phys., 88, 1277 ( 1996 ).
4. Ageev, S. Z., Man, P. P., and Sanctuary, B. C., Detection of double and four
quantum coherences for spin 7/2 excited by spin lock pulse sequences, 1997, 1 Magn.
Reson. ( submitted).
5. Ageev, S. Z., and Sanctuary, B. C., Analytical solutions for spin 3/2 off - resonance
line intensities in solid state NMR, Chem. Phys. Lett., 255, 71 ( 1996 ).
6. Ageev, S. Z., Isbister, D. J., and Sanctuary, B. C., Composite pulses in nuclear
quadrupole resonance, Molec. Phys., 83, 193 ( 1994 ).
7. Ageev, S. Z., and Sanctuary, B. C., Selective excitation of single and multiple
quantum transitions for spin 7/2 in NQR, Molec. Phys., 87, 1423 ( 1996)
In particular, papers 1 through 4 are presented in Chapters 1 to 4 of Part 2.

Paper 5 is included in Part 3. Part 4 incorporates papers 6 and 7.
The author of this thesis certifies that he initiated all of the above studies,
performed all analytical computations and wrote the manuscripts himself Dr. P. P.
Man was primarily responsible for experimental results of paper 2 which were obtained
utilizing equipment provided by Dr. J. Fraissard. The author assisted Dr. Man with
iv
measurements and processing results of this paper. Dr. Man has checked the results of
computation presented in papers 3 and 4. Dr. D. 1. Isbister co-supervised the author in
a project whose results were published in paper 6 and taught the author about
'MAPLE'.
v
Table of Contest
Abstract 1
Resume 11
Acknowledgment III
Special Statement IV
Table of Contents VI
List of Figures IX
List of Tables Xl11
General Introduction 1
Part 1 Basic Theory 4
Chapter 1. Line Broadening Mechanisms
in Solid State NMR 4
Quadrupolar interaction 5
Dipole - dipole interaction 9
Chemical shift interaction 11
Chapter 2. Density Matrix Method 13
Dirac basis states for a single spin 13
Multispin states 14
Definition of the density matrix 15
Time evolution of the density operator 17
Chapter 3. Rotating Frame Transfonnation in NMR 21
Chapter 4. Density Matrix for a Spin 112 24
References 26
Part 2 Theory and Applications of Spin 7/2

Soft Pulses in Solid State NMR 28
vi
Chapter 1. Introduction 28
Chapter 2. Analytical Solutions for Spin 7/2
Line Intensities in Solid State NMR 30
Chapter 3. Detennination of Quadrupolar Spin
Coupling for Spin 7/2 using Two Pulse Sequences 42
Even and odd quantum detection 47
Rotmy echo type pulse sequence 49
Experimental 50
Conclusions 51
Chapter 4. Solomon Echoes for Spin 7/2 by
Soft Pulse Excitation 55
Summary of echoes 60
Conclusions 68
Chapter 5. Detection of Double and Four
Quantum Coherences for Spin 7/2
excited by Spin Lock Pulse Sequences 70
Density matrix after one pulse 72
Density matrix for spin locking sequences 74
Detection of MQ transitions 76
Conclusions 79
References 83
Part 3 Analytical Solutions for Spin 3/2 Off-resonance

Line Intensities in Solid State NMR 86
Chapter 1. Introduction 86
Chapter 2. Theoretical Consideration 88
Chapter 3. Conclusions 94
vii
References 96
Part 4 Theory ofrfPulses in NQR 97
Chapter 1. Composite Pulses in NQR 97
Single pulses 99
Composite pulses 106
Powder averages 108
Generalization for higher order spins 112
Results and discussion 115
Conclusions 116
Appendix A 118
References 121
Chapter 2. Selective Excitation of Single
and Multiple Quantum Transitions for Spin 712 in NQR 122
Introduction 122
Eigenvalue and eigenfunction problem for spin 7/2 124
Effect ofrf pulses 128
Transitions 1±3) ~ I±l) 130
On-resonance solution and the NQR 1C / 2 pulses 136
Comparison with corresponding spin 112 system 137
Transitions 1±4) ~ 1±2) 139
Conclusions 140
ApendixB 143
References 144
Contributions to Original Knowledge 146
Suggestions for Future Work 147
General Conclusions 148
viii
List of Figures
Figure 2.2.1. The intensity of the central line F 4S as the function ofthe rfpulse length.
Figure 2.2.2. The intensity ofthe first inner line p34 as the function ofthe rfpulse
length. Figure 2.2.3. The intensity ofthe second inner line p23 as the function ofthe rf
pulse length.
Figure 2.2.4. The intensity ofthe outer line p12 as the function ofthe rfpulse length.
Figure 2.2.5.,6. The sixteen amplitudes Lij1;j X 2j as the functions of wQ / WI'
Figure 2.2.7., 8. Log-log plots of the frequencies IWif I/ WI as the functions of wQ / WI
. Figure 2.3.1. SIV spectra of a single crystal BiV04 acquired with one pulse excitation
along positive x axis.
Figure 2.3.2.Fit of 51V experimental line intensities of Fig. 2.3.1 using eq.( 2.3.8 ,
2.3.3 ) with opposite sign.

Figure 2.3.3. SIV spectra of a single crystal BiV04 acquired with two -x pulses with
the time delay between them.
Figure 2.3.4. Fit of SIV experimental line intensities of Fig. 2.3.3 using eq.( 2.3.8 ,
2.3.3 ) with opposite sign.

Figure 2.3.5. SIV spectra ofa single crystal BiV04 acquired with the sequence
{ x }-T-{ x} - [acquisition(y)] -recycledelay
{ {-x }-'t-{x} - [acquisition(y)] -recycle delay
Figure 2.3.6. Fit of 51V experimental line intensities of Fig. 2.3.5 using eq.( 2.3.8 ,
2.3.3) with opposite sign.

Figure 2.3.7. 5IV spectra ofa single crystal BiV04 acquired with the sequence
{ x } -T - { x} - [acquisition (y)] -recycle delay
{ { -x }-T-{ x} - [acquisition ( -y)] -recycle delay'
Figure 2.3.8. Fit of SIV experimental line intensities of Fig. 2.3.7 using eq.( 2.3.8 >
2.3.10) with opposite sign.
ix
Figure 2.3.9. 51V spectra of a single crystal BiV04 acquired with the rotary echo type
pulse sequence {{-x} 1: {x}-[acquisition{y)] - recycle delay}.
Figure 2.3.10 Fit of 51 V experimental line intensities ofFig. 2.3.9.

Figure 2.4.1. Spin echo amplitudes E~2(tl't2' 1:1 =1: 1 3) and E!·2(tl't2 • tl =1: 1 3) as
the functions ofthe second pulse length.

Figure 2.4.2. Spin echo amplitudes E;·3(tl't 2.t l =t 12) and E;·3(tl't 2 ,'t l =1: 12) as
the functions of the second pulse length.

Figure 2.4.3. Spin echo amplitudes E~2(tl't2' tl = 2t 1 3) and E!·2(tpt 2•1:1 = 21: 1 3)
as the functions ofthe second pulse length.

Figure 2.4.4. Spin echo amplitudes E:.4(tl't2' 1:1 = t) and E;.4(tl' t 2•1: 1 = 1:) as the
functions ofthe second pulse length.

Figure 2.4.5. Spin echo amplitudes E;·3(tl't2.1: 1 = 1:) and E;·3(tl't 2•1:1 =: 1:) as the
functions of the second pulse length.

Figure 2.4.6. Spin echo amplitudes E;2{tpt2.tl =t} and E!,2(tl't2.t l = t)as the

Figure 2.4.7. Spin echo amplitudes E;,3{tl't Z .t 1 =3t/2) and E;·3(tl't 2.t\ =31:/2)

Figure 2.4.8. Spin echo amplitudes E~2(tpt2.tl =51:/3) and E!'2(tl't2,'tl =5t/3} as
the functions of the second pulse length.

Figure 2.4.9. Spin echo amplitudes E;.4(tl' 12, 1:1 = 2t} and E;.4(tI'/2' t\ =2t} as the

Figure 2.4.10. Spin echo amplitudes E;,z{t1 ./2 .t 1 = 2t} and E!·2(tl't 2 .1: 1 = 21:) as the

Figure 2.4.11. Spin echo amplitudes E;·3(/1' 12 , 1:1 =51: / 2) and E;·3(tl' 12 , tl =5t 12)
Figure 2.4.12. Spin echo amplitudes E:.4(tpt 2 .t 1 = 3t) and E;.4(/l't2.t1 =3t} as the
x
Figure 2.4.14. Spin echo amplitudes £;.4(tl't2'''C 1 = 5"C) and £;.4(tl't2'''C 1 = 5"C) as the

Figure 2.4.15. Spin echo amplitudes £:.4 (tl' t 2 • "C 1 = 6"C) and £;.4 (tl' t2' "C 1 = 6"C) as the

Figure 3.2.1 a, 3.2.1 b The amplitudes KIej , KIej = LIejZk~ , as functions of 00 Q /00 1 for
the case of L\oo = 00 1 .

Figure 3.2.1 c. The frequencies oolej /00 1 as functions of ooQ /00 1 for the case of
L\oo = 00 1 ,
Figure 3.2.2 a. The amplitudes KIej , KIej = LIejZk~ , as functions of ooQ /00 1 for the
case of L\oo = O.
Figure 3.2.2 b. The frequencies oolej /00 1 as functions of ooQ /00 1 for the case of
L\oo =00 I' Figure 3.2.3 a. The line intensity ~32 as the function of the pulse length t in
units of 1/00 1 for various ratios of ooQ /OOl for the cases of L\oo = O.
Figure 3.2.3 b The line intensity F:2 as the function of the pulse length t for various
ratios of ooQ /00 1 for the cases of L\oo = 00 1 •
Figure 3.2.4 a. F:2 and F',;32 as the functions of L\oo /00 for ooQ = 0 and t = 1.5/00
1 1
,
Figure 3.2.4 b. F:2 and F;2 as the functions of L\oo /00 for ooQ = and t =1/00 1 (0)
1
,
Figure 3.2.5 a. The imaginary part of F{oo) ( eq. ( 3.2.18 )) as the function of 00 / 00 1
for the situation, 7; = 10 /00 1 and L\oo = o.
Figure 3.2.5 b. The imaginary part of F(oo) ( eq. ( 3.2.18 » as the function of 00 / 00 1
for the situation, 1; =10 / 00 1 and L\oo =00 1•
Figure 4.1. 1. The orientation ofthe rf field with respect to the QPAS system ofthe
crystal in the powder sample.

Figure 4.1.2. The composite1l'-pulse of 78 0 -195 182 - 318 12 ,
Figure 4.1.3. The composite1l"-pulse of 860 -160180 -260 9 ,
xi
Figure 4.1.4. The composite1r-pulse of 2600 -161 190 - 854 ,
Figure 4.1.5. The composite 1l" -pulse of 83 90 - 233 191 ,
2
Figure 4.1.6. The composite 1l" -pulse of 38590 - 320270 - 2590 ,
2
Figure 4.2.1. The frequencies of single and multiple quantum transitions IDQas
functions of 11.
xii
List of Tables.
Table 2.1.1. The matrix representation of il / Ii , Al and HAl/ii 38
Table 2.2.2.2. The submatrices in eq. (2.2.6) 40

Table 2.2.3. The components ofthe density matrix ~t) (eq. (2.2.7» 41
Table 2.3. 1.The (I~i) functions used in eq. (2.3.3,2.3.4,2.3.9,2.3.10) 52
Table 2.3.2. The functions C;(t z ) used in the equations
(2.3.3,2.3.4,2.3.9,2.3.10) for the density matrix component P54(tl' 't,tJ 53

Table 2.3.3. The functions A;(tJ used in the equation( 2.3.4 )
for the density matrix component P43 (t l' t, t z) 54
Table 2.4.1. Values Kl with a. = ZliXZi and f3 = 1";;Yzi 69
Table 2.5.1. The density matrix of spin 7/2 after the rf pulse applied
along the y axis 80
Table 2.5.2. The functions D;(t2) and V;(t z) used in the eq. (2.5.9) 81
Table 3.2.1. The matrix form of the Hamiltonian H(O) 95

Table 3.2.2. The components [I,m] located at the I th row and m th
column ofthe density matrix Po(t) (eq. (3.2.11 » 95
xiii
General Introduction
Over last two decades, Nuclear Magnetic Resonance ( NMR ) has become one of
the most powerful analytical tools of the natural sciences with applications ranging from
solid state physics to chemistry, biology, and medicine. With the availability of pulse and
Fourier transform ( FT ) methods, an enormous number of ingenious pulse techniques
have been created to probe the microscopic dynamics of nuclear spins and provide
valuable information on molecular structure. The advent of fast and affordable computers
and the superconducting magnets has led to the development of sensitive and precise
NMR spectrometers. The efficient data processing systems have made it possible to
manipulate nuclear spins in prescribed fashions.

The most numerous NMR are on spins of 112, mainly 1Hand 13e. In multiple pulse
NMR involving such systems the vector model is usually used to visualize the nuclear
magnetization as a vector in 3D space. This model fails to provide an adequate description
for spins of higher magnitude as these require more than two ( 21+ 1 ) quantum mechanical
states to be considered. Moreover, such nuclei are subject to quadrupolar interactions,
which induce rapid relaxation and significantly broaden experimental lines. However, these
can serve as unique probes into electronic environments in solutions and crystals and are
increasingly attracting attention of researches. The object of the problem is to extract
physical parameters and relate them to experimental spectra especially in cases when lines
are broad and featureless. The present thesis partially addresses this problem.
Of the wide variety of theoretical developments, the quantum mechanical density
operator is of significant importance as practically all recent experiments deal with a
statistical mixture of nuclear spins and require consideration of multiquantum coherences
among all possible spin levels. The spin dynamics problem is then shifted to the solution of
the quantum Liouville equation of motion.
1
In this thesis, a spin density formalism is adopted for describing the spin dynamics
of selected spin systems. Part 1 briefly reviews basic interactions in NMR of solids and the
density matrix method which are required for theoretical treatments of experiments
described in Parts 2 to 4.
In Part 2 the following considerations are developed:
( i) The density matrix of spin 7/2 subject to the first order quadrupolar interaction and
excited by an rf pulse is calculated, ignoring rapidly oscillating terms. The results are
studied for a variety of conditions which scale as the ratio of the quadrupolar coupling
constant to the rf pulse amplitude. The line intensities ofthe central and satellite transitions
are also investigated.
( ii ) The calculations are extended further to two in-phase pulse experiments where
pulses are separated by time delay t. The results obtained are valid for any ratio of the
quadrupolar coupling to the amplitude of the rf pulse. It is shown that single and multiple
quantum coherences developed during the first pulse can be detected at the end of the
second pulse. Other pulse sequences with various phase cycling are also discussed. The
results are applied to 51V nuclei in a single ferroelastic crystal of BiV04 .
( iii ) Solomon echoes are calculated for spin 7/2 in solids taking into account the first
order quadrupolar interaction while the pulses are on. The computation is performed using
the algebraic computer program 'MAPLE'. Fifteen echoes are predicted and the amplitude
of each echo is calculated. Each satellite transition produces five echoes whereas no echo
is detected for the central transition. Among these echoes, six are "forbidden" which are a
result of the refocusing of exclusively multiple quantum coherences which are developed
during the first pulse. These echoes cannot be predicted by a calculation based upon
"hard" pulse excitation. The results are valid for any ratio of the quadrupolar coupling to
the frequency ofthe rffield ( (i)Q / (i)1 ).
( iv) The response of a spin 7/2 in solids ( subject to a first order quadrupolar
interaction) to a spin lock pulse sequence is calculated. The results are valid for any ratio
2
of quadrupolar coupling, ro Q ' to the amplitude, ro 1 , of the rf pulse. It is shown here that
the measurements of the central line intensity as a function of the second pulse length can
be used for the determination of quadrupolar parameters using various phase cycling
techniques. It is proven that DQ coherences developed during the first pulse can be
selectively detected.
In Part 3, the density matrix of a spin 3/2 subject to a first order quadrupole
interaction in solids and excited by an off-resonance rf pulse, is calculated analytically
using the algebraic computer program 'MAPLE'. The results are valid for arbitrary values
of the rf amplitude, ro 1 ,quadrupolar coupling, ro Q and offset, Aro and applied to the
analysis of cases where the offset term is none negligible. The treatment can be readily
modified to include a second order quadrupolar interaction during the pulse.
In Part 4,
( i ) it is shown that composite pulses optimized for NMR can also be used for NQR
experiments. Specifically identical composite pulses can be used in both cases and are
most effective for single crystal NQR experiments. A number of composite pulses used for
NMR are calculated exactly for the NQR case and shown to be as good as or better then
existing pulses designed exclusively for NQR.
( ii ) The dynamics of spin 7/2 in pure NQR is considered. The orientation of the applied rf
field is assumed arbitrary and nonzero asymmetry parameter i.e. 11 *- O. For this situation
various selective pulses are considered and analytical results obtained. Shaped pulses are
also treated. A key feature of this work is to construct the interaction representation from
which analytical calculation, using the computer program 'MAPLE', can be obtained in a
tractable form. The results presented are general for all half integer spins of magnitude
higher than 1/2.
3
Part 1
Basic Theory
Chapter 1.
Line Broadening Mechanisms in Solid State NMR
The NMR of liquids is sensitive to two major interactions including chemical
shifts and indirect spin - spin ( J) couplings which can provide valuable information
about molecular structure and conformation. The linewidths in this case are quite
narrow ( I-10Hz) in contrast to solids where the lines are significantly broadened by
other dominant interactions involving mainly quadrupole interaction ( for spin
magnitude>112 ), dipole - dipole interactions and the chemical shift anisotropy. The
quadrupolar interaction in general has values from zero up to tens of MHz whereas the
dipole - dipole interaction ranges from 10 to 100 kHz depending on the nucleus under
observation. The chemical shift anisotropy can be in-between 10 and 100 ppm [ 1 ].
The dipolar interaction itself in solids can mask the chemical shift information
( -1 kHz) and wipes out completely J coupling ( -0.01 - 0.2 kHz ).
The overall form of the Hamiltonian describing a solid is
(1.1.1)
where Hz is the static Zeeman Hamiltonian, Hif is the radio frequency ( rf )
Hamiltonian, and Hint is the Hamiltonian responsible for the internal interactions,
including the quadrupolar interaction, the dipole - dipole interaction, the anisotropic
chemical shift, the spin - rotational coupling, and the J coupling. The first three
contributions to Hint as the most significant, are discussed below.
4
Quadrupolar interaction
The nuclear charge distribution about the center of gravity of the nuclei with
spin magnitude higher than 112 is not spherically symmetric, This gives rise to electric
quadrupolar moments which interact with the electric field gradients ( EFG ) created
by the distribution of charge in the lattice, The electric quadrupolar interaction is found
to be sensitive to local electronic environments and vanishes in the case of cubic

symmetry,
The quadrupolar Hamiltonian in terms of raising and lowering spin operators
for an arbitrary set of axes is given by [ 2 ]
where
Vo =V:z
V±l =Vzr ± iVzy
V±2 = ~(Vxr-V»,)±iVl;)I
The functions Va.jj are defined as Va.1l = a:.~ V , (a, ~ = x,y. z) and V is the electric
r=0
potential due to external sources; the origin is chosen to be the center of mass of the
nucleus,
It is always possible to bring the tensor Va.Jl to its principle axes where Va.p = 0
if a :t= ~, At the same time the potential V is subject to the Laplace equation which
reads at the origin,
(1.1.3 )
5
This introduces the quadrupolar principle axes system ( QPAS ) where the
quadrupolar Hamiltonian takes the form
( 1.1.4)
Here the asymmetry parameter T) and the field gradient q are defined by
(1.1.5)
The principal axes x, y, and z are always labeled in such a way that Vzz is the largest
component of EFG and 1;;::: T) ;;::: O. In the case of the axial symmetry of the field, Vxx is
equal to Vy'vand T) =O. If the nucleus occupies a site of cubic symmetry, then
Vxx =VY.V =r:z which alone with eq. ( 1.1.3 ) makes all three derivatives equal to zero
and the quadrupolar interaction vanishes.
The axially symmetric quadrupolar Hamiltonian is diagonal in the II, M)
representation ( II, M) is the eigenfunction of I z with the corresponding eigenvalue
equal to M). The quadrupolar energy levels then are
( 1.l.6)
It is seen from eq. ( 1.1.6 ) that these levels are doubly degenerate except for M=O in
the case of integer spins. The presence of asymmetry in the quarupolar Hamiltonian
does not lift the degeneracies oflevels of half integer spins due to Kramers' theorem
6
[ 2, 3 ]. This theorem, however, does not apply to integer spins with 11 '# 0 where
degeneracies of energy levels are completely removed. The application of the rf field
to solids with the quadrupolar interaction being dominant can produce resonant
transitions between the quadrupolar energy level. This is called Nuclear Quadrupole
e 2 qQ
Resonance (NQR ). The quadrupolar coupling constant-- can be a single, well
li
defined value, or can be an average over many sites in the heterogeneous
environments. For example, strains and dislocations, as well as other defects in
crystals, can distort the intermolecular interactions causing the distribution of EFG at
the nuclear sites. Moreover, molecular motion, such as torsional motions, can shift the
resonance frequency due to EFG fluctuations [ 4 ]. In general, NQR frequencies are
sensitive to temperature changes and the derivative of the quadrupolar coupling
constant with respect to the temperature is negative. This effect is used in NQR
thermometers [ 5 ]. In addition, the presence of chemically inequivalent sites often
produces multiple lines in NQR spectra.
As was mentioned above, the quadrupolar coupling constant e

2
:Q can be
anything from zero to tens of MHz depending on the nucleus and the crystal in
question. In the case when the quadrupolar coupling is weak in comparison with the
external static Zeeman interaction ( so called "high field case" ), the static magnetic
field determines the direction of quantization and the energy levels of the system are
essentially Zeeman perturbed by the quadrupole. The first order quadrupolar
Hamiltonian which is the secular part ofthe quadrupolar Hamiltonian is given by
nQo) = 8/(21
e2qQ (2(i . 2(i -)[
-1) 3cos .., -1 +l1 sm ..,cos2(i 3/z
2 ( )]
11+ 1 ,
(1.1.7)
where the Euler angles a: and 13 describe the static magnetic field with respect to the
7
( QPAS ). Usually in NMR of quadrupolar half integer spins, only the central
transition, (1/2,-1/2), is observed. The central powder line is quite narrow and can
be selectively observed due to zero quadrupolar line broadening in the first order. The
satellite transitions are shifted by the first order quadrupole and often produce broad
lines of low amplitude. It is useful to note that the first order quadrupolar interaction
does not shift the center of gravity ofthe energy [ 2 ] i. e.
L[3M 2 -I(I + 1)] = o. (1.1.8 )

M
Moreover, the central transition frequency IS not affected by the first order
quadrupole.
In addition to the dipolar interaction, the central transition can be shifted and
broadened by the second order quadrupole which is written as
,...(2)
nil
rL4/(21 -I) J limo 2V
eQ l2 1 { -
Z.-1
- [( ) 2 ] - - [( )
f';./z 41 1+1 - 8/z 1 + 2f';.-2f';.2Iz 21 1+1 - 21z -1
2 ']}
( 1.1.9)
where ~.i> (i = O,±I,±2) are the spherical tensor components of the electric field
gradient in the laboratory frame and m0 is the Larmour frequency produced by the
external static magnetic field. In QPAS, the tensor components are [ 6 ]
P20 = JIeq , P2.±1 = 0, PZ.±2 = eqTt /2 ( 1.1.10)
The spherical tensor components in ( 1.1.9) are chosen because these are transformed
by Wigner rotation matrices [ 7 ] between various coordinate frames. For example,
8
(1.1.11 )
Equation ( 1.1.7 ) can be rewritten in terms of spherical tensor components as
(1.1.12)
The second order quadrupole retains all the information about quadrupolar interaction
parameters. However, it can lead to severe overlapping of lines from different nucleus
sites. It is seen from eq. ( 1.1.9 ) that second order effects decrease with an increase of
the static magnetic field, however for large values of quadrupolar coupling constant
these should be taken into account.
Dipole - dipole interaction

The dipole - dipole interaction is the coupling through space of a spin with the
local field of its neighbors and in general, is given by
(1.1.13)
where m J and m k are magnetic moments ofj th and k th nuclei respectively; r Jk and rjle
are the vectors pointing from the j th to the k th nucleus and its length respectively.
The total number of nuclei is assumed to be equal to N . Quantum mechanically,

magnetic moments m J and m k are expressed as
(1.1.14)
9
where Yj and Yk ; I J and Ik are the gyromagnetic ratios and spin operators ofj th and k
th nuclei respectively. The magnitude of the dipole - dipole interaction depends, as it

seen from eq. ( 1.1.13 ), upon the magnitude and the orientations of dipolar moments
and the distance in space between them. The dipolar Hamiltonian can be expressed in
terms of rising and lowering operators in a spherical coordinate system. This
transformation results in six terms [ 8]. Two of these terms cause zero change in the
spin quantum number (.1m =0), and, to first order, results in splittings of Zeeman
energy levels into many closely spaced levels. This, in tum, produces a distribution of
resonance frequencies and a broad line. The remaining terms cause .1m =±1,±2
transitions. These are of negligible amplitude [ 2 ]. Significant terms are combined
together to give the van Vleck formula [ 9 ] which is the truncated secular Hamiltonian
used to describe the homonuclear spin coupling (Yj = Yk ) ,
Hdhomo
1
= -Ii? ~
2
y (1-3cos
J
2
e.)( 3/. /
Jie - I j . Ik
) (l.1.15)
4 £... r3 JZ fa
}.k=1 ik
i'i"k
It should be noted that the Hamiltonian of eq. ( 1.1.15 ) commutes with the Zeeman
Hamiltonian. Here ejlc is the polar angle ofthe internuclear vector r jk in the coordinate
system where the z axis is defined to be parallel to the static field.

In the case of heteronuclear coupling, (Y j *Y k ), the dipolar Hamiltonian is
further simplified, since the heteronuclear flip-flop term, I j · I k , no longer commutes
with the Zeeman Hamiltonian. Thus the heteronuclear coupling can be described by
the following Hamiltonian,
(1.1.16)
10
Equations ( 1. 1.5-l. 1. 16 ) are the products of spatial and spin parts. For rapid motion
of molecules in liquids with frequencies much higher than those corresponding to

dipolar linewidths, the term 1- cos 2 ejk is averaged to zero making the effective
dipolar interaction vanish. In contrast to liquids, the molecules in solids are restricted
to rotation and vibration modes around their mean position thereby producing nonzero
effective dipolar interaction and consequently, broad lines. The heteronuclear dipolar
interaction of a given spin with others can be described by the single spin Hamiltonian
[ 10 ]
( 1.1.17)
r .(1- 3 cos 2
B.. )
where , = 1'1. ri L}
2
3 JI S jll . Here it is assumed that the system consists
j=l Tjl
of two types of spins namely spins I and S .
Chemical shift interaction

The chemical shift interaction is a second order effect which results in a shift in
the resonance frequency away from the Larmour condition, due to the presence of the
electrons associated with an atom which partially screens the nucleus in question from
external fields. This interaction depends upon the orientation of the molecule with
respect to the static magnetic field applied to the sample and includes both isotropic
and anisotropic parts. The secular part of the chemical shift Hamiltonian in the
laboratory frame is given by
(1.1.18)
11
Here the isotropic chemical shift term, U iso , and the asymmetry parameters of the
tensor,1')a and6 a are given by
and O'xx,O'», andO'zz are the principle components of the chemical shift tensor; the Euler
angles aand ~ describe the transformation from the principle axis frame to the
laboratory frame. For an axially symmetric shielding tensor, 1]0' is zero, whereas for
the spherically symmetric case both 1')a and6 a are zero. During calculations involving
the quadrupolar spins, the isotropic chemical shift is usually included in the Zeeman
term.
The quadrupolar, heteronuclear dipolar, and chemical shift Hamiltonian can be
treated as a "single" spin Hamiltonians. This can significantly simplifY the theoretical
calculations for the systems where these interactions dominate.
12
Chapter 2.
Density Matrix Method
Dirac basis states for a single spin

The angular momentum operators of a single spin Ix,!yandlz are defined by
the following commutation relationships
( 1.2.1 )
where p,r,s can take the values x,y and z are 8 prs is the third rank, totally
antisymmetric Levi - Chivita tensor. The units of Ii == 1 are used for angular
momentum operators,for convenience. Equation ( 1.2.1 ) defines a basis for SU ( 2 )
Lie algebra [ 11, 12] and the angular momentum operators serve as generators for the
corresponding SU ( 2 ) and SO ( 3 ) groups. The Casimir invariant for this SU ( 2 )
algebra (scalar) is the square ofthe total angular momentum 12,
( 1.2.2)
which commutes with each component of the angular momentum vector. Dirac basis
states are defined as the set of simultaneous normalized eigenvectors II, M) of 12 and
Iz . This set is orthogonal, i.e.
( 1.2.3 )
and
Izl/,M) == MI/,M) ( 1.2.4 )
121/,M) == 1(1 + l)I/,M) ( 1.2.5 )
13
The matrix elements of rising and lowering operators, 1± '
( 1.2.6 )
can be defined using the Condon - Shortley phase convention [ 13 ] as
( 1.2.7)
The values of 1 can be integers or half integers and for fixed values of 1, there are
21+ 1 corresponding quantum numbersM i. e. M=-/' -1+1, ..., I. Thus for a given spin
I the states 11,-1),11,-1 + 1),... ,11'/) span a (21 + lr dimensional Hilbert space in
such a way that the single spin operators act entirely inside this space. Equations
( 1.2.3-1.2.7 ) completely define an irreducible representation of the angular
momentum operators for any spin 1. For example, any single spin operator A can be
represented explicitly by a matrix such as
A = L:(I,MIAI1,M')11,M)(I,M'1 ( 1.2.8 )
M,M
where the matrix elements (I,MIAI1,M') can be deduced using eq. ( 1.2.4-1.2.7).
Multispin states
The basis for an arbitrary number of states can be constructed as the
orthonormal direct product of the basis states of each individual spin. For example, for
two spins of magnitude I) and 12 the composite system states are III' M) )1 12 , M2 ) .
14
These are eigenstates of I{ ,I; ,1z1 andlz2 . The total angular momentum operator is
defined as 1= 11 + I z whose components obey to eq. ( 1.2.1 ). This basis

is (2/1 + 1)(2/2 + 1) dimensional. A strongly coupled basis can be obtained using the
Clebsch - Gordon reduction [ 11, 12].
Definition of the density matrix.

The spin density matrix gives the most general physical description of a spin
system state. The density matrix in quantum mechanics plays a role similar to that of
the distribution function in classical statistical mechanics. The purpose of spin
dynamics studies is to calculate the density matrix. Here only some general aspects of
the density operator description relevant to the present thesis are addressed. More
complete review ofthe density operator can be found for example, in ref [ 14 ].
Consider a multilevel spin system described by a complete set of eigenstates

IUk ) of its Hamiltonian. Then the dynamical state of the system is given by a statistical
mixture ofpure states with the following normalized wavevectors
( 1.2.9)
where c/(t) are the probability amplitudes. Each of these pure states can be assigned
a nonnegative time independent statistical weight Pi which is the probability of the
system being in a particular state. The density matrix ( or density operator ) is then
introduced in the ket - bra Dirac's notation as
.o(t)= Lpllv/(t))(v/(t)1 ( 1.2.10)

i
Some important properties of the density matrix are listed below.
15
1) The density matrix is normalized i.e. Tr{p(t)} = 1;
2) The density operator is Hermitian and its each diagonal element is nonnegative,
Pnn ~ 0 ;
3) In general, Tr{p2(t)} 5; 1. For pure state in which all the spins are in the same state,
the density operator is idempotent, i.e. p(t) = p2 (t) with trace Tr{p2(t)} =1;
4) The average value of any observable A is given by the quantum trace
(A) = Tr{ Ap(t)} .
The thermal equilibrium density matrix, ~ 0), is defined by the Boltzman
distribution of the populations of energy levels. In the presence of the large static
magnetic field these levels are described by the Zeeman Hamiltonian. Assuming that all
spins are identical, obtain
( 1.2.11 )
where T is the temperature. The off-diagonal elements in ( 1.2.11 ) vanish according
to Random Phase Approximation ( RP A ) [ 15 ]. In the high temperature limit,
(1.2.12 )
For the spin dynamics calculations it is customary to introduce the reduced thermal
equilibrium density matrix as
( 1.2.13 )
16
Time evolution of the density operator
The time evolution of the density matrix is given by Liouville - von Neuman
equation which reads
dp
ih =[H p] ( 1.2.14)
dt '
where H is the Hamiltonian of the system. Equation ( 1.2.14 ) is a direct consequence
of the Schrodinger equation and is the quantum mechanical form of Liouville theorem
which describes the rate of change of density at a fixed point in phase space. The
formal solution to eq. ( 1.2.14) can be written as
( 1.2.15)
where U(t) is subject to the following differential equation of Schrodinger type
dU(t)
ih--=HU(t) ( 1.2.16)
dt
with the initial condition U(O)=O.
For time independent Hamiltonian H eq. ( 1.2.15 ) gives
~t) = exp(-iHt / h)~O)exp{iHt / h) ( 1.2.17 )
Equation ( 1.2.17 ) can be solved using BCH formula [ 16 ] as
17
( 1.2.18 )
The difficulty with eq. ( 1.2.18 ) is that an infinite number of nested commutators is to
be summed. For some cases computer programs are used to generate commutator
algebra [ 17,18]. Alternatively, the propagator U, U = eXJI...-iHt / Ii), can be evaluated
if the eigenvalue - eigenvector problem is solved for the Hamiltonian H. If R is a
matrix whose columns are eigenvectors of Hand E is the corresponding diagonal
matrix of eigenvalues then
eXJI... -iHt / Ii) =ReXJI...-iEt) R-1 ( 1.2.19)
Note that in eq. ( 1.2.19) the exponent ofthe diagonal matrix is easily computed.
In general, [H(t),H(t')];-t: 0 and the solution to eq. ( 1.2.16) is represented by
( 1.2.20)
Here T is Dyson time - ordering operator [ 19 ]. Equation ( 1.2.20 ) in fact, is a series
expansion ( Feynman - Dyson series) which often cannot be summed exactly.

Truncation of this series leads to the loss of the unitary character of the propagator
U(t). This series converges for nonsingular Hamiltonians, but the convergence is rapid
only for short evolution times or small spectral norm of the Hamiltonian H. However,
Feynman - Dyson series is useful for commonly encountered situations where the
Hamiltonian consists of a dominant time independent part Ho and relatively small time
dependent term H,. (t). The interaction representation is introduced as
( 1.2.21 )
18
Then eq. ( 1.2.14) in this representation reads,
iii t/jJ =
dt
[ii p-]
1>_
( 1.2.22)
where
( 1.2.23 )
Employing disentangling fonnulas for exponential operators, the solution for U(t)
includes increasing order perturbation terms of HI (t). Truncating this series gives
quite accurate results as these are equivalent to those given by time dependent
perturbation theory. A similar type of analysis is used in non - equilibrium statistical
mechanics.
The Magnus expansion [ 20 ] is designed to maintain the unitarity of the
propagator U(I}. The propagator U(I} in this case is sought ofthe fonn
IV
U(t) =exp(!l:t)) , !l:t) = I:n(k)(t) ( 1.2.24)

k=O
with the initial condition

0(0)=0. (1.2.25)
The first two terms in Magnus expansion are [ 19 ]
nO)(t) = -i / lit H(t')dt';

n(1)(t) = -1/ 2(i / Ii) 2 tdt'£'[H(tll),H(t')]dt ll ( 1.2.26)
This expansion is the extension of the BCH fonnula. The Magnus expansion has been
successfully employed in high resolution NMR. In particular, it was used to optimize
19
multiple pulse [21,22] and composite pulse sequences [ 23 ]. However, the Magnus
expansion is effective for the predictions of the dynamics of the spin system at
relatively short times. For longer time scales it leads to convergence problems ( for
example, the Magnus paradox ) [ 24 ] which is the case in the analysis of multiple
pulse spin locking experiments on dipole coupled many body spin systems [ 25 ]. In
general, it can be proved that ~t)exists only in the neighborhood of t=O and the
solution is not global. Moreover, Maricq [26] has shown that the transformation
U{t) 4- ~t) can introduce singularities into the solution for ~t). In general the
Magnus expansion is valid globally only for a Hamiltonian which commutes with itself
at different times. It should be noted that the Magnus expansion provids the theoretical
basis of averaged Hamiltonian theory ( AHT ) [ 21, 22 ] which was recognized as an
effective tool in NMR. [ 22, 27 ]. The essence of AHT is the use of the rf interaction
frame which introduces time dependence into the internal interaction Hamiltonian. The
averaged Hamiltonian theory can simplifY the calculation, however it is limited to
stroboscopically viewing the system at specific times related to a cycle time which
depends upon the pulse sequence under consideration. The information about the
system at any other time is thus lost [ 26 ]. These problems can be solved to a certain
extent by using Krylov - Bogolubov - Mitropolsky approach ( KBM ) [ 28 ] and
Floquet theory [ 29 ] which can serve as alternatives to Magnus expansion. In
particular, the KBM method proved to be useful for the analysis of slow motion during
multiple pulse sequences [ 30 ]; Maricq's [ 26, 31] utilized Floquet theory in the study
of pulsed spin locking phenomenon resulting from the application of multiple pulse
sequences. It is an important consequence of Floquet theorem that the averaged
Hamiltonian expansion exists for the periodic Hamiltonians [ 26].
20
Chapter 3.
Rotating frame transformation in NMR
Consider an isolated spin subject to a static Zeeman field and rotating rf field
given by the Hamiltonian Hif '
( 1.3.1 )
where 00 1 ' 00 and cP are the amplitude, carrier frequency, and phase of the pulse
respectively. All these are assumed to be constants.
The rotating frame is introduced by the operator in analogy with eq. ( 1.2.23 )
to give
Hif =exJA.iootlJHif exp(-iootIJ

( 1.3.2 )
= Ix cos(cP) + Iy cos(cP))
-0)1 (
The equation for the evolution ofthe density matrix in this representation
( see eq. ( l.2.14 ) reads
iii dP
dt =[--]
H,p_
( 1.3.3 )
where the effective Hamiltonian H is
( 1.3.4 )
21
Note that the rotating frame transformation is different from the interaction picture
introduced by eq. ( 1.2.21 ). These, however, do coincide when ID = IDo .
The effective Hamiltonian of eq. ( 1.3.4 ) is time independent and the density
matrix p(t) can be solved using the results of preceding section. Equation ( 1.3.4) is
easily modified to accommodate single spin interactions such as first and second order
quadrupoles, anisotropic chemical shifts, and heteronuclear dipole - dipole interactions

as all these commute with the operator I z and thus are not affected by a
transformation to a rotating frame.
Excitation of spin systems in NMR is often achieved by virtue of linearly
polarized rf fields. Thus the single spin Hamiltonian for a field applied along the x-axis
1S
( 1.3.5 )
Equation ( 1.3.5 ) can be broken into two components corresponding to two
oppositively rotating rf fields as
Hif =H+ +H ( 1.3.6 )
where H+ is given by eq. ( 1.3.1 ) and
( 1.3.7)
Transforming eq. ( 1.3.7) into rotating frame one can write
( 1.3.8 )
22
The component H- is often neglected as it rotates at twice the ro frequency in the
opposite direction in the rotating reference frame ( eq. ( 1.3.8 ». This ( counter
rotating ) component produces negligible effect on the absorbency and if retained,
leads to Bloch - Siegert shifts [ 32 ]. Thus the effect oflinearly polarized field in NMR
is effectively described by its rotating component.
23
Chapter 4.
Density Matrix for a Spin 1/2
Consider an isolated spin 112 described by the Hamiltonian of eq. ( 1.3.4 ) in
the rotating frame. The density operator, in general, is ofthe form
_ = 1(
p(t)
Mz
2 M +iM (1.4.1)
x y
where M = (Mx,My.Mz) is the magnetization vector. It is assumed here that the
equilibrium density matrix is given by eq. ( 1.2.13 ) and M; + M; + M; = 1 for

convenience. The equation for the evolution of the density matrix is equivalent to the
system ofBloch equations ofthe form
~x = -(co -coo)My -COl sin{cj))Mz

~y = (CO -coo)Mx -COl cos{cj)}Mz ( 1.4.2)
~z =COl sin(cj))Mx -COl cos{cj))My
The system ( 1.4.2 ) is easily solved for the rf field ofthe constant amplitude and phase
[ 8 ]. The evolution ofthe spin system under the Hamiltonian which is a linear function
of spin operators gives rise to rotations. Thus the propagator in this case can be
written in the form of a Wigner rotation operator [ 33 ]. This approach leads to the
well known Euler geometric equations of a rigid body [ 7 ] for the corresponding
Euler angles. This system of.equations can be further reduced to a Riccati equation
[ 34 ]. The rotation operator approach is similar to the Wei - Norman technique [ 35 ]
for the Hamiltonian of the form
24
m
H(t) =La"(t)O,, ( 1.4.3 )

"=1
where a" (t) are nonsingular functions of time and the set of operators 0" forms a
closed Lie algebra then the evolution operator can be written as
U(t) =TIexp(c,,(t)O,,) ( 1.404 )

"=1
It should be noted that the equations for Euler angles are identical for spins of any
magnitude I. In fact, the density matrix of eq. ( 104.1 ) can be generalized for a spin of
arbitrary magnitude if the Hamiltonian is a linear function of spin operators as
( 1.4.5 )
where Ms are governed by Bloch equations ( 104.2 ). However, if a spin I is subject to
the interaction which is non-linear in spin operators beyond the rf field, such as a first
order quadrupolar interaction, then the spin dynamics problem must be solved for each
spin separately.
25
References.
1. Jezzard, P, Attard, J. J., Carpenter, T. A, and Hall, L. D., Progress in Nucl. Magn.
Reson., 23, 1 ( 1991 ).
2. Stichter, C. P., Principles ofMagnetic Resonance, Springer - Verlag ( 1990).
3. Das, T. P., and Hahn, E. L., Nuclear Quadrupole Resonance Spectroscopy,
Academic Press, N. Y. ( 1958 ).
4. Baeyer, H, Z. Physik, 130,227 ( 1951 ).
5. Nakamura, D., Ikeda, R, and Kubo, M., Coord. Chem. Rev., 17,281 ( 1975 ).
6. Samoson, A, Kundla, E., and Lippmaa, E., Joum. Magn. Reson., 49, 350 ( 1982).
7. Edmonds, A R, Angular Momentum in Quantum Mechanics, Princeton University
Press ( 1974 ).
8. Abragam, A, Priciples ofNuclear Magnetism, Oxford University Press, Oxford
( 1961 ).
9. van Vleck, J. H., Phys. Rev. 74, 1168 ( 1948).
10. Man, P. P., J. Magn. Reson., 100, 157 ( 1992).
11. Wigner, E. P., Group Theory, Academic Press, N. Y. ( 1959).
12. Hamermesh, M., Group Theory and Its Applications to Physical Problems,
Addison - Wesley, Readings, Mass., ( 1962 ).
13. Condon, E. n, and Shortley, G. H, The Theory of Atomic Spectra, Cambridge,

( 1935 ).
14. Fano, n, Rev. Mod. Phys., 29, 74 ( 1957).
15. Goldman, M., Spin Temperature and Nuclear Magnetic Resonance in Solids,
Oxford University Press, Oxford ( 1970).
16. Baker, H. F., Proc. London Math. Soc., 35, 333 ( 1903 ).
17. Studer, W., J. Magn. Reson., 77,424 ( 1988).
18. Isbister, D. 1., Krishnan, M. S., and Sanctuary, B. C., Molec. Phys., 86, 1517
26
( 1995 ).
19. Blum, K., Density Matrix Theory and Applications, Plenum Press, N. Y. ( 1937).
20. Magnus, Commun. Pure Appl. Math., 7, 649 ( 1954).
21. Haeberlen, U, High Resolution NMR in Solids: Selective Averaging, Academic
Press, N. Y. ( 1976 ) ..
22.Mehring, M., Principles of High Resolution NMR in Solids, Springer - Verlag, N.
Y. ( 1983 ).
23. Tycko, R, Phys. Rev. Lett., 51, 775 ( 1983 ).
24. Abragam, A, and Goldman, M., Nuclear Magnetism: Order and Disorder, Oxford
University Press, N. Y. ( 1982 ).
25. Cantor, R S., Ph. D. Thesis, MIT, Cambridge, MA ( 1979).
26. Maricq, M. M., Phys. Rev. B 25, 6622 (1982).
27. Haeberlen, U, and Waugh, 1. S., Phys. Rev. 175,453 ( 1968).
28. Bogolubov, N. N., and Mitropolsky, Y. A, Asymptotic Methods in the Theory of
Non - Linear Oscillations, Gorden and Breach, N. Y. ( 1961 ).
29. Barone, S. R, and Narcowich, B. A, Phys. Rev. A 15, 1109 ( 1977).
30. Aibinder, N. E., and Furman, G. B., Sov. Phys. JETP, 58, 575 ( 1983 ).
31. Mariq, M. M., Phys. Rev. B 31, 127 ( 1985 ); Phys. Rev. B 33, 4501 ( 1986);
Phys. Rev. Lett., 56, 1433 ( 1986 ).
32. Bloch, F., and Siegert, A, Phys. Rev. 57,522 ( 1940).
33. Zhou, 1., Ye, C., and Sanctuary, B. C., 1. Chern Phys., 101, 6424 ( 1994 ).
34. Warren, W. S., and Silver, M. S., Adv. Magn. Reson., 12, 247 ( 1988 ).
35. Wei, 1., and Norman, E., 1. Math. Phys., 4, 575 ( 1963 ); ibid., Proc. Am. Math.
Soc., 15, 327 ( 1964 ).
27
Part 2
Theory and Application of Spin 712 Soft Pulses in Solid State NMR
Chapter 1.
Introduction
The majority of elements in the periodic table possess nuclei with half integer
quadrupolar spins. The NMR spectrum of a spin of the magnitude greater than 112 in
many solid compounds features several transitions with the frequencies determined by
the static external field and by the interaction between the nuclear electric quadrupolar
moment of this particular spin with the electric field gradient of the surroundings. This
is the case of relatively large (10 2 kHz - IMHz) quadrupolar coupling which dominates
NMR of such systems. The chemical information provided by NMR of quadrupolar
nuclei is not limited to chemical shifts data and also includes the parameters which
describe the quadrupolar interaction. These parameters are sensitive to local symmetry
and carry direct structural information. However, the evolution due to the quadrupolar
interaction should be taken into account while the rf pulse ( typically about 50 kHz ) is
on. These type of pulses are called "soft" pulses. Soft pulses do not give rise to
rotations and require a separate treatment for each spin magnitude.

These quadrupolar parameters can be determined by NQR In the absence of
the external fields, the energy levels of the system are described by the quadrupolar
interaction. The frequencies of the transition between quadrupolar energy levels
provide the quadrupolar parameters. However, the sensitivity ofNQR is rather low
( especially for small quadrupolar couplings ) and it is necessary to cover a large
frequency range in order to search for possible resonances. These problems can be
solved to a certain extent by solid state NMR nutation experiments as nutation of the
central transition magnetization around the strong rf field is modulated by the
quadrupolar interaction [ 1 ].
28
In subsequent chapters the dynamics of spin 7/2 subject to the first order
interaction is discussed in detail. In particular, a number of two pulse sequences with

various phase cycling techniques is calculated. The results are applied to the
determination of the quadrupolar parameters. In addition, all possible Solomon echoes

arising from soft pulse excitation are predicted.
29
Chapter 2.
Analytical Solutions for Spin 7/2 Line Intensities in Solid State NMR
Since the seminal work of Samoson and Lippmaa [ 1 ]. a variety of theoretical
approaches have appeared in connection with solid state NMR dealing with half
integer spins subject to a first order quadrupolar interaction. Spin 3/2 has received the
most attention [ 2-8 ] . For spin 512 attempts to extract structural and dynamical
information in have been made [ 4,9]. Some important results can be found in
[ 10. 11 ]. Recently, Man [ 12] has extended the analytical calculations from spin 3/2
to spin 512. That is while the pulse is on the first order quadrupole is retained. He also
treated Solomon echoes of a spin 5/2 under the same conditions [ 13 ] . In the present
Chapter, the density matrix and spin 712 line intensities after a single rf pulse are
calculated. The results suggest simple experiments for obtaining line intensities as
functions of the rf pulse length which can be used to determination the amplitude of
the first order quadrupolar interaction [ 12 ].
The Hamiltonian for the problem in question is
(2.2.1 )
whereOJo is the Larmour frequency of the central transition,
with the quadrupolar coupling 00 Q '

2
3e qQ ( - -)
) 3cos ~ -1 + l1 sin2 ~cos2a ,
2
ooQ = (
81 21 -Iii
30
is the first order quadrupolar Hamiltonian. Euler angles (i and ~ describe the strong
static magnetic field with respect to the quadrupolar principal axis system ( QPAS ).
The rf amplitude, lU) , the carrier frequency, lU , and the phase, fjJ , describe a pulse.
Transforming into the rotating frame associated with the central transition and
neglecting the offset and high frequency terms results in
(2.2.2 )
where we assume that fjJ = 1&'.
The dynamics of a spin 7/2 system under a pulse along the negative x-axis
(i.e. fjJ = 1&') is described by the density matrix
-tilt tilt
~t) = e II- ~ O)e11 ( 2.2.3 )
with initial condition r1-0) = I z .
The matrix representation of jj is given in Table 2.2.1. It can be diagonalized
by a two step procedure described elsewhere [ 1,12 ]. The transformation Al brings

the Hamiltonian ( 2.2.2 ) to the block - diagonal form H A, -= A/HAl ( Table 2.2.1 )
where each block is a 4 x 4 symmetric matrix. The eight eigenvalues of H A,, given by
the solution oftwo fourth order equations, are
Ell =(~ +DJ 12-rol 12 E21 -= (Rz + D2) 12 + ro l 12

El2 =(RI -D1)/2-ro 1 12 En = (Rz - D2 ) 12+ro l 12
(2.2.4 )
E13 =-(R) -SI) 12-ro) 12 En -= -(~ -S2) 12+ro) 12
E)4 = -(RI +S)) 12-ro l 12 E24 =-(~ +S2)/2+ro) 12
where
31
and
in which
with
The values X Yand Z depend on the interaction parameters ofthe Hamiltonian

J
(2.2.2) as
X1.2 =4(±lCKn Q(l)1 +42ro~+lCKn;)
~.2 = ~+8ro~(l)1 - 32ro Q(l); +640>~) (2.2.5 )
Z],2 = 1~105ro~ + 9<0: ± 74o>~(l)1 + 59(l)~(l)t ± 18ro Q(l)i)
The subscripts 1 and 2 refer to the upper (4 x 4) block and lower block respectively.
Moreover, the upper sign in ( 2.2.5 ) refers to block 1 and the lower sign refers to
block 2. We have put Ii = 1 (frequency units), for convenience.
The components of the normalized eigenvectors associated with these eight
eigenvalues are
32
(2.2.6 )
(2.2.7 )
where i= 1 , 2 , 3 and 4 , the values QIi and Q2i are chosen to normalize each
eigenvector and
-J7 c=J15
A=-OJ
2 1 2 OJ 1
The transformation matrix which diagonalizes HAl and the matrix of eigenvalues are
T= (~° Pz0) ' E

=(E10 E0)
2
(2.2.8 )
33
where the corresponding blocks are defined in Table 2.2.2. The equation for the
density matrix ( 2.2.3 ) becomes
(2.2.9 )
The calculation ofthe density matrix components was performed using 'MAPLE' [ 14 ]
for UNIX on a SUN SparcStation. The results are given in Table 2.2.3. From these
one can determine the intensity of lines ofthe central and three satellite lines.
The total line intensity is given by
(2.2.10 )
Following [ 15 ] , the intensities of the various lines can be obtained using the
operators I';' where
[ y = .J7([12
y
+ [78)
y
+2..fj([23
y
+ [67)
y
+ .J15([34
y
+ [56)
y
+ 4[45
y (2.2.11 )
and I, m take values from 1 to 8 . Hence the intensity ofthe central line is
4 (2.2.12 )
= 2 LLijl;/X2j sinroijt ,
i.j=1
where
The nutation frequencies ro jj are defined as
34
8
....
t.I 2
O.S
Figure 2.2.1. The intensity ofthe central line p.s as the function ofthe rfpulse length in
• (0 (J (0 (J (0 (0
umts of 11 (01 for: 1) - = 0 ; 2) - = 0.5 ; 3) ~ = 1 and 4) ~ = 3.5 .
(0, (01 (01 (01
Figure 2.2.2. The intensity ofthe first inner line pM as the function ofthe rfpulse length
(0 (0 (0 (0
in units of 11 (01 for: 1) ~=O; 2) ~= 0.5; 3) ~= 1 and 4) ~= 3.5.
(0. (01 (01 (01
Figure 2.2.3. The intensity of the second inner line F13 as the function ofthe rfpulse
lU lU lU lU
length in units of 1I tV! for: 1) ~ = 0 ; 2) ~ = 0.5 ; 3) ~ = I and 4) ~ =3.S .
tV! lUI tV. tV)
."k
I
Figure 2.2.4. The intensity ofthe outer line Fil as the function ofthe rfpulse length in
• tVe tVe t» tV
urnts of 11 tV! for: 1) - = 0 ~ 2) -=0.5; 3) ~= 1 and 4) ~= 3.5.
tVl tV! tVa tV1
(2.2.13 )
Note that the sums run over i and j from 1 to 4 only. Referring to Table 2.2.2, the
energies associated with the upper block are Eli and those of the lower block are E 2j
so each block has 4 energies. The operators t: only couple elements between the
upper and lower blocks.
In a similar way the corresponding intensity ofthe first inner satellite line is
(2.2.14)
while that ofthe second satellite inner line is
(2.2.15 )
The intensity ofthe outer satellite line is
(2.2.16 )
The total line intensity for spin 7/2 is simply
F = F 45 + 2F 34 + 2F 23 + 2F12. (2.2.17 )
The line intensities are thus sums of all sixteen frequency terms, each of
different amplitude. These amplitudes and frequencies vary with the relative strength
(j)
of (j)Q and the ratio -2L. Figures 2.2.1-4 show the intensities of the central and
(j)}
35
41
0.4
~
~
£....;':l 0.2
~~
0
-0.2 fiJQ I 411
-0.4
-0.6
0.8
2.5
-0.6
Figure 2.2.5.,6. The sixteen amplitudes LfjruXlj as the functions of OJQ / OJ], The
numbers relate to the subscripts i andj of LijruX2j' Only non-negligible amplitudes are
shown.
41
10.
7• .1.---
-e:::'
_4.
2.
0.1 O.S
Figure 2.2.7., 8. Log-log plots ofthe frequencies 1m, I/ mt as the functions of m" /m
1
.
The numbers relate to the subscripts i and j of 1m, I / m1 .

satellite lines as functions of the pulse length. The intensity of the central line
approaches the following limits
45
F = -16sin{i) j t when {i)Q«{i)t
(2.2.18 )
4sin4{i)l t when {i)Q »{i)j
which show that the amplitude and position change in agreement with [ 1 ]. The factor
of 4{i) l t follows from the general treatment of selective pulses where
(2.2.19 )
with /=712 and M=-1I2 [ 16 ] .

There is a region in Fig. 2.2.1 for t s 0.25 where the intensity of the central
{i)j
transition is a linear function of the pulse length. This is useful for treating powders
[ 12 ]. The other line intensity decreases to zero when (i) Q increases and do not show
similar linear regions, in contrast to the situation for the central transition.
The equations ( 2.2.12-16) can be investigated in the frequency domain. The

Fourier transform of these give sets of sixteen lines located at (i)1j . The amplitudes of
these lines are seen in eq. (2.2.12-16) and vary depending on the ratio {i)Q / {i)l . For
example, for the central transition, the amplitudes LIj1;.;Xzj are shown in Fig. 2.2.5,
2.2.6 as the functions of {i)Q / {i)j . These show the relative importance of the various
(i)
contributions from the nutation frequencies. For example, for ~ ~ 00 only one is
(i)]
(i)
relevant, as expected, whereas for ~ s 0.5 all but seven are negligible. To relate to
{i)l
the previous work [ 1, 12 ], Figures 2.2.7, 2.2.8 show log-log plots of I{i)ij I/ {i)j vs
(i) Q / (i)]. For (i) Q / {i)j S 0.5 seven non-negligible lines are located near (i)]. As the
ratio {i)Q / {i)j increases, these plots show that some lines become less important while
36
others more important. Eventually for (f) Q / (f)l ~ 10 the only one line is important
which is located at 4 WI .
A treatment similar to [ 12 ] for /=5/2 is presented here for spin 7/2. It

provides the starting point for spin echo studies [ 11 ] and a detailed analysis of the
nutation frequencies and intensities for such systems.
37
Table 2. L L The matrix representation of Ii / h , Al and HAl / h.
o o o o o o I
o o o o I
o o o o I
o 2ml o o I
o o o
o I
o
o
o
o
o
o o
-3m Q Jim 1
;1
Troll
o o o o o 7ro Q J
(1 0 0 0 0 0 0 1\
0 1 0 0 0 0 1 0
0 0 1 0 0 1 0 0
~~ ~I ~ 0
0
0
0
1
-1
1 0 0 01
1 0 0
01
10 0 -1
~J
0 0 1 0
'0 -1 0 0 0 10
l-1 0 0 0 0 0 0
38
(
1 700 Q o 0 o 0 o o I
1J7 1
Il-rl'O
2 ..... 1 ..firol o 0 o o I
I 0 o 0 o o I
I I
HAl =' 0
o o 0 o o I
11 I 0 o o 0
.Jl5 o o I
-5ro Q +2ro 1 -2-001
1
I .Jl5 I
I 0 o o 0 TOOl -3roQ ..firol o I
I J71
I 0 o o 0 o "J3ro 1 TOOl I
l 0 o o 0 o 0 700 Q j
39
Table 2.2.2.2. The submatrices in eq. ( 2.2.6 ).
(xu X l2 X I3 X 14 1 (X21 X 22 X 23 X 24 1
~l ~2 ~3 ~41 I 1;1 1;2 1;3 1;4

II
ll-lZ
-'
II Z12 Z13 Z14 J 1\ =l ZZl
7;.)
ZZ2 Zn
Tn
Z24
T24
J
7;1 7;2 7;3 7;4 7;.2
o 01 o 01
;J o
:'J
I
40
Table 2.2.3. The components ofthe density matrix ~t) ( eq. (2.2.7». It is assumed
here that each term has to be multiplied by Lij / 4 and summed over i and j.
Because the density matrix is Hermitian, only its lower triangle is reported.
v
P42 =.lJi
X 2je wyt + T.Ji Z2je -i(J)U t
P44 =2T X 2j COSIDil

li
PS5 = -P44
PS2 =-P71* P8:, = *

-P61 PH:' =-P61* P84 =-PSt*
P86 = -P;l PS7 = -P21• PS8 = -Pn
41
Chapter 3.
Determination of Quadrupolar Spin Coupling for Spin 712 using Two
Pulse Sequences
Quadrupolar spin 7/2 nuclei are often used as NMR probes for the
investigation of local environments in inorganic compounds containing for example,

59Co[17-20], SIV[2I-24],or 139La[25, 26]. The physical parameters such as
quadrupolar coupling constants, asymmetry parameters and the true chemical shifts are
usually obtained from line shape analysis [ 6, 26-28 ] when the second order
quadrupolar interaction dominates. In other cases ID nutation experiments can be
employed. These experiments consist of a series offree induction decays acquired with
increasing pulse length. Fitting of the line intensities to the analytical expression allows
for the determination of the quadrupolar coupling in the single crystal or the
quadrupolar coupling constant and asymmetry parameter in the case of powders.
In the present Chapter the response of a spin 7/2 to two in-phase rf pulses
separated by a time delay is studied following to a second pulse. The analytical

expressions for the central line and three satellites are developed using the computer
algebra program, 'MAPLE' . This represents a step forward towards the understanding
of Solomon echoes [29] for spin 7/2. The results here are verified by a variety of
pulse sequences applied to the nuclei 51V in a single crystal of BiV04 . It should be
noted that the two pulse response of a spin 7/2 as the function of the second pulse
length has more extrema than that of one pulse which makes the fitting procedure
more accurate.
The dynamics of a spin 7/2 excited by two -x pulses with the delay t between
them is given by the density matrix ~ t1.t. t 2) ,
42
~t), 1:,t2 ) =exit.-iiit2 ) exp{-iHg) 1:) eXIA,-iiitJri" 0)
(2.3.1 )
x eXIA,iiit)) exp{iJtd\) exit.iiit2 )
where
and t) and t2 are the durations ofthe first and the second pulse respectively.
The equation for the density matrix ( 2.3.1 ) can be written as
~tl' 1:, t2 ) = Al TeXIA,-;Et2)( Al T) t exp(-iHg)1:),:{t1 ) exp(iHgl1:) AJ TeXIA,iEt2)( Al T) t

(2.3.2 )
where matrices Al ,T,E and ~tJ are given in Chapter 1 and [ 30 ]. The components
of the density matrix ~ t J) can be written for convenience as line intensities using the
(r;-n) notation [ 31 ]. There are 32 independent elements which fully describe the
density matrix ~ t}). Only line intensities of odd quantum coherences change sign
when the -x pulse is changed to a +x pulse and vice versa. Concurrently the
polarizations (I;·n) and the even quantum coherences remain unchanged when the
0
pulse phase is changed by 180 •
The calculation of the density matrix was performed using 'MAPLE'. Since
only single quantum coherences are detected by the experiment only these components
of the density matrix ( 2.3.2 ) are reported below. The component PS4(tl'1:,tJ which
corresponds to the central line is given as
43
+(/;·4 + 1;·5)( C9 sin2ro Q't +CJO cOS2roQt) +(/;.4 - 1;.5)( -Cll sin2ro Qt +C12 cOS2roQt)
+(/;·3 + 1;·6)( Cl3 sin4ro Qt +C14 cos4ro Qt) +(/;·3 - 1;·6)( -CIS sin4ro Qt + C16 cos4roQ't)
+( 1;.4 + 1;·5)( CI7 sin6ro Q't +C 18 cos6ro Q't) +(/;.4 - 1;,5)(-C19 sin6ro Q't + CzO cos6ro Qt)
+(/~·2 + 1~·7)( C21 sin6ro Q't +C22 cos6roQ't) + (I!.2 - 1~'7)( -CZ3 sin6ro Q't +C24 coS6roQt)
+(/!·3 + 1!·6){C25 sin 1OroQ't +C26 cos 1OroQ't) + (I!.3 - 1!·6)(-C27 sin 1000Q't +C28 cos 1Oro Qt)
+(/;.4 + 1;·5)( CZ9 sin 12ro Q't + C30 cos 12ro Q't) + (/!.4 - 1!,5)( -C3I sin 12ro Q't + C32 cosl2ro Q't)
+( 1:.4 + 1;·5)( C cos2mQ't - C
9 lO sin2ro Q't) +(1:.4 - 1;·5)( -CII cos2ro Qt - C12 sin2m Qt)
+(1:,3 + 1:,6){ C13 cos4roQ't-CI 4sin4m Q't) +(1:,3 - 1:,6)( -CIS cOS4roQ't - CI6 sin4ro Qt)
+(/;.4 + 1;·5)( C17 cos6roQt - CI8 sin6ro Q't) +(I;.4 - 1;·5)(-CI9 cos6ro Qt C20 sin6ro Q't)
+( 1~,2 +1~'7)( CZI cos6roQ't - Czz sin6ro Qt) +(!~.2 - 1~'7)( -C23 cos6roQt -C 24 sin6ro Q't)
+(/~,3 + 1~·6)( C2S cosIOroQ't - C26 sin I OroQ't) +(/;3 - 1~,6)( -C27 cos 1OroQ't - CZ8 sin 1OroQt)
+(/;4 + 1~·5)( C29 cos12roQt- C30 sin 12roQ't) +(/;.4 - 1;5)(-C31 cos12mQt -C32 sin12m Qt)}
(2.3.3 )
Note that P54(tl't.tz) is pure imaginary. The subscripts 5 and 4 in P54(/l't.tz ) locate
the density matrix component at the 5th row and the 4th column of the matrix
representation of the density matrix P (I}> 't, t2)' In eq. ( 2.3.3 ) \ ) are omitted but
assumed for each I';.n . These are given in Table 2.3.1. The functions C j
( see Table
2.3.2) depend on the second pulse length while the quantities (I';-n) are the functions
of the first pulse lengths. The expressions for X's, Y' s, Z' s, T s and mjj I s are reported
in Chapter 1 and [ 30 ].
The component ofthe density matrix P43(t1 , t,tJ which corresponds to the
first satellite line is expressed as
44
P43(tl' 1:,t2 ) =_1:,7 (Al +iB1) - 1;,6(Az +iB2)- 1;,S{A3 +iB3)- I;ll( A4 +iB4)
+1;,6(~ +iBs) +1;'7(~ +iB6) +1;,8(~ +iB7)+1:,S(Ag +iBs)
+(1;.4 + 1;,5)(( ~ +iB9)cOS2CO Q1: -( AlO +iBIO ) Sin2co Q1:)
+(/;.4 _ 1;,5)(( All +iBll ) cos2co Q1: +( AJ2 +iB12 ) sin2co Q1:)
+(/;,3 + 1;,6)((AI3 +iBJ3) cos4co Q1: -( A14 +iBIJ sin4co Qt)

+( 1;,3 - 1;,6)(( AIS + iBIS) cos4coQ1: +( Al6 +iB16 )sin4co Qt)
+(/;,4 + 1;,s)((A17 +iB17 ) cos6co Qt -(AI8 +iB18 ) sin6co Qt)
+(/;.4 - 1;·5)({A19 +iBI9)cos6coQt +(Azo +iB20 )Sin6co Qt)

J
+(/;,2 +1;,7)(( AzI +iB2 cos6co Q1:-(Azz +iB22 ) sin6co Q t)
+(/;,2 _/!·7)((~ +iB23)cos6coQt+(~ +iB24 )Sin6co Qt)
+(/!·3 + 1;·6)((Azs +iB25 )cosl<Ko Q1: - (Az6 +iB26 ) sin 1<KoQt)
+(/;3 _ 1!·6)((Az7 +iB27 )cosl<KoQt +(Azs +iB28 ) sin 1<KoQt)
+(/;4 + 1!,5)((Az9 +iB29 )cos12roQt -( A30 +iB30 ) sin 12roQt)
+(/!,4 - 1;,5)(( ~l +iB3 JcoS12co Q1: +( ~2 +iBn) sin12ro Q t)

-t{/;.4 + 1;·5)(-(A9 +iB9)sin2co Qt -(Alo +iB1o ) cos 2ro Qt)
-t{/;.4 - 1;·5)(-(All +iBll )sin2co Qt +(A12 +iB12 )cos2coQt)
-t{/:·3+ 1~·6)( -(Au +iB13 )sin4co Q1:-( Al4 +iBI4 )COS4ro Qt)
-t{/:,3 - 1:,6)(-( Als +iBIs) sin4co Qt +( AI6 +iBI6 )cos4co Qt)
-t{ 1:,4 + 1:,5)(-( A17 +iBI7 )sin6co Q1: - (Au + iBIS) cos6co Q1:)
-t{/:.4 - 1:.5)(-(A19 +iBI9 )sin6co Qt +( Aw +iB20)cos6coQt}
+{/;'z + 1;·7)(-( Azl +iB21 ) sin6co Qt -(A22 +iB22 )cos6coQt)
+{/;2 _ 1;7)( -(~ +iB23 )sin6co Qt +( Az4 +iB24 ) cos6co Q1:)
45
+{I;3 + 1;·0 )(-( A25 + iB2S ) sin l()nQ't -( Au; +iB20)cosl()nQ't)
+(1;·3 - 1;·0)(-( A27 +iB27 ) sinl()nQ 't + (AZ8 + iB 2S ) cosl()ng 't)
+(/;4 + 1;5)(-( A29 +iB29 )sin12ro g't -( A30 +iB30 )cos12ro g't)
-r( 1;4 - 1;·5)(-(A3l +iB3J sin12ro Q't +( A32 +iBn) cos12ro Q't)
(2.3.4 )
The A's in eq. (2.3.4 ) are given in Table 2.3.2. The functions Bj are obtained from
~ by making the following changes
4j ~ Nij = ZJjX2j -1;);j ,

sinroijt2 ~ - cosroijt2 (2.3.5 )
cosroijt2 ~ sinro il 2
The expression for P32 (tl' 't, t2) can be written in the same fashion as
eq. (2.3.4 ) but using
Lij = r;i¥;j + ZJj Z2j ,

(2.3.6 )
N if = r;i¥;j - Zli Z2j
for A'sandB's respectively. Similarly, for PZ)(tp 't,t2 ) one can use eq. (2.3.4) with
the substitutions
Lij = X li Z 2j + 1';; 1;j ,

(2.3.7 )
Nij = X UZ2j - 1';;1;j
for A's and B's respectively.
From eq. ( 2.3.3-7 ) the intensities of the central line and three satellites are
readily obtained. For example, the intensity of the centrallineF45 (t 1 , 't,tJ is
46
(2.3.8 )
Here 1;5 is a spin operator [ 8 ], not a line intensity. The brackets ( ) it used in
eq. (2.3.3-4) are not present in eq. (2.3.8 ).

It should be noted here that MQ coherences denoted by (I'tn) are not
detectable with a one pulse experiment but can be observed indirectly during a two
pulse experiment at the end of the second pulse measuring lQ coherences. Moreover,
the measurements of the central line intensity as the function of the second pulse
length 12 allows for the determination of the quadrupolar coupling ro g for a single
crystal and the quadrupolar coupling constant and 11 for powders. Note that changing
the phases ofthe two pulses along -x axis by 180 reverses the sign ofthe component
0
of the density matrix given by eq. ( 2.3.3 ). This is verified by substituting -(01 for ro l
in all equations.
In most cases only the central line is observed in experiments on powders.
Thus in the subsequent section we focus on line intensity studies of this central
transition.
Even and odd quantmn detection

The central line intensity ( eq. ( 2.3.3, 2.3.8 ) ) is a function of even and odd
quantum coherences. As noted above, odd quantum coherences developed during the
first pulse change sign if the first -x pulse is replaced by a +x pulse. On the contrary,
even quantum coherences do not change sign if the first -x pulse is replaced by a +x
pulse. This is the case for all half integer spins. Thus alternating the phase of the first
pulse without changing the phase of the receiver cancels odd quantum coherences. For
its purpose the following pulse sequence can be employed
47
{ - x } -'t - { - x} - [acquisition (y)] - recycle delay
{ {x } -'t - { -x} - [acquisition (y)] - recycle delay .
As the result the central line intensity is given by eq. ( 2.3.8 ) with
PsAtl' 't,t2 ) = 2i{/;·6CS + 1;·7 C6 + 1;·8C7 + 1:·5CS

+1;·s((C9 +Cll )sin2ro Q 't+(CIO -CI2 ) COS 2c> Q't)
+1;·6((C13 + CIS) sin4ro Q't + (C14 - C16 ) COS4ro Q't)

+1;.4 ((C17 - C19 ) sin6ro Q't +( CIS +C20 ) cos6ro Q't)
+1;·7(( C21 +C23 )sin 6ro Q't +( C22 - C24 )COS 6ro Q't)
+1;·3((C25 - C27 ) sin lOco Q't +( C26 + C2S ) COSIOcoQ't)
+1;·5((C29 +C31)sinI2C>Q't+(C30 -C3JcoS12c> Q 't)
+1;·5(( C9 + Cll )cos2c> Q't -(CIO - C1J sin2C>Q t)
+1:.6 ( (CJ3 + CIS) cos4ro Q't -(C14 - Cl6 )sin4ro Q't)
+1:.4((Cl7 - C19 ) cos6roQ't - (CIS + C20 ) sin6ro Qt)
+1;·7(( C21 + C23)cos6ro Q't - (C22 C24 ) sin6ro Q't)
+1;·3((C2S - C27 ) coslOcoQ't - (C26 + C2S ) sin HkoQ't) (2.3.9 )
J
+1;·5((C29 + C31 )cos12roQt - (C30 - C3 sin12c> Q't)}
Conversely, alternating the phase of the first pulse and the receiver simultaneously
results in cancellation of polarizations and even quantum coherences. Thus the
sequences
{ -x }-t-{ -x} - [acquisition (y)] -recycle delay

{ {x } - t - { - x } - [acquisition ( - y)] - recycle delay
48
can be used for the detection of odd quantum coherences. In this case the central line
intensity is given by eq. (2.3.8) with
P,54(t p 't,tJ = 2i{1:'1CI +1;·6C2 +1;·5C3+1;8C4
+1;.4 ((C9 - ClI ) sin2m Q't +( C10 +C12 ) COs 2m Q't)
+1;·3( (C13 - CIS) sin4m Q't +( CI4 +C16 ) COS4mQ't)

+1;·5((CI1 +C19 ) sin6m Q't +( CIS - C20)cos6m Q't)
+1~·2(( C2l - C23 ) sin6m Q't +(C22 +C24 ) cos 6m Q't)
+1~·6(( C25 +C21 ) sinlOmQ't +( C26 - C2S ) cOSIOmQ't)
J
+1;·4((C29 - C3 sinl2m Q't +( C30 +C32 ) cosl2m Q't)
+1;.4((C 9 - ClI)cos2m Q't -( C10 +C12 ) sin 2m Q
't)
+1:. 3 ( (C13 - ClS ) cos4mQ't - (C14 + C16 ) sin4m Q't)
+1:.5 (( CI7 +CI )cOS6mQ't - (CIS - C20 ) sin6m Q't)
9
+1;·2(( C21 - C Jcos6m Q't - (Cn +C sin6m Q't)

2 24 )
+1;·6(( C25 +C27 ) coslOm Q't - (C26 - C2S ) sin 1OmQ't) (2.3.10)
+1;.4(( C29 - C31 ) cosl2mQ't - (C30 + C32 ) sinl2m Q't)}
Note that eq. ( 2.3.9 ) and ( 2.3.10 ) contain fewer terms than eq. ( 2.3.3 ) and this
results in more efficient data fitting. Moreover, the replacement in these pulse
sequences the pulses along the negative x axis by pulses along the positive x axis and
vise versa results in a simple change of PS4 --t -P54 .
Rotary echo type pulse sequence

Consider the pulse sequence {{x} - 't - {-x} - [acquisition(y)] - recycle delay}
which consists of two pulses of opposite phase and the same duration t I , with time
delay t . The line intensity is given by eq. (2.3.8, 2.3.3 ) with the following changes.
Firstly, II is assumed to be equal 12 . Secondly, due to the fact that the first pulse is an
49
x -pulse, the signs of odd quantum line intensities in eq. ( 2.3.3 ) are changed to
opposite values. This pulse sequence also represents· a way to determine the
quadrupolar coupling.
Experimental
The static 51V NMR spectra were obtained on a Bruker ASX-300 multinuclear
spectrometer operating at 78.9 MHz. The high power static probehead was equipped
with the standard ( 10 mm diameter and 30 mm length) horizontal solenoid coil. The
aqueous solution of NaV03 in a glass tube of the same size as the solenoid coil was
used to determine the length of 1t 12 pulse. The corresponding to 1t 12 pulse ro if 121t
was determined to be 41.7 kHz. Bismuth vanadate crystal, used for the experiments,
is a monoclinic ferroelastic system ( space group, 12/a) with point group 21m at room
temperature. Each spectrum was obtained with a recycle delay of 100s , 8 scans, a
spectral width of 1 MHz and the dead time of the receiver of 8J,.1.s. For the one pulse
experiment (Fig. 2.3.1 ) the pulse length was increased by IJ,.1.s steps from I to 1OJ..I.S.
Two pulse experiments were performed using II = 2J.1S,'t = 10J.lS and 12 was increased
from 1 to 10 J.lSby IJ.1S steps. For rotary type echo experiment the pulse length 11 was
increased from 1 to 10J,.l.s by IJ..1.S steps. The experimental value of ro g 121t determined
from the spectra using two consecutive lines was found to be 55.0 kHz. Figures 2.3.3,
2.3.5,2.3.7 represent the 51V spectra for two pulse experiments corresponding to
eq. (2.3.3,2.3.9,2.3.10). The results for rotary echo type experiment are given on
Fig. 2.3.9. Due to the large spectral width, only the central line in all spectra was
properly phased. Figures 2.3.2, 2.3.4, 2.3.6, 2.3.8, and 2.3.10 show the integrated
experimental central line intensities corresponding to Figs. 2.3.1, 2.3.3, 2.3.5, 2.3.7
and 2.3.9 fitted by theoretical curves. The sets of values (roif 1 21t,ro g 121t)
determined by these five sequences are ( 51.0 kHz , 54.5 kHz) ,

( 51.8 kHz , 54.5 kHz) , (52 kHz , 55.0 kHz ) , ( 50.6kHz , 55.0kHz ) ,
50
.. .. -
BiYO. acquired with one pulse excitation
Figure 2.3.1. 'ly spectra ofa single crystal
along positive x axis.
12
~ -0.5
~ -1
, 2.3.~ )
intensities ofFig. 2.3.1 using eq.( 2.3.8
Figure 2.3.2.Fit of Sly experimental line
with opposite sign and the parameter
s
11 =t =O.CDIf 121t= 5LOkHz and Q

CD 121t =5405kHz.
~I
~.r
....- "'ri"
l af
:I
Ul!~
III .~
~
Ir" :1 .... r
1 :J I'f L'"
.-
.J .A
1
1r-1.,.. ...." r
.J.
_I
I ...ar ~ I" -
aT
J- b- .1
--.""
•
.J-.:::.",.J
.... , •
'"
-r 4r
S
Figure 2.3.3. SIV spectra

•• '.
... -iiliD' ......
ofa single aystaI BiV04 acquired with two -x pulses with the
time delay between them.
"'"":
::s
---...
as
........
--
...;
....
'It
r.-t..
,e-.
10
Oil
is
.-
C
u
10 12
-.5
U
-
ta
....
c
u
·10 .
second pulse length t 2 (J,ls)
FtgUre 2.3.4. Fit ofSlV experimental line intensities ofFig. 2.3.3 using eq.( 2.3.8,2.3.3 )
with opposite sign and the parameters
1\ = 2J.1S, t = 10J.1S,(D" / 2" = 518kHz and (Do / 2" = 545kHz.
- • ... -- ....
Figure 2.3.5. Sly spectra ofa single crystal BiYO. acquired with the sequence
{ x }- t - { X } - [acquisition (y)] - recycle delay - .
{ { -x } -t-{ x} - [acquisition (y)] -recycle delay
12
second pulse length I z(IJS)

51
Figure 2.3.6. Fit of y experimentaJ line intensities of Fig. 2.3.5 using eq.( 2.3.8.2.3.3 )
'I =2J.1S, t =1000000tD.,- 121£ =52.0kHz and tD{l 1211: = 55.0kHz.
-.
Figure 2.3.7. SIV spectra ofa single crystal BiV04 acquired with the sequence
{ x }-'t-{x} - [acquisition (y)] -recyclede1ay
{ { -x }-t-{ x} - [acquisition ( -y)] -recyclede1ay'
•
second pulse length t 2 ( IJ.S)
·20
Figure 2.3.8. FIt ofS1V experimental line intensities ofFig. 2.3.7 using eq.( 2.3.8,2.3.10)
'I =2J.lS. 't =1000000m" 121t =SO.6kHz and CD Q 121t =SS.OkHz.
.. .. i. ... -tliit ... ...
Figure 2.3.9. $Iy spectra ofa single crystal BiY04 acquired with the rotary echo type
pulse sequeoce {f-zl-t-tx}-[acquisitiou(y)]-recyde delay}.
-2
Figure 2.3.10 Fit of sly experimental line intensities ofFig. 2.3.9 using
t = l~mif 121t= 520kHz and mg 12K = 5503kHz.
( 52.0 kHz , 55.3 kHz ) respectively. From these the average value of {OQ /27t is
calculated to be 54.86 kHz ( standard deviation 0.35 kHz) which is less than the value
determined directly from the spectrum by 0.14 kHz. The amplitude of the pulse
determined by the fitting procedure is approximately 20% higher than the experimental
value obtained using an aqueous solution. This can be explained using the fact that the
size of the crystal is significantly smaller than the size of the solenoid coil. Secondly,
there is a difference due to a quality factor between the crystal and the aqueous
solution. The values of (1) Q obtained from the fitting of the experimental data show
good agreement with the value provided directly by the spectrum. Thus our technique
represents a way of determination of {O Q when the satellite lines are not resolved. This
also applies to powders when the MAS method is not applicable. In addition, based on
our technique, the contribution of second order quadrupolar shift to the chemical shift
can be calculated [ 22 ].
Conclusions
The response of spin 7/2 to the variety of two pulse sequences taking into
account the first order quadrupolar interaction is presented. It is shown that single and
multi quantum coherences developed at the end of the first pulse are detected at the
end of the second pulse through single quantum coherences. The results are valid for
any ratio {OQ / {Oif • Fitting to the theory of experimentallD nutation results allows the
determination of the quadrupolar coupling constant. However, the interpulse delay
must be short in comparison with the duration ofFID. In addition, it should be noted
that the ID experiment is technically simpler than the line shape analysis. Finally, our
results represent an important step towards Solomon echo studies for spin 7/2.
51
Table 2.3. 1.The (/:/) functions used in eq. (2.3.3,2.3.4,2.3.9,2.3.10 ). It is assumed
here that each term has to be multiplied by Lfj /4 ( see eq. ( 2.2.12 ) of Chapter 2 )
and summed over i and j .
(/;·8) = 2X 1'zj COSaJil}

jj
(/;·7) =2~iZ2j cosaJfjt}

(/;·6) = 2Zjj 1';j cosaJfjt}
(/;.5) = 2TJ.jX2j COSaJfjt}
(/;,8) 2X}j1; j sinaJfjt)
(/~·1) = 2~;Z2j sin aJfjt)
(/~·6 ) = 2ZJ.jY2j sin aJfjt}
(/;,5) = 21;;X2j sinaJfjt}
(/;·2) +i(/;·2) =(X}jZlj + ~i1'zJCOSaJift) +i(XliZZj - ~J;j ) sin aJfjtl
(/;·3)+i(/;·3) = (XJj1';j +ZIi1;j)cosaJfjt) +i(X}j1';j -Zli1'zj)sinaJfjt1
(/;.4) + i(/;.4) =(XliXlj + 1;; 1;j )COSaJift1 + i( Xli X lj - 1;; 1;} )sinaJijtl
(/;,5) + i(/;.5) = (XJjXlj - 1;; 1;j )cosaJjl1 + i( XjjXlj + 1;;1;j) sinaJfjtJ
(/;·6) +i(/;,5) =(Xl/1';j - ZJj1;j )COSaJ;/I +i(XI/1';j + Zli1;j )sinaJi/ 1
(/;·7) +i(/;·7) =(XUZlj - ~;1;j )cosaJijtJ +i( X Z2j + ~i1'zj )sinaJij.tJ
U
(/;,3) +i(/;,3) = (~)';I + Z\iZ2/)cosaJ!A +i(~);j - Zu Z2JsinaJ;A

(/;,4) + i(/~.4) = (~IX2j + 1;jZ2i )COSaJiftl + i(~iX2i - 1;jZli) sinaJijt}
(/;,5) + i(/~'S) = (~jX2j 1;jZ2j ) COS aJijt} + i(~iX2j + 1;;Zlj ) sin aJfjtl
(/;,6) + i(/~·6) = (~j1';j - ZIiZli )COSaJfjtl + i(~t1';i + ZliZ2i) sinaJfjt}
(/;.4) + ;(/:.4 ) = (ZliXlj + 1;11';j) COS aJij tl + i(Z}1 X 2i - 1;;1';i) sin aJ;ll
(/;·5) +i(/~·5) = (Zli X 2j - 1;;1';j )cOSaJijtl +i(ZlIX2i + 1;11';j )sinaJijtJ
52
Table 2.3.2. The functions Cj ( / 2 ) used in the equations ( 2.3.3, 2.3.4, 2.3.9, 2.3.10 )
4
for the density matrix component PS4 (II' t, 12)' The symbol L
l.j=1
in front of each term is
omitted but assumed.
CI = r;,1;jX2jZ2j COSID,l2 Cs = Zli1;j X 2j J;j sinIDijtz

C2 = Zli1; j X 2j 1';j COSIDijl2 Co =r;j1;; X 2j Z 2j sinmyf2
C3 =1;;1;,X2j X 2j COSIDij/2 C7 = XJj 1;; X 2j 1;j sinIDij/2
C4 Xli 1;; X 1j 1;j cosmijl2 Cg = 1;j1;jX2jX2j sinmij/2
C9 = Zli1;,X2j X 2j cosmijl2 CIO =Zli1;jX2jX2j sinID j/ 2
CI1 = 1;i1;,X2j 1';j cosmij/2 CI2 = 1;; 1;1 X 2j 1';j sinmijtz
Cn = r;,1; jX 2j 1';j COSIDijt2 CI4 =r;i1; j X 2j1';j sinIDijt2
CIS = Zli1;jX2jZ2j COSIDijlZ CJ6= ZJj1;jX2jZ2j sinmijt2
Cl7 = r;i1;,X2jX2j cosm l/ 2 CIS = .r;i1;;X2j X 2j sinmijt2
Cl9 = 1;,1;jX2jZ2j cosmijl2 C20 = 1;;1;;X2jZ2j sinm i/ 2

e21 = Xli1;;X2jZlj cosmijlZ C22 = Xli 1;1 X 1j Z 2j sinm i/ Z
C23 =r;i1;,X2j 1;j cosIDij/2 C24 = .r;i1;j X 2j 1;j sinID j/ 2
C25 = XJj 1;1 X 2j 1';j cosmijt2 C26 = XJj1;; X 2j 1';j sinIDij/2
CZ7 = Zu1;,X2j 1;j COSIDijt2 C28 = ZliI;.,Xzj 1;j sinID;/z
C29 = Xli1;,X2jX2j cosm,/2 C30 = Xli 1;;X2j X 2j sinmij/2
C31 = 1;; 1;,X2j 1;j COSIDijll C32 = 1;,1;;X2j 1;i sinm;lz
53
Table 2.3.3. The functions ~(tz) used in the equation( 2.3.4 ) for the density matrix
1 4
component P43(tl' "C,tJ. The symbol 2'~ in front of each term is omitted but
I,J=I
Al = LijY;jZZj sinro;/z As =LijZli~j cosroijtz

A2 = LijZIj~j sinmijtz A6 = LijY;jZzj cosroijtz
A3 =Lij~iXZj sinro;/z A7 = LijX}j1;j cosmijt z
A4 = LijXli1;j sinroijtz As = Lij~jXzj cosmij12
A9 = LijZ]jXzj cosroijtz AlO = LijZJjXzj sinroij12
All = Lij~j~j cosmijtz Al2 = Lij~l~j sinro i/ 2
AI3 = LiiY;;~j cosmijt z Al4 = 4Y;j~j sinroij12
A15 = LijZliZ2j cosro;/2 Ale = Lij Z li Z2j sinroij12
An = LijY;jXzj cosro j/ z Au LijY;jX 2j sinroij12
A19 = LijT]jZ2j cosroijt2 A20 = Lij~iZ2j sinroijtz
A21 LijXJjZ2j cosmijt z A22 = LijXljZ2j sinroijtz
AZ3 = LijY;j 1;j cosroijtz A24 = LijY;j 1;j sinroijlZ
A25 = LijXIiYzj cosroijt z Aze = LijXli~j sinm j/ 2
A27 LijZ]j 1;j COSroijtl A28 = LijZJj 1;j sinroijtl
A29 = LijXljXzj cosmijtz A30 = LijXli X 2j sinroijt2
A31 = Lij~i1;j cosroijt z A32 = Lij~j1;j sinroyiz
54
Chapter 4.
Solomon Echoes for Spin 712 by Soft Pulse Excitation
Recently, use of computer algebra has made it possible to calculate Solomon
[ 13 ] and Hahn [ 33 ] echoes for soft pulse excitation for higher spin values ( e. g. up
to 5/2's ). The calculations retain the first order quadrupolar interaction while the pulse
is on. Six Solomon echoes are predicted for spin 5/2. Two of these are the result of
refocusing exclusively multiple quantum ( MQ ) coherences generated by the first
pulse. There are no echoes predicted for the central transition. Here we extend this
type of calculation to Solomon echoes for spin 7/2. The Solomon echo pulse sequence
includes two rfpulses separated by a time delay [29]. This sequence can be applied to
static powder samples to generate the spectrum due to satellite transitions which are
usually lost in the dead time of the receiver during one pulse experiment. The time
delay must be short in comparison to the FID of the central transition in contrast to
Hahn echoes [ 34 ]. Considering soft pulse excitation allows the prediction of the
fifteen echoes for spin 7/2. If the first order quadrupolar interaction is ignored, only
nine echoes are obtained. These arise from "hard" pulse excitation and are therefore
called allowed echoes. The six others, are a result of the refocusing of multiple
quantum coherences developed during the first pulse and depend upon the presence of
a first order quadrupole. We call these IIforbiddenli • Allowed echoes are bell shaped
whereas forbidden ones are the derivatives ofbell shaped curves being sine shaped
[ 4 ]. The origin offorbidden echoes can be understood using the Zassenhaus
[ 35 ] formula for the propagator while the pulse is on as
55
The first exponential term on the RHS gives rise to rotations whereas the second one
is responsible for the quadrupolar evolution. The remaining terms which are in fact
multiple commutators produce forbidden echoes [ 35 ]. Here Hgl and Hif are rotating
frame time - independent quadrupolar and rfHamiltonians respectively. The use of the
spin echo pulse sequence in NMR is considered in detail in [ 36 ]. Solomon echoes
arise from the refocusing of off-resonance single and MQ transitions. They are one
dimensional in contrast to MQ MAS NMR [ 37 ] where multiple pulse sequences
employing complex phase cyclings are used to refocus on-resonance MQ transitions
generating powder spectra due to the central transition. These techniques are
essentially two dimensional.
The components of the density matrix necessary to study echoes are calculated
here analytically starting from the equilibrium density matrix ~O) = lz . From these the
amplitude of each echo can be determined as a function of both pulse lengths, the time
delay and the ratio of ro Q / ro 1 • The results are valid for any ratio of ro Q / ro I and thus
the measurement of any echo amplitude as a function ofthe second pulse length allows
the determination ofthe quadrupolar coupling constant.
The spin echo pulse sequence in question consists oftwo soft rf pulses of
durations t 1 and t 2 applied along the negative x axis with an interpulse delay 't
followed by the evolution period tl .While the pulses are on, the system is described in
the rotating frame by the Hamiltonian ofeq. (2.2.2.). During the periods t and tl the
system evolves under the Hamiltonian Hgl . The effects of dipole - dipole and second
order quadrupolar interactions and relaxation are not considered in the present
Chapter.
The starting point for the calculation of Solomon echoes is the density matrix
components related to lQ coherences i. e. PU-l(tl't,t2 ,tJ wherek runs from 5 to
2. The subscripts k. k - 1 in PU'-l (t l' t, t2 ' t I) denote the k th row and k-l th column of
the density matrix PU-I (tl' t,t z , tJ

56
The density matrix for the two pulse experiment with the time delay between the
pulses P (tl' "t,t2 ) has been calculated in Chapter 3 and [ 38 ] taking into account the
first order quadrupolar interaction while the pulses are on. The evolution ofthe density
matrix during the acquisition time "t J is described by
(2.4.2 )
The required components ofthe density matrix P (t1''t, t2 ' 't 1) deduced from
eq. ( 2.4.2 ) are
P54(tl'''t,t 2 ,''tJ = P54(tp "t,tJ,

P43 exp( 2iCD t I)'
(t p "t,t2 , "tt) = P43 (tjl t, t 2 ) Q
(2.4.3 )
P32 (tl' t,t tJ = P32 (tl' t.tJ exp{ 4iCD t1)'
2• Q
P21 (tl' t.t2' tt) = P21 (tl' "t. t exp( 6iCD Qtl)'
2)
The intensities ofthe central line and satellites F:· k - 1 and Fyk •k - 1 can be defined as
(2.4.4 )
Note that the central line intensity ( see eq. ( 2.4.3 ) ) does not depend on tl and
F::4 •5 (tl' t,t2 , t l ) = 0 . Practically this means that the echo cannot be predicted for the
central transition and the interpulse delay 't must be much shorter than the time of the
free induction decay of this transition. This experimental condition is general for all
half integer spins [ 13 ].
Equations ( 2.4.3 ) are lengthy when developed using 'MAPLE'. Thus only the
contributions to the line intensities related to echoes are reported below. For the first
inner satellite line ( II-S ) the equation for the intensity is given by
57
F;.4(I\, 1:,12,'[\) +iF:.4(lp 1:,12,1:\) =:
{(I:.4 +1;·S}K1 +(I;.4 1;·5)K2-(1;.4 +1;")K3 +(1;.4 -1;·5)K4}exp{2iID Q{1:\ -1:))

-i{(I;.4 + 1;·5)K (/;.4 - 1;·5)K4+ (/;.4 + 1;·5 )Kl + (/;.4 - 1:" )K2 }eXp{2iID Q{1:) - 1:))
3 -
+{(I;.3 +1;·6)K, +(1;,3 - 1;,6)K6- (1~,3 +1~·6 )K7 +(1~.3 1~·6)Kg }exp(2iIDQ{1:1 - 21:))
-i{(I;,3 +1;·6)K7 _(I;.3 -1;·6)Kg+(1~.3 +1:·6)Ks +(1~.3 _/~·6)K6}eXp{2iIDQ(1:1 -21:))
+{(I;.4 +1;·5)K9+(1;.4 -1;·5)KIO _(1~.4 +/~·5)KI1 +(1~.4 _/~·5)K12}exp{2iIDQ(1:1 -31:))
_i{(I;.4 +1;.5)Kll -(1;.4 -1;·5)KI2 +(1~.4 +/~·5)K9 +(1~.4 _/~·5)KJO}exp(2iIDQ(1:1 -31:))
+{(I!.2 +1;·7)K 13
-/!·7)KI4 _(1;.2 +1;·7)K15 +(/;·2 _1;7)KI6}exp{2iOJ Q(1: 1-31:))
+ (I!.2
-i{(I!·2 +1;7)KI5 _(1;.2 -/!·7)K)6 +(/;·2 +1;·7)KI3 +(1;.2 -1;·7)K14 }eXp(2iID Q(1:\ 31:))
+{(/1.3
x
+/1.6)K
x 17
+{/x1•3 _ /x1•6)K18 _{/1.3
y
+/)·6)K
Y 19
+(/\·3
Y
-/l.6)K
Y 20
}exp(2iOJ (t I -51:))
Q
_i{(/x1. +/1.6)K
3
x 19
_(/1.3_/1,6)K
x x 20
+{/1.3+/1.6)K
y Y 17
+(/1,3_/1.6)K
Y Y 18
}e--{2iOJ
kp\
(1: 1 -5t))
Q
+{(/x1A + 11.5)K
x 21
+(/1.4
x
_11.5)K
x
_(/104
22y
+ 11,S)K
Y
+(/1.4
Y 1!,5)K24 }exp{2iID Q{1:1 - 61:))
23
_i{(/1.x 4+/x1,S)K23 _(/1,4_/1.5)K

x x 24 +{/y1.4+/1.5)K
Y 21 +(/I.4_/l.S)K
Y Y 22 }eX.J2iID
f'\. (1: 1-61:))
Q
( 2.4.5)
In this equation, for the sake of compactness, the average brackets, ( ), are dropped.
Hence each r;·n should read (r;-n) and are functions which depend on the first pulse
length II and are given in Chapter 3 and [ 38 ]. The values K[ , 1=1, ... ,24, are
functions of the second pulse length I z. These are reported in Table 2.4.1. Note that
KI = K z and K3 = K4 .
For the second inner satellite line (21-S ) eq. ( 2.4.5 ) can be used for the line
,'tJ
intensity F:.2•3 { Ii' t,12 , t l ) +iFyz•3 {11' t, 12 with the following substitutions
K) ~ F; ,1: 1 ~ 2t l · The values F; are identical to the Kit s in Table 2.4.1 except that
ex.and ~ are given by ex. = r;i1'; j and ~ = ZJjZ2j . Note that Fs = F6 andF, F'g.
58
Similarly, for the outer satellite line ( o-S ), F;,2(/1' 't,t2,'tJ +iF;,2(tp 't,t 2 , 'tjL
one can use eq. ( 2.4.5 ) changing K, ~ G" 't J ~ 3't J and substituting
a. = XJjZ2j and ~ = 1';/1";j for the values in Table 2.4.1. Note that in this case
G13 = G J4 and GI5 = Gu; .

The echoes corresponding to Ftl,k(tl' 't,/2•'tJ +iF;-l,k(/1''t,12,'t J ) with k=3
are seen to twice as broad as those with k=4, whereas echoes with k= 2 are three times
as broad as those with k=4. This is seen from eq. ( 2.4.3 ).
In quadrature detection mode one can observe fifteen echoes in both channels.
The echoes described by F:· 4 (/1''t,t2 ,'tJ+iF:.4(tl''t,t2,'t 1) are located at 't l ='t,
't J = 2't, 't 1 = 3't, 't I = 5't and 't 1 = 6't. The echoes which correspond to
F;,3(t p 't,/z' 't l ) +iF;,3(tl' 't,/2,'t l ) are detected at 't l = 't /2, 't l 't, 't 1 = 3't /2,
't I = 5't / 2and 't 1 = 3't . For the outer satellite line with the line intensities
~t.2(/1''t,t2,'tl)+iF;·2(/1''t,t2,'tI) the echoes are at 't J ='t/3, 't J =2't/3, t1 =t,
't J =5't /3and t J =2't. These echoes are the result of the refocusing of off
resonance coherences developed during the first pulse. Each echo includes two
contributions responsible for the shape of the echo. The first one is related to the sine
function which gives sine type shape echoes while the second one is related to the
cosine function which gives symmetrical bell shape echoes. The contributions of the
sine type are equal to zero at the positions where bell shape contributions have
maximums and thus the maximums of bell type echoes are not modified by sine type
echoes. However, the shapes ofthe echoes can vary ranging from a symmetrical bell to
a sine type shape depending on the experimental conditions.
The equations for the amplitude of each echo are given below. As a rule [ 13 ],
measuring the amplitudes of the echoes as the functions of the second pulse length
allows the determination of the quadrupolar coupling ()) Q for single crystals or
2
e qQ
quadrupolar coupling constant -ft- and asymmetry parameter 11 for powders.
59
Summary of echoes
1) The amplitudes ofthe first echo located at t 1 =t 1 3 are given by
E1•2 (t t2' t 1 =t 13) =_{{/x3.4 + 1x3•S )G3 - {Ix3.4 - 1x3•5 )G4 + (/3.4 + 13•S )G1 + (/3.4 - 13,5)G2 )
y 1> y Y Y Y
(2.4.6 )
and
Ex1•2 {t l' t2' t 1 =t/3) ={Ix3.4 +/x3•S )G1 +(/3.4_/

x x
, )G _{/3.4+/ ·')G +{/3.4_/ • )G
3S
2
3
3 y
3S
Y 4 Y Y
(2.4.7 )
Figure 2.4.1 shows graphs of E;·2(t p t 2,t 1 =t/3)andE;2(tl't2 ,t 1 =t/3} as
2) The amplitudes of the second echo located at t J =t I 2 are given by
35 3S 1
Ey2•3 {t 1> t 2' t =t/2)=-({/3.4+/3.5)F _(/3.4_/ • )F +(t3.4+/ . )F, +(/3.4 l y3.5)F2)
1 x x 3 x x 4 y Y Y
(2.4.8 )
and
Ex2•3 {t P t 2' t J =t 12) = (/3.4

x
+ 1x3•5 )F,1 + {Ix3.4 - 1x3•5 )F2 _{/3.4
y
+ 13•5 )F3 +(/3.4 r.s)F4
Y Y Y
(2.4.9 )
Figure 2.4.2 shows graphs of E;·3(tl't 2 .t 1 =t/2)andE~·3(tl't2,tl =t/2) as
60
0.4
,
M
...
II
0.2
...
lii
-8
.-c.2
i0
..c -0.2
H
til
I
0
-0.4
Figure 2.4.1. Spin echo amplitude E;·2(11)12,''(1 =1: 13), CD I 121f. =50kHz, as a
function of the second pulse length ' 2 (fJS) for 1, 'I =5fJS'0)(2121t = OkHz;2,
I) =3fJS'O)QI21t=50kHz;3, Spin echo amplitudeE!·2(/1";p1: 1 =1:/3).
-I 121f. = SOl:Hz. for tl =3fJS.CDQ 121t = 50/cHz .
,
N
... 0.4
II
...
-0.4
Figure 2.4.2. Spin echo amplitude E:'lt p / 2 ."C 1 =1:/2}. CD l /2fC=50kHz, as a
function of the second pulse length '2 (fJS) for 1, 'I = 5fJS.CDQ 121t = okHz ;2,
Spin echo
0)1 121t = 50kHz, for '1 =3fJS,CD(2 121t = 50kHz.

These two echoes are the result of refocusing the 1I-8 (3 B 4)and 2Q(3 B 5)
coherences.
3) The amplitudes ofthe third echo located at 't J :::: 2't 1 3 are given by
(2.4.10)
and
£1,2(t
x
t 't :::: 2't 13) :::: ([2,3
I> 2' I x
+ [2,6)G
x S
+([2,3 _ [2,6)G _{[2.3 + [2,6)G +([2,3 _ [2.6)G
x x 6 y Y 7 Y Y 8
(2.4.11 )
This echo is the result of refocusing the 21-8 (2 B 3)and 4Q (2 B 6) coherences.

Figure 2.4.3 shows graphs of £!.2(t p t 2.'t J =2't/3)and£~'2(tl't2,'tl =2't/3) as
4) At the time 't 1 = 't three overlapping echoes of different widths are detected.
The amplitudes ofthe echo related to the l1-S are given by
(2.4.12 )
and
(2.4.13 )
This echo is the result of refocusing the lI-S (3 B 4)and 2Q (3 B 5) coherences.
61
0.8
t"'l
' 0.6
t..t
~
II 0.4
ti
4) 0.2
"0
.a
1
.B -0.2
~
o
CIl
-0.4
~O.8
Figure 2.4.3. Spin echo amplitude E;2(/»/2.1:1 =2t/3), ml /21C=SOkHz, as I
function of the second pulse length '2 (~) for 1. II =5~.coe 12ft =Okllz;2,
Spin echo
ml 121t =50kHz, for 'I =3J.LS'.me 121t =50kHz.
1
t..t
II
t..t
ti
4)
"0
.-.a
J
.Bu
4)
CIl ~O.S
-
~
Figure 2.4.4. Spin echo amplitude E;·4(t../2 .1: 1 =t}. 0)1/21t=SOkHz, as a function
of the second pulse length '2 (J.LS') for I, I] = 5J.LS'.(i)Q 121t = OkHz;2,
I] =3~'O)QI21t=50kHz;3. Spin echo amplitudeE!·4(/1'/2 ,1:] =1:), G)\/21t=50kHz.

for '1 = 3J.LS"O)Q 12K = 50kHz.
the second pulse length.
5) The amplitudes ofthe echo related to the 2Iw8 are given by
(2.4.14)
and
(2.4.15)
This echo is the result of refocusing the 21-8 (2 ~ 3) and 4Q (2 ~ 6) coherences.

Figure 2.4.5 shows graphs of £;.3(tl' t 2 , 't 1 't) and £;.3(tl't2' 't 1 = 't) as functions of
6) The amplitudes of the echo related to the 0-8 are
£1.2{t
Y
t 't = 't)
l ' 2' 1
=_((/2.4
x:r
+ /2.5)G 11 _ (/2.4
:r
/2.5)G
x 12
+ (/2.4
y
+ /2.5)G
Y 9
+ (/2.4
Y
/2.5)G)
Y 10
(2.4.16 )
and
£1.2{t t 't ='t)=(/2.4+j2.5)G +(j2.4_/2.5)G _(j2.4+/2.5)G +(/2.4 j2.5)G

x l' 2' 1 x:r 9 :r x 10 y Y 11 Y Y 12
(2.4.17)
This echo is the result of refocusing the 0-8 (1 ~ 2), 6Q (1 ~ 7), 2Q (2 ~ 4) and
3Q(2 ~ 5) coherences.
62
...1/
...
1iii
.e--8
~ 2
.8
~
c;f,) -O.S
...!.
N
Figure 2.4.5. Spin echo amplitude E;3(tp t Z ,t 1 = t), CD 1 1271: = 50kHz, as a function
of the second pulse length tz ().IS) for 1, II =5j.1S,coQ I 2x =OkHz;2,
II=3).1S.co 12x =50kHz;3, Spin echo amplitudeE;·3(tp l z,t) = 't), CD) 12x =50kHz,
Q
for II =3j.1S,CD 12x =50kHz .

Q
...
II
...
1iii
-8
a
1
f
CI( -0.5
o
Figure 2.4.6. Spin echo amplitude E~·z(tl'lz' 't J ='t). COl 1271: = 50kHz. as a function
of the second pulse length 12 ().IS) for 1, =okHz ;2,
tl = 5).1S.co Q 1271:
'I =3j.1S,co Q 12x =50kHz;3, Spin echo amplitudeE~·z(tptz' t\ = 't). COl 1271: =50kHz.
for 1\ =3j.1S,Q)Q 1271: =50kHz.

7) The amplitudes of the seventh echo located at 't I = 3't / 2 are given by
E Y2 ,3(tl' t 2' 't 1 = 3't /2) = \ '(/x2,4 + 1x2•5)F.II _(/x2,4 _ /x2,5)F.12 + (/ 2'4 + 1Y2'S)F9 +(/Y2'4 _ /Y2.5)F.10 )
y
_ ((/ x1,2 + 11.1)F.

x 15
_ (/1.2
x
+ II.,)F.
x 16
+(/y1,2 + 11.1)F.
Y 13
+ (/1.2
Y
_/Y1,1)F.14 )
(2.4.18)
E x2.3 {tl ' t 2' 't 1 = 3't /2) = (/2.4

x
+ 1x2,5)F9 +(/x2•4 _/
2
x
,S)F.10 _(/2.4
y
+ 1Y2,S)F.11 +(/2.4
Y
_/Y2•s )F.12
+{/x12
• + II,1)F.
x 13
+(/x1'2 + 11.7}F.
x 14
_{/1.2
Y Y IS +(/Y • _ /Y• )F.16
+ 11.7)F. 12 1,
(2.4.19 )
This echo is the result of refocusing the o-s (I B 2), 2Q (2 B 4), 3Q (2 B 5) and 6Q
(1 B 7) coherences.
Figure 2.4.7 shows graphs of E;·3(tl't 2,'t 1 =3t/2)andE~,3(tl't2,'t1 =3't/2) as
8) The amplitudes of the eighth echo located at 't l =5t / 3 are given by
E y1•2 {t II t 2' 't 1 = 5't / 3) = _((/x + 11.6)G
13
•
x 19
- (/x _/1.6)G
x
13
20
•+(/y1•3 + 11.Y 6)G17 +(/ Y1.3 11.6)G)
Y 18
(20 )
E x1•2 {t I' t 2' 't 1 =5't / 3) =([1.3x + [1.6)G

x 11
+ ([1,3
x
_ [1.6)G _ {[1.3 + [1.6)G + (/1.3 _ [1.6)G
x 18 y y 19 Y y 20
63
N
"- 0 .•
...
f"")
...
1/
0.2
lit
-8
j
i
.8 -0.2
~
-
CIl
N
I
-0 .•
-0.6
Figure 2.4.7. Spin echo amplitude E;.3(lp I2 .t l =3t/2), (Ot/27t=SOlcHz. as a

function of the second pulse length I" (,..IS) for I, 'I =SJ.1S.co QI =
21t 0 kHz ;2,
Spin echo
(01 121t =SO/cHz, for 11 =3J.1S.(Og 12ft =SOlcHz.
f"")
"
J::" -0.002
... -0.004
1;;
-8 -0.006
.a
:::: -0.008
~ -0.01
~
6 -0.014
-0.016
-0.018
Figure 2.4.8. Spin echo amplitude E;.2(llJ/l.tl =StI3), COl i27t=50IcHz, as a

function of the second pulse length 12 (J.IS) for I, '1 =3J.1S,(I) Q I27t=SO/cHz; 2, Spin
echo amplitudeE!,2(/I'/". tl =5t 13), (1)1 I 21t =50kHz, for I} =3J.1S.(I)Q 121t =SO/cHz.
(2.4.21 )
This echo is the result of refocusing the 2Q (1 'H' 3) and 5Q (1 'H' 6) coherences.
Figure 2.4.8 shows graphs of E!,2(tl't 2.t 1 =5t/3)andE;,2(tp t2.t l =5t/3) as
9) At the time tl = 2t two overlapping echoes of different widths are detected.

The amplitudes of the echo related to the 11-S are given by
2t) _((]2'3
x
+ ]2.6)K
x
_ (]2,3
7' x
_ ]2,6)K
x 8
+ (]2,3
y
+ ]2,6)K
Y 5
+ (]2.3
Y
_ ]2.6)K
Y 6
)
(2.4.22 )
and
E3.4{t
x
t t =2t) =(]2,3
I' 2' 1 x
+ ]2,6)K
x 5
+ (]2,3
x
_ ]2,6)K _(]2,3 + ]2,6)K +(]2,3
x 6 y Y 7 Y
]2,o)K
Y 8
(2.4.23 )
This echo is the result of refocusing the 21-S (2 'H' 3) and 4Q(2 'H' 6) coherences.
Figure 2.4.9 shows graphs of E;.4(tl't2.tJ = 2t)andE:A(t p t 2.t l = 2t} as functions
ofthe second pulse length.
10) The amplitudes ofthe echo related to the O-S are given by
EI'2{t
Y
t t = 2t) = '"\'(]lA
I' 2' I x
+ ]1,5)0
x 23
- (]1.4
x
_ ]1,5)0
x 24
+ (]1.4
y
+ ]1,5)0
Y 21
+ (]I.4
Y
_ ]1.5)0
Y 22
)
(24 )
64
0.4
N
...
..." 0.2
1ii
-8
=
.t=
l
i
--
CIl
I
-0.2
-0 .•
Figure 2.4.9. Spin echo amplitude E;·4(t).t2 ,t) = 2t), (D1/2,.; = 50kHz, as a function
of the second pulse length 12 (fJS) for 1, I) = 5fJS.(J)e/2,.; = OkHz; 2 •
3, Spin echo
CD 1 12,.; = 50kHz, for I} = 3fJS.(J)e/2x = 50kHz.
0.002
1ii
-8
.i -0.002
'a
i
f
CIl
-0.004
o•
-0.006
Figure 2.4.10. Spin echo amplitude E;·2(/

p /2 .t 1 =2t). CD,/2,.;=50kHz, as a
function of the second pulse length 12 (fJS) for 1, 11 =3fJS,(J)e 12,.; =50kHz; 2, Spin
echo amplitudeE;·2(tp / 2 ,'t l = 2t), CD l 12,.; = 50kHz, for '1 = 3fJS.(J)e 121t = 50kHz.
EJ,2(t
x
t t =2t)=(1x1.4+1x1•5 )G21 +(1J,4_11.5)G
J. 2' 1 x x 22
_(1y1.4+11.5)G
Y 23
+(11.4
Y
11.5)G
Y 24
(2.4.25 )
This echo is the result of refocusing the 3Q (1 ~ 4) and 4Q (1 ~ 5) coherences.
Figure 2.4.10 shows graphs of E~·2{tJ>t2.tl =2t)andE;2(tl't2 ,t J =2t) as functions
11) The amplitudes ofthe eleventh echo located at t 2 = 5t I 2 are given by
E y2•3 (t l' t2' 't I =5't/2)=_((1x1,3+1xJ·6)F.19 _{11.3

x
1x1•6)F.20 +(IJ,3+11,6)F.
y Y 17
+(11'3_ll,6)F.)
Y Y 18
(2.4.26 )
13
2
Ex ,3{t t t = 5t 12) = (11,3 + 11.6)F. + (1 • - II,6)F. -
l ' 2' 1 x x 17 x x 18
(1y
13
• + II,6)F.
Y 19
+ (11,3
Y
- I1,6)F
Y 20
(27 )
This echo is the result of refocusing the 2Q (1 ~ 3) and 5Q (1 ~ 6) coherences.

Figure 2.4.11 shows graphs of E;,3(tp t 2 ,t j =5t/2)andE:·3 (t p f 2 ,t\ =5t/2) as
12) At the time tl =3t two overlapping echoes of different widths are detected.
The amplitudes ofthe echo related to the 1I-S are given by
65
0.03
N 0.025
"
...
lI')
II
... 0.02
'=t
.---I 0.015
i 0.01
i
-
c:Il
N
I
0.005
Figure 2.4.11. Spin echo amplitude E;.3(/,,/2 .1: 1 =StI2). CD. 121t=50kHz, as a
function of the second pulse length '2 (JJS) for 1, '. =3JJS,co Q 121t= 50kHz; 2, Spin
echo amplitudeE;.3(tI'/2 .1: 1 =51: I 2), for
I) =3JJS,CDQ I 21t =50kHz.
0.3
...
('f")
0.2
II
... 0.1
'=t
«)
"0
.-.a
Q.
i -0.1
0
1i
c:Il -0.2
-
...!.
-0.3
Figure 2.4.12. Spin echo amplitude £;.4(/.. 12,1: 1=31:), COl 121t =50kHz, as a
function of the second pulse length '2 (JJS) for 1, II =5JJS'COQ 127t =OkHz;2,
Spin echo
COl / 21t = 50kHz, for I) =3JJS.coQ I 27t = 50kHz.
E3.4(t
Y
t "C = 3"C) =_(([2.x 4 + [2.5)K
l> 2' 1 x 11 - ([2.4 _
x
[2.5)K
x 12 + ([2.4
y
+ [2.5)K
Y 9 + ([2.4
Y
_ [2.5)K )
Y 10
(2.4.28 )
E3.4(11
x
t =3t)=(/2.4+/2.5)K
I' 2' I x x 9 +(/2.4_/2.S)K
x x 10 _{/2.4+/2,S)K
y Y 11 +(/2,4
Y
/2.5)K
Y 12
+(/1.2
x
+ /1.7)K
x 13
+(/1.2
x
_ /1.7)K
x 14
_(/1.2
y
+ /1.7)K
Y 15
+ (/1.2
Y
_ /1.7)K
Y 16
(2.4.29 )
This echo is the result of refocusing the O-S (1 B 2), 2Q (2 B 4), 3Q (2 B 5) and 6Q
(1 B 7) coherences.
Figure 2.4.12 shows graphs of E;.4(/1' 12, 1:1 = 3t) and E:A(/l' 12 , t, = 3t) as functions
of the second pulse length.
13) The amplitudes ofthe echo related to the 21-S are given by
EY2.3(1 I' 12' t 1 = 3t) = _((/1.4

JC
+ /1.5)p,
X 23
_ (/1.4
x
/1.5)p,
JC 24
+(/1.4
Y
+ /1.5)p,
Y 21
+(/1.4
Y
/1.5)p,)
Y 22
(2.4.30)
E x2•3 {11 t =3t)=([I.4+/1.5)p,

I' 2' I x x 21
+(/lA_/l.5)p,
x x 22
_(/I.4+/1.S)p,
y Y 23
+(/1,4_/1.5)p,
Y Y 24
(2.4.31)
This echo is the result of refocusing the 3Q (I B 4) and 4Q (1 B 5) coherences.
Figure 2.4.13 shows graphs of E;·3(/1' 12 , tl =3t} and E~·3(/1'/2' t} = 3t) as functions
66
0.01
M
...
II 0.008
...
1iii 0.006
j
Q.
0.004
i 0.002
.8
~
C"I'l
~
N
0.002
-0.004
-0.006
Figure 2.4.13. Spin echo amplitude E;'(t p t2 • tl =3t), CD I 12ft =50kHz, as a

function of the second pulse length 12 hIS) for 1, tt = 3/JS.CDQ 12ft = 50kHz; 2, Spin
echo amplitudeE;·3(/p t 2 • tl =3t), CD I 12ft =50kHz. for t. =3/JS.0l,/2ft =50kHz.

0.014
0.012
0.01
~
II 0.008
-0.006
-0.008
Figure 2.4.14. Spin echo amplitude E;·4(/\J/2 .t 1 =5t), Q)

1 /2ft=50kHz, as a
function of the second pulse length '2 (/JS) for 1, 'I =3/JS.CD Q 12ft =50kHz~ 2, Spin
echo amplitudeE;·4(/l't2 .t l =St)' CD) 12ft =50kHz, for 11 =3/JS.CI)Q 12ft =50kHz.
14) The amplitudes ofthe fourteenth echo located at 't l = 5't are given by
E3.4{11
y
't =5't)=-((/1,3+/1,6)K
l' 2' I x x 19
_(/1,3_/1,6)K
x x 20
+(/1,3+/1,6)K
y Y 17
+(/1.3
Y
11.6)K)
Y 18
(2.4.32)
E3,4{t
x
1 ... = 5't) = (/x1,3 + 11,6)K
I> 2' "I x 17
+(/x1,3 - 1x1,6)K18 (/ y1,3 + 1Y1,6)K19 +(/Y1,3 - 11.6)K
Y 20
(2.4.33 )
This echo is the result ofrefocusing the 2Q (1 -H> 3) and 5Q (1 -H> 6) coherences.
Figure 2.4.14 shows graphs of E;.4(/l't2,'t1 = 5't)and E;.4(/1'/2,'t'\ =5't) as functions
15) The amplitudes ofthe fifteenth echo located at 't 1 = 6't are given by
E y3.4{tI' 12' 't I =6't)=-((/IA+/I.5)K

x x 23
_{/IA_/I.5)K
x x 24
+{/I.4+/1.5)K
y Y 21
+(/I.4_/1.5)K
Y Y 22 )
(2.4.34 )
15
E3,4{11 't =6't)
x l ' 2' 1
(/IA+/l,s)K
x x 21
+(/1.4
x
11.5)K
x 22
_(/IA+/
Y Y
• )K +(/I,4_/1.5)K
23 y y 24
(2.4.35)
This echo is the result of refocusing the 3Q (l-H> 4) and 4Q (l-H> 5) coherences.
Figure 2.4.15 shows graphs of E;.4(/1't2 ' 't 1 = 6t) and E;.4(II' t2 ,'t 1 = 6t) as functions
67
0.004
..." O.OO~
1if
-8
.-ja
0
-5 -o.oo~
--
I)
fIl
I
-0.00.
Figure 2.4.15. Spin echo amplitude E;,4(/1'/2 ,t 1 =6t), al I /21C=50kHz, as a

function of the second pulse length '2 (~) for 1, 'I =3J.1S,COg I 21C =50kHz; 2, Spin
echo amplitudeE;'·(/1'/2 .t l = 6t), all 121C = 50kHz, for '1 =3~,Ci)g 121C= 50kHz.
The echoes discussed in 1 to 15 above are functions of the first and second
pulse lengths. In order to optimize the echo amplitudes, it is necessary to choose both
pulse lengths suitably. It should be noted that Figs. 2.4.1-15 are plotted for the case of
a single crystal only. In addition, it is seen from eq. ( 2.4.6-35 ) that x components of
the echoes are important for the medium values of ID Q and these tend to zero when ID Q
is much greater than or much less than ID).
It should be noted here that the SIX echoes at

't) = 5't I 3,'t 1 =2't(O-S),'t) = 5't 12,'t) =3't(2I-S),'t 1 =5't,'t j =6't are "forbidden"
in the sense that these cannot be predicted with the calculation based upon "hard"
pulse excitation. These echoes result from refocusing of exclusively multiple quantum
transitions. The expressions for the echoes for "hard" pulse excitation are easily
obtained from eq. ( 2.4.6-19, 2.4.22-23, 2.4.28-29) by neglecting multiple quantum

coherences and substituting ID Q = 0 in the equations for X's, r s, z' s, T's and IDij IS
(see Chapter 2 and [ 30 ] ). The immediate result of these changes is that all x
components ofthe echoes are zeroes, as expected.
Conclusions
Here Solomon echoes are considered for the "soft" pulse excitation for spin
7/2. Nine echoes are predicted to arise if the pulses are "hard". In addition, six
forbidden echoes are calculated which arise exclusively from the refocusing of multiple
quantum coherences developed during the first pulse. The amplitudes and the shapes
of these echoes depend upon both pulse lengths and the ratio of 0) Q 10)1 . Our results
are valid for any ratio of 0) Q 10)1 .
68
Table 2.4.1. Values K z with a =ZUX2j and (3 = 1;;¥Zj . It is assumed here that each
term has to be summed over iandj The functions X's,Y's,Z's,T'sandroij's are
reported in Chapter 2.
Kl = (3Zli X 2j cosroijt2 K2 = a1;;1";j cosroijt2

K3 = (3ZJjX2j sinro t/ 2 K4 = a1;t1';j sinroijt2
Ks = (3~/¥Zj cosro/l2 K6 =aZUz2j cosroijt2
K7 = (3~t1';j sinroijt2 Ks = aZ1i Z 2j sinroijtz
K9 = (3~iX2j cosroijt2 K10 = a1;jZ2j coSroijt2
Kll = (3~;X2j sinroijt2 K12 =a1;;Z2j sinroijt2
K13 = (3XlI Z2j cosroijt2 KI4 = a~l1;j cosroijt2
K]S =(3XIi Z2j sinroijt2 K16 =a1';/1;j sinro tl 2
K17 = (3Xu1';j cosroijt2 KI8 = aZlI 1;j coSroijt2
KI9 = (3Xli 1";j sinroijt2 K20 = aZt; 1;j sinro tl 2
K21 = (3Xli X 2j COSroijt2 K22 =a1;j1;j cosroijt2
K23 (3XIi X 2j sinroijt2 K24 =a1;/1;j sinroijt2
69
Chapter 5
Detection of Double and Four Quantum Coherences for Spin 7/2 Excited
by Spin Lock Pulse Sequences.
Investigation of multi quantum ( MQ ) coherences has received a great deal of
interest in NMR [ 39-41, 15, 42-44 ]. Spins of value higher than 112 with quadrupoar
splitting are multilevel systems with unequal spacing between the energy levels. As a
result, a single rf pulse acts selectively being resonant to a certain transition. The pulse
which is applied to a single quantum (SQ) transition, may excite other off resonance
SQ transitions as well as MQ coherences. However, MQ transitions are not detected
with this type of experiment. These can be studied by using various pulse techniques.
Particularly, double quantum ( DQ ) transitions in spin 3/2 systems are well
investigated by a combination of multipulse techniques with the two dimensional
representation [ 45 ]; a combination of spin lock and rotary echo pulse sequences [ 6 ];
and spin echo pulse sequences [ 46 ].
Application of computer algebra eases the calculations for spins higher than 3/2
as it allows efficient matrix manipulations. Recently, explicit analytical expressions
were reported for spin 5/2 [ 47 ] and spin 7/2 ( Chapter 3 and [ 38 ] ) for two pulse
experiments with time delay between pulses. These solutions indicate that under soft
pulse excitation condition, which retains the first order quadrupole, MQ coherences,
developed during the first pulse, are detected at the end of the second pulse via SQ
coherences. Also Solomon echoes were calculated analytically for spin 5/2 [ 13 ] and
7/2 ( Chapter 4 and [ 48 ] ). Some of these echoes ( "forbidden" echoes) arise
exclusively from the refocusing ofMQ transitions.
In this Chapter we use the same pulse sequences employed for spins 3/2 and
5/2 in [ 49 ] to investigate MQ transitions for spin 7/2. These two pulse sequences
include various phase cyclings as well as a combination of spin lock and spin echo
70
pulse sequences. The first order quadrupole is retained throughout and, in particular,
in the calculations of pulses. Thus our results are applicable to symmetrical ( i. e.
featureless ) lineshapes. However, the measurement of the integrated areas of the
central line as a function of the second pulse length permits the determination of the
quadrupolar parameters and the true chemical shift of the line. The response of half
integer quadrupolar spins to two pulse sequences as a function of the second pulse
length has more extrema than that of one pulse. The shape of this curve is determined
by the ratio of quadrupolar coupling, co Q' to the applied rf field in the case of a single
crystal or by the ratio of quadrupolar coupling constant, e :Q

2
,to the applied field in
the case of powders. Various two pulse sequences produce different shapes of the
curve in question. This makes the fitting procedure more accurate as the results for
quadrupolar parameters should be consistant for each experiment.
Our approach is different from MAS as we deal with the single static magnetic
field and as a result of this with featureless spectra to determine quadrupolar
parameters.
The Hamiltonian of the system excited by an rf pulse in the rotating frame
associated with the central transition, and neglecting the offset and high frequency
terms is
(2.5.1 )
where H2) is the first order quadruPQlar interaction Hamiltonian and H~) ,
71
is the rf Hamiltonian. Here the rf amplitude ro l and phase 4> describe a pulse. The
quadrupolar coupling ro Q can be defined experimentally as one half of the frequency
difference which separates two consecutive lines in the spectrum of a single crystal.
Angular frequency units are used and the effects of relaxation and second order
quadrupole are ignored.
The dynamics of a spin 7/2 excited by two rf pulses of arbitrary phases is given
by the density matrix P+1'+2 (I) ,(2 ),
(2.5.2 )
where
and 11 ,4» and 12,q,2 are the durations and phases of the first and the second pulse
respectively.
Density matrix after one pulse

Consider the dynamics of a spin 7/2 excited by an rf pulse with an arbitrary
phase q, . Note that
(2.5.3 )
where H~), H~) = -m.Ix , is the Hamiltonian of the rfpulse applied along the x-axis.
Thus using eq. (2.5.1 ) one can write
(2.5.4 )
72
Then the density matrix for the system, p. (t 1)' after an rf pulse of arbitrary phase 4> is
given by
(2.5.5 )
where Po(tJ is the density matrix ofthe system excited by an +x-pulse. At this point it
is easy to relate the density matrices after a pulse along the +x axis, Po (t l ), and after a
pulse along the -x axis, PlI(t1)(q, 71:), viz.
(2.5.6 )
Furthermore, the matrix I z is diagonal in the II,M) representation which simplifies the
computation of eq. ( 2.5.5, 2.5.6 ) giving,
(2.5.7)
are matrix elements of matrices
p.(t1) and Po(tl ) located at the k th row and the j th column of their matrix
representations. The immediate result which follows from eq. ( 2.5.6, 2.5.7 ) is that
only line intensities of odd quantum coherences change sign when the -x pulse is
changed to a +x pulse and vice versa. Concurrently the polarizations and the even
In
0
quantum coherences remain unchanged when the pulse phase is changed by 180 •
general, shifting the pulse phase by q, changes the phase of m-quantum coherences by
mq, but polarizations remain unchanged ( see eq. ( 2.5.7 ) ). The matrix elements of
PlI(t1) for spin 7/2 have been calculated previously ( Chapter 2 and [ 46, 28 ] ).
73
Making use ofeq. (2.5.7) allows the calculation of P1tI2(tJ the matrix form of which
is given in Table 2.5.1, in notation described in Chapter 3 . In the next section we use
the results of Table 2.5.1 to calculate a response of the system in question to spin lock
pulse sequences. However, it should be noted that the results represented by
eq. (2.5.1-7) are valid for all half integer spins.
Density matrix for spin locking sequences

The spin lock pulse sequence is comprised of two rf pulses without time
delays. The first pulse is applied along the y axis whereas the second pulse is applied
along the -x axis. Symbolically this sequence can be written as (y) (- x). The numbers
1 2
1 and 2 indicate that y-pulse is the first to apply followed by -x pulse. The density
matrix after a spin lock pulse sequence is given by
(2.5.8 )
The diagonalization procedure for H(1t) has been discussed in detail in Chapter 2. Thus
eq. ( 2.5.8 ) can be developed using computer algebra program 'Maple'. However, in
most cases only the central line is observed in experiments on powders. Below we will
focus on line intensity studies of this transition. The density matrix component
P1tI2.1t(tl't2 )[5,4] which corresponds to the central transition is calculated to be
74
PlIIZ,1I(t\Ot z)[5,4] = _(/~S XDI+ Dz)+ (/;,7 XD 3+ DJ - (/:,6 )(Ds + D6)
+(/:,S)(D7 +Ds) _(/~'2XV; VJ (/~,4XV3 -V4)+(/~,6XVS +V6)
_(/:,3 XV7- Vs) - (/~'S XV9 + V;O) (/~,4 )(V;I V;Z) - (/!,2 )(V;3 + V;4)
_(/!·4XV;S +V;6)-(/!,6)(V;7 +V;S)_(/;·3XV;9 +V2o )-(/';,SXV;1 +V 22 )
_(/;.4 XV;3 + VZ4 )

+;( (/;·6)C + (/;·7)C6+ (/;·8)C7+ (/:,S)C -(/;'SXC -C + (/;,6XC -C
S g 10 1Z
)
14 16
)
-(/;.4 XCIS + C20 ) - (/!·7)(CZ2 - C24 )- (/;3)( CZ6 + C2S ) + (/!·5)( C30 - C3J
-(/:,SXC9+Cll ) + (/;,6XC13 + CIS) - (/:,4 XCI7 C19 ) (/~'7)( C21 +C23 )
_(/;·3XC -CZ7 )+ (/;SX CZ9 +C
25 31 ))
(2.5.9 )
Here functions Cj (i = 1, ... ,32) were detennined previously and can be found in
Chapter 3. The equations for D'sand V's are given in Table 2.5.2. Note that
C' s.V' sand D's are functions of the second pulse length tz whereas line intensities
(I';-n) are the functions of the first pulse length t 1 • In general, the density matrix of
spin 7/2 excited by an rf pulse is defined by 32 independent components ( denoted by

(I';-n) ). All ofthese are present in eq. ( 2.5.9 ).
The profile ofthe central transition is defined as
(2.5.10 )
In the next section certain pulse sequences are considered which allow selective
detection of some multiquantum transitions.
75
Detection ofMQ transitions.
The density matrix component Pnl2,n(ll'lz)[S,4] is complex ( eq. (2.S.9) ) and
thus the FlO can be detected by quadrature in both x and y channels ( eq. ( 2.5.10 ) ).
It should be noted that odd quantum coherences, developed during the first rf pulse,
are detected in the x channel, whereas even quantum coherences and polarizations are
detected in the y channel. Consider the following pulse sequence ( Sequence 1 )
{y)(-x){y}recycle delay (-y)(-x){-y}. (2.5.11 )

1 2 3 4
Here inside of the curly brackets, {...}, the phase of the receiver is shown. Applying
a spin lock pulse sequence ofthis type ( where the phases of the first and third pulses
as well as the phase of the receiver after the second and fourth pulses are different by
180· ) results in the cancellation of the y component of the FlO. However, the
equation for the x component is simply deduced from eq. ( 2.5.9 ) being twice as big
as that given by eq. (2.5.9,2.5.10). Thus the density matrix component of the central
transition for the Sequence 1 is
(2.5.12 )
Sequence 1 can be modified such that the phase ofthe receiver after second and fourth
pulses remain unchanged. This new sequence ( Sequence 2 ),
{y)(-x){y}recycle delay (-y) (-x){y}, (2.5.13 )

1 2 3 4
76
results in the cancellation of the x component of the FID. At the same time the y
component is twice as big as that produced by the sequence which consists of the
pulses 1 and 2 only. The equation for the density matrix component of the central
transition becomes
(2.5.14)
From our previous results of Chapter 3, the density matrix component of the central
transition for the sequence ( Sequence 3 )
(-x) (-x){y}recycle delay {x) (-x){y} (2.5.15)

I 2 3 4
which designed for so called even quantum detection ( rotary echo type ) is given by
pSeq.3(tp tJ[5,4] =2i{(I;,6)Cs +(1;'7)C6 +(1;,8)C7 +(1:,5)Cs

+(I;'sXC C12 )+(1;·6XC C +(1;.4)( C +C
lo - I4 - 16 ) 18 20 )
+(I;·7XC -C24 ) +(1;·3XC +C +(I;"XC C

22 26 28 ) 3o n ) (2.5.16 )
+(1:")(C +C +(1;·6 XC +C +(1~,4 XC C

9 J1 ) l3 1S ) 17 - 19 )
+(1;")(C21 +C23 ) +(1;·3)( C25 - C27 ) +(1;·5)( C29 +C31 )}
Sequences 2 and 3 can be combined into Sequence 4,
(y)(-x){y}recycle delay ( -y)( -x){y}recycle delay (-x) (-x){y} recycle delay (x) (-x){y}
12 34 56 78
(2.5.17)
such that the resultant magnetization vector is aligned with the z axis and the equation
for the density matrix component ofthe central transition for the Sequence 4 is
77
pseq.4(tp tJ[5,4] = 4i{(/;,6)CS+(/;,1 )C6+(/;,8 )C1+(/:.5)CS
+(/;·6XCI4 -CI6 ) +(/~·5XC30 - C3J+(/;·6 XC13 +C1s )+(/;,'XC29 +C3J}'( 2.5.18 )
Thus 4 quantum coherences, developed during the first pulse, are detected. However,
if Sequences 2 and 3 are combined into Sequence 5 in the following manner
(y)(-x){-y}recycle delay (-y)(-x){-y}recycle delay (-x) (-x){y}recycle delay (x) (-x){y}

12 34 56 18
(2.5.19)
( the phase of the receiver in Sequence 2 is changed by 180· ) then double quantum
coherences developed during the first pulse are detected,
pseQ.5(tl'tJ[5,4] = 4i{(/;'sXC1o -C12 ) +(/;.4 XC +C +(/;,3 XC26 +C

18 20 ) 2S )
+(/;.5 )(C9 +Cll ) +(/;.4 )(C17 - C19 )+(/;,3 XC25 -C27 )} (2.5.20)
It should be noted here that the response of the system in question becomes
less complicated as the number of pulses increases. However, at the same time the
signal-to-noise ratio decreases. On the other hand these pulse sequences allow the
determination of quadrupolar coupling in a single crystal, and the quadrupolar
coupling constant and asymmetry parameter in powders. These values, obtained using
the variety of pulse sequences, should be consistent and this improves a level of
confidence in our results. The important feature of Sequence 5 is that it drastically
reduces spurious signals [ 50 ] generated by the NMR probe for nuclei with small
gyromagnetic ratios. When used in NMR with no quadrupolar splitting, this sequence
places the magnetization vector along the z- axis and causes no response. However for
78
quadrupolar nuclei, the signal is detected due to multi quantum coherences developed
during the first pulse.
Conclusions
The response of a spin 712 to a spin lock excitation and to a variety of two
pulse sequences involving the combination of rotary echo and spin lock sequences is
presented. The first order quadrupole is retained throughout the calculation. Our
primary focus is on the central transition as it represent the main signal detected in
powders. Also calculated. the density matrix ofany half integer spin after an rf pulse of
arbitrary phase. This shows that when the phase of the pulse is shifted by q" the
phases of m-quantum coherences change by mq,. However, these are not detected
except for the case of m= 1. Use of two pulse experiments allows the detection of multi
quantum coherences through SQ coherences after the second pulse. It is proven here
that DQ coherences can be selectively detected using appropriate phase cycling. In this
case, the central line intensity is a function of both pulse lengths, the quadrupolar
coupling and the amplitude of the rf pulse. Fitting curves representing the central line
intensities as the function of the duration of one of the pulses, allows the determination
of quadrupolar parameters. Finally, we remark that our analytical solution for the
eigenvalue - eigenfunction problem [ 30 ] can be used in the analysis of [ 51 ] for spin
7/2.
79
Table 2.5.1. The density matrix of spin 7/2 after the rf pulse applied along the y axis. The
functions (I;i) are reported in Table2.3.l of Chapter 3.
( I
;(I;·2) +(I~) -(I~)+i{I;l) _1(1;4) _(1;4) (1!")-i(I;") i(I~") +(I;' ) _(I;·7 )+;{I;') -111.·\
I (I:") \ Y / I
1-;(1;:)+ (I~) (I:1.1 );(1;1)+(/;3) _(/;·4)+1(1;·4) -/(I;')-{I;") (1;')-;(1;') (1;·7) (11.7)
, +/{I, /\ II
17
_(I;3)_;(I;.3) -t{1;3) +(I~) (I;') 1(1;4)+ (1;4) -(I;") +i(I:.,) -(I;') _(/,1.6)_/(1;') -i(/~") +(/;·6) I
1{/;4} _(/;4) _/{/;4) +(1;4)
_(1;/')_/{1;4) (I:") (I;") (1;")+;(1:) ;(1;")-(/;') -(/;'}-I(/;') 1
(/;')+i(/~ ) I{/;')- (I;')
-(/;")-/(I;') (I;" ) -(I:") _/(/;,4)+ (/;4) (l.t4) +1(/;4) ;(1;4)_ (/;14) I
-/(1;')+(/;') (I;") +/(1;') -(I!") (/;')_;(Iy3.5) 1(1;,4) +(1:,4) -{/ U \
: f -/(1;')+(1;.3 ) (I.1.3)+/(I;.1) 1
1-{1;')-t{I;") \1;') _(1;')+/(1';6) -/(/;')-(1;.5) (1;4) _1(1;4 ) i(I;.3) +(I;3) -(1;,7) -/(1;2 )+(1;'2) I
l -(/;1) (1;')_1(1;') I{I;') +(I;") -(I;") +I{I;') _/(1~'4)_ (1;4) (1~)_i(/;3) 1(/;,2)+ (I!'l) _{I;·I} J
4
Table 2.5.2. The functions D i (t 2 ) and V;(t 2 ) used in the eq. ( 2.5.9 ). The symbol L
;,j=\
In
1 +
D\ = 2 Xli X Jj T,.J'1j cos'Vij
1
D2 = 2 1'z;1'zjX2iXlj cos'Vij
1 +
D3 = 2 ~)~jT,.iT,.j cos'Vij
1
D4 = 2 Z2iZ2jX2iX2j cos'Vif
1 +
Ds = 2 ZJjZ\jT,.iT,.j cos'Vij
1
Do = 2 Yz;YzjXliX2j COS'!I;
D7 = 2:1 T,.i T,.jT,.; T,.j COS'!Iij+

1
Ds = 2 X2iXljX2iX2j COS'!I;
V; = X)j~jT,.j T,.j cos'!I~
Vl = Z2;1'z j X 1;Xlj cos'!Iif
~ = 1'z;X2jX2;X2j cos'Vij
V4 = X\iT,.jT,.;T,.j COS'!I;
~ = ZJjX\jT,.iT,.j COS'!I;
V6 = 1'ziYz j Xli X 2j COS'Vij
V7 = Z\ir;jT,.iT,.j COS'!I;
Vg = YzjZ2jXljX2j COS'!Iii
81
Table 2.5.2 (con'd).
~ = 1';; 1;/1;,1;j COS\jI;

V;o =Z2iX2jX2iX2j COS\jIij
V;l = 1;,Zjj1;,1;j COS\jI;
V;2 = J;,X2jX2,X2j COS\jIij
V;3 = X Ii 1';j1;,1;j sin\jl;
V;4 = Tz,Z2jX2iX2j sin\jl'ij
V;s = 1;i X lj1;,1;j sin\jl;
V;6 = TziX2jX2iX2j sin\jl'ij
V;7 = ZIi X lj1;jI;j sin\jl;

V;8 J;iTz j X 2;X2j sin\jlij
V;9 = Yli Z1j 1;i1;j sin\jl;
V20 = J;iZ2jX2;X2j sin\jlij
V21 = ~i1;jr;ir;j sin\jl;
V22 = X2iX2jX2iX2j sin\jlif
V23 = ZJi1;jr;;1~i sin\jl;
V;4 = X2iJ;jX2iX2j sin\jlif
82
References.
1. Samoson, A, and Lippmaa, E., 1. Magn. Reson., 79, 255 ( 1988 ).
2. Samoson, A, and Lipprnaa, E., Phys.Rev. B 28,6567 ( 1983 ).
3. Goldman, M., Adv. Magn. Opt. Reson, 14, 59 ( 1990).
4. Sanctuary, B.C., and Halstead, T.K, Adv. Magn. Opt. Reson., 15, 79 ( 1990).
5.Man P.P., Molec. Phys., 72,321 ( 1991 ).
6.ManP.P, Molec. Phys., 76,1119 (1992).
7. Harris, R K, and Leshitt, G. J., J. Magn. Reson.,78, 245 ( 1988 ).
8. Wokaun, A, and Ernst, RR, 1. Chern. Phys.,67, 1752 ( 1977).
9. Lee, N., Sanctuary, B. C., and Halstead, T. K, 1. Magn. Reson.,98, 534 ( 1992 ).
10. Fenzke, D., Freude, D., Frohlich, T., and Haase, 1., Chern. Phys. Lett., 111, 171
( 1984 ).
11. Haase, J., Freude, D., Pfeifer, H., Lipprnaa, E., Chern. Phys. Lett., 152, 254
( 1988)
12. Man, P. P., Molec. Phys.,78 (2),307 ( 1993 ).
13. Man, P. P., Z. Naturforsch., 49a, 89 ( 1994 ).
14 Char, B. W., Geddes, K 0., Gonnet, G. H., Leong, B. L., Monagen, M. B., and
Watt, S. M., Maple V, Language Reference Manual; Maple V Library Reference
Manual, Springer - Verlag ( 1993 ).
15. Wokaun, A, and Ernst, R R, 1. Chern. Phys., 67, 1752 ( 1977).
16. Vega, S., J. Chern. Phys., 68,5518 ( 1978 ).
17. Eguchi, T., Nakayama, H., Ohki, H., Takeda, S., Nakamura, N., Kernaghan, S.,
and Heaton, B. T., 1. Organornetal. Chern., 428, 207 ( 1992 ).
18. Hirschinger, J., Granger, P., and Rose, 1., 1. Phys. Chern., 96, 4815 ( 1992 ).
19. Eaton, D. R., Buist, R J., and Sayer, B. G., Can. 1. Chern., 65, 1332 ( 1987).
83
20. Chung, S. C., Chan, J. C. C., Au - Yeung, S. C. F., and Xu, x., J. Phys. Chern.,
97, 12685 ( 1993 ).
21. Eckert, H, and Wachs, I. E., J. Phys. Chern., 93, 6796 ( 1989).
22. Skibsted, J., Nielsen, N. c., Bildsoe, H, and Jakobsen, H J., Chern. Phys. Lett.,
188,405 ( 1992).
23. Lapina, O. B., Mastikhin, V. M., Shubin, A A, Krasilnikov, V. N., and Zarnaraev,
K. I., Progr. NMR Specrosc., 24, 457 ( 1992 ).
24. Ocelli, M. L., Maxwell, R. S., and Eckert, H, J. Catalysis, 137, 36 ( 1992).
25. Dupree, R., Lewis, M. H, and Smith, ME., J. Am. Chern. Soc., 111,5125
( 1989).
26. Herreros, B., Man, P. P., Manoli, J. M., and Fraissard, J., J. Chern. Soc., Chern.
Commun., 464 ( 1992).
27. Trokiner, A, Le Noc, L., Yakubovskii, A, Mykhalyov, K. N., and Verkhovskii,
Z. Naturforsch, 49a, 373 ( 1994 ).
28. Dec, S. F., Davis, M. F., Machiel, G. E., Bronnirnann, C. E., Fitzgerald, J. J., and
Han, S. S., Inorg. Chern., 32, 955 ( 1993 ).
29. Solomon, I., Phys. Rev., 110,61, (1958).
30. Ageev, S. Z., and Sanctuary, B. C., Molec. Phys., 84, 835 ( 1995 ).
31. Man, P. P., and Tougne, P., Molec. Phys., 83, 997 ( 1994 ).
32. Lipprnaa, E., Samoson, A, and Magi, M., J. Am. Chern. Soc., 108, 1730 ( 1986 ).
33. Man, P. P., J. Magn. Reson., Series A 114, 59 ( 1995 ).
34. Hahn, E. L., Phys. Rev., 80,580 ( 1950).
35.Ahragam, A, Priciples ofNuc1ear Magnetism, Oxford University Press, Oxford
( 1961 ).
36. Pulsed Magnetic Resonance: NMR, ESR and Optics, Ed. D. M. S. Bagguley,
Clarendon Press, Oxford ( 1992).
37. Frydrnan, L., and Harwood, J. S., J. Am. Chern. Soc., 117, 5367 ( 1995).
84
38. Ageev, S. Z., Man, P. P., Fraissard, 1., and Sanctuary, B. C, 1997, Molec. Phys.
(in press ).
39. Bodenhausen, G., Prog. NMR Specroscop, 14, 137 ( 1981 ).
40. Vega, S., Shattuck, T. W., and Pines, A, Phys. Rev. A 22, 1980 ( 1980).
41. Vega, S., Shattuck, T. W., and Pines, A., Phys. Rev. Lett., 37, 43 ( 1976 ).
42. Stoll, M. E., Wolf, E. K, and Mehring, M., Phys. Rev. A 17, 1561 ( 1978 ).
43. Hatanaka, n, and Hashi, T., Phys. Rev. B 27, 4095 ( 1983 ).
44. Hatanaka, H., and Takahama, Phys. Rev. B 47, 3213 ( 1993 ).
45. Rooney, W. D., Barbara, T. M., and Springer, C. S., Jr., 1. Am. Chern. Soc., 110,
674 ( 1988).
46.Newitt, D. C, and Hahn, E. L., J. Magn. Reson., A 106, 140 ( 1994 ).
47. Man, P. P., SoL State Nucl. Magn. Reson., 2, 165 ( 1993 ).
48. Ageev, S. Z., Man, P. P., and Sanctuary, B. C., 1., Molec. Phys. 88, 1277
( 1996).
49. Man, P. P., 1995, J. Magn. Reson., A 113, 40 ( 1995 ).
50. Gerothanassis, I. P., 1987, Prog. NMR Specrosc., 19, ( 1987).
51. Ding, S., and McDowell, C. A, 1. Magn. Reson., A 112, 36 ( 1995 ).
85
Part 3
Analytical Solutions for Spin 3/2 Off-resonance Line Intensities in Solid
StateNMR
Chapter 1.
Introduction
The quadrupolar coupling constant and the asymmetry parameter of

quadrupolar nuclei at a nuclear site can be detennined from nutation experiments.
These parameters provide valuable information about local electronic environments
and are of importance for molecular structure and dynamics studies. However, useful
information from nutation spectra can be obtained only when the ratio of the
quadrupolar coupling constant to the frequency of the rf field lies in the certain region
which depends upon spin magnitude 1. For example, the application of the rf field of
50 kHz to spin 3/2, with the quadrupolar coupling constant greater than 1 Mhz, causes
difficulty in determination of quadrupolar parameters as the main features of spectra
concentrated at 100 kHz. In the off-resonance nutation experiment, the spin system is
irradiated with an off-resonance rf field and the response is recorded on-resonance in
order to avoid detection problems. This extends the applicability of nutation NMR to
systems with large quadrupolar coupling constants as the resonance offset is the easily
controlled parameter, and it is possible to obtain nutation spectra with different
effective fields on one particular spectrometer with the same probehead. However, the
application of large offsets leads to severe loss of the signal·[ 1 ]. This problem can be
solved by using, for example, the frequency - stepped adiabatic half - passage [ 2 ].
The density matrix of a spin 3/2 subject to the first order quadrupolar
interaction while the rf pulse is on was calculated by Man [ 3 ]. Later he extended his
calculations to two pulse sequences retaining the secular part of the heteronuclear
interaction during the evolution period between the pulses[ 4 ]. His results are valid for
86
any ratio of quadrupolar coupling to the applied rf field amplitude, co Q / co I' and can be
used for the determination of quadrupolar parameters from ID nutation experiments
[ 3 ]. However, his calculations are valid for on resonance excitation conditions only.
His treatment uses the fictitious spin 1/2 operator formalism [ 5 ]. Here, based on the
application of computer algebra and simple matrix methods, we calculate the response
of a spin 3/2, taking into account the resonance offset and the first order quadrupolar
interaction during the pulse. Our results are valid for any ratios of co Q / COl and
~co / COl' The offset term is relevant for accounting for the secular part of the
heteronuclear interaction and / or the static magnetic field inhomogeneity while the
pulse is on [ 4 ]. Moreover, our results are readily modified to include the second
order quadrupolar interaction during the pulse. The conclusions of our studies should
be useful for Raman NMR studies of a spin 3/2 [ 6 ] and for off-resonance nutation
experiments performed by Kentgens [ 1 ].
87
Chapter 2.
Theoretical Consideration
The Hamiltonian for a system excited by an rf pulse in the rotating frame
associated with the central transition, and neglecting high frequency terms, is
(3.2.1 )
where
and the rf amplitude ro 1 , phase cI> and offset .tlrodescribe a pulse. Here H2) is the first
order quadrupolar Hamiltonian. Angular frequency units are used and the effects of
relaxation and second order quadrupolar interaction are ignored in the present
Chapter.
The dynamics of the spin 3/2 system under a pulse along the positive x-axis
(cI> = 0) is given by the density matrix
-ilf Olt ilf Olt

Po(t) = e-1i-~O)e (3.2.2 )
with the initial condition ri. 0) = I z • The matrix representation of H{O) is given in Table
3.2.1. The four eigenvalues of H<O) can be obtained using results of [ 7 ] and are found
to be
E) =(R+D)/2
E2 =(R-D)/2
( 3.2.3 )
E3 = -(R-S) 12
E4 = -(R+S) 12
88
The various constants are
(3.2.4 )
and
112 )112
R = 6-1I2 [ X + (X2 + 12Z) cOs(6 13) (3.2.5 )
where
(3.2.6 )
Here X, Y and Z depend on the interaction parameters as
X =-4{-51 2am 2
-
2
2ro/ -512ro 1 ).
Y = -4(-4am 2ro g + 2ro g ro l 2 )

Z= 1~9 18Aro 2ro l 2 +ro g 4 - 5 12am 2 ro g 2 + 9 1 1<:,ro I 4 + 9 1 16Aro 4 + 11 2ro g 2ro J2 )
(3.2.7 )
are the generalization of the result of ref [ 5 ] and valid for any ratios of
ro g 1 ro 1 and am 1 ro 1 .
The components of the normalized eigenvectors associated with these four

eigenvalues are
Xi =11 Ni
1"; =-albl N j
(3.2.8 )
Zj =(b-calb)/ro l 1 N j
1; =b(b-calb)1 gl N j
89
where i = 1,2,3 and 4, the values N are chosen to nonnalize each eigenvector and
j
a=3/2A<D+<D Q -Ej
b=-J3<D 1
c = 1I 2A<D - <D Q - Ej (3.2.9 )

d = -1/2A<D-<D Q -Ej
g= -3/2A<D +<DQ -Et
The transformation matrix which diagonalizes H<O) and the matrix of eigenvalues are
(Xl Xl X3 X4 1 (E 1 0 0 01
1
-I ~ 1'; 1; ~I Jo El 0 01
T -lZ' 1;
Zl
1;
Z3
~
Z4j
~
E
l~
0
0
E3
0 ;J (3.2.10 )
respectively and the equation for the density matrix ( 3.2.2 ) becomes
(3.2.11 )
The calculation of eq. ( 3.2.11 ) is performed using 'MAPLE' [ 8 ] and the result is
given in Table 3.2.2. From these one can determine the intensities of the central and
satellite lines.
The central line intensities F;l and F;2 are defined by
(3.2.12 )
In general, F;l is the sum of sixteen terms whereas F;2 is the sum of twelve terms
( four ofthe sixteen terms with j=k are zero ) i. e.
90
4
F;2 = LLkjZk~' cosrokjt (3.2.13 )
k,j=1
4
F:2 LLkjZk~ sinrokjt

=- ;;,j=1 (3.2.14)
where
(3.2.15 )
and
(3.2.16)
In a similar manner, the intensity of two other possible transitions ( satellites)
can be found using Table 3.2.2. In contrast to the case of Aro =0, these have different
offsets of Aro + 2ro Q and - Am + 2ro Q respectively when Am * O. The line intensities are
the functions of ro I ,ro Q and Aro.
Below, we concentrate on the central transition as this is the only one which is
usually observed in powders. Equations (3.2.14) can be investigated in the frequency

domain. The Fourier transform of these give sets of twelve lines located at ro kj . The
amplitudes Kkj , Kkj = LkjZk~' are shown in Fig. 3.2.1 a, b as functions of ro Q / ro 1
for the case of Aro = ro l' These demonstrate the relative importance of the various
contributions from the nutation frequencies. All twelve amplitudes are shown. Figure
3.2.1 c depicts six corresponding values of mkj .
Consider the case Am = O. The x component of the line intensity F;2 is

reduced to zero for any value of ro Q whereas F:2 is given by
(3.2.17 )
It should be noted that for this case the matrix of amplitudes is antisymmetric i. e.
Kkj = - Kjk and the K13 = K24 == O. The latter originates from the fact that the
91
a)
0.0&
0.02
b)
-0.02
Figure 3.2.1a, 3.2.1b The amplitudes Kkj , Kq = LkjZt~ , as functions of O)Q /0)1 for the
case of AO) = 0)1' The numbers relate to the subscripts k and} of Kkj
Figure 3.2. Ic. The frequencies (1)if / (1)1 as functions of (1)e / (1)1 for the case of ~(O =(01'
The numbers relate to the subscripts k and j of (0 if / (01 .
Hamiltonian given in Table 3.2.1 can be block diagonalized [ 9 ]. The nonzero
amplitudes are plotted in Fig. 3.2.2 a as the functions of co Q / co I' The corresponding
values cokj are shown in Fig. 3.2.2 b. In this case our results are in complete
agreement with those of Man [3 ].

Figures 3.2.3 a and b show the line intensity F:2 as the function of the pulse
length, t . for various ratios of co Q / co I for the cases of Aco =0 and Aco =co I
respectively. It is seen from this figure that in general, the duration of a "x / 2 II pulse is
a function of COl ,co Q and Aco. We define the duration of a x / 2 pulse as the position at
which the response to a +x pulse reaches the first maximum when plotted as the
function of time. Figures 3.2.4 a and b depict the off-resonance profiles of a pulse for
the cases co Q =0 and co Q =co I' The duration of pulses is chosen to be the duration of
"X / 2" pulse. As the value of co Q increases the on-resonance amplitude decreases and
the effective width of the curve increases. In the case that co Q »co 1 , the pulse becomes
selective, and the on-resonance amplitude is 112 of that in the case when co Q =0 . In
addition, the width of the curve is twice as broad as in the case co Q = 0 because for a
selective pulse the rf frequency COlis replaced by the effective frequency co 1ejf ,
co lejf =leo l [ 10 ].
The FT spectra produced by the system in question can be simulated using eq.
( 3.2.12 ). Since satellite transitions are usually lost in the dead time of the receiver,
they are not retained. The line broadening is assumed to be Lorentzian, hence the FT
ofthe eq. ( 3.2.12) is
(3.2.18 )
The imaginary part of eq. ( 3.2.18 ) is plotted in Fig. 3.2.5 a and b for the cases that
Aco =0 and Aco = COl' In particular, it is seen from Fig. 3.2.5 that the spectra exhibit
92
a)
se of
es Kk j , KIrj = Lk jZ i~ • as func tions of mg / ml for the ca
Figure 3.2.2a. The amplitud
j
Th e nUmb ers reJate to the subscripts k and} of Kk
Am = O.
b)
for the case of &co ='",.

Figure 3.2.2b. The frequen
~
cies '" / '", as function s of ," . / "',
bscripts k and} of m~i / COl
•
The numb ers relate to the su
.)
Figure 3.2.3 a. The line intensity F;2 as the function of the pulse length t in units of
1/ COl for various ratios of co Q / COl for the cases of &0 =0 : I, COg / CD l =0; 2,
CD g / COl = 0.5 and 3, COg / COl = 1.
b)
Figure 3.2.3b The line intensity F;,32 as the function ofthe pulse length t for various ratios
of COg / COl for the cases of &0 = coli, CD g / COl = 0; 2, COg / COl =0.5 and 3, COg / COl =1.
a)
Figure 3.2.4a. F;2 ( 1 ) and F;2 ( 2 ) as the functions of &.D / COl for co Q = 0 and
t = 1.5 / COl'
b)
Figure 3.2. 4b. F;2 ( 1 ) and F: 32

( 2) as the functions of &.D / COl for co Q = COl and
t = lIco 1 •
Figure 3.2.5a. The imaginary part of F(CD) (eq. (3.2.18» as the function of CD I CD I for the
situation, T; = 10/ CD 1 and ~CD =0 : 1, CD Q / CD I =0; 2, CD Q / CD} =0.5 ; 3, CD Q I CD! =1; 4.
Q)Q I Q)1 = 1.5 and 5, CD Q / CD 1 =2.
) 1
Figure 3.2.5b. The imaginary part of F(CD) ( eq. ( 3.2.18 » as the function of Q) / Q)t for the
situation, T; =10 / CD 1 and ~Q) = Q)!: 1, Q)Q / Q)} = 0; 2, CD
Q
/ Q)I = 0.5 ; 3, Q)Q I Q)j = 1; 4,
Q)Q / Q)I =1.5 and 5, Q)Q / Q)} =2.

more characteristic powder features in Fig. 3.2.5 b than those in Fig. 3.2.5 a for higher
co Q I COl ratios. This is in agreement with Kentgens [ 1 ].
Finally. we note that the density matrix p.(t)after an rf pulse of an arbitrary
phase is [ 11 ]
(3.2.19)
where the components of the density matrix after the +x pulse) Po(t)[k,j] , are given
in Table 3.2.2. This completes the calculation of the response of spin 3/2 to a pulse of
arbitrary amplitude, phase and offset while the first order quadrupole is active.
93
Chapter 3.
Conclusions
Here we have presented the analytical calculations ofthe response of a spin 3/2
subject to the first order quadrupolar interaction and resonance offset while the pulse
is on. Our results are in complete agreement with Man's on-resonance analytical
computation [ 3 ] and Kentgen's numerical work [ 1 ]. Moreover, inclusion of
resonance offset in the calculation allows for a detailed analysis of Raman multi
quantum NMR experiment for spin 3/2. It should be noted that practically all" single II
spin interactions can be included into the calculation simply by modifying are the
constants X, Y and Z in eq. ( 3.2.7). Currently most nutation experiments are carried
out with Magic Angle Spinning (MAS ). It is straightforward to modify our results to
include second order quadrupolar effects. Thus for calculations, such as those in
[ 12 ], which use MAS, the CPU time can be dramatically reduced if the second order
interaction is important [ 13 ]. Moreover, using computer algebra, it is possible to
extend our calculation to at least two pulse sequences with time delay between the
pulses.
94
Table 3.2.1. The matrix form ofthe Hamiltonian H(O) •
r31 2.1ID + ID Q -1/2..J3m 1 0 0 1

I
(0) I -1 1 2..J3m 1 1/ 2.1ro -IDQ -ro l 0 I
H =, 0 -ro 1 -1/2.1ro -ID Q -1/2..J3m\ I
l 0 0 -1/2JiID 1 -3 1 2.1ro +IDQ)
Table 3.2.2. The components [ I,m] located at the I th row and m th column of the
density matrix Po(t) ( eq. ( 3.2.11 ) ). It is assumed here that each term has to be
multiplied by Lkj , LItj =1/2(3XkX j +.r.~ - ZkZj - 3~Z;) ,and summed over k and j .
Here IDkj = Ek -Ej •
[II] =2XkX j COSIDkjt ,[22] =2.r.~ COSIDrc/ ,[22] =2ZkZj cosIDItjt.[44] =2J;I; COSIDltjt
[21] =YgXj exp{-iIDltjt) ,[31] =ZkXj exp{-iIDltjt) ,[32] =Zk~ exp{-iIDltjt),
[41] =~Xj exp(-iIDltjt), [42] = ~~ exp(-iIDltjt) ,[43] = ~Zj exp{-iIDkj.t)
95
References.
1. A P. M. Kentgens, 1. Magn. Reson. A, 104,302 (1993 ).
2. Kentgens, A P. M., 1. Magn. Reson., 95, 619 ( 1991 ).

3. P. P. Man, 1. Magn. Reson., 67 , 78 ( 1986); ibid 77, 148 ( 1988 ).
4. P. P. Man, 1. Chim. Phys., 89, 335 ( 1992 ); P. P. Man, 1. Magn. Reson., 100 , 157
( 1992); P. P. Man, Solid State NMR, 1, 149 ( 1992).
5. A Wokaun and R R Ernst, J. Chem. Phys., 67, 1752 ( 1977 ); S. Vega, 1. Chern.
Phys., 68, 5518 ( 1978 ).
6. G. 1. Bowden, W. D. Hutchison, andF. Separovic, 1. Magn. Reson.,79 , 413
( 1988 ); C. S. Yannoni, R. D. Kendrick, and P. K. Wang, Phys. Rev. Lett., 58, 345
( 1987).
7. Creel, R.B., J. Magn. Reson. 52, 515 ( 1983 ).
8. Char, B. W., Geddes, K. 0., Gonnet, G. H., Leong, B. L., Monagen, M. B., and
Manual, Springer - Verlag (1993 ).
9. A Samoson and E. Lippmaa, 1. Magn. Reson., 79 , 255 ( 1988 ).
10. P. P. Man, Appl. Magn. Reson., 4, 65 ( 1993 ).
11 S. Z. Ageev, P.P. Man and B. C. Sanctuary, J. Magn. Reson, ( 1996 ), submitted.
12. Ding, S., and McDowell, C. A, 1. Magn. Reson. A, 112, 36 ( 1995 ).
13. Sun, W., Stephen, 1. T., Potter, L. D., and Wu, Y., J. Magn. Reson. A, 112,36
( 1995 ).
96
Part 4
Theory ofrfPulses in NQR
Chapter 1.
Composite Pulses in NQR
A composite pulse is a sequence of closely spaced radio frequency ( rf) pulses

of rectangular shape which are phase shifted with no time evolution between pulse
elements in the sequence. Composite pulses have been extensively studied for NMR
and have found numerous applications [ 1 ]. Simple examples ofcomposite pulses are
1t-pulses and 1tI2-pulses which are insensitive to frequency spread. Particularly,
composite pulses are useful for non-linear situations where the intensity of rf field
differs from one region to another ( rf inhomogenity ) and where it is necessary to
excite spins over a range of chemical shifts ( resonance offset ). A variety of methods
have been developed to produce composite pulse sequences. These include the
Magnus expansion [ 1-3 ], similarity transformations [ 4 ], recursive expansions
[ 1,5,6] and the multipole approach [ 7,8 ].
Recently several reports on NQR pulses have appeared in the literature[ 9-11 ].
Pure NQR of solids is characterized by broad lines of order of a few kilohertz due to
various internal interactions such as electric field inhomogenity ( EFG ) and dipole
dipole interactions present in the system, which create multiple spectral lines. These
interactions are not averaged as in the NMR case and thus the spin - spin relaxation
time ~ and spin - lattice relaxation time 1; are much shorter then those in NMR case.
Consequently, long pulse sequences are not desirable for NQR. A single short rf pulse,
however, does not excite the broad NQR spectra uniformly. It is possible to overcome
this problem by increasing the rf power, but this, in turn, can cause difficulties in
practice. In addition, the response to a single rf pulse is sensitive to rf inhomogenity.
These problems can be partially resolved by using composite pulses of short durations.
97
Up to date the design of composite pulses for NQR has relied heavily on the Magnus
expansion. This approach offers both advantages and disadvantages. First, composite
pulses can be constructed regardless of the initial conditions of the density matrix. On
the other hand the Magnus expansion converges slowly and the extension of the range
of compensation is difficult without increasing the length of the sequencer 11 ]. In
addition, a separate numerical optimization of pulse lengths and phases must be
performed if pulses have rotation angles different from 90 and 180 degrees.
Ramamoortby's approach to the design of NQR composite pulses is based on the
Magnus approximation. One of the major results of our approach is to recognize that
NQR composite pulses for half integer spins basically reduce to the study of spins of
112. As such we can use composite pulses derived for NMR. We find, in fact, that such
composite pulses work well for NQR crystal studies, but less well for powders.
In actual calculation ofNQR composite pulses, Ramamoorthy [ 10 ] essentially
uses spin 112 operators. Our approach is to recognize that a pulse in NQR is selective,
being between two pairs of levels. These are treated as rotations since selective
excitation between neighboring pairs of levels has zero effective quadrupole while the
pulse is applied [ 12,13 ]. In addition, we are able to solve much of this problem by the
use of matrix computer algebra, particularly the program 'MAPLE'. This allows us
easily diagonalize the quadrupole Hamiltonian when Tf ¢ 0 and to calculate the Baker
Campbell - Hausdorff formula ( BCH ) analytically for integer and half integer spins.
In this Chapter it is shown that NMR composite pulses are directly applicable
to pure NQR problem for half integer spins in the case of single crystal. The effects of
composite pulses of short durations are obtained using the multipole approach in NMR
[ 8 ] and applied to the NQR case. Powders are also considered.
98
Single pulse
The pure quadrupole Hamiltonian for spin I in the quadrupole principal axis
system ( QPAS ) is given by
(4.l.1)
where e2 qQ is the nuclear quadrupole coupling constant and 71 is the asymmetry
parameter. Much interest has centred on the evolution ofthe density matrix (1](t) as it
responds to a pulse sequence of 1, 2 or 3 pulses applied to a system ofisolated
spins I These spins are initially prepared at thermal equilibrium as characterised by
temperature T at time t=O. The equilibrium density matrix can be expressed in the high
temperature limit as
(4.1.2)
where /3= 11 kT is the Boltzman factor and E] is the (21 + 1) x(21 + 1) identity
matrix.
We shall illustrate our analysis for (1](t) with 1=3/2 as an example, using it to detail the
various steps in our approach which then can be followed for other half integer spin
values. The equilibrium density matrix (13/2(0) for spin 312 can be expressed in the
customary 11M) basis in the following matrix form
99
o Js °llI
11
-1 o -II
.J31~
-I o II
II
1 JJ
(4.1.3)
The quadrupolar Hamiltonian of eq. ( 4.1.3 ) can be diagonalized by means of the
unitary transformation defined by S where
( cose
I 0
0
cose
sine
0 .e'\l
0
-sm 1
s=1
l-s:e 0
sine
cose
0 co~eJ
(4.1.4 )
and tan2B= Js .
In this new basis ItP), I'P) =SlIM), the equilibrium density matrix is
a~2(0) = ~ {E3I2 - f3H~} (4.1.5 )
where
fl 0 0 ol
1
'Y _ limQ I 0 -1 0 01
HQ - 2 10 ( 4.1.6)
0 -1 01
Lo 0 0 d
is used and
100
(4.1.7)
The basis vectors in the 1lJ1> basis are 1+3/2),1+112),\-1/2) and 1-3/2). If a
radiofrequency ( rf) pulse is applied along the x-axis of the QPAS, the rf Hamiltonian
in the 11M) basis is customarily written as
{ 4.1.8 }
where £()l is the rf amplitude, OJ is the carrier frequency and ; is the arbitrary phase
associated with the rf pulse. In the I 'P) basis, Hr f can be written in matrix form as
(0 A 0 B1
H;, =-11m, ={mt-.t1 o Col ( 4.1.9)
~ ~ ~j
where
A =( 2J1I2
1+!L
3
B=l+(l+ ~r
1-11
(4.1.10 )
Following Pratt [ 14 ], the interaction picture of the evolving density matrix
allows significant simplification enabling one to follow the von Neumann equation in
this frame. This is achieved by the transformation operator n defined as
101
(4.1.11)
Transforming HT~ of eq. ( 4.1.9 ) into the interaction picture and omitting high
frequency terms [ 14 ] , lengthy but straightforward algebra leads to a dramatic

simplification of the radiofrequency Hamiltonian H:; in which only the rf phase
variable , plays a leading role,
~+
e-~ 0 01
( oI
1iro A, e
l
0 0
-'1' __ 1 I
l~
(4.1.12)
0 0
H,r 2
0 e-i+ e:J
The total Hamiltonian containing the quadrupole interaction of eq. ( 4.1.6 ) and rf field
of eq. ( 4.1.12 ) can therefore be written in the interaction representation as
(4.1.13 )
where the block matrices "" and ~ consist of off-resonant constant terms (ro e ro)
and elements depending only on the rfpulse characteristics (rol,cP andTJ), viz:
(4.1.14 )
102
In the interaction representation the equilibrium density matrix at 1=0 can be
similarly decomposed into 2 x 2 block form as
(4.1.15 )
where
(4.1.16)
and
(1 0)
~(O) = 0 -1 ,A(O) = 0
(-1 0)1 (4.1.17 )
Note that the Hamiltonians iiI and ii2 of eq. ( 4.1.13 ) and the density matrices ~ (0)
and 152 (0) commute.
The evolution of the system under the Hamiltonian iiJ in the interaction
representation is given by the density matrix
(4.1.18)
where
(4.1.19)
Note that ~ and ~ are SU ( 2 ) Hamiltonians for the elementary spin 112 problem.
Under a single rf pulse applied along the x-axis of the QPAS, the 11/2) ~ 13/2) and
1-1/2) ~ 1-3/2) transitions are uncoupled and independent. The only approximation
103
used so far involves the omission of the nonsecular terms in Hy~ of eq. ( 4.1.12 )
which, if retained, would lead to corrections ofthe order ofBloch - Siegert shifts
[ 15 ]. Equations ( 4.1.12-4.1.19 ) can be interpreted as follows. The transition
/1/2) ~ 13/2) is caused by the Hamiltonian h,. with 1-1/2) ~ 1-3/2) being caused
by the Hamiltonian ~. This means the applied rf field is effectively decomposed into
two components rotating in opposite senses [ 16 ]. The transformation into the
quadrupole interaction picture shows that h,. and ~ arise from the rotating and
counter-rotating components of the applied rf field respectively. In contrast to the
treatment of the counter-rotating field in ordinary NMR ( where it is dropped in
theoretical calculations) both rotating and counter rotating components are significant
in NQR The transformation to the quadrupole interaction picture in NQR is similiar to
the rotating frame tranformation in NMR.
jlltt i"zt
The propagators e -A' and e-11 can be written in terms of the rotational
operators [ 17 ] 3l(a., ~, y) ,
illtt
e-t; :;: 31(0.+ +4>,~ ,0.+ -4> +1t)
ihzt
(4.1.20 )
e-" =~o.- -4>,~ ,0.- +4>+1t)
where (o.:t ± 4>, ~ ,o.:t +4> +1t) are the Euler angles. These Euler angles are found to be
[ 15 ]
104
cosJl = ~(ro21eff COs(Ot) + Aro 2)
Q= ~Aro2 +ro 21eff ( 4.1.21 )

Aro=rog-ro
3+11
ro leff =( 2)112 ro 1 .
3+11
Consider the application of a single rf pulse of duration tl on an isolated spin

3/2 system. Using the properties ofWigner rotation matrices ~ of eq. ( 4.1.20 ) it can
be shown that the density matrices are
(4.1.22 )
in the quadrupolar interaction frame. To find the free induction decay we define two
magnetization components 1:0, which are "in-phase", and, 1:'00 "out-of-phase" with the
rffield:
1':0 =t: cos(rot)

I'!o = r: sin(rot) (4.1.23)
Transforming eq. ( 4.1.23 ) to the interaction picture and truncating high frequency
terms as is in the derivation of eq. ( 4.1.12 ) leads to
105
· (4.1.24)
The response ofthe system in the interaction picture is then defined by the averaged
"in phase" and "out-of-phase" quantities W; and W; respectively, where
W; = tr{lx~~~2(t)}
(4.1.25 )
W; =tr{lx':o~~2(t)}
and tr{... } indicates the trace ofthe quantity in {...} brackets is to be taken. Using
eq. ( 4.1.18,4.1.22,4.1.24 ), W; and W; can be found for a single rfpulse as
W; =-sinf3 co.s{a.+ -cf»

(4.1.26)
W; = sinf3sin(a.+ -<I»
f3hro g A
where units of 8 are used. It should be emphasised that the NQR spectrometers
are sensitive to actually measuring the responses W; and W; of the quadrupolar spin
system in the interaction picture.
Composite pulses
We now are in a position to solve the problem for the application of composite
pulses in NQR. To achieve this goal the 11s in eq. ( 4.1.20) are replaced by products
of11s, whose arguments are defined by the new values of a±. p and (j)lejJ of
eq. ( 4.1. 21 ) associated with each new pulse. In general ~ (t) and 152 (t) in
eq. ( 4.1.18 ) for any number of pulses are
106
(4.1.27 )
Then the response ofthe system, as measured by W; above, is
(4.1.28)
For example, for a two pulse experiment we have
W; = ~ {-2COS~1 sin ~2 co~a.; -CP2) - sin ~l( 1- COS~2) co~a.; + 2CPl -CPl)}
+ ~ sin~l(l+cos~Jco~a.; +2a.; -CPJ .
(4.1.29 )
The response of the system is clearly analogous to that in the NMR spin 1I2-case
evolving under the Hamiltonian ~ or h2 [ 8 ]. An important difference between the
NMR and NQR cases is seen in the vanishing of (I:) for the NQR case,
(4.1.30 )
whereas the NMR counterpart is nonzero. Thus we need a different definition for the
inversion pulses to contrast to the NMR case: this is provided by the energy
(4.1.31 )
which according to eq. ( 4.1.27 ) is simply
107
(4.1.32 )
. umts
In . 0f - f3(1ko4 g f . This defini'
tion 0 fth· . pu1'
e InverSIOn se 1S chosen so that t he
NQR result agrees with that ofthe corresponding spin II2-system.
Powder averages
In powders, crystals are randomly oriented. Assuming that the rf coil is
oriented in the QPAS-system as is shown in Fig. 4. L 1., the rf Hamiltonian for the
crystal is
(4.l.33 )
In the frame defined by the operator Sin eq. ( 4.1.4 ), this becomes
r (0 0 B')
A ( 0 -iAl 0 iBJ ')
~ I A 0 C o II I iA o I
l Afsm ml
0 -iCj
.•. e I I
14 =-lieD, cos(mt-+) cos. L sine 'l~ C 0
I
-~IJ
LS 0 iC\ 0
0 A o -iBl 0 iAJ
(-D 0 E 0')1
I 0 -DI 0 l~
+ coset ! 0
-E
D)
0
-E
~jJ
(4.1.34)
108
where A,B, C are given by eq. ( 4.1.10) and
AI = ( 2J1I2
1+.:L
3
B, =1-(l:i),n
1+1)
C, =1+(1+ ~r
D=-l- (l+~r
2
Dl = 1- ( 1}2 J1I2
1+
3
( 4.1.35 )
E=
2J1I2
(1+.:L
3
Transforming Hr~ of eq. ( 4.1.34 ) into the interaction picture defined by the operator
II in eq. ( 4.1.11 ) and neglecting all non-secular terms as was done in deriving eq.
( 4.1.12 ) results in
109
a-iJ3 y o I (0 -ia- J3 -iy o I I
o 0 -y I Iia-
'RjCOS++ .
J3 0 0 . II
-IY ~
o Jain' J
0
l
a-I.., ry 0 0 . Jl
la;
-y a+iJ3 o 0 iy -ia+J3
(4.1.36 )
where
a= Asin8Lcos;L'
p= Al sin 8Lsin;L'
r=cos8L E .
We now define new operators Q and P such that
(4.1.37 )
J
where {))leff = {))l a Z +f/ + y . The total Hamiltonian can be then represented as
( 4.1.38)
where
( 4.1.39)
Note that the operators P, Q and R obey the standard SU ( 2 ) algebra commutation
relations. The solution for the propagator is
( 4.1.40)
110
where the Euler angles are given by eq. ( 4.1.21 ). The only difference is the new (J)leff
specified in eq. ( 4.1.37) which reflects the arbitrary direction ofthe rf coil( see eq.
(4.1.33) ).
At this point it is straightforward to recover the block-diagonal fonn of the total
Hamiltonian ( contrast to eq. ( 4.1.13 ) ) by means ofthe following change of the basis
(4.1.41)
where
( y a-if) ')
o o
~(a2 +f}2 +y2) I
o 1 o I
y a+if}
o o
o 1 J
such that
Hr = Tilr l (4.1.42 )
and
(1 0 0 01 (0 e- +
t
0 01
_ _li(m Q -m) 10 -1 0 oIL lim leff· I l41 e 0 0 01
HT - 2 l~ 0 -1
0 0 ~J
2 l~ 0
0
0
e- + e;J
i
(4.1.43 )
Comparison of eq. ( 4.1.43 ) and eq. ( 4.1.13 ) shows that these are identical and the
equilibrium density matrix is given by eq. ( 4.1.15 ). Thus all arguments of the
previous section can be directly applied to the problem ofarbitrary orientation of the rf
field with respect to QPAS of the crystal. The response of the system is simply given
111
by ~ using the appropriate effective frequency of the rf field, or in tenns of the
operator Q in the I tF) basis
(4.1.44 )
It is clear that in the case of an arbitrary directed rf coil, the response ofthe system is
analogous to the responce of a spin 112 system responce in the rotating frame. If the
distribution ofthe crystals is assumed to be random and " = 0 then the averaged signal
ofthe system is [ 9-10 ]
(4.1.45 )
For the inversion signal ofa single crystal defined as
(4.1.46 )
the averaged signal for a powdered sample ( " =0 ) then is
(4.1.47 )
The responses ofthe system to the various pulse sequences is discussed after treating a
spin 5/2 system.
Generalization for higher order spins

I
The quadrupolar Hamiltonian matrix for a spin 5/2 in the I M) basis is [ 16 ]
112
( 10 0 l1JiO 0 0 0 l
, 0
-2 0 3.Ji11 0 0 ,I
pe 2 qQ l1JiO 0 -8 0 3.Ji11 0
,
HQ = (4.1.48 )
40 , 0 3J211 0 -8 0 11M I'
l~ 1~ J
0 3.Ji11 0 -2
0 0 11M 0
Assuming that an rfpulse is ofthe form given in eq. ( 4.1.8 ), in the 'F)-basis we I
obtain
(0 A B 0
0 Dl
IA 0 C 0 E 01
H; =-21iml co.s{cot-,)I ~ C 0 F 0
0 F 0 C
BI0, (4.1.49 )
'0 E 0 C 0
lD 0 B 0 A ~j
where A,.., F are specified in terms of the transformation elements Uij as ( A6 ) in
Appendix A. The energy splittings between pairs of degenerate levels are unequal and
a single pulse acts selectively affecting only a single transition. If the carrier frequency
of an rf pulse aJ I I
is close to the frequency of the ±3 / 2} ~ ±5 / 2} transition, then it is
reasonable to define a reduced matrix for the quadrupole and rf Hamiltonians which
specifically includes those pairs of states. Other states are simply not affected by the
pulse. Thus the problem reduces from a 6 x 6 to a 4 x 4 problem. This is realized as

H'Q ' defined as
(4.1.50 )
113
Using the actual frequency of the transition under consideration as
co Q = e;~; (2E1 + E z) eriables eq. ( 4.1.50 ) to be written as a sum oftwo matrices
(1 0 0 0, (1 0 0 0,
r _ 1tco Q ! 0 -1 0 0 11_ heZqQ E2110 1 0 01
l~
(4.1.51 )
~J l~ ~J
Ho - 2 0 -1 40 0 1
0 0 2 0 0
whenceforth the second contribution will be dropped as it merely resets the energy
reference state. The reduced rf Hamiltonian associated with eq. ( 4.1.49 ) is similarly
obtained,
(0 A 0 D,
IA 0E
H;, =-21i<D, cru{mt-.t~
01
(4.1.52 )
~J
E 0
0 A
where A, D and E are defined in Appendix A

The interaction picture is now defined by the operator n which uses the
reduced Hamiltonian H~ of eq. ( 4.1.11 )
Hr
-i~t
1kD
TI=e Q (4.1.53 )
Transforming the total reduced Hamiltonian into the interaction picture and omitting
the high frequency terms gives
114
01 (0 01
01 i+ o oI
Lnco 11e0 I (4.1.54)
I o
1J 0 o e:J
This is the same form as the total Hamiltonian for spin 3/2 in eq. ( 4.1.13 ) provided
that O)lejf = 2AtDI . In the case of an arbitrarily directed of rf coil ( with angles
(e L ,cp J in the QPAS frame) the total reduced Hamiltonian is given by eq. ( 4.1.38 )
with the following changes
a =2Asin (}L COStPL'

p= 2AI sin(}L sintPL' (4.1.55 )
r= 2COS(}L E
where A is given by eq. ( A6 ) of Appendix A and
1 3 1
Al = 2 J'SUl1USS - J"iUsSUI3 + 2 U3SU31 - J"iU3sUSI + 2 J'SU1SU51>
( 4.1.56)
5 1 3
E =2 UlIU1S +2U31U35 - 2 U51USS •
Thus the response of a spin 5/2 system to a single rf pulse is of the same form as a spin
3/2 system with a different effective frequency of rf field. In the case of 11 = 0, A
reduces to J5 and tD 1ejf = J5tD I sin 8L .

2
Results and discussion

Here we discuss the performance of the composite pulses designed by various
NMR methods in the NQR case. As explained in the previous sections, that NQR
pulses for half integer spins can be described by rotation matrices. The multipole
approach to composite pulses uses the fact that a product of rotations is a rotation
115
itself Recently , several pulse sequences of short durations were designed using this
method [ 8 ]. In general, for spin I in the case of 17 = 0 the effective frequency of the
r.f field for a transition M ~ M -1 is
( 4.1.55 )
Thus to obtain a universal description for spin I system, the signal is plotted as a
function of offset divided by the factor .J(I + M)(I - M + 1) . One example [ 8 ] is the
composite 7r-pulse of 780 - 195182 - 318 12 , In this notation each pulse is represented by
8 t where 8 = COl t and <I> is the phase of the rf pulse. The performance of this sequence
is shown in Fig. 4.1.2 for a single crystals and powder averages for 17= O. Figure 4.1.2
reveals the fact that this composite pulse sequence has an offset range between ±1.5
where the responce is relatively flat. This is good not only for single crystals but also
for powders. Other pulse sequences designed for NMR are shown in Figs. 4.1.3, 4.
Although they work similarly well for single crystals they are less effective for the case
of powders.
Figure 4.1.5 shows a 7r composite pulse designed by the multipole approach
2
[ 8 ]. We report a pulse sequence which works well not only in the case of single
crystals but also in powders. This is 38590 - 320270 - 2590 designed using Magnus
expansion[l] (Fig. 4.1.6). It should be noted that the pulse sequences in Fig. 4.1.2, 6
work better then those presented in [ 10 ], using the flatness about zero offset as the
criterion. We conclude that composite pulse sequences designed for NMR, perform
well for NQR in the case ofsingle crystal but less well for powders.
116
..
Coil
I-----+--- Y
4.1.1. The orientation ofthe rf field with respect to the QPAS system of the crystal in the
powder sample.
1
~
I
I I
'"
..
0.5 1-"
I
,/
,,,,",
'-...,,/\ \
,..'"
,,
.
.§
; 0 .
.~ ,'"\ i ,r
~ \\1' ,
I
I
I
I
I
I
(\J/
\
\,/\ I •,I l\)
I
-'0.5 ,, ,••
\
,, , I
,,• ,,
I \
-1 \II . I L U
-3 -1.5 o 1.5 !Jm 3
0)1 ~(I + M)(I - M + 1)
4.1.2. The composite1l'-pulse of 78 0 -195 112 - 318 12 .The solid line represents a single
crystal, the dashed line is used for powders.

0.5 \ l\ .
(
/ I \"\'
,'\,\
I \
J O~-+--____ - 4_ _~
o 1.5 3
Cl)1 ~(J + M)(J - M + 1)
4.1.3. The compositen-pulse of 860 -1601• - 260,.The solid line represent~
. a single
.

lr-~-r-'~~~~~~-
'\ f\ .
I \\
/ .......\
,
I \
j
I
"
.... ""
"
V ,,
\
i Or---~----------~--~
-1.5 o 1.5 Am 3
ml~(J + M)(J - M + 1)
4.1.4. The compositen-pulse of 2600 -161 190 - 8S...Tbe solid Hne represents a single
aystal, the dashed Hne is used for powders.

I""'..~' \
.,
\ ,
\ I
\J
.J 0 Hr--~~--------~--~~
\t
\
\,
\
~
1.0 &> 3'

O>lJ(I +M)(I - M + 1) .
4.1.5. The composite; -pulse of 83 90 - 233191 ·The solid line represents a single crystal,
the dashed line is used for powders.

1
~. N ( I , fV'.
,J
f"..., ...../'
,,,f"
~'\
..,
""\...
,,
, "
I
t ,J
0.5 I- , \ ,, 1
,
• ••,, ,,
,
I
I
, i ,,
I
I , I
f
I
I
,
I I
I,
, I
t
I •
I ,
I
,I I
I
\• •, .
I
-0.5 l- ,I
r ,
,
I
I

.
-1 • I ~ . I • I
-3 -1.5 0 &D 31.5

0>1 J(I + M)(J - M + 1)
4.1.6. The composite; -pulse of 38590 - 3202'10 - 2590 .The solid line represents a single

Conclusions
We have considered in detail the effect of composite pulses in the case of pure
NQR neglecting relaxation for half integer spins using spin 3/2 and 5/2 examples.
Mathematically the difference between NMR and pure NQR cases can be understood
in the terms of the corresponding Hamiltonians. In NMR a large static Zeeman field
causes the energy splittings to be almost equivalent between any two adjacent levels.
The application of an r.f pulse excites all transitions simultaneously. In contrast in
NQR., wth no large Zeeman field, the quadrupole interaction causes unequal energy
splittings. Thus pulses in NQR act selectively. The effect of a rf pulse on pairs of
adjacent levels can be described by an SU ( 2 ) Hamiltonian and thus the responses of
two different physical systems are found to be similar. The theoretical conclusion that
NMR composite pulses can be used with similiar effects for NQR should be useful in
practice. This conclusion, to date, has not been tested experimentally.
117
ApendixA
The energy eigenvalues of HQ of eq. ( 4.1.48) are doubly degenerate and obtained as
the solution to a cubic secular equation. These are given in detail in [ 16 ]. In summary
the energy eigenvalues are E} ,E2 and E3 where
(AI)
and
R21 : 80 2(1_1'12)2
'1
+ 64 C2
81} ,
C12 =3(3431'}6 + 3871'}4 + 146611'}6 + 6561), (A2)
C}
lana: 90( 1-1'} 2)"
=
In the case of 1] 0 the eigenvalues of the quadrupole Hamiltonian reduce to the
simple results E} =10,E =-8,E =-2

2 3 in units of e:~Q .The matrix which
diagonalises the quadrupole spin 5/2 Hamiltonian [ 18 ] is given as:
(U}) 0 U31 0 USI 0)

I
I0 Uss 0 U3S 0 U15 I
U33
s= IU I
0 0 US3
13
o (A3)
1 0 US3 0 U33 0 U131
l~5
U3S Uss
~J
0 0
US! 0 U31 0
where
118
31
U = N [(E
1 1 +2)(El +8}-1811 2 ]'
311
2
J20
US) = N1[(E1 +2){El +8) - 1811 2
]'
U -
13 -
rt50
N 2 (E2 -10) ,
1
U 33 =N 2
'
U _ 3rt1i
S3 - N 2 (E 2 +2)'
3112 J20
U - - --:--------:-"'-----:-:------:-=
IS N3[lOrt2 -(E3 -1O)(E3 +8}] ,
3r(10 - E3}.J2
U---:----'-;-----'-~--:-::-
35 - N3[lOrt2 -(E3 -1O)(E3 +8)] (A4)
1
US5 =
N3
The constants N J ,N2 and N3 are defined by normalizing each coloumn of the S
matrix to unity. The S matrix reduces to the identity matrix when T/ is equal to zero.
In this new representation I '1? H Q takes the diagonal form with elements
E] ,E3 ,E2 ,E2 ,E3 and El .

Assuming an rf pulse is applied along the x-axis in the QPAS such that
(AS)
then in the 1lf1-basis
119
(0 A 0 B 0 Dl
IA 0 C 0 E 01
c
=-2ml co~)I 0 0 F 0 BI
rot - q, I B
'II
Hr.!. (A6)
0 F 0 c 0 1
10 I
~j
E 0 C 0
lD 0 B 0 A
where
1 3 1
A = 2./""SUssUn + .J2US,U3l + '2US3U31 + .J2U3SUS1 + 2 ..[5U1SUS1 '
1 3 I
B= 2..[5us3 Un +.J2U53 U31 + 2 U33 U31 +.J2U33 Us1 + 2..[5U5I UI3 •
1 3 1
c= 2 ..[5US5U13 +.J2U33 U" + 2 U33 U3' +.J2U3,U'3 + 2..[5U'3U1S'
D=..[5US1 Ull +2.J2US1 U31 + ~U;l'
E =$USS U15 +2.J2U3,USS + ~ U;5'

F =$U53U 13 + 2.J2US3U33 + ~ U;3 .
120
References.
1. Levitt, M.H., Prog. Nucl. Magn. Reson.Spectrosc., 18,61 ( 1986). Freeman, R,

Levitt, M.H., 1. Magn. Reson., 33, 473 ( 1979).
2. Tycko, R, Schneider, E., Pines, A, J. Chern. Phys., 81 , 680 ( 1984 ).
3. Tycko, R, Phys. Rev. Lett., 51, 775 ( 1983 ).
4. Blumich, B., Spiess, H.W., 1.Magn. Reson. 61, 356 ( 1985 ).
5. Waugh, 1.S., 1. Magn. Reson. 49... 517 ( 1982).
6. Shaka, AI., Keeler J., Fremkiel, T., Freeman, R, J. Magn. Reson., 52, 335
( 1983 ).
7. Sanctuary, B.C., Cole, H.B.R, J. Magn. Reson., 71, 106 ( 1987).
8. Yang, x., Keniry, M.A, Sanctuary, B.C., J. Magn. Reson. (in press).
9. Ramamoorthy, A, Narasimhan, P.T., I. Molec. Structure, 102, 333 ( 1989 ).
10. Ramamoorthy, A, Chandrakumar, N., Dubey, AK., Narasimhan, P.T., J. Magn.
Reson., Series A 102,274 ( 1983 ).

11. Sanitha Bai, N., Ramokrishna, M., Ramachandran, R., J. Magn. Reson., Series A
104, 203 ( 1993 ).
12. Wu, D., Sanctuary, B.C., J. Magn. Reson., Series A 103, 171 ( 1993 ).
13. Sanctuary, B.C., Molec. Phys., 49 ( 4 ), 785 ( 1983 ).

14. Pratt, I.C., Molec. Phys., 34, 539 (1977).
15. Furman, G.B., Izv. Vysh. Uchebn. Zaved., Radiofiz., 27, 667 ( 1984 ),
( in Russian ).
16. Krishnan, M.S., Temme, F.P., Sanctuary, B.C., Mol. Phys., 78 (6), 1385 ( 1993 ).
17. Zhou, J., Gao, H., Sanctuary, B.C., I. Magn. Reson., Series A 101, 119 ( 1993 ).
18. Krishnan, M.S., Notes (unpublished).
121
Chapter 2.
Selective excitation of single and multiple quantum transitions for spin 7/2
inNQR
Introduction
NQR has the simplifying advantage that the nuclear electric quadrupolar
interaction usually dominates the dipole - dipole interaction. This means that the spin
dynamics can be treated, to a good approximation, as an average over a single spin
with 21+ 1 states. Pulses and, via internal interactions, developments of spin states can
be calculated in a straightforward manner providing a description from which various
physical constants can be extracted, or a framework from which relaxation can be
studied.
On the other hand, the study of spin dynamics in NQR causes difficulties. For
one, non-zero asymmetry can lead to intractable calculations. Another problem is that
the rf field is randomly oriented in powders. Moreover, rf pulses are not "hard" with
the consequence that the quadrupole cannot be ignored while a pulse is on. In NQR,
however, methods to handle these difficulties have been developed for spins of low
magnitude. In 1977 Pratt led the way by introducing an interaction representation. He
found that the response of the quadrupolar spin 3/2 nuclei is similar to that of a spin
112 NMR [ 1 ], thereby enabling a variety of approaches used in NMR to be carried
over to NQR.
From a mathematical view point, these approaches have relied on operator
formalisms such as fictitious spin 112 [ 2 ], spherical tensor [ 3 ] and projection
operators [ 4 ] techniques. However, the algebraic complexity and corresponding
lengthy tables of commutation relations is the common drawback of these methods.
Obviously calculation involving spins higher than 3/2 become even more cumbersome
and manual calculations, become more error prone. Such approaches also often mask
122
the physical picture of the quadrupolar phenomena in progress. An attempt to clarify
the physical picture has been made by Krishnan et al [ 5 ], but those arguments are
applied only to half integer spins. Moreover extension to 11 ~ 0 is not obvious.
Recently, using simple matrix methods, we have extended spin dynamic

calculations for cases 11 ~ 0 involving spin 3/2 and 5/2 [ 6,7 ]. This is possible due to
developments of computer algebra which allow facile manipulation of matrices of

dimension needed for such calculations. Although computer algebra has allowed
significant extension so that higher spins can be routinely treated, its use requires some
care and strategy. Extension from spin 5/2 to 7/2 is not straightforward. Algebraic
equations can easily become too long to treat, and proper reduction techniques to
obtain matrices of manageable size must be invoked. To this end, it is not suprising
that no studies ofspin 7/2 exist in the literature. In this paper we consider this problem
and present spin 7/2 results with 11 ~ 0 as it responds to selective rf pulses. In tandem
with previous approaches, we assume that the pulses are selective [ 5 ] since in NQR
the zero magnetic field splittings are unequal, and excite only one specific multiple
quantum transition. We also do not restrict the rf field to one direction in space.
Consequently, field orientation angles can be averaged, thereby allowing application to
powder samples to follow without difficulties. Finally, we allow for shaped rf pulses
which are used in both NMR [ 8 ] and NQR [ 7 ].
The calculation is performed using the computer program MAPLE' [ 9]. It is
shown here that the excitation of these spin systems can be considered in the
interaction picture as being equivalent to the excitation of two uncoupled spins of 1/2.
At any time these spins 112 have equal x and y projections of the magnetization vector
whereas their z components are oppositively directed. The main important difference
with NMR is the appearance of an effective rf field frequency which depends on the
orientation of the crystal with respect to the applied rf field and transition in question.
In addition in this Chapter, we allow an arbitrary nonsingular modulation of the
123
applied rf field. It should be noted that double and higher quantum coherences are not
detected in a one pulse excitation experiment for t'l =0 as these are forbidden by well
known selection rules [ 10]. These can be analyzed by using multiple pulse sequences
developed by Hatanaka [ 11 ]. In particular, using this technique, double quantum
transitions have been investigated theoretically for spin 5/2 [ 12 ] and spin 7/2 [ 10 ].
For the case of spin 7/2 (t'l '* 0) it is shown in this paper that double and triple quantum
transitions can be excited by one pulse experiments, as these can be allowed transitions
according to the selection rules obtained.
One shortcoming of our approach is the restriction to selective excitation.
Although excitation in NQR is primarily selective, accidental degeneracies ( as a
function of t'l ) can arise. On the other hand, for particular special cases, when
experimentally more than one frequency is excited, our methods can be modified to
study such situations. We also impose one major approximation, well justified in
practice. This is we ignore high frequency ( non-secular) terms which give corrections
ofthe order ofBloch - Siegert shifts [ 13 ].
This work, then, extends previous computer algebra treatments of 1/2 integral
spins from /=3/2 and 5/2 to /=7/2 with the same level ofapproximations and accuracy
[6 ].
Eigenvalue and eigenfunction problem for spin 7/2

The quadrupolar Hamiltonian matrix for a spin 7/2 in the 1/M) basis is given
by
124
r 21 0 tW2i 0 0 0 0 0 l
I 0 3 0 f)J45 0 0 0 0
I
11W'2i 0 -9 0 rrJ60 0 0 0 I
tze2 q Q 0 TJJ45 0 -15 0 f)J60 0 0 I
Hg = 84 I I
0 0 rrJ60 0 -15 0 TJJ45 o I
I 0 0 0 rrJ60 0 -9 0 1W'2i 1
I
I 0 0 0 0 f)J45 0 3 o I
L0 0 0 0 0 1W'2i 0 21 J
( 4.2.1 )
The eigenvalues are doubly degenerate and the trace of HfJ is zero. Thus the
corresponding secular equation is of quartic fonn
(4.2.2 )
with the solutions given by [ 14 ]:
El =(R+D)/2
E2 =(R-D) 12
( 4.2.3 )
E3 =-(R-S) 12
E4 =-(R+S) 12
The various constants are
[1( Y)
D= 2 X+ R _R2 JII2
( 4.2.4 )
and
112 112 ]1/2
R = 6- [X + ( X 2 + 12Z) cos(o 13) (4.2.5 )
125
where
(4.2.6 )
Here X, Y and Z depend on the interaction parameters as
X =504(3+1'\2)
Y= 6912(1-1'\2) (4.2.7 )
Z = 15120(3 + 1'\2y
1ie 2 Q
The energtes EI'E2 ,E3 ,andE4 are in units of 8: and are ordered as
EI ~ E2 ~ E3 ~ E4 . In the case of 11 =0 the eigenvalues of the Hamiltonian ( 4.2.1 )

reduce to the simple results EI = 21,E2 = 3,E3 = -9,E4 = -15. The difficulty with
the quadrupolar spin 7/2 Hamiltonian (4.2.1) is that it is nondiagonal in 1'\. The matrix
which diagonalizes quadrupolar spin 7/2 Hamiltonian ( 4.2.1 ) is given by
(Ull 0 U13 0 U1S 0 U17 o1

I0 U22 0 U24 0 U26 0 U28 I
IU 31 0 U 33 0 U35 0 U 37 oI
I 0 U42 0 U44 0 U46 0 U4S1I
u=1
IU4S 0 U46 0 U44 0 U42 oI
I 0 U37 0 U35 0 U33 0 U31 I
I 0 U 26 0 U24 0 U22 oI
lU;a UI7 0 UIS 0 U13 0 ullJ
(4.2.8 )
where
126
1 ..,.J2I(( 15+ EJ(3- EJ + 45112)
Un = NI (El +9)((15+EJ(3-E1) + 45rJ2) +60.,2(EI -3)'
1 2112 .J3i5( 3 - E 1)
1S
U = Nl (El +9)((15+EI )(3-EJ +45112) +60.,2(EI -3)'
1 211 3 J14175
U -
17 - N} (El +9)((15+EI)(3-EJ+45T)2)+60.,2(EI -3)
(4.2.9 )
1
Un =-,
N2
1 1W'45((9+ E2 )(E2 - 21) - 21112)
U = N2 (15+E 2)((9+E2 )(E2 -21)- 21112)+6Or}2(21-EJ'
24
1 (E -21)112J27oQ 2
u
U = N2 (15+EJ((9+EJ(E2 -21)- 21112)+60.,2(21-E2)'
1 2113~14175
U
28 - N2 (15+E 2 )((9+E2 )(E2 -21)-21112)+60rJ2(21-E2)
( 4.2.10)
1 rrfij
U 31 = N 3 E 3 -21'
(4.2.11 )
127
( 4.2.12)
The constants N 1 , N 2 ' N3 and N 4 are chosen by normalizing each row of the U
matrix to unity. For 1'/ = 0, U reduces to the identity.
As an example, the doubly degenerate eigenvalue EI has eigenfunctions
( 4.2.13 )
These eigenfunctions are denoted by 1±1). The upper sign in eq. ( 4.2.13) refer to 1+1)
and the lower sign refer to 1-1) For the eigenvalues E2 ,E3 andE4 we similarly define
the eigenfunctions ofthe quadrupolar Hamiltonian as I±2) ,I ±3) and I±4).
Effect of rf pulses
The rfHamiltonian is given by
( 4.2.14 )
where the strength, CD 1 , the carrier frequency, co , and the phase, cP , describe the
pulse. The unit vector ii =(sin9 L coscPL,sin9L sin cPL ,cos9L) specifies the orientation
128
of the rf coil with respect to the quadrupolar principal axis system ( QPAS ). We
assume that Cl)l and cjl are nonsingular functions of time. The selection rules for spin
712 quadrupolar system are deduced by transforming the rf Hamiltonian ofeq.
( 4.2.14 ) into a representation where the static ( quadrupolar ) Hamiltonian ( 4.2.1 ) is

diagonal. These transformations are easily performed using 'MAPLE'. The matrix
elements ofthe angular momentum components in this new frame are lengthy and only
the elements ofinterest with respect to the problem in question are given below.
It should be noted that when 11 :;:. 0 not only "single quantum" transitions such
as 1±4} ~ 1±3) ,1±3) ~ 1±2} andl±2) ~ 1±1} are allowed but also "double
quantum", 1±4) ~ 1±2} ,1±3} ~ !±l) • and IItriple quantum", 1±4) ~ I±l}
transitions are possible.
The energy splittings between pairs of degenerate levels of spin 712 are in
general unequal and a single pulse acts selectively. If the carrier frequency of an rf
pulse is close to the frequency of the certain transition, this means that others are off
resonant and are not affected by the pulse. Thus the problem reduces to the study of
4 x 4 matrices, thereby considerably simplifying the problem. This is an example of the
"simple line" concept when the rf radiation influences exclusively one transition in a
spin system of arbitrary complexity [ 15 ]. The frequencies of the single and multiple
quantum transitions ofa spin 7/2 are shown in Fig. 4.2.1. It is seen from Fig. 4.2.1 that
accidental degeneracies may occur for certain values of 11. In particular, when 11 =0
the frequencies of the single quantum, 1±2) ~ 1±1}.. and double quantum,
1±4) ~ !±2) , transitions are degenerate. However, these cannot be excited
simultaneously as the DQ transition in question is forbidden by the selection rule
discussed below. For the other cases of degeneracies the treatment given below should
be modified.
129
3,__________
35
25
2
20~----====~3::::====____~4~__________________
1St
s
1
6
0.2 0.4 0.6 0.8
4.2.1. The frequencies of single and multiple quantum transitions co~as functions of 11:
1) co Q = E, - E4 (triple quantum transition), 2)co Q = E1 - E3(double quantum transition),
3)co Q = El - E4 (double quantum transition),4)co Q = E1 - El(single quantum transition},
S)co Q =E2 - E 3 (single quantum transition).6)co Q = E3 - E 4 (single quantum transition}.
Transitions 1±3) ~ I±l)
Consider, for example the "double quantum" transition 1±3) ~ I±l) . The
reduced quadrupolar Hamiltonian is given by the diagonal matrix with elements
E1 ,E3 ,E3 and E1 . The reduced rf Hamiltonian in the representation where the
quadrupolar Hamiltonian is diagonal is given as
J (0
la
a. 0 b+ ')
0 c. o I
( 0
I ia
-ia_ 0
0 -ic_
ib 'I
- I
oI
H if' = -2lim, Cru\OlI - op) sine cosop, l
L ~ c. 0 a.j+sine,sin·'lo ic_ 0 -i;_ )
l b. 0 a. o -4b_ 0 ia_
(-d 0 e 0,1
I 0 -d1 0
+cose'l ~ ~I}
~j
0 d1
-e 0
(4.2.15 )
(4.2.16 )
130
The interaction picture is introduced by the operator L as
L=~- ~imtR) (4.2.17 )
where R is a diagonal matrix with the elements of ( 1,-1,-1,1). Transforming the
reduced rf Hamiltonian eq. ( 4.2.15 ), into the interaction frame eq. ( 4.2.17 ) and
discarding high frequency tenns [ 6 ] gives
jj~ =_~l {Pcoscl> + Qcoscl>} (4.2.18 )
with
( 0 a-if3 y 0 1 ( 0 -ia- f3 -ty 0 1

Ia+if3 0 0 -y I
I _1I1a
. - f3 0 0 -iy I
I
P=l ~ '
a-I.~J Q- ty
J
0
-y
0
a+if3 0 l 0
0
ty
0
-ia+ f3
ia; f3
where
a =2a+ SinaL coscl>L '
f3 =2a_ sineL sincl>L' (4.2.19 )
Y =2ecose L ·
The values a+,a_ and e as the functions of'll are shown in Fig. 4.2.2.
The reduced quadrupolar Hamiltonian then can be written in a simple diagonal fonn
(4.2.20 )
where
(4.2.21 )
is the actual frequency of the transition and the contribution proportional to the unity
matrix is dropped as it resets the energy reference state. The total Hamiltonian in the
131
interaction picture is brought to the block-diagonal form by the unitary transformation
T [7],
(
I
ylr
0
o
o
0 (a-i~) Ir I
T=I
1 o I (4.2.22 )
l-(a+~p) IT Y~T J
-1 0
o 0
where
(4.2.23 )
to give
(
lir =.-n ilr "\
.../ro -ro) R'1rr+ =(h+ 0)
T l if +
Q
2 ) 0 h_
( 4.2.24 )
with
where the upper sign refers to h+ and the lower sign refers to h_ and
(4.2.25 )
It should be noted that the transformation T does not change the reduced quadrupolar
Hamiltonian ( 4.2.20) and the initial reduced density matrix is given simply by
~O)=R (4.2.26 )
132
The evolution of the reduced density matrix in the interaction picture is governed by
the Liouville equation
( 4.2.27)
Using eq. ( 4.2.24 ) we arrive at
(4.2.28 )
where jl(t) : ( ~ :J k labels + and - and the initial condition is given by eq.
( 4.2.26). The general solution to eq. ( 4.2.28 ) is
(4.2.29 )
Introducing f =0';2 / (1+0'11) to transform eq. ( 4.2.28 ) with k=+ is into a Riccati
equation [ 8 ]
( 4.2.30)
133
Equation ( 4.2.29) can be also written in terms of Euler angles, (XI ,131 and 11 ,
using the rotation operator approach [ 6, 16 ] and writing
(4.2.31)
which along with the initial condition of eq. ( 4.2.26 ), gives
0'11 =cos 131'

(4.2.32)
0'12 = - sin 13/1'1
Here I y and I z are spin operators for spin 112. Note that with this initial condition, the
result ( 4.2.32 ) does not depend on (XI' This fact can be used to establish the relation
between the Euler angles and the Riccati equation.
To assess the response of the system to an rfpulse, following [ 17], we define
the quantities
(I. iil =(Ir .ii)COSOlt ,

(I. iil = (I ii) sin
( 4.2.33 )
r • Olt
where the corresponding reduced angular momentum components lx' ,Iy ' andIz' of
spin 7/2 are in the representation where the quadrupolar Hamiltonian is diagonal.
These can be deduced from eq. ( 4.2.15 ). Transforming eq. ( 4.2.33 ) into the
interaction picture ( 4.2.17 ) and truncating high frequency terms results in
134
(4.2.34 )
In the same representation as the density matrix, see eq. ( 4.2.29 ), defined by the
transformation T, eq. ( 4.2.22), operators P and Qare ofthe form
(0 r 0 0) (0 -ir 0 0)
Ir 0
p=1 0
0 0\ Jir 0 0 01
~J ,Q l~
.( 4.2.35 )
l~ i~J
0 0 0
0 r 0 -ir
The response ofthe system can then be represented by the following quantities
w; =tr{~ pP(t)} ,
( 4.2.36 )
"i = tr{ ~ QP(t}}
Using the Euler angle parametrization ofthe density matrix ( 4.2.32 ) gives
W; = -r sin 13 1 cosy 1 '

(4.2.37 )
"i =r sin~l sinYl .
Ramamoorthy introduces the concept ofinversion of quadrupolar systems [ 18 ]. This

is defined in our notation for a quadrupole as
( 4.2.38 )
which gives
135
(4.2.39 )
W; is inverted for ~1 =1t. Note that the quantity W; does not represent the signal
actually measured in NQR spectrometers.
It is clear from eq. ( 4.2.28,29,32,36,37) that the response ofthe system is
determined by the sum of the two Pauli spin matrix type responses. We can thus
introduce two pseudospins 1/2 to illustrate the physical picture of the evolution. The
corresponding components ofthe associated magnetization vectors are
M,/ =-sin~l COSYl Mo

M/ = sin~l sinY l Mo (4.2.40 )
M/ =±cos~lMo
where Mo = rmo and mo is the constant.
On-resonance solution and the NQR 1t / 2 pulses

Consider now the application of the rf pulse to an on-resonance situation i. e.
Aro = 0 . Solving the Riccati equation we obtain Euler angles of
1t
Yl =--4>
2
( 4.2.41 )
~I =j ro1ejf(t')dt'
o
From eq. ( 4.2.40 ) we find that our system can be visualized as being composed of
two pseudospin 1/2 with equal Mx and My components and oppositely directed M z
components. The simple magnetization trajectories on-resonance are obtained from eq.
1t
( 4.2.41 ). It is clear that when ~l ='2 these two associated magnetization vectors are
136
aligned, thereby allowing an NQR definition of a x 12 -pulse. On the other hand when
~1 = x these are oppositely directed from each other and their initial direction. This
situation can be considered to correspond to a x -pulse. The duration of these pulses
are determined by the effective frequency of the applied rf field. In the case of the
single crystal the duration of x -pulse is twice as long as that ofax 12 -pulse. In the
case of powders the response ofthe system is represented by the integration of
eq. ( 4.2.37 , 39) over polar and azimutal angles
(4.2.42 )
where p=1,2,3. The duration of the xandx/2 pulses in this situation can be
determined, for example, by setting 4> =O. In this case a x 12 -pulse corresponds to the
maximum of~ as a function of time whereas a x-pulse corresponds to the minimum
of ~ . The application of composite and shaped pulses to such systems can be
considered in a similar manner to [ 6,7 ]. The main feature of the rectangular pulse is
the extensive side band excitation beyond resonance conditions. Shaping of the pulse
can produce complete, localized and uniform excitation profile [ 8 ]. The shape of the
pulse determines the excitation profile, the effective width of which is controlled by
suitable pulse parameters [ 8, 19]. For a pulse to selectively excite a given transition it
is important that the off resonance profile of the pulse be considerably less than any
neighboring transition [ 19 ].
Comparison with corresponding spin 1/2 system

The response of the system in the laboratory frame can be found in the
following manner. Transforming the density matrix of eq. ( 4.2.29 ) from the
137
interaction picture to the laboratory frame and using the Euler angle parametrization of
eq. ( 4.2.32 ) only requires change
(4.2.43 )
Now we transform the density matrix of eq. ( 4.2.28 , 29 ) to the basis

11),1-3) ,13)andl-1) to obtain
( all o '\
,.j) II(0.+;/3)0'·12 Ir -(j
• I
121 r II
1-'\1 T =
=T+'";:./ }
f.II,l
0'·121 I r -0'11 (a. - ;/3)0'·12 I rJ'
o (a. + ;/3)0'12 I r all
(4.2.44 )
Then the response is given by
(4.2.45 )
or using eq. ( 4.2.32 , 43 )
( 4.2.46 )
The response of eq. ( 4.2.42 ) is similar to the response ofa corresponding response
112 system in the laboratory frame. Apart from the factor 2r this type ofthe response
is given by spin 112 with the following changes
( 4.2.47)
138
The corresponding response to eq. (4.2.46) for a spin 112 is given by
(4.2.48 )
where I xl /']. and P1I2(t) are the x spin 1/2 operator and the spin 1/2 density matrix in the
laboratory frame respectively. The equations ( 4.2.18-46 ) are general for 1/2 integral
spins. For example, substituting the Euler angles and co leff for spin 3/2 for a pulse of
the constant amplitude and phase [6] into eq. ( 46) recovers eq. (30) off 1 ].
Transitions !±4) ~ !±2)

In order to consider the "double quantum" transition, 1±4) ~ 1±2), we create
the reduced matrices ofthe fonn of eq. ( 4.2.15) and ( 4.2.20 ) where now the states
are 14}.1-2}.12)J-4}. The calculation of the pulse response is similar to the
1±3) ~ !±l) transition considered above. The main difference in the result is the
appearance ofthe different co I eff given by
CD I "" =CDlr, (4.2.49 )

r=~a.2+~2+y2
a. = 2a+ sinaL CQsch.

~ = 2a_ sinaL CQSch,
y=2eCQSaL
where aL and cP L are polar and azimutal angles specifying the orientation of the rf coil
with respect to QPAS ofthe crystal and a different CD Q' given by
(4.2.50 )
139
Here
a± =±]..,J7U
2 48 U22 +J3U22U46 ±]...J15U24U46 +2U24U44 ±]...J15U26 U44
2 2
+.J3U26U42 ±!,J7 U28U42 , ( 4.2.51 )
2
5 1 3 7
e=-U22U42 +-U24U44 --U26 U46 --U28U48 •
2 2 2 2
For the "single quantum" transitions and the "triple quantum II transition the
states can be included in the reduced rf and quadrupolar matrices in the similar manner
in order to apply the above formulation. For example, for the "triple quantum"
transition, 1±4)--+1±1), the 4x4 matrices are given in the 14),/I),I-l)andl-4)
representation. The values a±, and e required for the calculation of the effective
frequency are
and now
( 4.2.53 )
Conclusions
Note that when 1'\ =0 all effective frequencies of multiquantum transitions
vanish. For the single quantum transitions we obtain, ( for 1'\ *0 )
140
These as functions of 11 are shown in Fig. 4.2.3. Comparison with Fig. 4.2.1 shows that
in the case of double quantum transitions the effective frequencies are considerably
smaller ( and thus the pulses should be longer to achieve the same excitation) than in
the case of single quantum transitions.
In the case of T/ =0 the expressions ( 4.2.54-56 ) reduce to the simple results [ 6 ]
141
1.6
1.4
1.2 2
0.8
0.6
0.4
0.2
3
0.2 0.4 0.6 0.8
4.2.2.The values a+ ,a_ ande as functions of 11 for the double quantu

m transition
i±3) -+ j±l) : l)-a+ . 2)-a_ .3) - e.
0.3
0.25
0.2
0.15
0.1
0.05
0.6 0.8
4.2.3. The values a...a_ ande as functions of 11 for the single quantu
m transition
1±2) -+ I±l) : l)a... 2)a- .3)- e.
3 r;:;.
CO Q = 14 e2 qQ ,rold' = COl "1/7 sme L ;
CO Q =71 e qQ,co1d' =ro l "l/12sme
2 r:;;; .
L ; (4.2.57)
1 2 ~ •
CO Q = 14 e qQ.co1d' =ro(v15sme L •
The results of the paper are impossible to obtain without the aid of computer
algebra. Even with computer algebra, it is necessary to organize the treatment in the
way which lead to a curtailment of the long and unmanageable expressions that can
occur. This was not a trivial task and extensions to the higher spin cases is not routine.
The results here are practical and easily programmable to give numerical
results in special cases. We checked our analytical results for each transition
numerically using 'MAPLE' built in routine for the numerical solution of eigenfunction
- eigenvalue problems, an example of which is given in Appendix B. Moreover, using
similar techniques, it is straightforward to modifY the initial conditions ( 4.2.26 ) to
include other cases. Hence, a sequence ofpulses can be treated in NQR for spin 7/2.
Finally, we have been able to give a definition and understanding of 7t 12 and 7t
pulses in NQR. The physical interpretation is different from NMR, which is based
upon a vector, in that NQR requires two vectors which can be parallel or antiparallel
to produce a 7t 12 and 7t type pulses respectively. The length of these pulses depends
upon which transitions are being excited.
142
AppendixB
BifIINiilllfH##!J!!J!lIl1l1iJ1ll11"flUUIJIlf#h':'MtflfIfIlIUliiNIlff:'nHrumlllffJfNf¥IfIUUIJJ!J#III;'JJi#l#h~Nf#I:##I##
This Maple program numerically ealculates matrix fonns of spin components in the
representation where the quadrupolar Hamiltonian is diagonal for arbitrary '11.
NNlIlIIUII##HIJU}!{IiJUI/}/:tlNJ;'NII....IIJlJllllllllffYII#Uh·llltllllk·/JI#HHtlJliHlJltiIHIIIIlIllfll##JIII##!U#lIUUUI
llset the number ojdigits to J5 and spin magnihltie to 7/2

Digits:=IS;
i'=712;
II the next seventeen lines are used to calculate spin components
#1•• 17 , I. and i_q = i(i + 1)-3i_z2 in II.M} basis
1ISize:-2*i+l;
i..,PIus:=art'3)'(sparse, I..nsize.l .. nsizc): UnillUS:-amy( I ..nsize.l ..n.sizc):
Cx:-anay(Lnsize. Ln.sizc):
i .Y"'IlTIy( I .. nsize. LIISize): i_z:=amy(spIne, l..n.sizc, l ..n.sizc):
i_q:=anay(sparse, l ..lISize.l ..IISize):
i row:';1;
tOr m fJOm ·i 10 i·I do
i..,Plus(i.row.Crow+l):"'CVIIf{sqrt({i+m+l)'*(i-m))):
i_q(i.row.i_rowJ:-i·(i+ 1)-3·m**2;
i_zli.row,i.rowJ:= -m;
j_row:= Crow + 1;
Old;
i.zli.IOW-i.IOW]:= .j;
i_q(i.lOW,i.rowJ:=i*(i+I)-)*j**2;
i.minus;""e\I3lm(lJaIISp05e(iJllus»;
i_,,:=evalm(tplus + i..miDUS)I2;
iJ:=cvaIm«i"'plus • i_miDUS)l(2*1);
i:.set Tl:; 0.385 andca/culate RQ{units of Ae 2qQ /84)

eta:=O.38S;
Uleta:=cva!m(-i.q+etaI2*{i...PlUS&*i"'p1us+i_minus&·i..millUS»;
# solve eigerIVQlue and eigenfunction problemfor HQ and assign the matrix
# containing eigenfunctions to T . Note that Maple arranges eigerIVQluesfrom lowest
#10 highest and coloumns oJT contain corresponding eigenfunctions.
CYalf(Eigcnvals(i_ qeta,wcs»;
T:=cvaIm(wcs);
tl :""lraDSpOIle(;
II calculate spin components and quadrupolar Hamiltonian in the representation
:twhere the quadrupolar Hamiltonian is diagonal
Umllw:=e'\-aIm(t II:*i_x.&*T);
i.,YDeW:"'CValm(tll:*jJI:.T);
i_ZDeW:=e'\-alm(t II:*j_z&: *T);
Uletanew:"'CValm(tll:*Ulcta.t*T);
143
References.
1. Pratt, 1. C., Molec. Phis., 49, 539 ( 1977 ).
2. Vega, S., 1. Chern. Phys., 68,5518 ( 1978 ).
3. Ramamoorthy, A, and Narasimhan, P. T. Molec. Phys., 73, 207 ( 1991 ).
4. Zueva, O. S., and Kessel, A R, Sov. Phys. JETP, 46, 1136 ( 1977).
5. Krishnan, M. S., Temme, F., and Sanctuary, B. C., Molec. Phys., 78, 1385 ( 1993 ).
6. Ageev, S. Z., Isbister, D. 1., and Sanctuary, B. C., Molec. Phys., 83, 193
( 1994).
7. Ageev, S. Z., and Sanctuary, B. C., Chern. Phys. Lett., 225, 499 ( 1994 ).
8. Warren, W. S., and Silver, M. S., Adv. Magn. Reson., 12, 247 ( 1988 ); McDonald,
S., and Warren, W. S., Concepts Magn. Reson., 3, 55 ( 1991 ).
9. Char, B. W., Geddes, K. 0., Gonnet, G. H., Leong, B. L., Monagen, M. B., and
Manual, Springer - Verlag ( 1993 ).
10. Ramamoorthy, A, Molec. Phys., 72, 1425 ( 1991 ).
11. Hatanaka, H., Terao, T., and Hashi, T., 1. Phys. Soc. Jpn., 39, 835 ( 1975 );
Hatanaka, H., and Hashi, T., 1., Phys. Soc. Jpn., 39, 1139 ( 1975 ).
12. Reddy, R, and Harasimhan, P. T., 1. Molec. Struct., 192, 309 ( 1988 ).
13. Bloch, F., and Siegert, A., Phys. Rev., 57,222 ( 1940); Furman, G. B.,
Z. Naturforsch., 458,565 ( 1990).
14. Creel, RB., J. Magn. Reson., 52,515 ( 1983 ).
15. Wokaun, A., and Ernst, R R, 1. Chern. Phys., 67, 1752 ( 1977).
16. Zhou, 1, Gao, H., and Sanctuary, B. C., 1. Magn. Reson., AI02, 137 ( 1993 );
Zhou, J., Ye, C., and Sanctuary, B. C., 1. Chern. Phys., 101,6424 ( 1994 ).
17. Haeberlen, U, High resolution NMR in solids : selective averaging, Academic
Press, N. Y. ( 1976 )
144
18 . Ramamoorthy, A., Chandrakumar, N., Dubey, A. K., and Narasimhan, P. T., J.
Magn. Reson., Series A 102, 274 ( 1993 ).

19. Gaggeli, E., and Valensin, G., Concepts Magn. Reson., 4, 339 ( 1992 ) and
references cited therein.
145
Contributions to Original Knowledge
Part 2. The analytical solutions for the density matrix: of spin 7/2 excited by one and
two rf pulses are new. The first order quadrupolar interaction is retained while the
pulses are on. Using the symmetries ofthe density matrix after one pulse experiment, a
number of pulse sequences with various phase cycling is calculated. This include even
and odd detection sequences, a rotary echo type sequence and spin lock pulse
sequences which can be used for manipulation of MQ coherences developed during
the first pulse. In addition, fifteen Solomon echoes are predicted. The results are
applied to determination of quadrupolar parameters and true chemical shifts using ID
nutation spectroscopy and verified by the experiment.
Part 3. The density matrix: of spin 3/2 subject to a soft off-resonance rf pulse is
calculated analytically for the first time. The results are valid for arbitrary values ofthe
rf amplitude, quadrupoar coupling and offset. The simulated spectra are in perfect
agreement with Kentgen's experiment.
Part 4. NQR pulses are treated in a consistent way leading to expressions based upon
SU ( 2 ) algenra for non-zero asymmetry parameter f1 and arbitrary orientation of the
rf field. The definition and physical understanding of 1t / 2 and 1t pulses is given. It is
shown that composite pulses designed for NMR can be used for NQR A number of
pulse sequences is calculated for the case ofa single crystal and powders and shown to
be as good as or better then existing composite pulses optimized exclusively for NQR
In addition, shaped pulses are introduced to NQR The results are generalized for
selective excitation ofMQ coherences of spins ofmagnitude higher then 3/2.
146
Suggestions for Future Work
Part 2. The calculations can be extended to Hahn echoes taking into account the first
order quadrupolar interaction and heteronuclear dipole-dipole interaction during the

evolution and detection periods Furthennore, pulse sequences, discussed in Part 2 can
be solved numerically for spin 9/2.
Part 3. The results can be modified to include the second order quadrupolar
interaction during the pulse. Using computer algebra, it is possible to extend
calculations to two and three pulse sequences. In particular, these can be used to
analyze Raman multi quantum experiment for spin 3/2. Relaxation can also be
considered.
Part 4. The analysis of composite and selective pulses can be extended to integer spins
such as spin 1 and spin 3. For some special cases these can be treated as rotations.
However, for general field orientation and 11 ¢ 0 the problem becomes more
complicated as the double degeneracy of the energy levels is lifted. Relaxation

phenomena can also be studied.
147
General Conclusions
Two distinct cases are considered in the present thesis. The first is the solid state
NMR where the quadrupolar interaction dominates and is retained during the rf pulses.
The second is the case of NQR in the absence of external magnetic fields. A spin 7/2
system is chosen for on - resonance studies of solids. The response to one rf pulse is
shown to be a sum of sixteen terms. Usually, up to seven of these terms can be significant.
The shape of the curve representing the response as the function of the pulse length
depends upon the ratio of the quadrupolar coupling constant to the frequency of the rf
field. In the case of this ratio being large, we deal with selective excitation of the central
transition whereas if this ratio is small, the common NMR situation is recovered. A variety
of multiple pulse sequences for manipulation of MQ coherence developed during the first
pulse are calculated. Solomon echoes are predicted for soft pulse excitation. Six out of
fifteen echoes are the result of refocusing of exclusively MQ coherences. The results are
shown to be useful for the determination of quadrupolar parameters and true chemical
shifts in single crystals and powders. It should be noted that the true chemical shift of a
line can be related to the mean bond angle. In particular, a bismuth vanadate ferroelastic
crystal was chosen to test the theoretical predictions. The results show excellent
agreement between theory and the experiment. However, in the case of on resonance
nutation spectroscopy, the difficulty can arise for large quadrupolar coupling constants.
This problem can be resolved by using off - resonance spectroscopy. Here we use a spin
3/2 to calculate, analytically, line intensities in the case of non - zero offset. The response
for the central transition is shown to be the sum of sixteen terms in contrast to on
148
resonance spectroscopy where only four terms are present in the solution. Clearly, off
resonance spectra exhibit more features and this extends the applicability of nutation
spectroscopy. On the other hand, the results obtained can be used for Raman NMR. and
can be extended further to at least two pulse sequences.
The theory of NQR pulses is based upon the idea of selective excitation since the
zero magnetic field splittings are unequal and only one specific MQ transition is excited.
The orientation of the rf field is assumed to be arbitrary which allows us to treat powders
by averaging over space angles. The excitation of such systems can be considered in the
interaction representation as a sum of the responses of two uncoupled spins of 112 which
have equal x and y projections of magnetization vector but their z components are
oppositely directed. Thus the pulses are treated as rotations and examples of spins 3/2, 5/2
and 7/2 are presented all with non - zero asymmetry parameter. The important difference
from NMR. is the appearance of an effective rf field frequency which depends on the
orientation of the crystal to the applied rf field, asymmetry parameter and transition in
question. It should be noted that MQ coherences can be excited for spins greater then 3/2
with one pulse experiments provided that the asymmetry parameter is not zero. The
important consequence of our theory is that many NMR. pulse techniques can be applied
to NQR directly. In particular, it is shown here that NMR composite pulses can be used in
pure NQR. In addition, we allow an arbitrary nonsingular modulation of the applied rf
field, so that shaped pulses can be calculated. It can be of great importance to NQR of
powders as these much superior to composite pulses and allow insight into the electron
density surrounding nuclei.
149

1997 Ageev

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1997 Ageev

Uploaded by

Copyright:

Available Formats

APPLICATION OF DENSITY MATRIX METHODS TO

QUADRUPOLAR SPINS IN SOLID STATE NMR AND NQR

A thesis submitted to the Faculty of Graduate Studies and Research in partial

fulfillment of the requirements for the degree ofDoctor ofPhilosophy.

(:) 1997 by Sergei Z. Ageev

Acquisitions and Acquisitions et

Our file NoIre nffllrence

The author retains ownership ofthe L' auteur conserve la propriete du

appropriate phase cycling designed for detection of MQ coherences developed during

of quadrupolar parameters and true chemical shifts utilizing a ID nutation experiment.

magnitude greater then 3/2.

de pulsation. Des sequences de pulsation possMant des phases cyclees appropriees

parametres quadrupolaires et les relais chimiques reels pendant une experience de ID

quadrupolaire du premier ordre est retenue pendant la pulsation. La troisieme cas

It is my pleasure to thank to Professor B. C. Sanctuary for professional and

personal support and guidance throughout my graduate studies in Chemistry.

in studies of spin 712 dynamics in solid state NMR.

Valuable discussions with Drs. M. S. Krishnan, D. J. Isbister and F. P. Temme

are gratefully acknowledged.

a pleasant research environment.

The financial support of McGill University is also recognized.

In particular, papers 1 through 4 are presented in Chapters 1 to 4 of Part 2.

List of Tables Xl11

Part 2 Theory and Applications of Spin 7/2

Part 3 Analytical Solutions for Spin 3/2 Off-resonance

Figure 2.2.7., 8. Log-log plots of the frequencies IWif I/ WI as the functions of wQ / WI

along positive x axis.

2.3.3 ) with opposite sign.

the time delay between them.

2.3.3 ) with opposite sign.

2.3.3) with opposite sign.

2.3.10) with opposite sign.

Figure 2.3.10 Fit of 51 V experimental line intensities ofFig. 2.3.9.

the functions ofthe second pulse length.

the functions of the second pulse length.

as the functions ofthe second pulse length.

functions ofthe second pulse length.

functions of the second pulse length.

functions ofthe second pulse length.

as the functions ofthe second pulse length.

the functions of the second pulse length.

functions ofthe second pulse length.

functions ofthe second pulse length.

functions ofthe second pulse length.

functions ofthe second pulse length.

functions ofthe second pulse length.

the case of L\oo = 00 1 .

for the situation, 1; =10 / 00 1 and L\oo =00 1•

crystal in the powder sample.

Figure 4.1.3. The composite1l"-pulse of 860 -160180 -260 9 ,

Table 2.1.1. The matrix representation of il / Ii , Al and HAl/ii 38

Table 2.2.2.2. The submatrices in eq. (2.2.6) 40

Table 2.3. 1.The (I~i) functions used in eq. (2.3.3,2.3.4,2.3.9,2.3.10) 52

Table 2.3.2. The functions C;(t z ) used in the equations

(2.3.3,2.3.4,2.3.9,2.3.10) for the density matrix component P54(tl' 't,tJ 53

Table 2.4.1. Values Kl with a. = ZliXZi and f3 = 1";;Yzi 69

along the y axis 80

Table 3.2.1. The matrix form of the Hamiltonian H(O) 95

Fourier transform ( FT ) methods, an enormous number of ingenious pulse techniques

manipulate nuclear spins in prescribed fashions.

magnetization as a vector in 3D space. This model fails to provide an adequate description

states to be considered. Moreover, such nuclei are subject to quadrupolar interactions,

increasingly attracting attention of researches. The object of the problem is to extract