(Slichter, 1989) - Principles of Magnetic Resonance (3rd Ed.)
(Slichter, 1989) - Principles of Magnetic Resonance (3rd Ed.)
(Slichter, 1989) - Principles of Magnetic Resonance (3rd Ed.)
• To obtain values in 51 units multiply value in Gaussian units by the respective ratio
The effect of Ul",U- 1 for various U's and for a = X,Y,Z. The bottom line applies
for spin! only
1. I, 1,
U
e i91 % 1% I,oos8-/~sin8 I~ cos 8+1, sin 8
ei(Jl y /%cos8+I~sin8 I, /~ cos8-/% sin 8
ei91~ /% cos 8 -I, sin 8 1,0058+1% sin 8 I,
e-i(JI~S~ 1. c05(8/2) 1, 000(812) 1,
+ 1,(25, 5;0(812)) \ • • - 1.(25, ';0(p/2))
. ,
'.'
;,
:1 ','
... - ,
. .
,
.,
. ..
1 Springer Series in Solid-State Sciences
Edited by Peler Fuldc
Springer Series in Solid-State Sciences
Editors: M. Cardona P. Fuldc K. von Klitzing H.·J. Queisscr
c. P. Slichter
Managing Editor: H. K. V. Lotsch Vollllnt$ 1-49 a'C' listed at l~ md of the boot
The first edition of Ihis book was wrilten in 1961 when I was Morris Loeb
LeclUrer in Physics 31 Harvard. In the preface I wrote: "TIle problem faced by
a beginner today is enonnous. If he attempts to read a current aniele. he oftcn
finds that the first paragraph refers to an earlier paper on which the whole article
is based. and with which the author naturally assumes familiarity. That reference
in turn is based on another, so the hapless student finds himself in a seemingly
endless retrcat.
I have felt that graduate siudems or others beginning research in magnetic
resonance needed a book which really went into the details of ca1culUlions, yet
was aimed al the beginner rather than the expert."
The original gool was to treat only those topics that are essential to an
understanding of the literature. Thus the goal was to be selective rather than
comprehensive. With the passage of time, important new concepts were becoming
SO all-pervasive that I felt the need to add them. That led to the second edition,
which Dr. Latsch, Physics Editor of Springer~Verlag,encouraged me to write and
which helped launch the Springer Series in Solid-State Sciences. Now, ten years
later, that book (and its 1980 revised printing) is no longer available. Meanwhile,
workers in magnetic resonance have continued to develop startling new insights.
There are new topics which are so important that they must be included in a
book that is intended to be an intrOductory text. The original course was one
semester - the book had more topics than I taught in that amount of time. The
new edition clearly would need a full year.
Throughout the history of this book, I have considered it as a textbook.
Thus the aim was to explain in a rigorous but physical manner concepts which
are essential for the student of magnetic resonance. Rather than giving an ex-
haustive treatment of anyone topic, my intention has been to help prepare a
student to read the literature on that topic. The main additions include an en-
largement and modest rewrite of the topic of double resonance, explanations of
one- and two-dimensional Fourier tmnsfonn methods, of coherence transfer, of
multiple quantum coherence, and of important topics related to dipolar coupling
that underlie the method of spin-flip line narrowing. In the earlier editions, the
chapter on the density matrix focused almost exclusively on its use in analyzing
relaxation processes. The new edition explains its use in analyzing the effect of
rf fields.
v
The first edition of this book came oot shortly after the publication of Ana-
tole Abragam's classic-to-be Principles of Nuclear Magnetic Resonance. There Contents
were not many other books on the market Today there are many books on mag-
netic resonance, a large fraction written by some of the most illustrious founders
of the field and practitioners of the art. The styles of these books are as varied as
the scientific styles of their authors. As I have explored these marvellous books,
I have been struck by the opportunity that exists today to get the flavor of the
scientific thought of these great scientists.
When I wrote the first edition of this book at Harvard, I had no duties other
than preparing and giving the lectures. The third edition has been written on
1. Elements of Resonance .........................•.......... 1
evenings and weekends while I tried to carry on my nonnal activities. It would
1.1 Introduction ...................................•..•. 1
not have been possible without the expert, dedicated. and always cheerful support
1.2 Simple Resonance Theory ........................• _••. 2
of my secretary, Ann Wells. Her association with the book has been long, since
1.3 Absorption of Energy and Spin-Lanice Relaxation . 4
she helped type the original edition. I am also grateful for the help of Jamie
Froman in the final stages.
2. BasicTheory . 11
2.1 Motion of Isolated Spins-Classical Treattnent . II
Urbana, Illinois
2.2 Quantum Mechanical Description of Spin in a Static Field 13
June 1989 Charles P. SUchter
2.3 Equations of Motion of the Expectation Value . 17
2.4 Effect of Alternating Magnetic Fields . 20
2.5 Exponential Operators . 25
2.6 Quantum Mechanical Treatment of a Rotating Magnetic Field 29
2.7 Bloch Equations . 33
2.8 Solution of the Bloch Equations for Low HI . 35
2.9 Spin Echoes . 39
2.10 Quantum Mechanical Treatment of the Spin Echo . 46
2.11 Relationship Between Transient and Steady-State Response
of a System and of the Real and Imaginary Parts
of the Susceptipility . 51
2.12 Atomic Theory of Absorption and Dispersion ........•.•.. 59
Problems 557
Appendixes . 579
A. A Theorem About Exponential Operators ............••..... 579
B. Some Further Expressions for the Susceptibility ........•. , ... 580
C. Derivation of the Correlation Function for a Field
That Jumps Randomly Between ±ho , . 584
D. A Theorem from Perturbation Theory ...............••..... 585
E. The High Temperature Approximation ................• _.... 589
F. The Effects of Changing the Precession Frequency-
Using NMR to Study Rate Phenomena . 592
x
XI
1. Elements of Resonance
1.1 Introduction
2 3
basic interest in the theory of nuclear structures. they would take us rather far -1(2 _ _-,-,,---_ N_ FIg. 1.2. Energy levels rOT 1 =;
afield. However, a simple classical picture will enable us to make a correct yAHo
+1(2 N.
order-of-magnitude estimate of ..,.
Let us compute the magnetic moment and angular momentum of a particle
of mass m and charge e moving in a circular path of radius r with period T.
The angular momentum is men
21rr 2 ~+ : N_WH _(+) - N+W(+)_H (1.13)
J = mvr = m (1.10)
T Without as yet attempting to compute W(+) ..... (_} or W(_) ..... (+). we note a
while the magnetic moment (treating the system as a current loop of area A famous formula from time-dependent perturbation theory for the probability per
carrying current i) is second Po; ..... " that an interaction Vet) induces a transition from a stale (a) with
energy Eo to a Slate (b) whose energy is E,,:
~:iA (1.11)
Since i = (elc)(llT), we get p._, 2; l(blVla)I'.(E. _ E, _
= r",,) (1.14)
e 'lfr 2
~. -- (1.12) Since l(alVlb)1 2 :::: l(bIVla)1 2, we note that po; ..... " is the same as the rate
c T P" ..... o. Such' an argument describes many situations and leads 10 the condition
Comparison of the expressions for 1J and J therefore gives us "'/ = e/2mc. Besides
W(+)_H: WH_(+) " W.
enabling us to make an order of magnitude estimate of the expected size of "'/.
for our purposes the imporlant resuh of this fonnula is that large masses have
dN+ = W(N_ _ N+) (1.15)
low ",/'s. We expect about a factor of 1.000 lower.., for nuclei than for electrons. dt
In fact. for magnetic fields of 3,()(X) to 10,000 Gauss. electronic systems have a
resonance at w/21r = 10,000 MHz (the 3 em microwave region), whereas nuclear It is convenient to introduce the variable n:::: N+ - N_. the difference in pop-
systems are typically 10 MHz (a radio frequency). Of course one can always ulation of the twO levels. The two variables N+ and N_ may be replaced by n
change w by changing H o, but in most cases it is advantageous to use as large and N. using the equations
a magnetic field as possible, since the quanta absorbed are then larger and me (1.16)
resonance is correspondingly stronger.
In later sections. we shall comment somewhat more on typical experimental (1.160)
arrangements.
Substitution of (1.16a) inlo (1.15) gives us
1.3 Absorption of Energy and Spin-Lattice Relaxation dn (1.17)
-::::-2Wn
dt
We now wish to go a step further to consider what happens if we have a macro- the solution of which is
scopic sample in which we observe a resonance. For simplicity we consider a
system whose nuclei possess spin ~ (Fig. 1.2). Since there are many nuclei in n:::: n(0)e-2Wt (1.18)
our macroscopic sample. we shall specify the number in the two m states +l where nCO) is Ihe value of n at t =O. We note mat if inilially we have a populalion
and - , by N+ and N_. respectively. difference, it will eventually disappear under Ihe action of the induced ttansitions.
The total number of spins N is a constant, but application of an ailernating The rate of absorption of energy dEldt is given by computing the number of
field will cause N+ or N_ to change as a result of the transitions induced. Let us spins per second that go from the lower energy to the upper. and by subtracting
denote the probability per second of inducing the transition of a spin with m = +! the number that drop down. emitting energy in the process:
to a state m:::: -~ by W(+)-+(_). We shall denote the reverse transition by
W(-l ..... (+} . We can then write a differential equation for the change of the (1.19)
popu ation N+.
4 5
Therefore, for a net absorption of energy, n must be nonuro; that is. there By using (1.20). we find that the ratio of W 1 to Wr is not unity but rather is
must be a population difference. We see that when the upper state is more highly
populated than the lower, the net absorption of energy is negative - the system W! =e'Ylllo/kT (1.22a)
supplies more energy than it receives. This state of affairs is the basis of the os· .wr
cillators or amplifiers known as masers (microwave amplification by stimulated It is natural to wonder why the argument given to show the equality of W(+) _ (-)
emission of radiation) or lasers (for light amplification). and W(_)_(+) does not also apply here. The ~Iution of this paradox is th~t
We see that if the equations we have put down wen:: complete. the resonant the thennal transition requires not only a couphng but also another system 10
absorption of energy would eventually stop and the resonance would disappear. an energy state that permits a transition. We can illustrate by assuming that the
A more serious difficulty is seen if we assume W = 0 (that is. we do not reservoir has only two levels whose spacing is equal to that of the nuclear system.
apply the alternating magnetic field). Under these circumstances our equations If the nucleus and reservoir are initially in the states of Fig. 1.3a given by the
say that dN+/dt = O. The populations cannot change. On the other hand. if crosses conservation of energy is satisfied by simultaneous transitions indicated
we applied a static field to a piece of unmagnetized material. we should expect by the :m.ows. The nucleus may therefore give up energy to the lattice. On the
it to become magnetized. 11Je preferential alignment of the nuclear moments other hand, if both systems are in the upper state (Fig. 1.3b), the simultan~~us
parallel to the field corTCsponds to N+ being greater than N_. (N_ = 0 would transition cannot occur because it does not conserve energy. The rate of translUon
represent perfect polarization, a state we should not expect to find at temperatures of the nucleus will therefore depend not only on the matrix elements but also on
above absolute zero). The process of magnetization of an unmagnetized sample. the probability that the reservoir will be in a state that permits the transition.
therefore, requires a net number of transitions from the upper to the lower energy
,. aT
state. In the process. the spins give up energy - there is. so to speak. a heat Nucleus Reservoir
Nucleus Reservoir
transfer. Therefore there must be some other system to accept the energy. If we
ask how big a population difference will eventually be found, the answer must
2~
depend upon the willingness of the other system to continue accepting energy.
Speaking in thennodynamic terms, the heat flow will continue until the relative
b "
populations N_/N+ correspond to the temperature T of the reservoir to which (bl
the energy is given.
(al
FIg. 1.3. (a) A possible ~ransition. (b) A forbidden tnmsition
The final equilibrium populations N~ and N~ are then given by
6 7
That is, the nuclear levels will have the same relative populations as do those of dE W
the lattice. The nuclear population will therefore be in thermal equilibrium with dt = nliwW = n o1iw 1 +2WTt (1.34)
that of the lattice. Note. moreover, that for this simple model, we can compute
WI and W": We shall see later that W is proportional to the square of the alternating magnetic
(1.26) field. Therefore (1.34) tells us that we can increase the power absorbed by the
nuclei by increasing the amplitude of the alternating field. as long as 2WT1 <t:: 1.
so that Wl and W! arc seen to be unequal. However, once W is large enough so that W -II2TI, this statement is no longer
We now leave our special model and return to (1.21). By making the sub- true. The power absorbed levels off despite an increase in W. This effect is called
stitutions of (1.16a) for N+ and N_, we find "saturation". Provided one has enough information to compute W (a situation
dn often realized), one can measure T 1 by observing the saturation effect.
dt = N(W" - W)) - neW" + WI) (1.27) We have now seen several quantities that will be important in describing a
magnetic resonance. The quantity T 1 will clearly be related to the microscopic
which can be rewritten as details of both the nuclear system and the reservoir. We shall wish to consider
dn no-n what mechanisms may give rise to spin-lattice relaxation, and how to compute
----- where (1.28)
dt - Tl TI for any assumed mechanism. In the early work on nuclear resonance. it was
feared that the spin-lattice relaxation might be so slow that a population excess
Wl- WI) I might not be achieved within reasonable times. The famous Dutch physicist CJ.
no =N ( W!+Wl =(WI+W)) (1.29)
TI Gorter, who has made so many of the important discoveries and proposals in
connection with magnetic relax.ation, was the first person to look for a magnetic
Since the solution of (1.28) is resonance in bulk matter [1.1]. That he failed was probably due to his bad luck
n = no + Ae- t / 7i in having a sample which was easily saturated because of its long T I .
(1.30)
When Purcell et al. [1.2] first looked for a resonance of protons in paraffin.
(where A is a constant of integration), we see that n~ represents the thermal they allowed the nuclei to sit in the magnetic field Ho for a long time before even
equilibrium population difference, and TI is a characteristic time associated with attempting a resonance. They used a value of alternating field sufficiently low
the approach to thennal equilibrium. T I is called the "spin-lattice relaxation to allow them time to observe a resonance even though Tl were many seconds.
time". For example, if we deal with a sample that is initially unmagnetized, the Their effons, as with those of Bloch et al. [1.3]. were made independently of
magnetization process is described by an exponential rise to the equilibrium: Gorter's.
We have also seen that the rate of absorption is related to the transition rate
n=no(l_e- t / T1 ) (1.31)
W. An estimate of the size of the resonance absorption is basic to a decision about
That is, TI characterizes the time needed to magnetize an unmagnetized sample. whether or not a resonance might be observed. We shall wish to consider ho'w to
We may now combine the two rate equations for dn/dt to find the com- calculate W. Moreover, since no resonance line is perfectly sharp, we expect that
bined transition rate due to both thennal processes and transitions induced by the factors governing the width of [he spectral line will be of interest. Closely
the applied alternating field: related is the question of what magnetic field to use in the relation w = '"'tHo, for
the nuclei are never bare. There will be magnetic fields due to electrons as well
dn no - n as due to other nuclei. which must be added to the external field. These fields
-=-2Wn+--- (1.32)
dt T1 produce effects of greatest interest. such as the splitting of the proton resonance
of ethyl alcohol (CH3CH20H) into three lines of relative intensities 3:2: I. They
In the steady state. (1.32) tells us that
are also responsible for the fact that there is a nuclear resonance in ferromagnets
no even in the absence of an applied static magnetic field.
n = 71-:+-;2"W~T'-1 (1.33)
Therefore, as long as 2WT, <t:: I, n = no, and the absorption of energy from
the alternating field does not disturb the populations much from their thennal
equilibrium values. The rate of absorption of energy dE/dt is given by
8
9
2. Basic Theory
We begin our sludy of (he basic theory with a classical description of the motion
of a spin in an external magnetic field H, assuming Ihal H may possibly vary
with lime. H will produce a IOrque on the magnelic moment 11. of amount
Il x H. If we applied a magnetic field to an ordinary bar magnet, mounted
with bearings so thai it could turn at will, the magnet would attempt to line up
along the direction of H. If H were constant in time and if the bearings were
frictionless. the magnet would actually oscillale about the equilibrium direction.
If the bearings were nOI frictionless, the oscillalions would die out as the magnet
gave up energy to the bearings. until eventually it would be lined up along H.
When the magnet also possesses angular momentum, the situation is modi-
fied. since it now acts like a gyroscope. As we shall see, in the event of frictionless
bearings, the moment would remain al fixed angle with respect to H (providing
H is constant in time), but would precess about it. The conversion of energy
back and forth between potential energy and kinelic energy would nOI occur. It
would still be lrUe, however, thai if Ihe bearings possessed. friction, the magnet
would eventually become parallel to a static field H. As we shall see, the friction
corresponds to relaxation processes such as TI.
The equation of motion of the magnet is found by equating the torque with
the rale of change of angular momentum J.
dJ
dt = I' x H (2.1)
Since p. = "(J, we may eliminate J, getting
(2.2)
11
We can now readily solve for the morion of p. in a static field H = kHo
Fig. :1.1. Relation of /' to Jl
by choosing {} such that He = 0. That is, we lake {} = -"(Hok. Since in this
reference frame 6p./6t = 0, p. remains fixed with respect to i. i. and k. The
motion with respect to the laboralory is therefore that of a vector fixed in a set
of axes which themselves rotate at n = -"(Hok. In other words, JL rotates at
an angular velocity n = -,,(Hok with respect to the laboratory. The angular
frequency"(Ho is called the "Larmor frequency".
We are struck by the fact that the classical precession frequency {} is iden-
tical in magnitude with the angular frequency needed for magnetic resonance
most useful for our future work, however, to introduce a special technique: the
absorption, as found by elementary quantum theory. Let us therefore look more
use of a rotating coordinate system.
closely at the quantum mechanical description.
Consider a vector function of time F(t), which we may write in terms of its
componenls F:t(t). Fy(t). Fz(t), along a set of rectangular coordinates. In terms
of the corresponding unit vectors i, i, and k, we have
2.2 Quantum Mechanical Description of Spin in a Static Field
(2.3)
Ordinarily we think of i. i, and k as being constant in time. but we shall wish to We have seen that the quantum mechanical description of a spin in a static field
be more general. Since their lengths are fixed, they can at most rotate. We shall gave energies in terms of the quanlum number m, which was an eigenvalue of
assume they rotate with an instantaneous angular velocity {}. Then the component of spin 1z parallel to the static field Ho. The energies Em were
di
dt =
n .
x , (2.4)
Em = -"(hHom (2.9)
The corresponding eigenfunctions of the time-independent SchrOdinger equation
The time derivative of Fis therefore may then be denoted by u/,m' The time-dependent solution corresponding to a
dF . dF:t di .dFv dj dFz dk particular value of m is therefore
dt = '"""dt + F:t dt + J"""dt + F v dt + k dt + F z dt
WI ,m (t) = UI ,m e-(i/h)E",t (2.10)
.dF:I: .dFv dFz •
= .d-+ ,-+
tdtdt
k-+ n x (.F: +,'F, +kF) (2.5)
most~eneral
,F
=-+{}xF
l:VZ The
+I
time-dependent solution !li(t) is therefore
wet) =
"
where we have introduced the symbol 6FMt. representing the time rate of change
of Fwith respect to the coordinate system i, i, k. For example, when 6F16t = 0,
"L
m=-1
cm u I,m e-(i/h)E",1 (2.11)
where the em's are complex constants. We may compute the expectation value
the components of F along i. i, and k do nOI change in time. of any observable by means of !P(t). as we can illustrate with the x-component
By making use of (2.5), we can rewrite the equation of motion of JL in terms of magnetic moment:
of a coordinate system rotating with an as yet arbitrary angular velocity {}:
(2.12)
'I'
6't+{}xJL=p.x"(H or (2.6)
We have emphasized that the expectation value of !J:t. (!J:t) will vary in time by
'I' explicitly writing it as a function of time.
lit = I' x (,H + n) (2.7)
Equation (2.7) tells us that the motion of p. in the rotating coordinate system
obeys the same equation as in the laboratory system, provided we replace the 1 We write a variable of integralion dr in lhe expression for the expedation value, in
analogy to that which we would do for a spatial coordinate :1:, V, z or angular coordinates
actual magnetic field H by an effective field He: 6, .p. For spin, the nolation is to be thought of as a symbolic representation of the scalar
n product of the two funClions >it(L) and 11rl/i(L)
He=H+- , (2.8)
13
12
/
By using the fact that P~ :: 'rhI~, and that !V(t) is given by (2.11) we find In order to gain funher insight into tile physical significance of the general
expression for (,ux(t», (2.13), we now consider the fonn it takes for a spin of
(p~(t»:: L: 'rliC:n,cm(m'll",lm)e(i/h)(Em,-Em)1 (2.13)
!.By using the fact Ihat the diagonal matrix elements of I", vanish, we get
m,m'
where (,u",(t)) "" 'rll[ ci'/2c-1/2(! II", I - t)e -i"(l1ot
(m'II",lm) == J uimd",ulmdr (2.14) . 1/ 2Clj2 ('II
+ c_ -'2' ~ I''2' )ei"( o l
l1 t (2.19)
is a time-independent matrix element. Expressions similar to (2.13) would hold It is convenient to define a quantity wo '" 'rHo. As we have seen, wo is the angular
for any operator. We denole that the expectation value will in general be time frequency we must apply to produce resonance and is also the classical precession
dependent, will consist of a number of tenns oscillating hannonically and that frequency. By utilizing the fact that (!Il",l- t)
is the complex conjugate of
the possible frequencies ' (-! IIzl!), and using the symbol "Re" for "take the real part oC', we get
Em/-Em
h ~I~
(2.20)
are just those which correspond to the frequency of absorption or emission be. We evaluate the matrix element by means of (2.17) and (2.18), getting
tween states m and m'. Of course it was the assumption that observable properties
of any quan.tum sys~em had 10 be given by expressions such as (2.13), which
(W.I- ~)=~.
It is convenient at this point to express the c's in tenns of two real, positive
was .the baSIS of Heisenberg and Born's fonnulation of the quantum theory in
matnx fonn. quantities a and b. and two other real quanlities (which may be positive or
negative) a and p:
Since matrix elements (m'II~lm) vanish unless m' :: m± I, we see that
ia i
all ~e tenos of (2.13) ha:e .an angular frequency of either +-rHo or -'rHo. Ct/2 "" ae , C_lj2 = be /1 (2.21)
~lf sum. m~st also contam Just 'rHo. The expectation value (p",(t» therefore
OSCillates In tIme at the classical precession frequency. The nonnalization of the wave function gives us a 2 + b2 "'" 1. These give us
It is convenient at this point to introduce the famous raising and lowering
operators I+ and I-, defined by the equations (1'",(0) "'" 'rhab cos (a - fJ + wot) (2.22a)
Similarly we find
r+:: I", +iIy I-:: I", - iIy (2.16)
(Py{t) = -'rhab sin (a - p + wot) (2.22b)
We may express I", or I y in tenns of I+ and'r by solving (2.16), getting
(p,(t» = 1h(a' - 6')/2 .
I + _ I
1.=,([ +1) I V =2i(p-r) (2.17) We note that both (1'",) and (lJy) oscillate in time at the Lannor frequency
'rHo, but that (P:) is independent of time. Moreover the maximum amplitudes
The operalors are called "raising" or "lowering" because of the effect they of (1'",) and (Py) are the same. If we define
produce when they Operate on a function 'Ill ,m:
(,,) " i(p.) + j(p,) + k(p,) (2.23)
r+UI,m:: VI(I + I) - m(m + 1)UI,m+1
(2.18) and utilize the fact that (p",)2 + (py)2 "'" constant, a fact readily verified from
I-UI,m '" VI(I + I) - m(m -l)ltl,m_1 (2.23), we see that (I-') behaves as does a veClOr making a fixed angle with. the
z·direction, precessing in the x-y plane.
I+ turns Ul,m into a ~unction whose m value has been raised by one unit. We In teons of polar coordinates 6, t/J (see Fig.2.2), any vector A may be
see,. Iherefore, tha,t (m II+lm) vanishes unless m' = m + I, while (m'II-lm) written as
vams~es unless m = m - 1. Van Vleck [2.1) has characterized these as "sharper" A",=Asin 6 cos t/J
se.lec n~n rules than those of the operators Ix or ly, which may join a state Ul
WIt h eIther uI,m+l or UI,m_l. ,m Ay=Asin6sint/J (2.24)
A:=Acos6 .
14 15
: ."1g.2.2. Relationship of the components A"" A,o and A, of a It is useful to consider briefly what we should expect for the wave func-
vcctor A to the polar angles 6, 1/>, and the magnitudc A
tion if we took a sample of many noninteracting spins which were in thermal
equilibrium. There will be a wave function for each spin, but in general it will
not be in one of the eigenstates (m = +~ or m = -~); rather it will be in some
A, linear combination. For a given spin, there will be a particular set of values for
a, b, 0', {3. The values will differ from spin to spin. For example, we have a
o A
distribution of the quantity 0' - {3 that gives the spin orientation in the x-y plane
at t = O. If the spins are in thermal equilibrium, the expectation value of the total
magnetization must be parallel to the magnetic field. We expect, therefore, that
there will be no preference for anyone value of 0' - {3 over any other. That is,
the spins will have a random distribution of 0' - (3. On the other hand, since the
spins will be polarized to some extent, we expect to find a larger than b more
often than b is larger than a. That is, the average value of a must be larger than
the average value of b. Since an observable quantity can be expressed in the fonn
By means of algebraic manipulation one can show that of (2.13), we see that we can specify either the individual em's or the complex
"'(It . products c:n,c m , which we shall label Pmm/ for convenience.
(p~) = 2"" Sill () cos ¢
h
(J1.y) = "'(2 sin () sin ¢ For our example:
,Ii Pl/21/2=a2 P_I/2_1/2=b2
(J1.z) = 2"" cos () (2.25)
Pl/2-1/2 = abe i (Q'-,8) P- 1/ 21 /2 = abei(f.l-Q')
provided
We may consider the Pmm,'s to be the elements of a complex matrix P. Notice
¢ = {3 - 0: - wot that the diagonal elements (m = m') give the probabilities of occupation of
2 l+cos(} the various states, while the off-diagonal elements are closely related to the
a= ~~
2 components of magnetic moment perpendicular to the static field. We shall make
One may look on (2.26) as a formal change of variables, of course, but the results use in a subsequent section of the average of the matrix P over a statistical
of (2.25) tell us that there is a simple physical significance; the expectation value ensemble. The statement that in thennal equilibrium the magnetization will be
of the operator It acts as does a vector of length "'(nfl, whose direction is given parallel to the field amounts to saying that (he average over the ensemble of
by the spherical coordinates (), ¢. If the orientalion is specified at any time, it can Pmm , for m l '" m is zero, whereas the average for m = m' is the Boltzmann
be found at future times by recognizing that it precesses at angular velocity wo factor describing the probability of finding the state occupied.
in the negative rP direction. The orientation may be specified quite arbitrarily (by (Of course, in the quantum theory, even for a number of spins with identical
specifying a or band (3 - a). We emphasize that an arbitrary orientation can be wave functions, any experiment that COUIllS the number of spins in the various
specified, since sometimes the belief is erroneously held that spins may only be m states will find a statistical distribution not related, however, to temperature.)
found pointing either parallel or antiparallel to the quantizing field. One of the
beauties of the quantum theory is that it contains features of both discreteness
and continuity. In terms of the two quantum states with m = ± ~ we can describe
2,3 Equations of Motion of the Expectation Value
an expectation value of magnetization which may go all the way from parallel
to antiparallel, including all values in between. Thus a wave function with a = b
has an expectation value corresponding to a magnetization lying somewhere in TIle close correspondence of the classical and quantum mechanical treatments
the x-y plane (that is, with vanishing z-component). Just where in the plane it is made particularly clear by examination of a differential equation relating the
points is given by the complex phase a - {3, as well as the time at which we lime variations of the expectation values (Pz), (Jly), and (J1.z). The equation is
wish to know the orientation. based on a well-known fonnula whose derivation we sketch.
16
17
Suppose we have a pair of wave functions !li(t) and w(t), both of which are
solutions of the same SchrOdinger equation:
That is to say, dF/dt does !lot mean to take the derivative of F with respect to Therefore, since Jl- ::: -yliI, we have the equation for the expectadon value of
t. Such a derivative vanishes, since F does not contain the variable t. Rather magnetization,
dF/dt is a symbol that has the meaning of (2.30). By using dF/dt in this symbolic
sense, we have de,,> =
dt
(,,>
x 1H (2.37)
(2.31) which is just the classical equation. In words, (2.37) tells us that the expectation
value of the magnetic moment obeys the classical equation of motion. Equation
where [h, F] is the usual commutator HF - F'1l. We may use this fonnalism (2.37) was derived for the expectation value of a magnetic moment of a single
to compute the time derivative of the expectation values of /lz. /ly, and j1.z. We spin. If we have a group of spins with moments Jl-k' for the kth. spin, their total
define the X-, y-, z-axes as being fixed in space but with the z-axis coinciding at magnetic moment Jl- is defined as
an instant with the direction of the magnetic field. (In this way we include both
static and time-varying fields.) Then
,,= L>k k
(2.38)
If the spins do not interact with one another, it is easy to prove that (2.37)
(2.32)
also holds true for the expectation value of the total magnetization. Since, in
We shall wish to use the commutation relations for the components of an- practice, we measure the results of a number of spins simultaneously, the ex-
gular momentum, all of which may be obtained by cyclic permutation from perimental measurements of magnetization measure the expectation value of the
various components of magnetization. That is, the experimentally detennined
(2.33) bulk magnetization is simply the expectation value of the total magnetic mo-
Then ment. Therefore the classical equation correctly describes the dynamics of the
magnetization, provided the spins may be thought of as not interacting with one
another.
It is important to bear in mind that (2.37) holds true for a time-dependent
2 To p~ove (2.28), one must use the fad that F is an Ile~mitian operalor. (See dis- H, !lOt simply a static one. Therefore it enables us to use a classical picture for
cussion in Sed. 2.5). studying the effects produced by alternating magnetic fields. We turn to that in
the next section.
18
19
2.4 Effect of Alternating Magnetic Fields The time dependence of HI can be eliminated by using a coordinate system that
rotates about the z-direction at frequency w,.. In such a coordinate system. H1
The effect of an alternating magnetic field H~(t) = H%o cos wt is most readily will be static. Since the axis of rotation coincides with the direction of Ho. Ho
analyzed by breaking it into two rotating compon~nts, each. of amplitude HI. will also be static. Let us take the :t-axis in the rotating frame along HI. Then
one rotating clockwise and the other counterclockwise (see Fig. 2.3). (2.41) becomes
y
i: = ~x [k(w,. + "1Ho) + i"1Hd (2.42a)
(2.40) H.,
y y
which will give us either sense of rotalion, depending on the sign of w,..
We now ask for the equation of motion of a spin including the effects both
of HI (t) and of the static field Ho = kHo. x
(oj (b)
(2.41)
FIg. 2.4. (a) Eff«.tive field. (b) Motion or the moment Jl in the rotating coordinate system
20
21
We notice that the motion of the moment is periodic. If it is initially oriented lowing the pulse, the excess magnetization will be perpendicular to Ho and will
along the z·direction, it periodically returns to that direction. As it increases precess at angular frequency "'IHo. As a result, the moments will produce a flux
its angle with the %-direction, its magnetic potential energy in the laboratory through the coil which will alternate as the spins precess. The resultant induced
reference system changes (in the laboratory system the magnetic energy with emf JIlay be observed.
respect to Ho is much larger than that with respect to HI, so we customarily What we have suggested so far would indicate that the induced emf would
neglect the laller). However, all the energy it takes 10 tilt JL away from HO is persist indefinitely, but in practice. the interactions of the spins with their sur-
returned in a complete cycle of JL around the cone. There is no net absorption roundings cause a decay. The decay may last in liquids for many milliseconds,
of energy from the alternating field but rather alternalely receiving and returning but in solids it is more typically 100,IS. Even during that shan time, however,
of energy. there are many precession periods. The technique we have described of observ-
Note thal if No is above resonance (thai is, Ho >wh), the effective field ing the "free induction decay" (that is, decay "free" of HI) is a commonly used
has a positive %-component, but when Ho lies below that resonance (Ho <w!"(), technique for observing resonances. It has the great virtue of enabling one to
the effective field has a negalive %-cornponent. study the resonance signal in the absence of the voltages needed to produce HI.
If the resonance condition is fulfilled exactly (w = "'IHo), the effective field Since oscillators always generate noise. such a scheme may be advantageous.
is then simply ill I. A magnetic moment that is parallel to the static field initially One interesting application of the rotating rererence frame is to prove the
will then precess in the y-z plane. That is, it will precess but remaining always following theorem, which is the basis of another technique for producing reso-
perpendicular to H,. Periodically it will be lined up opposed to Ho. If we were nance signals. Suppose we have a magnetic field [-[0 of fixed magnitude whose
to tum on HI for a short time (Ihat is, apply a wave train of duration t w ), the direction we may vary (no other magnetic field is present). Let the magnetization
moment would precess through an angle 8 = ..,Hlt w . If t w were chosen such that M be parallel to Ho at t = O. We may describe the changing direction of Flo
8 = r, the pulse would simply inven the moment. Such a pulse is referred to in by an angular velocity w. Then the theorem states that if
the literature as a "180 degree pulse". If 8 = 1fn. (90 degree pulse), the magnetic
moment is turned from the %-direction to the y-direction. Following the turn-off ..,Ho>w
of HI, the moment would then remain at rest in the rotating frame, and hence the magnetization M will tum with Ho. always remaining aligned along Ho as
precess in the laboratory, pointing nonnal to the static field. Flo turns.
These remarks suggest a very simple method of observing magnetic reso- To prove this theorem, let us assume w to be a constant in the z-direction.
nance, illustrated in Fig. 2.5. We put a sample of material we wish 10 study in a We can take it perpendicular 10 110, since a component parallel to [-[0 produces
coil, the axis of which is oriented perpendicular to Flo. In thennal equilibrium no erfect. The relationships are shown in Fig. 2.6 at t = 0, with M and Ho
there will be an excess of moments pointing along Flo. Application of an alter- taken parallel to each other and pointing in the X -direction in the laboratory. If
nating voltage to the coil produces an alternating magnetic field perpendicular we choose a reference frame x, y, %rotating at angular velocity nR = w, Ho
to Ho. By properly adjusting Ht and t w ' we may apply a 90 degree pulse. Fol- appears static. but we must add an effective field nRh. Choosing the z- and
Z-axes as parallel, and x to coincide with X at t = 0, the effective fields and
magnetization at t = 0 are shown in Fig. 2.7.
The effective field in the rotating frame is static and given by
II. II, II,
nR
,, Ii , ,
W
Hclf=Ho+-=Flo+-, ,
f I - - - -
_.1
-+ --
I - - - - - z
" \ \ \ w
M
M
• II,
II,
l,) lb)
)'
24 25
similarly, where F is now an operator. We shall be particularly interested in the
!li(t) = exp [ - (i/h)( -")"IiHoI:)t]w(D)
function
= exp (iwotI:)!li(O) (2.50)
iF . (iF)2 (iF)3
e = I +IF+~+3!+ ... (2.45) where Wo :: ")"Ho.
We know that H 0 produces a rotation of the magnetic moment at angular
By using the series expansion, one can show that if F is Hennitian, exp (iF) is velocity a given by n = -")"Hok. We shall call such a rotation "negative",
nol. In fact since the component of angular velocity along the z-axis is negative. It is log·
ical to suppose, then, that !li(t) must correspond to the function w(D), referred,
J(e iFtJi)*lJidr = J p*e- iF !lidr (2.46) however, to axes rotated in the negative direction through an angle wot. Thus
exp(-iI:4')w(O) should correspond to a function identical to w(O) referred to
The exponential function of operators obeys some of the same algebra as does axes rotated through the positive angle 4'. If we compute the expectation value
the function of ordinary number, but as usual with operators, care must be taken or matrix elements of, for example, Ix., we find
whenever two noncommuting operators are encountered. Thus, if A and E are
two operators, one can verify by means of the series expansion that J w*(t)Ix.w(t)dr = J [eiwotl·i[i(O)j*I:r.ei""Otl·!li(O)dr
We shall now use the exponential operators to perfonn the quantum mechanical
equivalent of the classical "rotating coordinate" transfonnation. We shall consider
(2.54) a magnetic field HI, which rotates at angular velocity Wz, in addition to the static
field kHo. The total field H(t) is then
Therefore
H(t):= iHI cos wzt + JB I sin wzt + kHo (2.57)
f(¢):= A cos ¢+B sin ¢
where we must evaluate the constants of integrations. (As we shall see, the and the SchrOdinger equation
"constants" are actually operators.) Clearly, A '" f(O), but from (2.51), f(O):= I~.
Likewise, B:= t(O):= I y , using (2.53). In this way we get
8w
-:=-iw . • t1,alJi'
Ie- .Iw • '1 EIJiI+e- 1W
at zz
_
at (2.61)
" We may substitute (2.60) and (2.61) into (2.58), multiply both sides from the left
by cxp (iwztI z ), and obtain
x'
h alJi'
• x
Fig.2.9. Relation of axes "', Y to ",', y' and thO)
angle tP
-Tat:= -[hew: +,Ho)Iz +,hHIIz]1lI (2.62)
In (2.62) the time dependence of HI (t) has been eliminated. In fact we recognize
it as representing the coupling of the spins with an effective static field
k(Ho+~)+iHI
28
29
the effective field of our classical equations. The spins are therefore quantized Substituting in (2.68) we get
along the effective field in the rotating coordinate system, the energy spacing
being "'{fiHeff' {Jlz(t» = -{,JII(O» sin wit + {Jtz(O» cos wit (2.70)
The wave function .pI given by (2.60) is related to the function tJi by a If the- magnetization lies along the z·axis at t = 0 so that {J.ly(O» = 0, we get
coordinate rotation, the "forward" motion of Iz; relative to a stationary l/J having
been replaced by a stationary I z and "backward" rotating tJi'. As usual, resonance (2.71)
occurs when Wz ::;:::: - ",(Ho. If we define the transfonned Hamiltonian 11.' by Thus the z-magnetizalioo oscillates in time at ...,H 10 corresponding to the preces-
sion of {po) about HI in the TOtating reference frame. It is important to note lhat
1t = -[(Ttw z + "'{ IlH o)Iz + "'{II HI I:a:] (2.63) in lhis picture. which neglects all interactions of spins with one another or the
we can formally solve (2.62): lattice. the magnetization continues oscillating between +{1l:(0» and -{J.l:(0»
indefinitely. This behavior is very different from that which we should expect
tJi'(t) =e-(i!i)'H'IW'(O) (2.64.) from a time-independent transition probability such as we assumed in Chapter 1.
The time-independeOl transitions occur only if some physical process spoils the
whence, using (2.60). coherent precession about H t in the rotaling reference frame.
Another approach to solving (2.62) enables us to demonstrate a very in-
!li(t)::: e- iw,I1'e-(i!i)1t'I!Ji'(O) (2.64b) teresting and fundamental property of spin ! particles, their so-called spinor
=
{Note that at t O. !li(0) lli'(O)j. = nature. Formally, this property describes what happens to a wave function under
Equation (2.64b) gives us a particularly compact way to express the solution a rotation. One can discuss the problem by fonnal mathematical melhods. as in
of SchrOdinger's equation when a TOtating field is present. group theory (see below). Here. however. we display the mathematical result by
We can illustrate the use of the wave function of (2.64b) by computing the physically generating a rotation with an alternating magnetic field tuned exactly
time dependence of the expectation value of Jlz. Of course we know already to resonance so that (Ho + w:h) = 0 in (2.62).
what the resuh must be. since we have proved that the classical picture applies. Let Utj2 and U_Ij2 be the eigenslales of the spin operator I z • They are
Let us for simplicilY assume that HI is applied exactly at resonance. Then, from independent of time. Since they form a complete set for a spin ~ particle, we
(2.63), can express any function, such as I/J', as a linear combination of Ulj2 and U_lj2
with coefficients a and b. If I/J' is itself a function of time, ,p'(O, the coefficients
1t = -...,IlH lIz (2.65)
a and b must also be functions of time:
Then we have. using (2.64b) and (2.65),
,p'(0 ::: a(t)u Ij2 + b(Ou_Ij2 (2.72)
(p,(l» = J~'(l)I,,~(l)dT (2.66)
= "'(It J[e-i"", tf• e i "l'll1 I.. ttP'(O>t I z [e -i"",LI. ei"l'Il\/"ttJi(O)JdT. (2.73)
If we define WI.
Muhiplying by U;j2 from Ihe left, integrating over spin space, and utilizing the
WI ="'(H] (2.67)
f.ct that (11[.1~) = (- !II.I-1) =0, we get
and lise the fact thul I z and I z are Hermitian, we get
il da
="'(Ii Jv*(O)e -i""l t1 rei"", LI. Ize -i"".t1 ze i""l tI" !P(O)dT
1
{,tz(t» T dt = 1IiHtb(~I[.I-~) (2.74)
= "'(Ii J.p' (O)e -i""l tlz Izei""l If.0 tJi(O)dT (2.68) Similarly we get
Ii db
By using (2.55), we can write
7
I
-
dt
= 1"Ht a ( - !II.lt)
(2.69) Utilizing the fact that {!II:,.,j -!)::: i and {- illzl!) =! we can solve these
two simultaneous differential equations to find
30
31
aCt) "" a(O) cos (WI tn.) + ib(O) sin (WI t(2.) (al
(2.75) x x'
b(t) :::: ia(O) sin (WI t(2.) + b(O) cos (WI tn) front
"
y
",
y'
where
{bl - y'
(2.76) Bark p;f ·1 ~~
is the classical precession frequency about HI in the rotating frame.
The spinor properly is revealed by considering a 211'" pulse. We recall that
« I
such a pulse causes (he expectalion value of the magnetization vector to undergo
a rotation about HI which returns it to its initial value. Denoting the pulse ".
'wisled
. ".
y'
strip
duration as t2 .... we have. then X·
defining the quantities X' and Xii. By using (2.90) and (2.93), we get
I ,.
a=-+-y10I (2.88) Xo ...... (wo - W)T2
T, - I
X = -WO.L2 (2.93,)
Therefore 2 1 +(w - wo)2Ti
M+ = Ae-o·t + i-yMoHI
1/T2 + i,ho
(2.89) II
X = -w,p,
Xo......
2
I
1+(w-wo)2Ti
If we neglect the transient term and subsulUte Mo = xoHo. and define Wo :: 7Ho.
Wz = -w, we get It is convenient to regard both Mx(t) and Hx(t) as being the real parts of
complex functions M~(t) and H~(t). Then, defining the complex susceptibility
(2.90)
X by
I . /I
X=X -IX (2.94)
I
Mil "" XO(WO T 2) 1 + (W -Wo )'T'
2
H) and writing
Hi(i) = H XOeiWl (2.95)
Equations (2.90) show thai the magnetization is a constant in the TOtating
reference frame. and therefore is TOtating at frequency w in the laboratory. In we find
a typical experimental arrangement we observe the magnetization by studying
the emf it induces in a fixed coil in the laboratory. If the coil is oriented with M~(t) = xHi(t) or (2.%)
its axis along the X -direction in the laboratory, we can calculate the emf from
Icnowledge of the time-dependent component of magnetization MX along the MX(t) == Re HxH xoeil.Jt)} (2.%,)
X -direction.
Although (2.92) and (2.96a) were arrived at by considering the Bloch equations,
they are in fact quite general. Any resonance is characterized by a complex sus-
y
ceptibility expressing the linear relationship between magnetization and applied
field.
Ordinarily, if acoil of inductance Lo is filled with a material of susceptibility
xo, the inductance is increased to Lo{l + 47fxo), since the flux is increased by
the factor I + 47fXo for the same current. In a similar manner the complex
x susceptibility produces a flux change. The flux is changed not only in magnitude
but also in phase. By means of (2.93-96), it is easy to show that the inductance
"' Fig.2.1l. Rotating
X, Y
axes~, II relative to lll.borillory llXetl
at frequency w is modified to a new value L, given by
(2.97)
By referring to Fig. 2.1 I, we can relate the laboratory component Mx to
the components M~ and My in the rOlating frame. Thus where X(w) = i(w)-ixl/(w). It is customary in electric circuits to use the symbol
j for yCT. However, in order to avoid the confusion of using two symbols for
Mx = M z cos wt + My sin wt (2.91)
the same quantity, we use only i.3
If we write the magnetic field as being a linear field,
3 In practice, the sample never complC!lely fills all apace, and we must introduce the
Hx(t) = Hxo cos wi 2HI = H xo (2.92)
~fil1illg factor~ q. Ita calculation del>ends 011 a knowledge of the spatial variation of the
then we see that both M z and My are proponional to H xo, and we can write alternating field. Then (2.97) bccomCl!l
Ro
Low.
-L1R = - - 1TX "
Ro
:::: • 1rX"Q (2.99) HI = lOPQ vV
(2.106)
where we have introduced the so-called quality factor Q, typically in a range of with [Hd: Gauss; [P]: W; [II]: MHz; [V]: cm J .
50 to 100 for radio frequency coils or 1,000 to 10,000 for microwave cavities. The particular functions X' and XII, which are solutions of the Bloch equa-
Assuming unifonn magnetic fields occupying a volume V, the peak stored tions, are frequently encountered. They are shown in the graph of Fig. 2.12. The
magnetic energy produced by an alternating current, whose peak value is io, is tenn Lorentzjan line is often applied to them.
(2.100)
I H'xoX "V
P :::: ~w (2.102) Fig. 2.12. X' and X" rrom the Bloch equations ploued versus z "" (wo - ",,)T2
This equation provides a simple connection ,between the power absorbed, X", and
the strength of the alternating field. We shall use it as the basis of a calculation At this time we should point out that we have computed the magnetization
of X" from atomic considerations, since the power absorbed can be computed in produced in the X -direction by an alternating field applied in the X -direction.
tenns of such quantities as transition probabilities. Since X' and XII are always Since the magnetization vector rotates about the Z-direction, we see that there
related, as we shall see shortly, a calculation of XII will enable us to compute will also be magnetization in the Y -direction. To describe such a situtarion, we
Xl. Moreover, we recognize that the validity of (2.102) does not depend on the may consider X to be a lensor, such that
assumption of the Bloch equations.
C(t)-
M Ct' H (toe i""t 0: = XYZ
, , o:l=X,Y,Z
Another useful fonnula can be obtained from (2.100) and (2.101), relating - XCl"Cl'
the average power dissipated in the coil resistance, Pc, to the strength of Hxo. In general we shall be interested in Xxx.
Since Pc is half the peak power dissipmcd in the coil,
(2.103)
2.9 Spin Echoes
Solving for (~)iij and substituting into (2.100) we get
Just after finishing graduate studies, Erwin Hahn burst on the world of science
Hl ov : : PcQ (2.104) with his remarkable discovery, spin echoes [2.10]. His discovery provided the
8, w
38 39
key impetus to the development of pulse methods in NMR. and must therefore , ,
be ranked among the most significant contributions to magnetic resonance. What
are spin echoes and why are they so remarkable?
M,
Suppose one applies a 1fn. pulse to a group of spins to observe the free in-
duction signal which follows turn-off of the pulse. According to the Bloch equa-
}---y
tions. the free induction signal decays exponentially with a time constant T2' For
solids. T2 is a fraction of a millisecond, corresponding to line widths of several
Gauss. For liquids, the line widths are typically much narrower. corresponding x x
perhaps to times of several seconds. Such lines are a good deal narrower than (b) I"" 0'
the usual magnet homogeneity. As a result, the inhomogeneity-induced spread
in precession frequency causes the spins in one portion of the sample to get out , ,
of phase with those in other JXlnions. The free induction signal arises from the
sum total of all portions of the sample. As the separate portions get out of step,
the resultant signal decays. The decay lime is of the order of l/(-y.dH) where
tj,H is the spread in static field over the sample.
M!
/
,
----
0
--,
.""7'----i-- y
--,,
fJalm made the remarkable discovery that if he applied a second 7rn. pulse
a time T afler the first pulse, miraculously there appeared another free induction
---
" -- --'
/
signal at a time 21' after the initial pulse. He named the signal the "spin echo". x x
To produce a signal at the time of the echo, the spins must somehow have (d) I'" T+
gotten back in phase. The great mystery of the spin echo was what made the
spins get back in phase again? Was the echo a challenge to basic concepts of ,
irreversibility? Was there a Maxwell demon at work producing the refocusing?
fJalm discovered spin echoes experimentally, but was soon able to derive their FIg. 2.13a-<. The formation of a spin echo by means of a 7t /2-
existence from the Bloch equations. This solution showed that as one varied T. 7t pulse sequence viewed in the rotating reCeren<:e frame. (I)
At t = 0- the magnetizalion, Mo is in thermal equilibrium
the echo amplitude diminished exponentially with a time constant T2. Thus the Y lying along lhe :-direction. (b) shows lhe magnelizalion
echo provided a way of measuring line widths much narrower than the magnet immediately after lhe 71:/2 pulse. In (c) an element of
inhomogeneity. Understanding the physical basis of echo formation has led to magnetization, 6M, has precessed an utra angle' owing to
x the magnelic. field inhomogeneilY. (d) shows the effect of the
much deeper insight into resonance phenomena in general and pulse work in 7t pulse on 6M. In (e) we !lee that at time 2T all elements of
particular. (t') r= 2T magnetization hilove refocused along the +y-direclion
The essential physical ideas of refocusing can be most easily seen by con-
sidering a pulse sequence in which the first pulse produces a rotation of 7I:n.. the
second a rotation of 1f. Such a sequence we denote as a 7rn.-7r pulse sequence. By proper adjustment of the pulse length t p , we can generate a 'lrn. pulse. In our
It was invented by Carr [2.11] based on a vector model proposed by Purcell for discussion we consider H} to be sufficiently strong that t p is negligibly short.
the 1rn.-7f/2 echo [2.10]. and was first described in a famous paper by Carr and We designate the time just before or just afler the initial pulse by 0- and 0+,
Purcell [2.12). Consider a group of spins initially in thennal equilibrium in a respectively.
static magnetic field H in the z-direction. The thermal equilibrium magnetization It is convenient to view the behavior of the spins in the reference frame
Mo then lies along H as shown in Fig. 2.1301. We assume there is a spread in that rotates at,Ho. with the x-axis defined as lying along HI. Let the 7r/2 pulse
magnetic fields over the sample, and take the average value of field to be Ho. rotate Mo to lie along the negative y-axis (Fig.2.13b).
We first analyze what happens if we can neglect the effect of Tl and T2. This sense of rotation corresponds to a spin which has negative 'Y, such as
We apply a rotating magnetic field HI at t = 0 with frequency w. tuned to an electron. For a positive 'Y, the rotations are positive in the left-handed sense.
resonance at the average field Ho. Thus We give the laller case as a homework: problem for the classical derivation of
the echo, and treat it quantum mechanically in the next section. If there were
w=,Ho (2.107) not inhomogeneity in the static field, all spins within the sample would precess
at 'YHo, so that the magnetization of every portion of the sample would remain
40 41
oriented along the -V-axis. The existence of inhomogeneity leads to a spread in It/, pulsl: 11 l'ulsl:
precession rates and dephasing. Consider what happens during a time interval,.
In any small region of the sample, the magnetization oM
will remain in the x-V
plane since we are neglecting T, processes. But at the end of" the direction of Helm
"'rl:l: ;",1"CI;(/II
oM within that plane will advance from the -v-direction by some angle which
we call 6, given by
sigl/a/
'"
6='1oH, where (2.108) ,
o 2,
(2.109) FIg.2..14. The deeay of the echo from I = 2T onward ill the same function of time as the
6H=H-HO decay of the free indudioll signal from j =
0+ onwards. Note that the buildup before
t = 2T is the mirror image in time of the decay aner 2T. Nole furl her that no free
represents the inhomogeneity in H. The situation is shown in Fig.2.13c. (Note induction signal is produced immediately following the ... pulse
that oH may be either positive or negative, so that the "advance" may be either
positive or negative.)
Let us assume we can control the phase of the oscillating voltage in the develop exponentially with T,. The 'II" pulse will invert the z-component which
second pulse so that the H t again lies along the +x-direction in the rotating has developed, so that it does not contribute to the component of oM
in the :.:-y
frame. 4 Suppose now we apply a 'II" pulse at t =', again of negligible duration. plane existing at t = ,+. During the next interval" the comp:ment oMin the
We denote the time just before and just after the pulse by t = , - or t = ,+. :.:-y plane will continue to decay via the T2 tenns in the Bloch equations. As a
respectively. The situation just after the'll" pulse is shown in Fig.2.13d. Noting result, the size of the magnetization producing the echo signal M(2T) wiU obey
the orientation of oM, we immediately see that during a second time interval "
M(2T) '" Moe- 2T / T1 (2.110)
oM will again advance through the same angle 6, which will bring it exactly
along the positive y-axis at t = 2T. The argument applies to all spins. no matter If the second pulse is a 1fn pulse, as in Hahn's original experiment, the pulse
what oH they experience, because the result does not depend on the angle of at t = , will put some of the z-component of oM
into lhe :.:-y plane. One might
advance. then wonder whether or not the echo contained an exponential depending on T•.
Though all the spins are in phase at t = 2" they get out of phase again A delailed analysis shows that lhat does not happen. This result is most easily
owing to the field inhomogeneity, so the free induction signal decays. Note that seen by pretending the magnetization developed along the z-axis arose from an
its fonn as a function of time during the dephasing from t = 2, onward must be independent set of spins which possess no net x-y magnetization. With respect
identical to the fonn of the decay following the initial 1fn pulse (see Fig. 2.14). to the new spins the situation just prior to lhe second pulse is like Fig.2.13a,
The buildup of the echo signal just prior to t = 2r is the mirror image in rime with oM lying along the +z-direction. The 1f/2 pulse puts the magnetization into
of the decay after t = 2,. the :.:-v plane. The situation at t = 2T is identical to that of Fig.2.13c, in which
lt is easy to see now what will be the effect of the T , and T2 tenns of all the spins have dephased.
the Bloch equations. During the first time interval" the components of oM Thus the Tl induced magnetizalion does not affect the echo. To observe the
in the x-v plane will decay exponentially with T2' and a z-componem will signal, one would need to refocus - which requires still another pulse. Thus with
three pulses one could see T I effects.
4 With so-called coherent pulse apparatus, a highly stable oscilll\lor generl\tes l\ HO/111 in fact found such echoes. He observed that if he applied a third pulse
steady-state alternating voltage which is fed into a circuil which amplifies the rf voltllge at time T (hence T-, after the second pulse), he produced an echo at 2(T-,)
only during the lime thalli gl\le voltage is applied, lhereby producing a strong rf signal after the second pulse, or at t = 2T-, after the first pulse. He found, in addition,
coincident with the gale pulses. The output voltage is fed to the sample coil to generate echoes at T + T, 2T - 2T, and 2T. The various echoes induced by a third pulse
lhe lf l . rr lhe slable oscillator is tuned exactly to ~nance,..., = ""0, the phasc of the
Ifl is always tIle same in the ""0 reference frame. This method of operalioll, together are commonly referred to as "stimulaled" echoes.
wilh so-called ph~e coherent deledion of the free indudion decays, WI\S first introduced In liquids, the diffusional motion pcnnits a nucleus to move between dif-
independently by Solomon [2.131 and by Spokm and Stichter [2.'",15). In 12.15]' there ill ferent parts of the sample where the precession rates may differ. As a result,
a dCS<:.ription of the rel\SOns for using phase-coherent detection. In a famous paper on
nuclear relaxation in alkali metals, lheir colleagues, I/o/comb and Norberg 12.161 introduce during a spin echo the dephasing during the first interval, may differ from the
lhe use of signaillveraging to improve the signal to noise ratio in pulse experiments, lind rephasing during the second interval T, and the echo is diminished. The effect
point out lhal coherent detection is necessary to achie\'c the full potential of II noise is of great practical utility as a way of measuring diffusion rates in liquids. This
integration.
42 43
Fig. 2.16. The effec~ of l\ finite Til on lhe precession of l\
was discovered by Hahn, and reported in his first spin echo publication [2.10].
component of magnetization, 6M, which is off resonance by
Although Hahn gives correctly the expression for the effect of diffusion on the a ",agnetic field 6//. //1 lies along the z-axis. The ",'-axis,
decay of the transverse magnetization following a single pulse. there is an error , which liCli along the effective field H~ff, makes an angle .d6 =
in his expressions for the effect of later pulses, corrected by Carr and Purcell
,. 6Hitt.O tan-I (6// / I1 t ) with the z-axis. Thus, iJ.M precesses about H~ff
in a COlle of allgl~ ./2 - D6
[2.12] in a paper which also showed how the effect of diffusion could be elim-
inated (Sec!. 8.2). Carr and Purcell showed thai diffusion led to a decay of the
echo peak magnetization AI, given by
" 'IIn-A9
as is the case for a magnetic field of axial symmetry about the z-axis. We derive
this resuh in Appendix G. As a preliminary, in Appendix F we treat a simpler case
in which bodily motion affects the struClure of a nuclear absorption spectrum.
In an actual experiment, since HI is not infinite, the HI pulses have a along Hi> and the x'-axis along Herr. The magnetization contribution oM will
nonzero duration. Let t./ 2 be Ihe duration of a 7(0. pulse if one were exactly at precess about Herr. in a cone which makes an angle 1fo. - .69 with respect to
resonance (,H,t./ 2 '" 7(0.). We call this the "7(0. pulse" in what follows even %' where tan.69 '" SH/HI. With this as a starting point, one can then work
though it is not exactly so for spins which are off resonance. We then ask what through the details of the precession. If HI is much bigger than the line width,
influence the nonzero duration of the pulse has on the time al which the echo then .69 <: I, and one can reexamine the derivation given for the echo. We put
occurs. Indeed, one might ask whether or not the echo is fonned perfectly under the details of Ihe derivation as a homework problem. One finds, keeping only
these circumstances. To discuss the problem, we define the times T" and ~ shown corrections linear in .d9, that the "7fo. pulse" is a 1fo. pulse for rotations about x'.
in Fig. 2.15. We define T as the time between the end of the 7fo. pulse and the It puts the magnetization at an angle .69 = oH/H, with respect to the negative
start of the 7f pulse, T' as the time from the end of the 'If pulse to the peak of y-axis. Thus, at the Slart of the 1[" pulse the spin now makes an angle .d9+"'(oHT
the echo (assuming one is fonned!). with respect to the y-axis instead of SHT of (2.108). The 'Ir pulse can be shown
to introduce no elTor in the orientation of SM in the x·y plane. Thus at a time
T after the end of the 7f pulse when
, ,. -----< '
Echo
1\ Frn
(2.113)
,
induction
decOly
l-, _
f\ the magnetization lies along the +y-direction. This gives
o I Echo
'" I peak T''''T+--
I
(2.114)
Fig.2.tS. Defillition of ~he times .. lind ..' us~d to discuss echo for"'l\~ion when one con- ,HI
siders the effect of lhe finit.c si1.e of //1, h~nce lhe nonzero length of lhe 1r/2 lind 11" which is independent of 6fI. Thus, the oM's from all parts of the resonance line
pulses
are collinear, and an echo is fonned at this lime. Equation (2.114) can also be
written as
To analyze this problem, let us suppose Ihat Ihe resonance is broadcned
symmetrically by a field inhomogcncity and that HI is tuned exactly to the T''''T+~t1["/2 (2.115)
center of the resonance, W '" ,Bo. Consider an elcment of magnetization, oM,
which is off resonance by H - Ho '" oH. Then the effective field in the rotating This expression was derived on the assumption that SH« HI. Numeri-
frame aCling on this magnetization is shown in Fig.2.16. We take the x-axis cal simulations for SH>HI' however, show that the correction tenn, IhH\,
remains accumte to within 10% even for lines much broader than H, [2.17).
44 45
2.10 Quantum Mechanical Treatment of the Spin Echo dN = NP(lto)dllo (2.119)
The spin echo can also be derived quantum mechanically. Of course, since (2.37) (N is the total number of spins). See Problem 2.3 for a classical study of the
effecJs of this distribution function.
shows that the expectation value obeys the classical equations of motion, one
immediately knows that the classical derivation is true for the quantum me- We can simplify things by assuming that, while HI is on, HI ;:;. Ito for all
chanical expectalion value. However, there is added insight from following the spins, so that we can approximate 'H during the pulses by
development of the actual wave function in time.
We consider,then, a sequence ofevems in time shown in Fig. 2. 17, which (2.120.)
shows H I versus time. At t = O. HI of frequency w is turned on along the x-axis
in the reference frame rotating at -kw. It remains on until time tit producing a In between pulses, when HI '" 0, we have
1(12 pulse. From tl to t2 the spins precess freely. A second pulse is applied from
t2 to t3, where the interval is chosen to make this a 11" pulse. We denOle t2 - tl
(2.12Ob)
as T. Since we will be working solely in the rotating frame, we will omit from
Although 1{ is time dependent, its time variation occurs only at four times
t/J and 1t the prime used earlier to designate quantities in the rotating frame.
(t = O. tl. t2, t3). In between, it is independent of time, enabling us to use
(2.64a) over the four lime intervals 0 10 tlo tl to t2, t2 to t3, t3 to t. Integralion
H, of (2.116) across the discontinuities in 1{ shows thai ¢(t) is continuous. We use
this fact to join solutions across the discominuities.
Thus, using (2.120a) and (2.64a)
h-
--
i
a"ot = 1e·¥' where (2.116) T(t. 110) generates the development of Ihe wave function during those times when
HI = 0 for spins which are off resonance by " 0 .
(2.117) As can be seen by referring to (2.55), XeD) is the operator we need when
we wish to rotate a componenl of spin through an angle D in the right-handed
Ito = Ho -w/-y (2.118) sense about the X -axis. Thus, if 10 becomes 10 , following such a rotation.
If , is positive, this will produce rotations about the effective field which are (2.123)
left handed. If, were itself negative, the rotations would be in the right-handed
sense, as assumed in drawing Fig. 2.13. For 0 = 1rn.
We do not set Ito '" 0 since we wish to represent the fact that owing to the
inhomogeneity in }fo the typical spin is not perfectly at resonance. We introduce X-I(1rI2)lyX(7r!2) = I z
a distribution function P(1I 0 ) to express the fact that the number of spins with X- I (7r/2)lzX(7rf2.) = -ly (2.124)
Ho between 11 0 and Ito + dho is X~I(7r!2)I:r.X(7rf2.)= l oe
46 47
lf8=1r X- I (1r)T- 1(t - T, ho)IyT(t - T, ho)X(1r)
= X- I (1r)T- 1(t - T)X(7l') X- I (1r)lyX(1r)
X- 1(1r)Iy X(1r) = -Iy , ... '--.-.-
I 2
X- 1 (1r)I.. X(1r) = -Iz (2.125)
X- 1 (1I')I z X(1r) = I z . x ,X- I (1r)T(t - T, h o)X(7l') . (2.131)
•
3
Let us further consider that HI is so large that we can neglect the time
intervals tl and t3 - t2 compared to T. Then, using expressions We now have three tenns identified by the three numbered braces in (2.131).
We deal first with term 2, using (2.125) to get
,,(II) = X(./2)"(0)
,,(12) = T(r, ho)X(.i2)"(O) (2.126)
We write teml 3 as
etc. we get ,pet) for times t which are after the second pulse as
(2.132)
,,(I) = T(I - r, ho)X(.)T(r, ho)X(.i2)"(O) (2.127)
This term has I z in the exponent. To deal with it, we call on a theorem (given
for a single value of ho. Since the NMR signal arises from the transverse mag- as Problem 2.4) that if R is an operator, R- l its inverse, and G some other
netization, we need to calculate the expectation value of I z and I y . We leave it operator
as a homework problem to show that if H I lies along the x-axis. the expectation (2.133)
value of I z is zero as in the classical picture. Instead. we here calculate (Iy(t».
For a single spin we get which is readily proved using the series expllllsion of the exponential and making
repeated insertions of R- I R.
(2.128) Using (2.133), we then get
where dTl stands for the "volume element" in spin space. Now, we must sum X-I(1r)exp(iho(t - T)I.. )X(1r)
over all spins, using (2.119) to express the fact Ihat the various spins experience =exp(iX- J (1r)I.. X(1r)ho(t - T»
different ho's. Thus the expectation value of the total y-component of spin is = exp(-i...,.ho(t - T)I.. ) (2.134a)
The integrand is just what we would have if we calculated (1",lolal) immediately Suppose, to avoid saturation, we deal with sufficiently small time-dependent
after the first 1r(1. pulse. Using (2.124), in fact, we have magnetic fields. The magnetic system may then be considered linear. That is, the
(2.139) magnetization produced by the sum of two weak fields when applied together is
equal to the sum of the magnetization produced by each one alone. (ytIe shall
where 0- refers to the time just before t = O. If we assume the system is in not include the slatic field Ho as one of the fields, but may find it convenient
thennal equilibrium before the r(1. pulse, this resull states that the echo arises to consider small changes in the static field.) In a similar manner, an ordinary
from the thennal equilibrium Mo. Note that the minus sign shows that the echo electric circuit is linear, since the current produced by two voltage sources si·
forms on the negative y-axis. This same result would follow from the classical multaneously present is the sum of the currents each source would produce if
argumenl for a nucleus with a positive.., for which the rouuions about HI are the other voltage were zero.
in the lefl·handed sense. Let us think of the magnetization Ll."f(t) produced at a time t and due to
Going back to (2.134) and (2.135), we see that the effect of the 1r pulse may a magnetic field H«() of duration Llt' at an earlier time (see Fig.2.18). As a
be thought of as having changed the Hamihonian after t = T from its old value result of the linearity condition we know that LlM(t)oc H(t'). It is also oc Llt'
of "Hold to a new value "H new . In this viewpoinl, Ihe Hamiltonian from tl to t2 as long as Llt' <: t - t' , since two pulses slightly separate in time must produce
is "Hold = -..,llholn and then at t = r changes to its negative, "Hn(~w = +-,1I11 01z the same effect as if Ihey were applied simultaneously.
for later times. We replace the 1r pulse plus subsequent evolution under the real
Hamiltonian with a developmenl under a new or effective Hamiltonian for t > T
H(t')
in which there is no 11' pulse. The effect of the pulses is equivalent to a situation
in which there aTe no pulses. bllt the Hamiltonian changes in time. This concept
proves very useful in magnetic resonance. II is the basis of many important pulse
sequences.
In the case of the echo, the combination of 'H new and "Hold is such that when Fig.1.18. Pulse or magnetic field
t = 2T, 'H new has completely undone the effect of 'Hold. Over the time interval
2T we have eliminated the effect of magnetic field inhomogeneities from the
•
Hamiltonian. The echo can be used to eliminate, over a particular time interval, Therefore we may express the proponionality by writing
the effect of any interaction which is equivalent to a spread in magnelic field.
For e)(ample, spin-spin couplings 10 different nuclear species (e.g. coupling of LlM(t) = met - t')H(t')Llt! (2.140)
protons to C 13 , when Ct3 is under observation) or the effects of chemical shifts
where met -t' ) is a "constant" for a given t and t ' , which, however, must depend
discussed in Chap. 4 all can be eliminated by use of echoes.
on how long (t - t' ) after the pulse of field we wish to know the magnetization.
By using Ihe explicit expression of (2. I34a), it is easy to show that apart
The total magnetization at time t is obtained by integrating (2.140) over the
from a sign change the development of (ly,lotat(t» for t > 2T is identical to
history of the magnetic field H(t'):
its development immediately after the first 11'!2 pulse. We give this as a home-
work problem. This result has a useful e)(perimental consequence. The signal ,
immediately after the 7r!2 pulse gives one the Fourier transform of the line shape
function P(ho), as can be seen from the results of Problem 2.3. But the signal
M(t) = J met - t')H(t')dt' (2.141)
-=
immediately follows a strong pulse. In practice this pulse may block the signal
amplifiers of the NMR apparatus, making it impossible to observe the signal until Note that met - t') = 0 if t' > t, since the effect cannot precede Ihe cause.
the amplifiers recover. On the other hand, the echo occurs later than a pulse by To understand just what met - t') is, lei us assume H(t ' ) is a 6-function at
a time T, giving the amplifiers time to recover. Thus, the echo is easier to see. t = O. Then the magnetization at t > 0 (which we shall denote by M6) is
It is therefore fortunate Ihat the echo reproduces the earlier signal.
50 51
,
M6(t)::: J met - t')E(t')dt' =- met) . (2.142) Those readers familiar with Green's functions will recognize that met) is a
-~ Green's function, and that what we have just done is one example of how one
That is, met) is the response to as-function al t ::: O. Knowledge of met) enables finds a Green's function .
us to detennine from (2.141) the magnetization resulting from a magnetic field =
.If a unit step were applied at t 0 (Fig. 2.19), we should have magnetiuttion,
of arbitrary time variation. which we shall denote as M slep :
Ordinarily one detennines met) by going 10 the basic lime-dependenl dif- , ,
ferential equation which describes the behavior of the system. We can illustrate
the approach by an example. Suppose we consider a system described by (2.81)
Mslep(t) = j met - t')dt' = j m(r)dr (2.149)
o 0
with Mo = XoHo· We ask what the effect would be if Ho were taken to be time
dependent in magnitude though tilted in direction. We replace Ho by H(t). This
_I_H(O_ _
silUation describes the famous experiments of CJ. Gorttr and his colleagues at
[.e;den [2.18).
We then have o , Fig. 2.19. Step runctKm
which gives
J m(r)dr (2.153)
m(O+) _ m(O-). <0 (2.147)
o
T, converges.
Since m(O-) = 0, and m(O+) = A [from (2.145)], we get A = XO/TI, or Suppose we apply an alternating magnetic field. We shall write it as complex
for simplicity:
met) = XOe-t/Tt (2.148)
T, H~(t) = H Xoeil<l! (2.154)
Then
52 53
,
M~(t) ::: J m(t - t')HXOeiwt' dt'
mer) = 211"
1
J
+00
x(w)ei,,",(dw (2.158)
-00
,
= HXOeiwt J met - t')eiw(tl-t)dt'
-00
That"h, m(r) and X(w) are Fouricr transfonns of each other. Knowledge of one
-00
completely delennines Ihe other. One may attempt to predict the properties of
J
00
= Hxoeilo!t m(r)e-I"'T dT (2.155) resonance lines either by analyzing the response to an alternating signal or by
analyzing the transient response. Kubo and Tomita [2.19J, for example, base their
o
general theory of magnetic resonance on the transient response, calculating the
Comparison with (2.96) shows that response of the system to a step.
Examination of (2.156) enables us to say something about X' and i' at both
J
00
X= m(r)e-i"'T dr zero and infinite frequencies. Clearly, X/l vanishes at w = 0, since sinO vanishes,
but x' does not vanish at w = O. Moreover, if mer) is a finite, reasonably
o
00 Jo
continuous function whose total intcgral oo m(r)dr is bounded, both X' and
,,
X" will go to zero as w ----+ 00, since the oscillations of the sin wr or cos wr will
x'= jm(T) cos wTdr (2.156)
"average" the integrand to zero. Actually we may pennit m(r) to be infinite at
o
r = O. We can see this by thinking of Jri
m(r)dr, the response to a step. We
J
00
XII = meT) sin wrdr certainly do not expect the response to a step to be discontinuous at any time
other than Ihat when the step is discontinuous (t = 0). Therefore m(r) can have
o at most an integrable infinity at t '" 0, since the response must be bounded. We
(see footnote~). It is simple to show, using the integral representation of the shall represent this by a a-function. Thus, if
c-funclion,
1 mer) = m,(r) + Clo(r) (2.159)
6(x) = 21r J
+00
eixtdt that (2.157)
where m,(r) has no a-function, we gel
-00
Jm,
00
J
00
•
M~(t) =
-00
J met - t')Hxoc i .." " e"' dt'
X(w) - X'(oo) =
o
m(r)e-i,,",T dr (2.161)
= lI xo e'W'c" J
•
met - t')eiw("-')c'{"-l)dt'
where now m(r) has no a-function part.
[Of course no physical system could havc a magnetization that follows the
_00
excitation at infinite frequency. However, if one were rather making a theorem
= Ilxoe(iw+.), J
00
m(r)e-(·+iwjT dr and
about penneabilitY!1-, p(oo) is not zero. We keep x'(oo) to emphasize the manner
in which such a case would be treated.]
• We wish now to prove a theorem relating X' and X", the so-called Kramers-
.-.
x("') = Jim jm(T)e-(.+'W)Tdr
The advantage of this definition is that it has meaning for the case of a ~lossless resonator"
Kronig theorem. To do so, we wish to consider X to be a function of a complex
variable z = x + iy. The real part of z will be the frequency w, but we use the
symbol x for w 10 make the fonnulas more familiar. Therefore
(magnetic analogue of an undamped harmonic oscillator), in which a sudden application
of a field would excite a transient that would never die out.
54 55
=
x(z) - i(oo) = J m(r)e- iZT d, The presence of the tenn exp (yr) tells us that
o
Ix(z) - i(oo)I--+ 0 as Y --I 00
= -
= J m(T)eYTe-i:l;T dT (2.162)
·We already know that
o
Ix(z)-i(oo)I-O as X--I±OO
Since an integral is closely related to a sum, we see that X(z) is essentially a
sum of exponentials of z. Since each exponential is an analytic function of z. so Therefore X(z) - i(=) is a function that is analytic for Y:5 0 and goes to zero
is the integral, providing nothing too bizarre results from integration. as Iz I -+ 00 in the lower half of the complex plane.
To prove that X(z} - x'(oo) is an analytic function of z, one may apply the Let us consider a camour integral along the path of Fig. 2.20 of the function
Cauchy derivative test, which says that if xCi) - x'(oo)
x(z) - X/Coo) == U + iv (2.163) z' w
where u and v are real, u and v must satisfy the equations By Cauchy's integral theorem this integral vanishes, since X(z) has no poles
inside the contour.
au au avail
ax = ay and ax = - ay (2.164) X'(z') ~ X'(oo)
j "--''--'-;---'''-'.:.'-'dz
I
=0 (2.167)
ZI_W
From (2.162) we have C
56 57
Solving for the real and imaginary parts, we find x" (a) I X
. I (b)
I I
1 +00 1/( ')
x(w) - x'(oo) = -p ~c/w' J (2.169)
-n
I
I
I
I
X"(w) = _2.. p
11" w'-w
~-n n!(
I
w
1f w'-w
-00
These are [he famous Kramers-Krollig equations. Similar equations can be worked fig. 2.11. (a) Absorption spectrum. (b) Corl"C$ponding dispel'!lion spedrum
out for analogous quantities such as the dielectric constant or the electrical sus-
ceptibility.
The significance of these equations is that there are restrictions placed, for
example, on the dispersion by the absorption. One cannot dream up arbitrary
,
x (W)-X(OO), =- c( 1
11"
---
f}-w -f}-w
I) c(- -I + -I)
=-
11"
-
fl-w fl+w
X'(w) and X"(w). To phrase alternately, we may say that knowledge of X" for all
(2.172)
frequencies enables one to compute the X' at any frequency. Note in particular where we have used the fact that 6(%) = 6(-%).7
that for a narrow resonance line, assuming x'(oo) = 0, the static susceptibility Of course, near resonance (w ";l! fl), only the firsl tenn is large. The function
is shown in Fig. 2.21.
Xo is given by
XO = X(O) = -p
•
I
J _,_w_<iw'
+00 ..fl( ')
2 I +00
-
- •-wo
- J"(
x ')" wuw (2.170)
We shall now tum to obtaining expressions for the absorption and dispersion
in tenns of atomic properties such as the wave functions, matrix elements, and
o energy levels of the system under study. We shall compute X" directly and obtain
The integral of X"(w l ) is essentially the area under the absorption curve. We see X' from the Kramers-Kronig equations.
that i[ may be computed if the static susceptibility is known. G We make the connection between the macroscopic and the microscopic prop-
As an example, suppose erties by computing the average power P absorbed from an alternating magnetic
field HrlJ cos wt From (2.102) we have
x"(w) = c[6(w - m- 6(-w - fl)] (2.171)
P -w"H'V
-"2X zO (2.173)
The first tenn corresponds to absorption at frequency n. The second tenn simply
makes x" an odd function of w. For this funclion, whal is X'(w)? in a volume V. It will be convenient henceforth to refer everything to a unit
x'(w) - X'(oo)
= 2.. p
+00
J c{6(w' - m- 6(-w' - n>jdw'
wI-w
volume. (We shall have to remember this fact when we compute the atomic
expressions in particular cases).
7l"
-00
On the other hand, Ihe alternating field couples 10 the magnetic moment Itzk
of the klh spin. Therefore. in our Hamiltonian we shall have a time-dependent
perturbation 1{t)er~ of
G or course, if we arc talking about a m... gnetic re>;onance exp·eriment with the static
field in ~he z-direc~ion and the alternating field in the z-direetion, we are discussing Xu. 7 To show that 6(z) = 6(-z) we consider the integrals II = I~oo f(z)6(z)dz and
Then x'(O) or (2.170) is x~...(O), whereas xo is usually thought or as relating the total I~ =I~::: f(z)6(-z)dz. We have irnmedia~ely that II = j(O).1b eval~ate 12, we chAnge
magnetization Alo to the field I/o, which produces it, and is thllS X~r{O). lIowever, a variable to z' = -z. Then we get
Slnll.lIstatie field IIr in the z-directioll simpl)' rotates M o , giving +-
J =/(0)
--
I/r , ) 12 = f( -z')6(z')dz'
Mr = M o-
II,
= x.. (0 lIr
(2.182)
provided certain conditions are satisfied: We do not ask for details that appear
Fig.2.22. Eigenvalues of energy on a time scale shorter than a certain characteristic time T. It must be possible to
find such a time, which will satisfy the conditions that 1) the populations change
only a small amount in T and 2) the possible states between which absorption
The states la) and Ib) are eigenstates of the Hamiltonilln. The most general can occur must be spread in energy continuously over a range LJ.E such thaI
wave function would be a linear combination of such eigenstates: LJ.E» hiT.
(2.177) These conditions are violated if the perturbation matrix element l(aIFlb)1
exceeds the line width, as it does when a very strong alternating field is applied.
where the ca's are complex constants. The square of the absolute value of Ca We can see this point as follows: The quantity LJ.E may be taken as the line
gives the probability p(a) of finding the system in the eigenstate a: width. We have, then, that LJ.E< l(alFlb)l. But under these circumstances one
can show that the populations change significantly in a time of order IiII(alFlb)j.
p(a)" 1,.1 2 Thus to satisfy the condition I that the populations change only a small amount
If the system is in thennal equilibrium, all stales will be occupied to some during T, T must be chosen less than hll(aIFlb)l. This gives us
extent, the probability of occupation p(a) being given by the Boltzmann factor h
l(a!Flb)1 < -
e- EA / kT T
,
= /
(2.189)
Therefore
Substitution of (2.189) into (2.185), together with recognition that E a - E b = hw,
I X"(w) =, L: [P(Eb) - p(E,)II(a!",lb)I',(E, - E, - hw) . (2.184)
owing to the o-functions, gives
Ea > Eb
As long as Ea > Eb.
only positive w will give absorption because of the 0- xl/(w) = ~;; L: e-Ea/kTI(alfl':rlb)120(Ea - E b - hw) (2.190)
function in (2.184). Removal of the restriction E a > Eb extends the meaning of Ea,Eb
xll(w) fonnally to negative w. Note that since lJ(Eb) - P{Ea ) changes sign when
a and b are interchanged, i'(w) is an odd function of w, as described in the Another expression for XI/(w) is frequently encountered. It is the basis,
for example, of Anderson's theory of motional narrowing [2.21]. We discuss it
preceding section:
in Appendix B because a proper discussion requires reference to some of the
X"(w) =, L: IJKEb) - p(E,)II(al,',lb)I',(E, - E, - I>w) . (2.185) material in Chapters 3 and 5.
Ea,Eb It is important to comment on the role of the factors exp(-Ea/kT). If
Assuming i(oo) = 0 for our system, we can easily compute i(w), since one is dealing with water, for example, the proton absorption lines are found
to be quite different at different temperatures. Ice, if cold enough, possesses a
+=
x'(w)=.!.P
1T
Jw lie I
XI w)dw '
-w
(2.186)
resonance several kilocycles broad, whereas the width of the proton resonance
in liquid water is only about 1 cycle. Clearly the only difference is associated
-00 with the relative mobility of the H 20 molecule in the liquid as opposed to the
J ,(E, -,Eo -
+00 ,
62 63
Evaluation of x!' by using (2.190) would require knowledge of the wave
functions and energy levels of the system. As we shall see, we rarely have
3. Magnetic Dipolar Broadening of Rigid Lattices
that infonnation, but we shalI be able to use (2.190) to compute the so-called
moments of the absorption line. We see that the only frequencies at which strong
absorption will occur must correspond to transitions among states between which
the magnetic moment has large matrix. elements.
3.1 Introduction
A number of physical phenomena may contribute to the width of a resonance
line. The most prosaic is the lack of homogeneity of the applied static magnetic
field. By dint of hard work and clever techniques, this source can be reduced to a
few milligauss out of 104 Gauss, although more typically magnet homogeneities
are a few tenths of a Gauss. The homogeneity depends on sample size. Typical
samples have a volume between 0.1 cc to several cubic centimeters. Of course
fields of ultrahigh homogeneity place severe requirements on the frequency sta-
bility of the oscillator used to generate the alternating fields. Although these
matters are of great technical imponance, we shall not discuss them here. If a
nucleus possesses a nonvanishing eleclric quadrupole moment, the degeneracy
of the resonance frequencies between different m-values may be lifted, giving
rise to either resolved or unresolved splittings. The latter effectively broaden
the resonance. The fact thaI T, processes produce an equilibrium population by
balancing rates of transitions puts a limit on the Iifctime of the Zeeman stales,
which effectively broadens the resonance lines by an energy of the order of lilT,.
In this chapter, however, we shall ignore all these effects and concentrate
on the contribution of the magnetic dipole coupling between the various nuclei
to the width of the Zeeman transition. This approximation is often ex.cellent,
particularly when the nuclei have spin! (thus a vanishing quadrupole moment)
and a rather long spin-lattice relaxation time.
A rough estimate of the effect of the dipolar coupling is easily made. If
typical neighboring nuclei are a distance r apan and have magnetic moment lA,
they produce a magnetic field Bloc of the order
I'
Bloc ::; r 3 (3.1)
By using,. ::; 2 A. and J-l ::; 10- 23 erg/Gauss (10- 3 of a Bohr magneton), we find
Hloc s::o I Gauss. Since this field may either aid or oppose the static ficld Ho, a
spread in the resonance condition results, with significant absorption occurring
over a range of B 0 ....., I Gauss. The resonance width on this argument is indepen-
dent of H 0, but for typical laboratory fields of 104 Gauss, we see there is indeed
a sharp resonant line. Since the width is substantially greater than the magnet
inhomogeneity, it is possible to study the shape in detail without instrumental
limitations.
64 65
z Flg.3.t. Relationship between rectangular coor-
3.2 Basic Interaction dina~es :1:, II, , (describing the position of nucleus
2 rdative to nucleus I) and the polar coordinates
The classical interaction energy E between two magnetic moments III and 112 r, 8, '"
;s
(3.2)
where r is the radius vector from III to JJ.2. (The expression is unchanged if r
is Iaken as the vector from 112 to PI.) For the quanlum mechanical Hamiltonian
we simply take (3.2). treating 1-'1 and P2 as operators as usual:
k--li-r7-- -" I'
1/
y
If we express 1)z and [151 in tenns of the raising and lowering operators r:- We shall diagram the appropriate matrix elemenlS and energy levels in
and I), respectively, and express the rectangular coordinates z, y, z in tenns of Fig.3.2. It is convenient to denote a state in which ml = m2 = -~ by+!,
spherical coordinates r, 8, tP (Fig. 3.1), we may write the Hamiltonian in a fonn the nOlation 1+ -). The twO states 1+ -) and 1- +) are degenerate. and both
that is particularly convenient for computing matrix elements: have B-l = O. The states 1+ +) and 1- -) have, respectively. -hwo and +hwo,
where wo = ,Ro as usual. We first ask what pairs of states are connected by the
/12
'}-{d = /11'2 (A + B + C + D + E + F) (3.6) various tenns in the dipolar expression. The tenn At which is proponional to
r' II.zh.z' is clearly completely diagonal: It connects Iml m2) with (mtm21. On the
where other hand, B, which is proportional to It 12" + II It,
only connects Iml m2)
A"" [1,[2,(1 - 3 cos 8) 2 to states (ml + I, m2 - II or (ml - I, m2 + II. A customary parlance is 10 say
that B simultaneously flips one spin up and the other down. B therefore can
B = -t(I{ 12" + I) [i)(I - 3 cos 2 8) join only the states 1+ -) and 1- +). The states joined by A and B are shown
C= -~(It h, + 1),Ii) sin B cos Be-i,p diagrammatically in Fig. 3.3.
(3.7)
D = -~(Il h, + IlrI2") sin 8 cos 8e i,p
E= -~I{Ii sin 2 8e- 2i ,p
+----- -----+ o
F -- -;{'1-1- . 2 ee 2i ,p
I 2 Sin
----++
As we have remarked, (/11'2h2)1r3 corresponds to the interaclion of a nu-
Fig. 3.2. Energy levels or two iden~iclIl spins
clear moment with a field of about I Gauss, whereas the Zeeman Hamillonian
66 67
,--A ... Fig. 3.3. States joined by matrix elements
A .. nd B. The dashed lines &0 between
where Un is the wave function corrected for the effect of the penurbation 'Hpert
( )
states thal are joined and where, of course, the matrix elements (n'j'Hpertln) are computed between
the unpenurbed states of u~, and u~.
A A
...-........
+- ( , ( ) ....-- ....
-+
Qy means of (3.9) we can see that the state 1+ +) will have a small admixture
-----_ ....
B ..... > of 1+ -), 1- +), and 1- -). The amount of admixture will depend on (n'I'Hpertln)
and En - En" The former will be -y 2h 2Jr3 multiplied by a spin matrix element.
,..-_ ....A
Since the spin matrix element is always of order unity, and since Hl oc = -yTl/r 3,
( '+ +
we can say (n/l'Hperdn) ':! -yIIHloc . On the other hand, En - En' = Ilwo =
'YfIHO, so that
Note that B has no diagonal matrix elements for [he m 1m2 representation, (n'I1ip<,d nl I (3.IOl
but it has off-diagonal elements between two stales which are degenerate. The
fael that off-diagonal elements join the degenerate states 1+ -) and J _ +)
I En - En'
tells us, of course, that they are not the proper zero-order states. B therefore corresponding to a very small admixture. Of course the admixture produces a
plays an important role in determining the proper zero-order functions. When second-order energy shift. As a second, and for us more important, effect, the
Ihe ~roper zero-order functions are delcrmined, B turns QuI to have diagonal admixture enables the alternating field to induce transitions that would otherwise
matnx elements. We shall return to Ihis point later. be forbidden. Thus the transition from 1+ +) to 1- -), which would be forbidden if
Since terms C and D each flip one spin only, they join states shown in these were exactly the states, can now take place by means of the small admixture
Fig. 3.4. all of which differ by liwo in energy. Finally E and F flip both spins of the states 1+ -) and 1- +). (See Fig. 3.6.)
up or both spins down, connecting states that differ by 2Tlwo (Fig. 3.5).
++--.~,
Fig. 3.6. The strong transition is in-
dicated by the double arrow. The
+- -+ +---- -"-----+ transition by ~he tight arrOW has non·
vanishing matrix clements due to
the dipole I\dmix~urcs
Fig.3.4. States joined by the terms C aud D
The matrix element is smaller than the normal one, for example, between
i 1+ +) and 1+ -), in the ratio H1oc1Ho. 1llerefore the intensity of the absorption.
lE,F
I + which goes as the square of the matrix element, is weaker in the ratio (HlocIHo)2.
+---- The transition occurs, of course, at w = 2wo. A funher consequence of the
I
I Fig. 3.5. States joined by the dipolar admixture of states is that a transition near w = 0 can be induced. [Actually this
I ++ terms E and F transition is forbidden for a pair of spins, each of spin!, because the eigenstates
of }.II = mt + 1n2 = 0 are of different symmetry under exchange of the panicle
labels ((he singlet and triplet states), whereas the penurbation is symmetric. If
, more than two spins are involved, the transition is permilled.]
.The terms C, D, E. and F therefore are off-diagonal. They produce slight
admixtures of our zero-order states into the exact stales. The amount of the The net effect of the terms C, D, E, and F is therefore to give the absorption
admixture can be computed by second-order penurbation Iheory. using the well- near 0 and 2w0, shown in Fig. 3.7. The extra peaks at 0 and 2wo are very weak
known expression for the correction of the zero-order functions U?I of zero-order and may be disregarded for our purposes. Since they are the principal effects of
energy En: the terms C, D, E, and F, it will be an excellent approximation to drop C, D,
E, and F from the Hamiltonian. For some of our later calculations we shall see
U = u O + '" (n'l1-lperdn) 0 that failure 10 drop these terms can lead us into erroneous results. The remaining
nnL..EEun' (3.9)
n' n n' dipolar term A + B may be combined to give what we shall call Jt3:
68 69
l_-"--,-,,fl_~L"":"'..~
3.3 Method of Moments'
Ab."pti,"
Before outlining the method of moments, we must return to our original expres-
sion for xl/(w):
70
71
~
So, we can rewrite (3.23) as
j(w - (w»n j(w)dw
~ 1 1
(&.,n) = ~O'---O~,, _ (3.24)
(3.19) j j(w)dw = 2h L(al";la) = 2h T, (";)
0- •
j j(w)dw
o where the symbol "TT" stands for "trace" or sum of the diagonal matrix elements.
The expression (3.19) for n = 2 is called the "second moment". Clearly (L1w 2 ) Another imponant theorem tells us that when we go from one complete set of
is of the order of the square of the line width, so that onhogonal functions 1,8) to an alternative one I() [1,8) can thus be expressed as a
linear combination of the IO's], the trace is unchanged. We may therefore choose
(3.20) any complete set of functions to compute the trace. In fact we shall choose a
set of functions that is simply the product of the individual spin functions of
The two moments of (3.18) and (3.19) are closely related. as may readily be seen quantum numbers mt> m2, m3 ... mN for the N spins. Therefore
for n:: 2 as follows: Expanding (w - (w»2 = w 2 - 2w(w) + (w 2 ), one easily
shows from (3.18) and (3.19) that
Jo
~ I
f(w)dw:: 2h L
ml,ml,m:s, ..
(mlm2 m J ···lp;l m lm2 m J .•• ) (3.25)
(&.,') = (w') _ (w)' . (3.21)
Therefore we can compute either (&.;2) directly or compule it from the calcu-
Now, since pz:: Ej Pzj,
lations of (w 2 ) and (w). (We shall do the laneL) pi:: L,pzjpzk . (3.26)
. To illustrate the general methods, we shall first campUle k
j(w)dw. which j,k
IS, of course, closely related 10 the area under the absorption curve. Then we
shall compute (w) and (L1w2). Since f(w) is an even function, There are two types of terms: j '" k and j :: k. We examine the first kind first.
Let us consider j :: 1, k:: 2. Then, holding m2' m3, m4 ... fixed, we can
i
~ +~
first sum over mt. Now,
j j(w)dw = j j(w)dw
o -~
I+~
so that summing over m 1 gives us
=2 j D al".lb)(bl".)a)6(E. - E. - hw)dw (3.22)
-00 G,b
[~(mIIPh;lml)] (m2Ip2zl m 2) (3.28)
The integrand picks up a contribution from the IS-function integral every lime
hw = E a - E b- But for any pair of states la) and Ib), there is some value of Now, Eml (mllplzlml) :: O. This may be seen by noting that when we lake
w between -00 and +00 which satisfies the condition tlW :: E a - Eb' (Note, if ml as eigenvalues of liz, all the diagonal elements of liz and Plz are zero. Or,
we had the integral from 0 to 00, we should have zero from states for which alternatively, one may let ml be the eigenvalues of liz' But for every +m value
En - Eb was negative. It is for this reason that we let the integral range from there is a corresponding negative one, so that
-00 to +00.) Thus, changing the variable of integralion from w to liw, we get
L,(mllplzlmd:: "(ft L,(mI11lzlml):: 0 . (3.29)
~
j j(w)dw = 2 n ~:::<all'.lb)(bl".la)
1 ml ml
(3.23)
Therefore the contribulion from tenns j :/: k vanishes. For j :: k, we get, taking
o n,b
j :: I,
DUl it is a basic Iheorem of quantum mechanics that for any complete set 1,8')
and for any operators A and B,
j f(w)dw = 21'l72h2 I(I; 1) N(2I + l)N (B2) The energies Ea and E" are the sum of dipolar and Zeeman contributions
(--yhHoM + En), as we have remarked previously. We shall assume that the
o
dipolar energy changes are always small compared with the changes in Zeeman
We tum now to a calculation of the effect of the dipolar coupling on energy and that the latter correspond to absorption near wo (our earlier discus-
the average frequency of absorption, (w). The existence of such a shift im- sion of the role of the tenns A. H, ... F shows us this fact). Therefore. since
plies that the local field produced by the neighbors has a preferential orien- Eo > Eb. we write
tation with respect to the applied field. Since such an effect must correspond
to a Lorentz local field .t!J.H, it must be of general order XIIHo. where XII is Eo = -"(TiHoM + En
the static nuclear susceptibility. X'I is given by the Langevin-Debye fomlUla: Eb = -"(IiHo(M + 1) + En' (3.35)
XII = N"(2h 2 I(I + 1)l3kT, where N is the number of nuclei per unit volume. Eo-Eo = flWo + En-Er./
If the distance between nearest neighbors is a. N ;:' Iht 3• we have, therefore,
that t1H S! (,lda3)(-rliHolkT) S! Hloc(,IiHolkT). Since the nuclear Zeeman By using these relations, we can write (3.34) as
energy 7hHo is very small compared with kT. we see t1H is very small com-
=
pared with the line breadth Hloc and is presumably negligible. Notice that the
physical significance of our expression for t1H is that the neighbors have a slight
J
o
wf(w)dw = 2"
I
L
" '"I ,n,a'
(MalJ-lzIM + 10")
preferential orientation parallel to the static field given by the exponent of the
Boltzmann factor ("(IiHolkT). Hl oc has a nonzero average to this extent. Since X (M + 100'1/lz1M a)(!iwo + EO' - EO',) (3.36)
few) of (3.17) corresponds to infinite temperature, it must lead to a .t!J.H = 0,
and (w) = woo We shaJl first discuss the contribution of hwo tenn in the parentheses. It is
74 75
~o L: (MaIJJ;!:jM + la')(M + la'IJJ;!:IMa) . (3.37) = EOI, / u Al'OI,PU,\I01 dr
it /11,0',0"
= Eo:,(M'o'jPIMa) . (3.42)
Were it not for the restriction to M + 1, (3.37) could be converted to a trace by Therefore
means of (3.23a). This restriction can be removed by using the properties of the
raising and lowering operators and by noting that L (MoIJJ-jM'a')(M'o'lp+jMo)(Ea - EOt')
/11 ,/If' ,CI',O"
• L (Moll1i~,~-lIM'o')(M'o'I~+lMa) (3.43)
M,/I1',&,«'
Thus
.T'{[1i~'~-)I'+) .
(3.38)
A detailed evaluation of this trace shows that it vanishes. Therefore, combining
Since Jl+ connects only states M ' and /11 in a matrix element (M' a'IJJ+JM (X) the results of (3.36), (3.39), and (3.43), we get
where M' = Ai + I, we can rewrite (3.37), summing over all values of M' as
f00
OOJ I
j(w)dw = 2h T, {~~}
(3.39)
o
Therefore
00
J wj(w)dw
where we have used the facts that Tr {Jl~} = Tr fJl~} and (w) =~ =wo (3.45)
= EO'(M'a'IPIMa:) . / w 2 f(w)dw
Each term is clearly of order (-1H{;,,/ where H~c is the contribution of the kth
(3.48) spin to the local field at spin j. The imponant point about (3.53) is that it gives
a precise meaning to the concept of a local field, which enables one to compare
We can expand, using H = 'Hz + 'H'~ to get a precisely defined meoretical quantity with experimental values.
So far we have considered only the second moment for a case where all
nuclei are identical. If more than one species is involved. we get a somewhat
different answer. The basic difference is in the terms of type B in the dipolar
coupling that connect states such as 1+ -) 10 J - +).
If the two states are degenerate. as in the case when the spins are identical.
B makes a first-order shift in the energy. On the other hand, when the states are
nondegenerate, B merely produces second order energy shifts and gives rise to
where in Ihe "cross-Ierm" involving {[1tZ. 1Ja:] and l1t3, J.lz] we have used the
basic relation true for any pair of operators A and B: weak. otherwise forbidden transitions. It is therefore appropriate to omit B when
the spins are unlike. 2
Tr {AD} = Tr{BA} The interactions between like and unlike nuclear spins may be compared
(3.50)
and me second moment readily obtained. If we use the symbol 1 for the species
which is readily proved by applying (3.23a). If the dipolar coupling were zero under observation. and S for the other species, the effective dipolar coupling
only the first lenn on the right would survive. and of course the resonance would between like nuclei is
be a 6"-function at w::;; WOo In this case (w 2 ) = wfi.
Therefore we sec that the first .. ,0
(rtJ)Il:::
I 2 2~(1-3coS26H)
,("h L. (3[zl.:lz1 ~ II.: ·1/) (3.54)
lenn must contribute w~ 10 (w 2 ). Explicit evaluation in fact verifies lhis result. x,1
3
rk/
The second. or '\,TOSS", leon vanishes, since every term involves factors such as
Trl {Pb-}' The lasl term, when divided by k
j(w)dw gives In computing the second moment for like spins, the terms II.: • II do not can·
tribute, since they commute with 1-'% [see (3.49)]. The coupling between unlike
~,'i ft2 1(1 + I) (~) L: (I - 3 c~s2 8jk )2
(3.51)
spins is
N i,I.: rjl.: 2 Van Vied: points out thlll omitting these terms for unlike spins lIS well as the terms
C, V, E, and F for like sl,ins is crucial in computing (.1",,2). The !'i:lISOn is lhlll in
Now, by (3.21), = =
computing (.6",,2), the ralher wellk satellite lines at "" 0 lind"" 2<.00 cocreipond to a
typical f~uency fcom the center of the N50nancc, which is 110/1110<: larger than t1Kl1;C or
{Llw') = (w') _ {w)' the main transition. The second momenl mcUUI'(!S lile square or the rrequency deviation.
Therefore, althoush the sal.ellitcs are down in intensity by (Jl loc / Jlo)', they contribute
Therefore, since (w)::: WO, we have an amount quile comparable to lhe second moment. Since we arc concerned with the:
width or the main lrllnsitiol'l, we do not wish to include the satellites. We must exclude
the terms thal produce lhem fcom the Hamillonian.
78
79
(3.55)
, Fig.3.9. Unit cell or the ben'tcnc crystal. Solid lines
represent molccutCll in the II = 0 plane; dashed lin~
, represent moJccul~ b/2 abo,'c
Equations (3.54) and (3.55) differ primarily in the numerical factor of the zz- ,, ,
term, {3.SS) being small by a factor of j. This numerical factor becomes ~ in ,
the second moment, giving for the final answer.
... •
(&.12)/5 = ~7h~h2S(S + 1)2. E (I - 3~s2 8jk)2 (3.56)
3 Nj,k rjk
NOIice that it is S(S + I), nOI 1(1 + 1), Ihal comes inlo (3.56) expressing the rough sketch of the crystal structure is shown in Fig. 3.9, looking down the b--axis
fact that the local magnetic field seen by nuclei 1 is proportional to the magnetic edge on to the plane of me molecules. The plane of Ihe molecules is represented
moment 'YshJS(S + I) of the other species. The total second moment of the by straight lines, solid for those atoms in the y = 0 plane, dashed for those
resonance line of spin I is given by adding the second-moment contributions of b/2 above the y = 0 plane. (Since the samples studied by Andrew and Eades
Hke nuclei 10 those of unlike nuclei. were polycrystalline, studies of the effect of the orientation of the magnetic field
relative to the crystalline axes were not possible.)
As we can see, there will be contributions to the second moment from nuclei
3.4 Example of the Use of Second Moments within the same molecule and from nuclei outside the molecule. In principle,
if one knew the location and orientation of all molecules, the only unknown
Since the pioneering work of Pake and Gurowsky, numerous studies of sec- parameter would be the distance R belween adjacenl protons in the ring. By
ond moments have been reported. A panicularly interesting example is provided using isotopic substitution, however, Andrew and Eades were able to obtain an
by the work of Andrew and Eades (3.1) on solid benzene. By studying the var- experime1l1al division of the total second moment into contributions within and
ious isotopic compositions in which protons were replaced by deuterons, mey outside. We can see this readily by noting that replacement of a proton by a
were able to measure the proton-proton distance between adjacent protons in the deuteron on any given site reduces the contribution of that site to the second
ring and to show that at temperatures above about 90 K, the benzene molecules moment by the factor 0':
are relatively free to reorient about the axis perpendicular to the plane of the
molecule. We shall describe their worle.. 0' = ~ i&lo(Io + I) (3.57)
The three isotopic species studied by Andrew and Eades are shown in 9 i~Ip(Ip + I)
Fig. 3.8. The structure of the benzene crystal is very similar to that of a face-
centered cubic crystal with me benzene molecules on the comers and face cenlers where the subscripts P and 0 stand for the proton and the deuteron. By using
of the cube. However, although the sides of the unit cell are perpendicular, they the facts that 10 = l, Ip = i,
(iDl2:lI") = 6.535 x 102 , (iPf17r) = 42.57 x 102,
are not equal in length, the a-, b--, and c-axes being, respectively, 7.44 A, 9.65 A, we have 0' = 0.0236. Thus, consider SI> the second-moment contribution from
and 6.81 A. All benzenes have their planes parallel to the cystalline b--axis. A nuclei oUlSide the molecule. For C6H3D3, any given lattice position is equally
likely to have a proton or a deuteron. Therefore the proton contribution to the
second moment is cut by a factor of two. If all the lattice sites were occupied
by deuterons, the second moment would be cut by the factor a, but since only
one-half the sites are occupied by deuterons, the deuterons contribute O'Stl2. The
total second moment 51, contributed by atoms outside the molecule is therefore
2 + a51
5', = 51 2 = (1 2 +a)5' (3.58)
60
81
Fig. 3.10. Posi~ions of protons and deuterons and the rela·
~
D __ -_'~~ 10 (o'ig.J.ll. Second moment in Calla as a function of temperature
live distances
,/R ~R \
H I ' D
, 2ll ,
.
\ I
\ I
" ~.
~/
D ---- H
Thus we have for the second moments of 4H303 and C6H6. respectively,
Lei the radius vector from j to k make an angle 1jk with the rotation axis. Then,
where a and 5 are known. Therefore, measurement of 5 and S' gives us 51 and as the molecule rotates, the angle ejk (between Ho and the internuclear vector),
52 the separate contributions from outside and inside the molecule. The data for which occurs in the factor I - 3cos2 gjJ. in the second moment, varies with
C&HsD provide an independent check. time. Since the frequencies of rotation are high compared with the frequencies
On the basis of such studies Andrew and Eades detennined the distance R of interest in the resonan<:e)Jt is the time average of 1 - 3 cos 2 ejk that affects
between adjacent protons in the ring to be 2.495 ± 0.018 A, which is consistent the second moment. Assuming the motion is over a potential well of threefold
with a prediction of 2.473 ±0.025 A based on the C-C spacing as detennined by or higher symmetry, this average can be shown to be independent of the details
x-rays, and an estimmed value for the C-H bond length. Of course one can com- of the motion, and
bine the x-ray and resonance data to obtain the C-H bond distance. In principle. (3.61)
observation of the CI3 resonance would even pennit a detennination of the C-H
bond length directly. Equmion (3.61) shows that if the axis of rotation is parallel to the internuclear
TIle data we have mentioned were measured at temperatures below about axis (1jk = 0). in which case the relative position of the two nuclei is unaffected
90 K. A second imponant result of Andrew and Eades was deduced by their by the rotation. the angular factor is unaffected by the rotation. On the other
studies of the temperature dependence of the second moment (Fig. 3.11). TIle hand, if 1jk = 1r/2,
rapid drop in second moment is due to the rotation of the benzene molecules
2
about their hexad axis. Let us discuss this effect. (I ~ 3cos 8jk )avg = -~(l - 3cos 2 g') (3.62)
The effect of rotation may be expressed very simply in tenns of the angles
defined in Fig. 3.12. We consider a pair of nuclei j. k fixed in a molecule. the axis In a powder sample. we find all orientations of the crystal axes with respect
e'
of rotation of the molecule making an angle with respect to the static field Ho. to Ho. For a rigid lauice. we must therefore average (1 - 3cos 2 jk )2 over e
82 83
the random crystal orientations. When motion sets in, we musl first average For the case of benzene, Andrew and Eades found Ihal the second moment
1 _ 3 cos 2 0 Ok over the molion, to obtain the second momenl for a given crystal of C6H6 from prolOns wilhin the molecule drops from 3.lOGauss 2 at low tem-
orientation. ~en we muSI average over crystal orientations. peratures 10 0.77 ± 0.05 Gauss 2 at high temperatures. The assumption Ihat Ihe
For interacting pairs, the cOnlribution 10 the second momenl of the rigid narrowing results from rotation about the hexad axis makes 'Yjk = 1r!2, since all
lattice {Llw 2 )RL Ihen goes as protons lie in a plane perpendicular 10 the hexad axis, and predicts that the sec-
ond moment should drop to 3.10/4 = 0.78 Gauss 2, in excellent agreement with
(3.63) the observed decrease.
where the "avg" indicates an average over rotalion, and the bar indicates an
average of the orientation of Ihe rotation axis with respeci 10 Ro. By using
(3.61), we get
2
{Llw )rot ex(l- 3cos 2 0')2 (
2
3COS ')"k
J
2
-I)' (3.65)
Since the cryslal axes are randomly oriented with respect to No, so are lhe
rotation axes, specified by 0'. As a result
,
{Llw )rot = {Llw )RL
2
,(3COS 'Yjk -
2
1))' (3.66)
If ijk = 1r/2 (a pair rotating about an axis perpendicular to the internuclear axis),
the contribution of the pair interaction to the second moment is reduced by a
factor of 4. 3
3 The justification for averaging 1 - 3 cos 2 (Jjk over the motion before squaring rather
than averaging (1- 3cos 2 (Jjk)2 may be seen also by referring to the exact expression for
X"(w), which was proporlional lo
0,'
The stales la) and 1fT) may be considered to involve both spin and rolational quantum
numbers. But, since Ea - Eb is chosen to be ncar the Larmor frequency, lhe slales In)
and 1fT) must have the same rotalional quantum numbers. Therefore, in computing the
second moment, lhe lraee will be over spin variables only, but lhe angular factor will be
a diagonal matrix clement in the "laUice" coordinales. But this means that we replace
the classical 1- 3cos2 (Jjk by f 1l1.(1 - 3cos2 (Jjk)uLdT, where tiL is a lattice (in lhis case,
rOlalion) stale. This procedure amounts to "averaging" 1 - 3cos2 (Jjk over the motion
prior to squaring.
84 85
4. Magnetic Interactions of Nnclei with Electrons
4.1 Introduction
So far we have ignored (he fact Ihal the nuclei are surrounded by electrons with
which they can interact. In this chapter we shall consider the magnetic interac-
tions. postponing until later the consideration of the strong eleclfOstatic effects
that may be found when a nucleus possesses an electrical quadrupole moment.
TIle magnetic coupling of the electrons to the nucleus arises from magnetic fields
originaling either from the motion of the electrical charges or from the magnetic
moment associated with the electron spin. The fonner gives rise to the so-called
chemical shifts; the latter, to the Knight shifts in metals and 10 a coupling between
nuclear spins.
Both the chemical shifts and the Knight shifls have certain features in com-
mon. The total Hamiltonian of the electrons and the nuclei may be written as a
sum of four teons:
'Hnz(H) + 'He(O) + HeZ(H) + Hen
where 'Hnz is the nuclear Zeeman coupling in the applied field H; He(O) is the
Hamillonian of the electrons (orbital and spin) in the absence of H; 'Hez(H)
is the electron Zeeman energy; and 'Hen is the interaction between the nuclear
spins and the electron orbital and spin coordinates.
If 'hen were zero, the nuclear spin system would be decoupled from the
electrons, and the nuclear energy levels would be solely the Zeeman levels in the
applied field H. The tenn 'hen corresponds to the extra magnetic fields the nuclei
experience owing to the electrons. In a diamagnetic or paramagnetic substance.
the average field a nucleus experiences owing to the electrons vanishes when H
vanishes. However, since the interaction 'hez(H) polarizes the electron system,
the effect of the electron-nuclear coupling Hen is no longer zero. We may say
that the nuclei experience both a direct interaction with H through Hnz(H) and
an indirect one through the interplay of 'hez(H) and 'Hen. The problem is very
similar to the calculation of the electric field in a dielectric, in which we must
add 10 the applied electric field the field arising from induced dipole moments
in the other atoms.
Systems such as ferromagnets possess electronic magnetization even with
H = O. For them, the contribution of 'Hen is nonzero even without an applied
field.
87
We shall consider the orbital effects first, starting with a review of the major 4.3 Quenching of Orbital Motion
facts about chemical shifts.
Classical elcctricity and magnetism tell us that a charge q moving with velocity
v produces a magnctic field H at a point r ' away, given by
4.2 Experimental Facts About Chemical Shifts
q v x r
H= - - -' (4.3)
c 1',3
The most famous and most quoted example of chemical shifts is ethyl alcohol,
If we choose rather to ask for the field at the origin of a set of coordinates due
CH;jCH20H (see references on "Chemical Shifls" in the Bibliography). The
to a charge at position r, then r ' = -r and (4.3) becomes
proton resonance consists of three lines whose intensities are in tile rulios 3: 2: I.
If one possesses a highly homogeneolls magnet, each of these lines is found to H= 1.. r x v =..J..... r x mv =..J.....!:... (4.4)
possess structure that (as we shall see) is due to effects of eleclIon spin. The three c 1'3 me,.3 me 1. 3
lines are clearly due to the three "types" of protons, three in the CH3 group, two
where L is the angular momentum of the particle about the origin. Equation (4.4)
in the CHz group, and one in the DB. Evidently the nuclei experience fields of
has a quantum mechanical counterpart, as we shall discuss. We see immediately,
local origin that are different for different molecular surroundings. A comparison
however, that for s-states, H = 0 aI the position of a nucleus, since s-states have
of the spacing in magnetic field between the lines as a function of the frequency of
zero angular momentum, whereas H 4- 0 for p, d, and other states of nonzero
the resonance apparatus shows that the splitting is proponionalto the frequency.
angular momentum. The magnitude of H is of order
If we attribute the splitting to the fact that the nuclei must see a magnetic field
L!JH in addition to the applied field Ho, we may say the resonance frequency w
(4.5)
obeys the equation
where p is the Bohr magneton (10- 20 erg/Gauss). For fluorine the average value
w = ,(H0 + L!J1I) (4.1)
of I/r 3 for the 2]) electrons is
where L!JH ex Ho. We may therefore define a quantity u, which is independent
of H, by the equation (4.6)
88 89
(x ~Y) fer) 'f(') , (x :;;y) f(r) (4.8) We can make a more precise statement by computing (L z ), the expectation
value of the z-component of angular momentum. For generality, we shall make
where fer) is a spherically symmetric function. The three functions of (4.8) are our proof for any wave function whose spatial part is reaL The operator for L z
is
eigenfunctions of L z , the z-component of angular momentum and the m-values
being, from left to right, 1, 0, and -I. The wave functions of (4.7) arc simply
lincar combinations of those of (4.8). As long as the atom is frce, either set of
;, = "-
1
(x!'"ay -y!...)
ax (4.10)
wave functions is equally good, but if a magnetic field is applied parallel to the Therefore, for any wave function Uo,
z-dircction, the set of (4.8) must be chosen.
(OILzrO)~ Juo~(x~ -Y~)ltOdT (4.11)
, ,,' , (b)
which, since tlO is real, may be written as
,
~" )
(OILzIO)""~ Juo(x~ ~Y:x)UO(LT (4.12)
'" , ., Since all the quantities in the integral are real, (4.12) shows that (L z ) must be
-'I
'" pure imaginary unless the integral vanishes. But the diagonal matrix elements of
any Hennitian operator are real. Therefore the integral vanishes and
Fig. 4.1. (a) Four ehargcs plaeoo near an atom. The atom is assumed to be at the origin
and the charges to b<l all equidistant from the origin. The eharges +q lie on the z-axis (OIL,IO)=O . (4.13)
and -q lie on the y-axis. (b) The wave function zf(r) is largest along the z-axis for any
given distance r from the origin It is clear that this proof holds for any component of angular momentum.
We say that when
If now we surround the atom by a set of charges in the manner of Fig. 4. 1,
and for the moment assume that no static magnetic field is present, the degen-
(OIL. 10) = (OIL,IO) = (OIL, 10) = ° (4.14)
eracy is lifted. The proper eigenstates are then those of (4.7), since a symmetric the angular momentum is quenched.
potential such as that of Fig. 4.1 will have vanishing matrix elements between any Under what circumstances will the angular momentum be quenched? Clearly,
pair of the functions of (4.7). On the other hand, the diagonal matrix elements what is needed is the possibility of choosing the eigenfunctions as real. In the
will be different, since the state xf(r) concentrates the electron on the x-axis, absence of a magnetic field (which means that no spins are allowed to act on
near to the positive charges, whereas the state vf(r) concentrates the electron the orbit!), the Hamiltonian is real. If, moreover, a state is nondegenerate, its
near the negative charge. Clearly, xf(r) will lie lowest in energy, yf(,.) will eigenfunction is always real, since it is the solution of a real differential equation
be highest, and zf(I') will be between (unshifted in fact in the first order). The (apart from an arbitrary complex constant factor, which clearly docs not affect
resulting energy levels are shown in Fig. 4.2. the expectation value). Therefore (OIL",IO) ~ 0 for such a state. We conclude that
The ground state xf(r) may be written as whenever the crystalline electric fields leave a state nondegenerate, the orbital
angular momentum of that state is quenched. The physical bliSis of quenching
xf(r) = ~ [(X :/zYl f(r) + (x :;;y) f(r)] (4.9) is that the external charges exert torques on the eleclron orbit, causing the plane
of the orbit to precess. When the plane has exactly turned over, the sense of
a linear combination of the m ~ + I and m ~ -1 states. The states m ~ + I and circulation is reversed. Crudely stated, the eleclron path has been changed from
m ~ ~l correspond to electron circulation in opposite directions about the z-axis. lying in a plane to being much like the path of the string in a ball of twine.
Since they occur with equal weighting in (4.9), we see that xf(r) corresponds Of course applicalion of a magnetic field will change things. We can see
to equal mixtures of the two senses of circulation, or to zero net circulation. intuitively that a magnetic field in the z-direction will cause one sense of circu-
lation to be favored over the other. The wave functions will readjust so that the
ground state has a slight circulation in the favorable sense (the m ~ -I state will
yf{r)
1 be favored). In temls of a small quantity e, this will make a new ground state:
zf(r)
"
j xf(r)
I<1g.4.2. Splitting of the three p-slates by the charges
of Fig. 4.1a ~o =
I [ ( X + i Y)
,ji (I - e) ,ji f(,-j + (I + ,)
(X-i y ) ]
,ji f(,-j (4.15)
90 91
As we can see, this change results from the admixture of a small amount of the
state yl(r) into the ground state xl(I·). As we shall sec, the amount f: mixcd in (4.19)
is proportional to Ho, giving rise to a circulation that is proponiollal to Ho·
We turn now to a closer look at the details of chemical shifts. In a shnilar manner the operator for angular momentum, r x mv, is
~..'L
magnetic field produced by these currents aI the nucleus. We shall work out the
theory for one electron. We start by considering the Hamiltonian of the electron. j(r) ':.("'.,,'" - ",,,,,,') _ LA",'", (4.20)
2m I me
To treat the magnetic fields, we must introduce two vector potentials, Ao and
We note that j(r) is a vector function of position, and il is real (Ihat is, has
An, one associated with the magnetic field B o, the other with the magnetic field
zero imaginary part). We recognize it as being q rimes the quanlum mechanical
H n due to the nucleus. In tenns of Ao and An we have
probability current. Explicit evaluation, first using'" and A and then using t/J'
Ho = '\! x Ao (4.16) and A', shows that jeT) is gauge-invariant. Moreover, by assuming that t/J is a
solution of SchrOdinger's time-dependent equation, one can show that
As is well known, there is more than one vector potential that will produce a
given field. Thus, if H = 'il x A, a new vector potential A' = A + '\!¢ (where d· .
lVJ+ fJt =
0 ae with f! == q¢"t/J (4.21)
q, is any scalar function), will give the same field, since the curl of the gradient
of any function vanishes. A transfonnation from A 10 A' is called a gallge That is, j obeys the classical equation of continuity. For stationary stmes, ¢"t/J is
transformation. We must be sure that the physical results of any calcul:uion are independent of time and div j = O. j acts much as a classical current density. As
independent of the choice of gauge, that is, aT: guage-invariant. The effect of we shall see shortly, such an interpretation is very useful in considering chemical
a magnetic field is introduced into the SchrOdinger equation by replacing the shifts.
operator (Mi)'\! by (M)'\! - (qle)A, where q is the charge of the particle (q is The Hamiltonian for our electron acted on by two magnetic fields is therefore
positive or negative, depending on the sign of the charge of the panicle).
The Hamiltonian then becomes l(q
1/=- p--Ao--An
2m e e
'f)' +V (4.22)
I ( p--A
1/=-
2m c
q)' +V (4.17) where V representS all potential energy, including that due to fields that may
quench the orbital angular momentum.
where p == (li/i)'\!. If one uses n different gauge, AI = A + '\!¢(r), the new II is convenient to define :l quantity 'If by the equlIlion
solution ¢' is related to the old one (t/J) by the (unitary) trunsfonn:uiOIl
'If=p-!!.Ao (4.23.)
",' = '" exp [ + (iqlt<c)~(r) I (4.18) c
If we compure the expectation values of (li/i)'\!, (ljJ, Ul/i)\1ljJ), lllld (ljJl, (h/i)\1ljJ'), Since both p and Ao are Hennitian operators, so is 'If. Then (4.22) becomes
we find that they are not equal. Therefore, since any physical observable must be 1 q ([2
2 2
independent of the choice of gauge, we see that (fili)'\! cannot be the momentum 1(.=-2• 2 An + V
- - 2 ('If.A n +A n ·'If)+-2 (4.23b)
m 1ne 1ne
operator mv. The operator for mv, rather, is (II/i)'\! - (qle)A, which is gauge-
We shall choose An to be
invariant. That is,
A _ J.1.XT
n - r3 (4.24)
1 See rderences under ~ChefT1icl'Il Shifts" in the Bibliography.
92 93
where ~ is the nuclear moment, since this vector pOlcnlial generates the field of
a dipole. Since J.l is very small compared with electron moments, we expect to
we get
Epert "" -~
IJ~xr .
----;J . 30(r)dT
be able to treat it as an expansion parameter, and accordingly we drop the A~ (4.32)
term in comparison with me term linear in An. We thcn have "" -Il' [~J X~D(r)dT]
l'
I , + V - --(:Jr'
1i = -11" q An + A )
n .:Jr . (4.25) where, as we have remarked, Il is really the operator 7'11, but 3D(r) is simply a
2m 2mc
vector function of position. II is important to bear in mind that 30(1') is indepen-
In the absence of a nuclear coupling, (l/2m);r2 + V is simply the Hamiltonian of dent of the gauge of AD, the vcctor potc",ial of the static ficld. Equation (4.32)
the electron in the presence of the stalic field. We shall treat the term involving is identical in form to the classical intcraction of a magnctic moment Il with a
An as a perturbation, computing the energy by using first-order perturbation current dcnsity jo(r), since thc quamity in the square brackcts is thc ficld H due
theory. to the current.
lei us consider, then, the fim-order change in energy of a state whose wave Thcre is a good deal of similarity to the expression for the magnctic momcnt
function 'I/J is the exact solution of the problem of an electron acted on by the M of the elec(tons;
potential V and the static field Ho. The energy perturbation E1>crl is then
95
94
4.5 Computation of Current Density Then
"0 = "0 + L'''O''n
n
(4.43)
We tum now to computing io(r). To do so, we need the wave function,p, which
describes the electron when acted on by both the electrostatic potentials and the which gives us for the current,
static field. We have, then, • .1.*n.l. .1. '" *) '" hq * *
= -!t(}
2. ('flO v 'flO - 'flO v,po + L.J - 2• (,po 'V,pn - 'l./Jn'V,pokno
1# = E" where (4.34)
Jo(r)
ml n m)
Although this involves a particular gauge, we can see that in any gnuge, Ao will Bu'
be proportional to Ho. In the expression for the current io(r), "0 = (
X + iY)
Ji J(.) (4.47)
Let us define
(nIJipcrd O) v(r) = !:... k x r
enO Eo En (4.42) (4.50)
= m (x 2 +y2)
96 97
which is tangent to a circle whose plane is perpendicular (0 the z-axis, so that Then, using (4.54), we have
1.(r)l- -h 1
(4.51) 1{pert = --'-(Ho x r)'p
m +y2 ./:£2 2me (4.59,)
= --'-Ho' (r x p)
This gives a z angular momentum of 2me
(4.52) We recognize that r x p is the operator for angular momentum in the absence
of Ro. It is convenient in computing matrix elements to use the dimensionless
in accordance with our semiclassical picture of the electron in an m "" + I state operator (l/i)r x 'V for angular momentum. Denoting this by the symbol L, we
possessing one quantum of angular momentum. can write (4.59a) alternatively as
We see, therefore, the close relationship in this case of the current density,
qh
the "velocity", and our semiclassical picture of quantized orbits. -2-HoLz
1{pert = (4.59b)
When the states tPo and tPn may be taken as real (quenched orbital angular me
momentum), we have J(r) = 0, and Had we chosen the gauge of (4.55a), we would have
11 llperl =-2~eHo·(r-R)XP
io(r) = 2 .q l)£nO -£;IO)(tPO'VtP" -ljJn'V.pO) -
Tnl n
LAotP6
Inc
(4.53) (4.59c)
qf'
= --HoL,(R) I
For (4.53) to be valid, it is actually necessary only that the ground state possess 2me
quenched orbital angular momentum, but for excited states, we have assumed where Lz(R) is the z-component of angular momentum about the point at R.
that the real fonn of the wave functions has been chosen. The choice of gauge therefore specifies the point about which angular momentum
Let us now proceed to look at some examples. We shall consider two cases, is measured in the perturbation. It is, of course, most natural to choose R"" 0,
an ,,-state and a p-state. It will turn out that the chemical shirts for s-States are corresponding 10 measurement of angular momentum about the nucleus, since in
very small but that, for p-states, the effect of the magnetic field in unquenching general the electronic wave functions are classified as linear combinations of .s,
the orbital angular momentum plays the dominant role, giving chemical shifts p, d (and so on) functions. When the electron orbit extends over several atoms,
twO orders of magnitude larger than those typically found for ,,-states. more than one force center enters the problem. The choice of the best gauge then
To proceed, we must now choose a particular gauge for A o. It turns out, as becomes more complicated. A closely related problem in electron spin resonance
we shall see, to be panicularly convenient to take involving the g-shift is discussed in Chap. 11.
Let us now consider an .s-state. Then the wave function is spherically sym·
Ao = !Ho xr = !Hok x T (4.54)
metric:
although an equally correct one would be ~.(r) = ~.(c) (4.60)
Ao = ,Ho x (r - R) (4.55') It is clear that, since L,tP. = 0,
where R is a constant vector, or (nl1tpertl¢,) = 0 . (4.61)
Ao z =0 , Ao:>: = Hoy , Aoy =0 (4.55b) Therefore tno is zero for all excited states, and the entire current io(r) comes
In tenns of the Ao, (4.54), we have from the last term of (4.44):
2 2
div Ao =0 (4.56) io(r) = _LAo!J;5 = --q-Hok x r!J;;(r) (4.62)
me 2me
Then we have, from (4.41),
The current therefore flows in circles centered on the z-axis. The direction is such
1{......rt = --q-[Ao' p+ (,), Ao) + Ao .p] (4.57) as to produce a magnelic moment directed opposite to Ho so thm it produces a
,.- 2mc diamagnetic moment. We see that the current direction will also produce a field
where (p' Ao) means p acts solely on Ao. But since p = Ul/i)'V. opposed to Ho at the nucleus (see Fig. 4.3).
It is interesting to n()(e that there is a current flowing in the s-state. There
h
(p·Ao) = ,
~('J ·Ao) = 0 (4.58) must certainly be an associated angular momentum, yet we customarily think of
98 99
y FigA.3. Diamagnetic curnlllt flow in an
/I, ,,-state atom, and the magnelic fields
yf(r)
produced by the current 1
-q y zf(r)
/I -I/-k /I
+q "j
., +q
xf(r)
-q
Fig.4.4. Crystalline field due to charges .<ig.4.5. Energy levels for a crystalline field
+q at:l: =±a, .'1= z =0; such as shown in Fig.·l.'!
s-stales as having zero angular momentum. We are confronted with the paradox: -q at y = ± a, :I: = Z = 0
If s-states have zero angular momentum, how can there be electronic angular
momentum in a first-order perturbation treatment if the first-order pcourbalion
uses the unperturbed wave function? The answer is that the angular momentum The tenn 1/;0 V't/J" - t/Jn \71/;0 of (4.53) is readily shown to be
operator has changed from r x (Ii/i)\! in the absence of a field to r x [(llIi)\7 -
(q/c)A] when the field is present. By using the changed operator, the unchanged t/Jo \71/;n - t/J1l V't/Jo = (xj - yi)f2(,,) (4.65)
,s-state has acquired angular momentum. The angular momentum is imparted It is conventional that the pMt of (4.44) associated with the excited stllles
to the electron by the electric field associated with turning on the magnetic is called the paramagnetic current jp, since (as we shall see) it contribllles a
field, since this electric field produces a torque about the nucleus. There is a paramagnetic magnetic moment. We call the last tenn of (4.44) the diamagnetic
corresponding back reaction on the magnet. We note that since A is continuously current jo. Then, using (4.44), (4.64) and (4.65), we get for our example
variable, we can make the angular momentum continuously variable. By using
1i2 2 Ho
typical numbers for H o and r, one finds the angular momentum much smaller jp;= 2m ~c Ll k x r f2(r) (4.66)
than h. Does this fact violate the idea that angular momentum changes occur in
units of II? No, it does not, since the electron is not free but rather is coupled and by using (4.44) and (4.54),
to the magnet. The complete system of magnet plus electron can only change q2 I
angular momentum by h, but the division of angular momentum between the 3D = - - -Hok x rlt/J1 2
me2
parts of a coupled system does not have to be in integral units of Ii.
We turn now to a ]>-state Xf(I') acted on by the crystalline field such as that
discussed in Section 4.3. We duplicate the figures for the reader's convenience
"
= - - 2 Ho(k
me
x r)x2 j2(r) (4.67)
(F;g.4.4). It is clear that both jp and jn flow in concentric circles but in opposite direc-
The energy levels are then as in Fig.4.5. Let us consider H o to lie along tions. However, although div jp = 0, the same is not true of jl). Since div j = 0
the z-direction. In contrast to the s-state, the p-state has nonvanishing matrix for a stationary state (j = jp + jo), there is clearly a discrepancy. It may be
elements to the excited states, corresponding to the tendency of the static field to tmced to the fact that the wave functions used to derive jp and jo are not exact
unquench the angular momentum. For this orientation of Ho, the matrix element solutions of the crystalline field, but are, rather, only zero-order functions. The
to Zf(l-) vanishes. That to yf(r) is charges that give rise to the crystalline splitting will also distort the functions.
For example, xf(,·), which points towards the positive charges, will presulllllbly
(nl1i pert 10) = --'-Ho~
2mc 1
J (x'!!'-oy -
yf(I') y..£..-)Xf(1")dT
ox be elongated somewhat, whereas yf(r) will be somewhat compressed. This will
result in current flow with a radial component that will, so to speak, supply
H0 "J ' dr the circular currents of the dilllnllgnetic tenn. However, radial currents will af-
= +-2
q 7 [yf(')1 (4.63)
me , fect neither the chemical shift (since they produce zero field at the origin) nor
iqflHo the atomic magnetic moment. Therefore we shall not search for better starting
=----
2me functions.
where we have used the fact that the function yf(r) is normalized. The division between diamagnetic and paramagnetic currents would change
By using (4.63), we find if we had chosen a different gauge for Ao. However, if our solution for jo(r)
were exact (to order No), the towl current jo(r) would be gauge-invariant. It
(nl1i per dO) . qhHo I
GnO= =1--- (4.64) is for this reason that our expression is so useful, since it holds regardless of
Eo E" 2mc Ll
100 101
the gauge of Ao used to com pUle the currents or of lhe division of io between The diamagnetic field. which is given by
paramagnetic and diamagnetic tenns.
It is imponanl to compare the relative magnitudes of jp llnd in- From Ho == -~Ho
2mc2
j T x (k1'3 x r) x2 f2(1') (4.75)
(4.66) and (4.67) we get
turns 'out also to be in the z-direction only:
. __ . m x 2 .1 = _ . Ll
3D - Jr ,~t.2 JP(r'f')
/1 Inx
(4.68)
Ho"" - - 2
k q2
zHo 3
j
(x 2 + y2) 2 2
x f (7") (4.76)
me r
where h 2/mx 2 has the units of energy (comparable to the kinetic energy of an
It is most convenient to average H o over all orientations of Ho with respect
electron whose de Broglie wavelength is x). By substituting numbers, we find
to the X-, y-, and z-axes. This can be shown to be equivalent 10 averaging HI)
. . x 2L1 for Ho parallel 10 the X-, y-, and z-axes, in tum giving
JD = - J p - (4.69)
8 I q' j(x2+y2)+(X2+z2)+(y2+z2)] 2
where x is measured in angstroms and .1 in electron volts. Thus, if .1 = 8eV Ho = --3 - 2zHo 3 x j(l')
me r
(a fairly typical .1 for chemical shift problems), we see that ip is larger for x (4.77)
less than I A but that io is larger outside. As we shall see, the distance that is = -LHO(~)
3mc2 r
most important for typical chemical shifts is about 0.25 A, so the paramagnetic
current dominates; however, for computing magnetic moments, a distance of I A Since Ho = -oDBO. where (10 is the diamagnetic contribution to the chemical
or greater is more typical, muking il difficult 10 assess which factor is the more shielding parameter (j, we have
important.
We can now compute the chemical shift fields Hp and HO duc to jp and uD = 3~c2 (~ ) (4.78)
jp an expression first derived by Lamb [4.2] to describe the shielding of dosed
HI' =.!.jrx dr
c ,,3 atomic shells.
(4.70)
= 1i.
2
L 2 Ho j r X (k x r) f 2 (1')dr We can likewise average HI'; however, we note here that I-Ip is zero when
Ho is parallel to the x-axis, since the penurbation gives zero when acting on the
2m .1 mc r3
cylindrically symmetric function xf(1'). It is also convenient to assume that yf(1')
Direct evaluation shows that the x- and y-components of FIp vlll1ish, leaving
and zf(1') are degenerate, both a distance .1 above xf(r), since this corresponds
only the z-component:
to the typical case of a chemical bond. Then we have
2
(4.71) Hp = ~ li L.2
Ho ( I ) (4.79)
3 m mc L1 1'3
Now, for any wave function t/J(,'), Ihe mean valuc of 1/1,3 is given by
and 0"1', the paramagnetic contribution to u, is
C~ ) j ,I,I,pI'dT
= , (4.72)
(TI~ = - } :~ ~:2 ~ C\) (4.80)
so that we see
If we take .1 = 4.3eV, and (lh- 3 ) = 8.891(l~, where ao is the Bohr radius
j x2 f2(r)dr = ( I ) =
1,3 1'3
j y2 f2(r)dr
,.3
(4.73) (values appropriale to the 21' electrons of fluorine, with the energy chosen to be
appropriate for the F2 molecule), we find 0"1' = -20 x 10-'1. uo is typically
Therefore we find 10- 5 • We see Ihat this value of Up is quite comparable to the changes in 0"
H - kh
"q --
Ho ( I ) (4.74)
observed for fluorine compounds, whereas 0"0 is much too small to account for
the effects. It is dear also why the range of fluorine chemical shifts is so much
p- ~mc2Ll 1'3 larger than thai of protons.
We note that HI' aids Ihe static field and is in fact proportional, in keeping with Physically, the large fluorine shifts come about because the magnetic field
the experimental data. leads to an unquenching of the angular momentum. The smaller .1, the more
effectively Ho can "unquench".
102 103
What can we say is the cause of the s-st3le shift? One simple picture is the most prominent force center in the problem. Since our wave functions will
to nOle that an s-state is a radial standing wave. Since the magnelic force is be approximate. the approximations will then at least be so primarily because
transverse 10 the radial motion. it produces a slow rol3tion quite analogous to the of failure to acCount for the small crystalline potential rather than for the much
manner in which the Corialis force causes lhe direclion of a Foucault pendulum large-r atomic (central) potential.
10 tum. When we are dealing with chemical shifts in molecules, we find il very
As we have seen, for any reasonable values of il, the paramagnetic shielding hard to deal wilh bonds. since two force centers, one at each nucleus (for a pair
lenn will completely overwhelm the diamagnetic tenn. What can we say about bond), are important. The only simple approximation is to treat the aloms as
M, Ihe electron contribution to the alomic magnetic moment? virtually isolated. laking the excitation energies as being those of excited bonds,
and including such effects as ionic character by using non normalized alomic
M=~JrXjodT (4.81) wave functions. Pople [4.3] has discussed Ihis problem. using a technique due to
2c
London. His resulls can also be obtained by using perturbation theory. A similar
We contrast this with the shielding field
problem arises when calculating g-shifts, and this is discussed in Chap. I I.
H=!J r xio dr (4.82) When one has ring compounds such as benzene, the interatomic currents
c ,3 are important. Then one chooses a molecular force center; that is, a gauge in
Clearly, the 1/r J factor makes H relatively much more senSitive to currents which L z is the angular momentum about an axis around which the molecule
close to the nucleus. In fact we can quickly convert our Cannulas for shielding to has roughly cylindrical symmetry. For benzene. this is the hexad axis. About this
Cannulas for average susceptibility X by recognizing that only the radial averages axis only a diamagnetic current results. This current, which flows all around lhe
differ. Thus the paramagnetic and diamagnetic currents contribute Xp and XD. benzene ring, produces chemical shifts al the proton positions. Of course. since
respectively. to the susceptibility X. where lhe diSlance of the protons from the ring is comparable to the radius of the ring.
the chemical shielding field is not accurately given by replacing the ring by a
X=XP+XD M=XHo (4.83) dipole.
We find (averaging over all orientations) On the other hand. when the currents j(r) are well localized in an atom or in
a single bond, the dimensions of which are small compared with the distance to
I h 2 q2 I I q2_ the nucleus under study, we can represent the effect of the currents by a magnetic
Xp =-- --- XD =-- --,'
3 m mc .12 6 mc2 dipole. When we average the result over the random orientations of a molecule
There are anisotropies that, for both Xp and XD' are a substantial fraction of the in a liquid. we find that the shift vanishes unless the atomic susceptibilily is itself
average value. If we compare Xp and XD, we have a function of the orientation of the magnetic field with respeci to the molecule.
One important contribution to atomic currents, which vanishes in :l liquid, is
h2 r 2 -;:2 .1
Xo = -XP~ 2.1 = -Xp 2"8 (4.84) the contribution of currents in closed shells on atoms other than thm containing
the nucleus under study. The result in the liquid can be shown to follow simply
where r is measured in angstroms and .1 in electron valls. In general we must because the current distribution in a closed shell is independenl of the orientation
expect (r 2)1/2 ...... tA.. and .1 . . . 8eV. Therefore it is clear that XD ':!' - Xp· of Ho with respect to the molecular axes. We should emphasize that an attempt to
We cannot decide which lenn is the larger without specific examination. NOIe compute the closed-shell contribution to shielding fields byapproximmc methods
in particular that the mere fact that jp is dominant in producing the chemical is dangerous, since one may find that the paramagnetic and diamagnetic COntri-
shift does fJ()1 mean that it will be the major factor in determining the atomic butions are large. Their algebraic sum (which is zero for an exact calculation
susceptibility. The susceptibility. since it depends on currents far from the origin. in a liquid) may be nonzero unless a very accurate computalion is made. It is
is much more strongly influenced by the diamagnetic currents than is the chemical therefore always safest 1) to judge the currenlS jo(r) on physical grounds, 2)
shirl because the diamagnetic currents are the more prominent at large distances always to choose gauges for each atomic current that puts ?-(p<,rl as proportional
from the origin. to the angular momentum about the most important atomic force center. and 3)
It is important to bear in mind that our particular choice of gauge has to exclude from the calculation any currents that will give an exactly zero resull.
made 1ipert dependent on the angular momentum about the origin. For an exaci Finally, we emphasize again the fact that the a priori judgment of whether
solution, this gauge is no beller than any other gauge. However, we rarely deal a distant atom has a paramagnetic moment or a diamagnetic moment is not
with exact solutions. There may Ihen be some physical preference in choosing a possible. but rather a detailed judgment of excitation energy and mean square
gauge that puts the perturbation in tenos of angular momentum measured about atomic radius is essential. Moreover. for example, a paramagnetic moment on
104 105
another atom can produce either diamagnetic or pammagnetic shielding fields,
depending on whether the anisotropic moment is largest when the inlcrnuclcar
axis is perpendicular or paraUelto the direction of Ho· (4.92)
All the expressions given involve one eleclIOn only. If there are N elec·
lIOns we generalize by adding subscripts "j = I to N" to the electron position
coordinate. Thus we define AOj by the equation
'Vj x AOj = Ho 'Vj x Anj = H .. (4.85)
Typically, (4.93)
x.,.' ~
Anj=~ AOj = !Ho x(rj-R) (4.86)
rj
'H..... rt = --q- "'(p .. AO' + Ao' .p.) (4.94)
where R is a convenient origin. We then define .. - 2mc ~ J J J J
)
"i J - -Ao'
"'.J = -'V. q J (4.87) Pj being (M)'V j'
C
To obtain eltplicit solutions, one must now assume reasonable N electron
so that the Hamiltonian, including both external magnetic field and the nuclear wave functions for !Po and wn . Ordinarily one will choose the functions to
field, is be products of one electron functions, or perhaps pair functions to represent a
covalent bond. Although j labels electrons, one can often rewrite (4.91) so Ihat
7i:-
2m j=1
1"(
E q
'7rj--A"j
c
)' +V (4.88) the sum over electron numbers is replaced by a sum over orbits, and in this way
one can distinguish closed shell electrons from valence electrons.
The formalism we have discussed is useful for obtaining a physical under-
If we define IJi to be the eltact solution of the N eleclIOn problem in the absence standing of chemical shifts. The final result can be expressed more compactly in
of the nuclear coupling, it obeys the equation a single formula such as has been given by Ramsey [4.4]. To do this, we express
( _1 ".' +
2m~ J
v) 'I' = E'I' . (4.89)
the magnetic field, using (4.91) as
(O'~AOk'Pk,n)(nlt'¥lo) (
nuclear coupling E pert is then
-/.1-';I ~ Jr.J
+
x Jo·(r·) (4.91)
= dTj
Epert
)
/.: J
J
1
J
En-Eo J
The wave function !P, which is needed to compute the currents, is then found by q'
- me' (01 Lrj X AOjIO) (4.96)
using perturbation theory. Defining the functions tfl o, tfln to be solutions of the
)
Hamiltonian 1tO in the absence of the external field, with eigenvalues Eo and
En' If we further assume that AOk = ~Ho x rk and thai No = HOk, we get
106 107
Suppose we average 1{ over an s-state electron wave function 11(1'), as we would
, q2 h2
(o[L:.f+!n)<nlL:L.kIOJ
j k
J
do to perfonn a first-order perturbation calculation of the coupling. There are a
number of lenns 10 (4.98), similar to the tenns A, B, C, D. E, and F when
H=HO- "L: E _~_ computing lhe rigid lanice line breadth (see p.66). Let us pick out a tenn A,
Zm e n \ nUU
+ (01 p.dn)(nl tft[O) I which depends on angle and distance as (I - 3 cos 2 U)/r 3. Then, apart from a
multiplicative constalll, the average of such a tenn will be
u2(r) 2 2
En - Eo
_LHO(O[,,[k(xj+yj) _ iXjzj _ 6]IO)
iY
J ~(l - 3 cos 8)r <ir dO (4.99)
108 109
We shall represent the nucleus by a charge q going in a circular path of It is convenient to express the coupling as a tenn in the Hamiltonian that will
radius a with velocity v. This. effectively, is a current loop of current (qlc)(l/T), give this interaction. This is readily done by means of the Dirac S-function:
where T is the period of the motion. We can express If z ~ the magnetic field in 8,
the z·direction, due to the nucleus, averaged over the electron orbital probability J:i = -3JLe . Pn6(r) (4.109)
density lu(r)1 2 - as
where r is now Ihe position of the electron relative to the nucleus. It is convenient
- J
Hz = Hz(r)lu(r)1 2dr (4.101) also to re-express (4.109) in temlS of the nuclear and electron spins I and S. For
the electron we shall use a gyromagnetic ratio IC' which is positive. but for the
where H z(r) is the field of the current loop. We shall lake z to be nannal to the nucleus. in is to have an algebraic significance, being either positive or negative.
loop. The olher components of H can be shown to vanish when averaged, since Then we have
lu(r)1 2 is spherically symmetric for an s-statc. If we draw a sphere of radius a
IJ. c = -"'IerIS IJ. n = "'In Ii I (4.110)
about the origin, we can express If :(J') by means of a scalar magnetic potential
either for r < a or r > a. It is straightforward to show that the contribution which give
from regions outside the sphere vanishes from the angular integrations. If we 8, 2
express the scalar potential inside the sphere as a Slllp of products of spherical ?1.=31c/r.r,/·S6(r) (4.111)
harmonics with radial functions, all contributions except the first term (the term
We notice that in (4.108), the radius of the nuclear orbit has dropped oul.
that corresponds to a unifonn field within r < a) vanish. We can evaluate this
Clearly, we should get the same answer for a volume distribution of circular
tenn simply, since it is the only one thm does not vanish at ,. = O. Therefore
currents. Moreover, since the smeared nuclear spin moment is equivalent to a
(4.101) may be rewritten as
volume distribution of current, we have also included the intrinsic spin of the
-Hz = J" Heu (r)dr 2 (4.102)
nucleons if we use, for "'In and /, Ihe experimental values. Equation (4.111) is
therefore quile general. We see also that if we may flot neglect the variations in
o u(,·) over the nucleus, the answer will be a bil different. Two isotopes that have
where He is the field at the center of the sphere. We may approximate (4.102) different current distribUlions will then have couplings which are /lot simply in
by recognizing that u(r) varies lillie over the nucleus as the ratio of the nuclear moments. This phenomenon is the source of the so-called
- 2 411" 3 hyperfine anomalies.
Hz = Heu (0)3a (4.103)
The treatment of the imeraction by the Dirac equation is somewhat more
The field at the center of t!le loop is simply involved. We shall sketch the important steps but leave the details to the reader.
The Dirac Hamiltonian for an eleclron (charge -e) is
H=!lrxv_!l'!!""k (4.104)
e c r3 - c a2 ?1. = -0:" '(cp+eA) - pmc2 + V (4.112)
But the magnetic moment IJ. n of the nucleus is i:rra 2, where i is Ihe "current", or where 0:" and p are 4 by 4 matrices, V is the electron potential energy, and
Qav A is the vector potential. We can express 0:" and p in lenns of the two by two
JL =!l..!..:rra 2k=k (4.105) Pauli matrices, (T. and the two by two identity matrix 1 as
neT 2c
Thus
H _ 2IJ.n
a= (~ ~) fi= (~ ~,) (4.113)
e-~ (4.106)
The wave functions 1Ji, which are solutions of (4.112), are represented by a
By substituting into (4.103). we find column matrix of four functions, but these are most conveniently expressed in
tenns of the functions !lit and IJiz, each of which is a column matrix wilh two
- 871" 2
kHz = 3IJ.nU (0) (4.107) elements:
The effective interaction energy E with an electron moment lJ. e is then (4.114)
8, 2
E=-3IJ.c 'lJ. nU (0) (4.108) The eigenvalues E of ?1. may be written as
110 111
E=E'+me 2 (4.115) The tenn (4.12Ia) can be rewritten as
is the energy measured above me2 , so that for a free panicle at rest,
where E'
E'=O. 2,. )' ,'r"
( 21' + 1'0 (2mc 2)2
u. [..':.
1'3
x ("" x
1.3
r)] (4.12Ib)
If we define
The tenn in the square brackets goes as Ih.'l. For 1':»1'0, (4.12Ib) can be
ll'=ep+eA (4.116) shown to be of order 1'0/1' times (4.120b); therefore, much smaller. However,
and define I/> as Vie, we have when T :5 1'0, the radial dependence becomes less strong, going over to the hann-
less 111'0 near I' = O. The tenn is t/lerefore well behaved. It also has the feature
(E + el/> + 2me2)!Pl + u - 1l'!P2 = 0
I
(4.117) that it does not average [Q zero over angle of s-states. It gives the answer of
(E' + el/»!P2 + u - 1l'!P 1 = 0 (4.108) for the magnetic interaction energy.
where I/> = ell' is the potential due to the nucleus. As is well known, PI is much We see that these two tenns are very similar to taking a S-function for s-
smaller than P2 in the nonrelativistic region. One customarily calls P2 the "large states and asserting that a finite size of the electron prevents the radial catastrophe
component". For s~states in hydrogen, P2 is also much larger than PI, even at of the conventional dipolar coupling of (4.98). For computational convenience
the nucleus. One can eliminate !PI (still with no approximations) to obtain a we may consider that the dipolar interaction of (4.98) should be multiplied by
Hamiltonian for rJi2, 'H' such that the function 21'/(21' + 1'0), to provide convergence.
We shall now turn to the study of some of the important manifestations
(4.118) of the coupling between nuclei and electron spins, considering first the effects
By tedious manipulation one finds that that are first order in the interaction and then effects that arise in second order.
Further discussion of first-order effects will be found in Chap. lIon electron
1
?-i'= EI I/> 2 2(e 2p 2+ e2A 2+2ecA_p_iecdivA spin resonance.
+e + me
enc
+ eflcu ° '\1 x A)+, '12
(E +el/>+2me) 4.7 Knight Shift'
x (ieEoA +ieE·p - eu' E x p- eu·E x A) (4.119)
lbe Knight shift is named after Professor Walter Knight, who first observed
where E is the electric field due to the nucleus. For our present needs, we focus
the phenomenon. What he found was that the resonance frequency of Cu 63 in
on two tenns only:
metallic copper occurred at a frequency 0.23 percent higher than in diamagnetic
1 CuCI, provided both resonances were perfonned at the same value of static field.
' "2 2encuo'\1xA and (4.120a)
E +e'f'+ me Since this fractional shift is an order of magnitude larger than the chemical shifts
1 2 among different diamagnetic compounds, it is reasonable to attribute it to an
(E/+e<b+2mc2)2e neu-Ex A (4.12Ia) effect in the metal. Further studies revealed that the phenomenon was common
to all metals, the principal experimental facts being four in number. By writing
For our problem, the nuclear coupling is given by introducing the vector W m for the resonance frequency in the metal, Wd for the resonance frequency in
potential A = JL n x r/r 3. Therefore \7 x A is simply the magnetic field of the a diamagnetic reference, all at a single value of static field, there is a frequency
nucleus computed by using the dipole approximation. As 10llg as el/>« 2mc 2, displacement .dw defined by
(4.120a) is exactly the same as (4.98) and goes as 1/1'3. If, however, T is so
small that el/>_2mc2 , the answer is modified. By writing el/> = e'lh', multiplying Will =Wd +.dw (4.122)
(4.120a) by T on top and bOllom, using e 2 /mc2 = TO (the classical electron The four facts are
radius), and neglecting E', we get for (4.120a):
1. .dw is positive (exceptional cases have been found, but we ignore them
( 21'+1'0
~)~Uo\7XA
2mc
(4. 120b) for the moment).
Now we no longer have an infinity from the radial integral in computing H:,
and it is clear that the angular average makes (4.120b) to be zero. 3 See references to UNuclear Magnelic Resonance in Melals" in the Bil>liography
112 113
2. If one varies Wd by choosing different values of static field, the fraclional necessary to specify the electron wave function, which is, of course, a formidable
shift .tJ.W/Wd is unaffected. task from a rigorous viewpoint, one thai has not in fact been c:lITied Ollt because
3. The fractional shift is very nearly independent of temperature. electrons couple to each other so strongly via the long range Coulomb interaction.
4. The fraclional shift increases in geneml with increasing nuclear charge Z. Therefore we are forced 10 an approximation. We shall consider the electrons as
The fact that metals have a weak spin paramagnetism suggests that the shift being noninteracting - or at least only weakly so. Bohm and Pines have shown
may simply represent Ihe pulling of the magnetic flux lines into the piece of Ihal this approximation has considerable theoretical justification. For a review
metal. However, the susceptibilities are too small (10- 6 cgs units/unit volume) of the Bohm-Pines theory, see [4.5]. By means of a canonical transfonnation,
to account for an effect of this size. As we shall see, however, the ordinary they show that the principal effect of the Coulomb interaction is to give rise to a
computation of internal fields in a solid that involves a spalial average of the set of collective modes of oscillation, the plasma modes. The basic frequency of
local field is not what is wanted, since the nuclear moment occupies a very excitation of the plasma is so high that we may ordinarily consider the system
special place in the lattice - in fact a place at which the electron spends, so to be in the ground plasma state. There still remain individual particle motions.
to speak, a large amount of time in response to the deep, attractive potential The residual interaction between particles, however, is very weak and falls off
of the nuclear charge. As we shall see, the correct explanation of the Knight nearly exponentially with distance. For low energy processes that do nOI excite
shift involves considering the field the nucleus experiences as a result of the the plasma modes, we may therefore treat the electrons as weakly interacting.
interaction with conduction electrons through the s-state hyperfine coupling. If We shall describe the system, therefore, wilh a Hamiltonian:
we think of the electrons in a metal as jumping rapidly from a10rn to atom, we (4.123)
see that a given nucleus experiences a m<lgnetic coupling with many electrons.
lllCrefore the coupling to the electron spins must be averaged over the electron where He describes a group of weakly interacting eleclrons, H n is the nuclear
spin orientations of many electrons. In the absence of an external field, there is Hamiltonian and includes the Zeeman energy of Ihe nuclei in the st:ltic field
no preferential orientation for the electron spins, and thus there is zero average Ho as well as the magnetic dipolar coupling among the nuclei, and where 'Hen
magnetic coupling to the nuclei. On the other hand, the application of a static is the magnetic interaction between the nuclei and the electron spins. We omit
field, Ho, polarizes the electron spins, giving a nonvanishing coupling. Since the coupling of the nuclei to the electron orbilal motion because it gives effecls
the s-state interaction corresponds to the nucleus experiencing a magnetic field comparable to the chemical shifts. (Of course Ihe electrons in the metal are free
parallel to the electron magnetic moment,4 and since the electron moment is so that the orbital effect is a bit different from that in an insulator.) It can be show~
preferentially parallel to Ho, the effective field at the nucleus will be increased. that the conventional dipolar coupling between nuclear and electron spins, (4.98),
Since the shift in frequency is proportional to the degree of electron spin polar- contributes nothing in a cubic metaL For noncubic metals it gives rise to Knight
ization, it will also be proportional to Ho or Wd. Moreover, since the electron shifts which depend on the orientation of Ho with respect to the crystalline axes.
polarization is temperature independent (the spin paramagnetism of a highly de- Since re.sonance in metals is usually perfonned on powders (to pennit adequate
generate electron gas is independent of lemperature), the shift will be temperature penetratIOn of the alternating field into the material, a problem that we may
independent. And lastly, the Z-dependence will follow, since the wave function call the. "skin depth" problem), the anisotropy manifests itself through a line
is larger at the position of a higher-Z nucleus, as is well known from the study broadening. In the interest of simplicity, we shall confine our attention to the
of free alOm hyperfine splittings. We can see from these considerations that the o-function coupling:
hyperfine coupling possesses the properties needed to explain the major facts. 8. 2",
Let us now look into the details. Hen'" 3"Ye"Ynh LJlj' S{o(r/- Rj) (4.124)
j,l
We consider a system of nuclear moments and of electrons coupled together
by the hyperfine interaction. The relative weakness of the hyperfine coupling where r/ is the radius vector 10 the position of the lth electron, and R· that to
enables us to treat it by a perturbation theory in tenns of Ihe states of the the position of the jth nucleus. J
electrons and the nuclear spins. We shall actually be able to avoid specifying . Equation (4.124) can be rewriuen in a highly useful manner. When we
the nuclear states, since we shall show that the effect of the interaction is simply diSCUSS electric charges in quantum mechanics we introduce a charge density
to add an effective magnetic field parallel to the applied field. However, it is operator e(op)(r) given by
114
115
In a similar way we introdoce 3 spin magnetization density operator of Ftg..4.6. Meh.t repr~lted by a box, with a
electron magnetization V, polential deplh Vo
' - -_ _--',_L.
L -1.1>5,o(r,- r)
WOp)(r) =
, (4.126)
The expectation value of this operator is the classical spin magnetization density. being x, the wave functions would be sin kx or cos kx, whcre only those values
Using this definition, we get of k are allowed that satisfy the proper boundary conditions aI x = 0 and x = a.
In order to describe a situation in which a current can flow, it is customary to
14" = - L 8; WOP)(R;H1"hlj ) (4.121) consider instead solutions exp(ikz), where now the allowed values of k are those
1 that make the wave function the same at x "" a as it is at % "" O. The periodic
boundary conditions for a three-dimensional box give solutions of the fonn
Since we treat the nuclei and eleclTOns as only weakly interacting, we may
write the complete wave function tP as a product of the (many panicle) wave !/J = e ik ' r (4.132)
functions tPc and VJn of the electrons and nuclei:
These solutions are modified in a very simplc way to tuke account of the
(4.128) fact that the real potential is very deep in the vicinity of the nuclei. The wave
(Of course this wave function would be exact if ?ten were zero.) We shall then functions, called BlochfiUlctions, are then of the fonn
perfoon a perturbation calculation of the energy E ell : !/Jk = uk(1')eik' r (4.133)
E en = J l/J*?tcn!/JdTedTn (4.129) That is, there is still a quantity k, the allowed values of which are given by
requiring periodicity of the wave function on the walls of a box, but the plane
where dTc and dTn indicate integration over electron and nuclear coordinates wave exp (ik· 1') is multiplied by a modulating function uk(r), which is a func-
(spatial and spin). Of course we shall wish to see the effect of (4.129) on lransi· tion possessing the lattice periodicity. A typical uk(r) peaks up strongly near a
tions of the nuclear system from one nuclear state!/JI! to another tPll'. Since the nucleus. The fact that we explicitly label u with a subscripi k points out that U
transitions are in the nuclear system. they leave the electron state tPe unchanged. will in general vary with k.
In computing the energy of the nuclear transition £en - E en,. we have to com- We shall need 10 add a spin coordinate as well, giving us finally a function
pute both E en and Een ,. BOIh energies involve the integral over the electron
./.
coordinates. It is convenienl for us to postpone a specification of the nuclear "'k. = uke ik· r.1.
"'.
(4.134)
states and to compute. therefore. the electronic integral where tP. is a spin function. The wave function for the N electrons, tPe, will
h~n = J!/J;1icn!/Jedrc (4.130)
then be a product of tPk. 's, properly anti symmetrized to take account of the Pauli
exclusion principle. We can do this readily in terms of Ihe permutation opcrator
which is simply the first step in computing (4. I29) on the assumption of a product P [4.61:
function. t/J = !/Je!/J1l or!/J' = tPe!/Jn'. We denote (4.130) by h~n to emphasize Ihal
the nuclear coordinates still appear as operators. ,p. = ~ L(-i)P p,p..(I),p.,.,(2),p", .,,(3) ... ,p••••(N) (4.135)
vN! p
We can re-express (4.130) using (4.127) to obtain
, ,,8.
hen = - ~ 3" M(R}). 'Yn hI} (4.131)
where the symbol (-I)P means to take a plus or minus sign, depending on
whether or not the pennutation involves an even or an odd number of inter-
1 changes. The factor l/.,fiiff is, of course, simply noonaliz:ltion.
where now M(R}) is the classical spin magnetization density at the posilion of Let us compute h'en)'. the contribution to (4.130) of the jth nuclear spin,
the jth nucleus. and choose the origin of coordinates at that nuclear site (Rj = 0). Then we have
The function !/Jc will itself be a simple product of one electron functions
if we assume that the eleClTOns do not interact among themselves - or at least Ken} = 8; 7e'Ynh2 Ij' J¢; L
I
S,6(1")¢edTe (4.136)
interact weakly. For the individual electrons we shall take the so-called Bloch
functions. We remind ourselves of what these are: 1f the eleclTOns were thought of Since the operator S I 6(1") involves only one electron, there are not contributions
as moving in a one·dimensional box of length a (Fig. 4.6), the position coordinate to (4.136) from teons in which electrons are exchanged, and therefore we get
116 117
r
translational energy of the electron). For example, Espin is the Zeeman energy of
the electron spin in the static field Ho. There may be other contributions to Espin,
(4.137) however, from the electrostatic coupling between electrons, which depends on
their'"'Telative spin orientation. If we consider a typical tenn in the sum of (4.139)
We now assume the electrons are quantized along the z-dircction by the external corresponding to a single value of k, there are two values of m s , giving us
stalic field Ho. Then the only contribution to (4.137) comes from 5 zl ' (We could
alternatively have kept juS! that pari proportional 10 [zj, assuming the nuclear (4.141)
spins to be quantized along HQ. the result being the same.) We can write the
results of (4.137) as As we can see, the quantity in the brackets is (apart from a minus sign, since
IJ- e = ~l'enS) the average contribution of state k to the z-component of elec-
871" -2 '"" 2
31'cln h Izj L-!Uk,s(O)1 1»sP(k, s) (4.138) tron magnetization of the sample. We shaH denote it by I'z k. The towl z-
k,8 magnetization of the electrons, /-lz, is then '
where the sum is over all values of k. s, and where p(k,s) is a factor that is
1 if k, s are occupied by an electron, zero otherwise. The factor 1»8 is the m
~=L/-lzk . (4.142)
k
yulue of the state "l/Jk,s; hence it is +~ or -~, and of course llk,s(O) is the wave
If we define the total spin susceptibility of the electrons, x~, by
function evaluated at the position of nucleus j.
If we average Ihis expression over a ScI of occupations P(k. s), which are Ilz = X~Ho (4.143)
representative of the temperature of the electrons, we can write for the effective
interaction with jth nucleus, and define a quantity xi: by
-/-lzk=Xk
'H0 (4.144)
81r 2" 2
31'c1'nf, Izj L.!Uk,s(O)! msf(k,s) (4.139)
k,s then (4.142) is equivalent to
where f(k, s) is the Fenni function. For the electrons at absolute zero, f(k, s) is X~ = Lxi: (4.145)
I for all k, 5, which makes the total (that is, spin plus spatial) electron energy k
less than the Fenni energy EF and is zero for energies greater than E F • At We can therefore write (4.141) as
temperatures above absolute zero, f(k,s) is modified within about kT of EF
(Fig. 4.7). Of course f(E) is - 8 1r l'n nIzj !u,.:eO) 12xi: H o (4.146)
3
f(E) = 1 (4.140) so that the total effective interaction for spin j is
1 + e<p [(E - E,,)/kT]
The notation f(k, s) means f(E), of course, where E is the energy of an electron
-31'n L. 1
8, H zj '" Uk(O) I" XkHO (4.147)
k
with wave vector k and spin coordinate s. Typically, E = Ek+Espin, where Espin
is the energy associated with the spin orientation and Ek is the sum of the kinetic Our problem now is to evaluate the summation. It would be simple to make the
and potential energies of an electron of wave vector k. (We shall call Ek the assumption of completely free electrons, but it is actually no harder to treat the
case of a material with a more complicated band structure. We shall do the laller,
since the resulting expression will enable us to make some important comparisons
f(E)f---_~ (aj f(E) '---=<-]l (b)
between experiment and theory.
We assume, therefore, that the energy of the eleclrons, apart from spin
effects, is detennined by its k vector. For free electrons, the dependence would
be
i\ (4.148)
£'1,. E E j ,' E
Fig. 4.7. (a) Fermi rundion feE) at absolute zero. (b) Fermi rundion at a temperature T so that in k-space, all electrons on a sphere of given k would have the same
above absolute zero
energy. That is, the points of constant energy fonn a surface, in this case a sphere
118
119
" " for any slates k having the same value of translational energy E1.:' Even when
we allow the electrostatic coupling between the electrons to affect the energy to
/
" tum over a spin, it may be reasonable to assume Ihm this modification depends
at most on Ek- We therefore assume that xl
is a function only of the energy
(0) (b)
"- "/
We Clln therefore rewrite (4.150) as
Fig.4.8a,b. Intersection of two surfaces of constant energy wilh the k. = 0 plane. (:1) L IllkCO)!2 Xt = JIUk(O)12Xs(E,,)g(Ek,A)dAdEk (4.152)
Circular section of a free elCdron. (b) A less syrnmetric seetion for a hypothetical "real"
substance •
If we have any funclion F of E, we define ils average value over the surface of
(Fig. 4.8). In general the effecl of the lattice potential is to distort the surfaces constant translational energy Ek' (F(k)}e • as -,
from spheres. We shall assume that lhe function IUk(O)1
2
xt
varies slowly as one
JJ F(k)9(E•• A)dA 1
moves the point k in k-space from one allowed k-value to the next, so that we
can define a density function to describe the number of allowed k-values in any
(F(k» E, =
g(E•• A)dA
= -(E)
9.
J
F(k)y(E•• Jl)dA (4.153)
region. Let us define 9(EI.:' A)dEkdA as the number of allowed k-values lying
within a certain region of k-space, defined as follows: It is a small cylinder lying ll1en we can set the integral ovcr dA in (4.152) equal to
between the energy surfaces Ek and Ek + dEk (Fig. 4.9). Its surface area on the
top or bottom surface is an element of dA of the constant energy surface. We J 1".(O)I'y(E•• A)dA = 9(E.)(I".(0)I') E, (4.154)
denote the particular coordinates on the surface also by the symbol A in 9(E, A).
giving
The total number of states dN between Ek and Ek + dEk is found by summing
the contributions over the entire surface: L 1".(O)I'xk = J(1".(0)1') E,X'(E.)g(E.)dE. (4.155)
(4.149) •
Now XS(E k ) is zero for all values of E k that are not rather ncar to the Fenni
energy, since for small values of E k the two spin stales are 100 percent populated,
We can use these functions to evaluate the summation by replacing it with an whereas for large Ek' neither spin state is occupied. XS(EIJ must look much
integral: like Fig. 4.10. XS(Ek) will be nonzero for a region of width about kT around the
L 1",(0)1' xl, = J1".(0)I'xk9(E•• A)dE.dA (4.150)
Fenni energy EF.
We are therefore justified in assuming (Itt~:<O)12},.; varies sufficiently slowly
•
Now xl: depends on the Fermi functions f(k,~) and f(k,-~) :ll1d thus on the
to be evaluated at the Fenni energy and taken outside ~f the integral,
energy E1.:, and on the difference in energy of a spin in Slate k with spin up versus L ",,(O)I'xk = (1".(0)1') E,· JX'(E.)g(E,)dE. (4.156)
that with spin down ('YeftHo for free electrons). Therefore xl
would be the same •
The integral remaining in (4.156) is readily evaluated in teons of (4.145), since
.,
120
Fig.4.9. Volume in k-space associated with dEdA
.'CE,JIL
A
L+.''---_--:O-
EJ! E k
Fig.4.10
Function XS(E",) versus E",
121
which, performing the integral over A, becomes PF. We shall use PA to denote the wave function density at the nucleus for
x::::: JXS(Ek)e{E,,:>dEk (4.157b)
a free atom. It is then convenient to discuss the ratio PF/PA for Li and Na.
A comparison of RYler's values, theoretical values, and values deduced from
Therefore, using (4.147), (4.156), and (4.157), we may say that the interaction combining SchwlIacher's measurement of X~ with measured Knight shifts is
with jth nuclear spin is given in Table 4.1. There is excellent agreement among all three values. For
Na, Kolm and Kjefdaas [4.9,10] find PF/PA = O.80±O.03. We can combine
8. ,.] (4.158) this value with the measured Knight shift to obtain a quasi-experimental value
-'Yn lllzj [ 3(1l1k(0)1 ) EpXeHo
of ~. Before presenting these results, we shall describe Schumacher's direct
This is entirely equivalent to the interaction with an extra magnetic field aH, measurement of X:.
which aids the applied field Ho, and is given in magnitude by the equation
Table 4.1
aH 8'11" 2 5
Ho = T(I"k(O)1 )E,X, (4.159) l\,/PA in Li
We see that this formula has all the correct propenies to explain the exper· Kohn and KjetdafU lheordical 0.'19 ± 0.05
imental results. Experimenlal ex:
ptus Knigh~ shift.) 0.45 ±0.03
Ryter (experimenlal) 0.442±0.015
I) It predicts that a higher frequency is needed for the metal than in the
diamagnetic reference.
2) The fractional shift is independent of w.
The fundamental problem in measuring x:
is how to distinguish it from
the other contributions to the total susceptibility, which (in a melal) are quite
3) Since both (luk(0)12) Ep and r
are independent of temperature, all/Ho
comparable in size. The method employed by SCluwlOcher is to isolate the spin
is as well. Since the larger·Z atoms will have a larger value of (llIk(0)1 2) Ep contribution by use of magnetic resonance. From the Kramers·Kronig relations
corresponding to the pulling in of their wave function by the larger nuclear we have for the electron spin susceptibility .t::
charge, the in<:rease of Knight shift with Z is explained.
- 2 J-
=
The Knight shift fonnula can be checked if one can measure independently
x:-- X~dw (4.162)
aH/Ho, (Itlk(O)12)Ep and.t:. There is only one case for which all three quan· '. w
tilies are known (Li metal). The spin susceptibility has been measured by Schu- o
macher J4.7] by a method we will describe shonly. Ryler [4.8} has measured where X~ is the imaginary pan of the conduction electron spin susceptibility.
(l"k(O)1 ) E by measuring the shift of the electron resonance by the nuclear For a sufficiently narrow resonance, we may neglect the variation in w across the
moments. This shift,LlHe, is given by absorption line, taking it out of the integrnl. We can then change the integration
from one over frequency to one over field, using w:::: "(H:
IJ.H, 8. (I Uk (0)1') EpXn
Ho :::: 3 • (4 . 160)
=
J "dw 1 'J= "IH (4.163)
X• :::: -2 - I Xe = -2 - Xe (
where X~ is the nuclear susceptibility of the U 7 nuclei. Denoling the number of e lI" W o 'll"WO
nuclei per unit volume by N, we have o 0
An absolute measurement of the area under the resonance curve will therefore
s N"(~1t2 I(I + I) enable one to determine X~. (Actually, the approximation of a narrow resonance
Xn :::: 3kT (4.161)
is not well fulfilled in Schumacher's case; however, the final formula can be
Thus, since X~ is known, mcasurement of LlHe gives (11l/.:(O)1 2 )L.. In order to shown still to be correct. For a discussion of these points the reader is referred to
enhance the size of the shift, RYler polarized the nuclei, using the so·callcd Over- Schumacher's paper [4.7]. We may also note that the resonant absorption of en-
hauser effect (see Chap.7). It is necessary then to modify the formulas slighdy, ergy at the cyclotron frequency occurs degenerate with that of the spin. However,
but the principle remains the same. the rapid electron collisions broaden it so much as to render it unobservable. One
may be confident thai it is only X~ that is being measured.)
For comparison with experiment it is convenient to compute (lu/.:(O)1 2) E. Absolute measurements of absorption are always very difficult. Schumacher
using wave functions normalized to the atomic volume, which we shall cafl circumvented them by making use of the nuclear resonance of {he Li7 or Na 2J
122 123
nuclei in the same sample for which he measured the conduction electron reso- Ho = O. as with ferromagnets or anliferromagnets. lei us discuss the case of
nance. For the nuclei, onc has a spin susceptibility X~I' given by (4.161) lind for ferromagnetism briefly.~
which Once again the electron-nuclear interaction will consist of the sum of the
00
convellfional dipolar coupling 1td and the 8-state interaction 1ts :
X
s=~
n
"Ynj X"dH
1T WO n (4.164)
o (4.166)
By choosing wo/2-;r So' 10 Me/s, Sc/llullacher could observe either the electron or We shall average this over the electron wave function .,pe, as with the Knight
nuclear resonances simply by changing HOI the remainder of the apparatus being shift, to get an effeclive nuclear Hamiltonian 1t~n' which will contain the nuclear
left unchanged. The nuclear resonance occurred at abollt 10,0000au55, whereas spins as operators:
the electron resonance was al only a few Gauss. Then, if we denmc the area
under the electron or nuclear resonances by A c or An, respectively, we have 1t~n = j .,p~(1td + 1ts).,pedre (4.167)
G
X~ = 'Ye A c (4.165) where dre stands for integration over all electron coordinates (spin and spatial),
X~l 1'n An
and where the symbol G signifies that the spatial integration goes over the entire
Since X~I can be computed, we are thus able to determine X~. Note that we can physical volume of the sample. By breaking G into the atomic cells G I, G2, ...
measure the "area" in any units we wish (such as square centimeters on the face GN of the N atoms of the crystal, we may fonnal1y interpret the contribution of
of an oscilloscope) as long as they are the same for bOlh resonances. We do not 1td to (4.167) as arising from the summation of the dipolar fields of the electrons "
even need to know how much sample we have, since it is lhe same for both on the various atoms. The wave function .,pe will provide a detailed picture of the
electrons and nuclei. spatial distribution of the electron magnetization in each atomic cell. Evaluation
The experimental values obtained are listed in Table 4.2 IOgether with vari- of this tenn is identical to computing the local field due to a volume distribution
ous theoretical values. The first column of experimental and theoretical numbers of electron magnetization. Suppose the magnetization is unifonn throughout the
[4.5] gives theoretical values based on non interacting electrons, blll an effective sample. If the lattice has cubic symmetry, 1td will contribute an effective field
mass has been inttoduced to take account of the lattice potential. The effective given by the Lorentz local field:
masses are computed by Harvey Brooks, using the quantum defect method. The 4, _
second column shows the theoretical values obtained by Sampson and Seitz, who -M - a ·M (4.168)
3
took into account the electron-electron coupling by means of an interpolation for-
mula of Wigner. The next column shows theoretical values due to Pines, based
upon the Bohm-Pines collective description. Next we give values obtained by
-
where M is the magnetic dipole moment per unit volume and 0' is a demag-
netizing factor (in general a tensor) that expresses the effect of the "magnetic
using the Knight shift and the Kohn-Kje/daas theoretical values of PFIPA . The poles" on Ihe outer surface of the sample. For example ;; = 471i3(ii + jj + kk)
last column shows Schumacher's results. (See [4.11] for subsequent resulls.) for a sphere.
The s-state lenn may be interpreted as contributing a magnetic field H&j at
the jth nucleus.
Tabte 4.1 x: (all values are 106 cgs volume unit..s)
Free Sampson Knight shift. and
8<, e h L
H&j = - 3 ."j'
WeSjo(r/- Rj)1/JedTe (4.169)
eledrons and Seitz Bohm-Pines theoretical Pp/PA Schumacher I
124 125
where "occupied" means that we include in the sum only those states 1m con- zation. Therefore the inner electrons must give a field of about 400,OOOGauss
taining an electron and we have omiued lhe subscript 1 from Sand r. By using opposed to the local magnetization [4.14].
the fact thai Ile = -"YenS, we have This phenomenon, called core polarization, was actually already known
from e.lectron magnetic rcsonan.::e of paramagnetic ions for which the 48 electrons
8.
H,j = 3" '"
L (PI".'(r - Rj)IP) (4.171) are missing. In principle the 3d electrons are incapable of giving an isotropic
lJoccupicd hyperfine coupling, since d-states vanish at the nucleus. However, the d-electrons
Since the matrix element involves coordinates of only one electron, the various are coupled to inner shell electrons electrostatically, the coupling for an inner
values of I now appear as Ihe values of p Ihal 3rc occupied. In a substance such as electron of spin parallel to the d-electron spin being different from that of an
iron, we may think of some values of {3 as corresponding to closed shells, some electron whose spin is opposed 10 that of the d-electron. Consequently the spatial
to the 3d band, and some to the 48 band. We shall discuss these contriblllions part of two wave functions such as the 38 are different for the two spin states.
shortly. The spin magnetization of the two electrons does not add to zero at all points of
The contribution of the tenn 7-{d is somewhat different for a ferromag- the electron cloud. We can see from (4.171) that if the 3s electron densities at
net than for a paramagnet. For the latter, the magnetization is tlllifonn in both the nucleus differ, there will be a nonzero contribution from the 3s electrons to
magnitude and direction for ellipsoidal samples, nnd the simple dcmugllclizing Hsj, even though their spins are opposed.
arguments follow. For a ferromagnet the magnetization within a domain is uni-
form, but the various domains have differing magnetization vectors. Thus, for a
soft ferromagnet in zero applied field, the magnetization averaged over a volume 4.8 Single Crystal Spectra "
large compared with the domain size is zero. The density of magnetic poles on the
outer surface therefore vanishes. Within the body of the ferromagnet, div M "" 0 We saw in Sect. 4.5 that the chemical shift for a p-state such as xf(1") depended
even at domain boundaries. If, then, we calculate the dipolar contribution to the on the orientation of the static field lIo, with respect to the bond axes. If Ho
magnetic field al a nucleus, we may proceed as follows. was either parallel or perpendicular to the bond direction, an extra field was
We draw a small sphere about the nucleus, of radius small enough 10 lie induced, acting on the nucleus, which was parallel to Ho. For an arbitrary ori-
within one domain. We compute the field due to magnetization on atoms within entation, one can resolve Ho into X-, y-, and z-components. Ho z , Hoy, Ho z ,
the sphere by a direct sum. The atoms outside the sphere are treated in the in general gClling an interaction of extcrnal field with nuclcar spin involving
continuum approximalion. For cubic symmetry, the atoms within the sphere give bilinear products such as Hozly etx. Blocmbergen and Row/alld [4.15] showed
zero tOial contribution. The atoms outside the sphere contribute as a result of theoretically and experimentally that the same thing is (nle for a Knight shift
the surface pole density on the inner sphere and the Ollter sample surface. The if one includes the conventional dipolar coupling (4.98) of the nucleus with the
former is the contribution 411" MI3, where M is the magnetization within the spins of the conduction electrons in addition to the Femli contact tcrm. Although
domain containing the nucleus. The latter contributes - ;; . M ' , where l\t! is the conventional dipolar tenn does not give rise to a net shift of the resonance
the magnetization averaged over a volume large compared with ;\ domain size. if one averages over all orientations of Ho with respect to the crystal axes, as
TIle total field seen by the jth nucleus HTj is therefore given by for a powder sample or as in a liquid metal, it nevertheless gives an orientation
4. _ dependence to the Knight shift.
HTj""Ho+ M- 0' 'M+Hsj (4.172) To deal with these anisotropies, one expresses the bilinear form with cou-
3 pling coefficients which are components of a tensor interaction. Then, writing
where lIo is an externally applied field. Although lIo and M ' vanish in zero ap- the Knight shift J( and the chemical shift CT as tensors using a dyadic notation
plied field, HTj does nol. Therefore we have a "zero field" resonance. Such a res-
onance was first observed by Gos.tard and Portis [4.12] in the face-centered cubic
fonn of cobalt. Using the Co59 resonance, the measured Hsj = 213,400Gauss.
where
- CT = iCTui + iuzyj + iCTzzk + jCTyzi + jCTyyj + jCTy:k
+ kCT;zzi + kCTzyj + kCTzzk (4.173)
In iron, Hsj is 330,000 Gauss. Hsj has also been observed by means of the
Mossbauer effect. It was discovered there that application of a static field H o and similarly for J( , we get a Hamiltonian
lowered the resonance frequency, showing the fIsj points opposed to the mag-
netization M.
The contribution from the 3d and 48 shells in iron is expected by Marshall
-
'H = -'YliHo'( I - ';; + J( )·1 (4.174)
[4.13] to give field of 100,000 (Q 200,OOOGauss parallel to the local magneti- where I is the identity dyadic
126 127
+(1-uzz+Kzz) az
2 2 ]1/2 (4.184)
=ii+ 11 + kk (4.175)
Explicit evaluation of u er /3 and I<':X/3 (a = x, y, z; fJ '" x, y, z) shows that Since the shift parameters ux X, Kxx, etc. are in general very small compared
both tensors are symmetric to I ,_and since 0').- +O'} +a~ = I, weean rewrite (4.183) to get
etc. (4.176) W = 'YHo[l + (Kxx - uxx)ak + (l{yy - Oyy )a~ + (Kz z - uzz)a~]
so that one can find principal axes. Denoting these axes by X, Y, Z with unit = wo +waak +wba~ +wca~ (4.185)
vectors i p , 1p ' and k p , we have The approximation that the shifts are small could in fact be introduced in
the original Hamiltonian. 1l1en we write 11. as a large part 1lo plus a penurbation
'; =ipuxxip+jpoyyjp+kpuzzkp (4.177)
~ -
and similarly for K. In general, K and '; do not have to have the same
'lip
H = Ho +Hp (4.186)
principal axes, but we will assume for simplicity that they do. Then with Ho = kHo, where k is a unit vector along the laboratory z-direction,
Then if Ho lies along a principal ax.is the resonance frequency is we get
126 129
1i = -')'llHoP + (J(XX - <Txx) sin 2 8 cos 2 q, Often in working with solids one has a sample consisling of many small
+ (Kyy - <Tyy) sin 2 8 sin 2 ? crystallites in the form of a powder. Then the orientation of the crystallites is
random with respect to Ho. The anisotropic nature of the chemical and Knight
+ (l( ZZ - <TZZ) cos 2 9]1, (4.194)
shifts then gives rise to a line broadening and to specrra which have quite distinct
This resuh can be expressed in another manner which is useful when dis- sha~ depending on the relative values of woo Wh. and We' Such absorption
cussing the techniques of magic angle spinning which we take up in Chapter 8. spectra are called "powder patterns" and are discussed in Appendix I.
Utilizing the trigonometric idenlities
cos 2 q, = !(I + cos 21$) 4.9 Second-Order Spin Effects - Indirect Nuclear Coupling
(4.195)
2
sin 1$ = 4(1 - cos 21$) ,
We have discussed the role of electron spin coupling to nuclei in paramagnetic
or ferromagnetic materials. Since, in a diamagnetic substance, the tOlal spin of
the electrons vanishes, Ihe nuclei experience zero coupling to the electron spins
in first order. Effects are found if one considers the coupling in second oroer,e
however. The coupling is manifested through an apparenl coupling of nuclei
(4.196) among themselves. the so.-called indirect coupling.
The indirect coupling was discovered independently by Nahn and Maxwell
Defining
(4.16] and by GUlOWSIey el al. {4.17]. The phenomena they observed are illustfUted
(j = !(<TXX + O'yy + <TZZ) by the case of PF3 , a molecule in which all nuclei have spin 4.
In liquid PFJ , lhe
2<TZZ - <Txx - <Tyy (4.197) rapid tumbling narrows the line. It is found that both the pJI and F19 resonances
<TLO = 3 = (<TZZ - iT) consist of several lines, as illustrated in Fig. 4.11. Since all the fluorine nuclei
(LO for "longitudinal") are chemically equivalenl, the splittings cannot be due to chemical shifts. (Fur-
themlOre there is only one phosphorus atom per molecule but four phosphorus
<TTR =: <Txx - <Tyy = (<TXX - (j) - (<Tyy - (j) frequencies.) The fact that Ihe individual lines themselves lire narrow shows that
(TR for "transverse") and correspondingly for the components of J(, we get the motion is sufficiently rapid to narrow the direct dipolar coupling. Moreover,
the splittings are found to be independent both of temperature and static field.
- - 0 - 1) The number and relative intensity of lines are as though each nuclear species
11. = - '111Ho [ I + (/( iT) + (l( La - <TLO) (300,22
experienced a magnetic field proportional to the z-compooent of the tOlal spin of
the other species. It was found that OWF = owp (see Fig.4.11), where OWf.'
+ ( I(TR ; <TTR) sin2 8 cos 2?] I, (4.198) and owp are the f-requency separalions of adjacent lines in the phosphorus and
fluorine spectra, respeclively. These facts indicated the coupling was somehow
The two angular functions are linear combinations of the spherical hannonics related to th~ nuclear magnelic momenls.
Y2.no Therefore, they average to zero over a sphere. Since in a liquid there is The original explanation proposed was that one nucleus induced curreniS in
generally rapid tumbling motion, such a spherical average is appropriate, giving the electron cloud, which then coupled to Ihe other nucleus. In a simple picture.
W =wO+!(Wo+Wb+WC> the induced currents are represented by an induced electron magnetic moment.
'( )+ (2wo-w.-w,)(3COS 22
0-1) ' (4.200)
Fig. 4.11. (a) The p31 resonance in 1'10'3. The lines are equlIlly SPllCOO all amount &..,., and
the intensilies are 1: J: J: 1. (b) The FU n'SOmmce in PF3
w=wo+
3 w o+ Wb+ Wc 3
& See rererfflccs to -Il"h Couplins" in the Bibliography.
130 131
If this moment were isotropic (as one changes the orientation of the molecule Atom A AtomB
with respect 10 the nuclear moment), the coupling to a second nucleus would
average to zero in a liquid, owing to the rapid random tumbling of the molecules. I I Stat,r
However, as we observed in connection with the chemical shifts, the induced --------------
moment is in general not isotropic. We can estimate the size of the coupling
between the two nuclei from second-order perturbation theory. The first nucleus
.I t Sta'" II
exerts a magnelic field I' l h(l/r 3 ) where (1/1'3) is the average of the inverse Fig. 4.12. Bc<:ause of the bonds betwc<:n atoms A and B, the electron wave function is
formed by an equal mixture of the state I, in which the electron Elloment on atom A
cube of the distance between the electron and the first nucleus that partially points up (that 011 B, down), and state II, in which the spin orientation is reversed
unquenches the orbital angular momentum, producing a fractional admixture of
excited state of Il/e1i2(I/r3)/LJE, where LJE is the energy to the excited state.
A complete unquenching would produce a magnetic field at the second nucleus In the absence of a nuclear moment the electron bond will consist of an
of leMR3, where R is the distance between the nuclei (we treat the electron equal mixture of the states I and II, shown in Fig.4.12. If now we put a nucleus
orbital magnetization as equivalent to a magnetic dipole). ll1erefore the order of on atom A with its magnetic moment pointing up, state I will be slightly favored
magnitude of the nuclear-nuclear interaction energy EI2 on this model is over state II. The electronic spin magnetic moment of atom A will have a slight
polarization up; that on atom B, down. Therefore a nucleus on atom B will find
I'IIe1i2(i"1r3) leh a nonzero field owing to its own electron. Since this field would reverse if the
El2 ~ LJE RJ 12 h (4.201)
nucleus on atom A were reversed, an effective nuclear-nuclear coupling results.
This formula fails by an order of magnitude or more in accounting for the facts. We can easily estimate the size of the coupling. The fractional excess of state I
However, it has the virtue of making it seem reasonable that the splittings in PF J over state II will be
were an order of magnitude larger than those in PH3, since clearly this mechanism ¥llIeh2JU(O)I~ = hyperflne energy
is closely related to chemical shifts that are always smaller for hydrogen than (4.203)
LJE electrostatic energy
for fluorine.
As was pointed out by Hahn and Maxwell, and GUlowsky and McCall, any where lu(O)I~ is the wave-function density of the electron lit atom A, and LJE is
mechanism such as we have described, and which will lead to a result that is the energy to an appropriate excited state. The coupling of the electrons on atom
bilinear in the two nuclear moments, must take a very simple fonn. Since the B to nucleus 2 are thus given by the proouct of the electron spin coupling if the
interaction is averaged over all molecular orientations, it can depend only on the electron spin is in one orientation only, (81r/3ht1'eIi2Iu(O)11, times the excess
relative orientation of the nuclei; hence it must be of the form fraction of the time the electron is in the favored orientation. Thus the coupling
is
(4.202)
where Al2 is independent of temperature and field. These workers also pointed (~llle1t2Iu(0)1~) (¥llle1t2Iu(O)11)
(4.204)
out that this panicular form would also explain the puzzling fact that, for example, c,E
there were apparently no splittings of f1uorines by fluorines in PF3 . We shall not This coupling turns out to have the correct order of magnitude. If the electron
give the proof here but physically the explanation is based on the idea that the funclions do not contain an s-part, we should instead use the ordinary dipolar
interaction energy, (4.202), which depends on the relative orientation of spins, is coupling between the electron and nuclear spins.
unchanged if bOlh spins are rotated through the same angle. For equivalent nuclei The extension of these ideas to solids was made independently by Bloem-
such as the three f1uorines in PF J , one cannot rotate one fluorine spin without bergen and Rowland [4.19] and by Ruderman and Kittel [4.20]. We shall discuss
rotating the others by an equal amount, since the alternating and static fields are the situation for metals, confining our attention to the coupling via the s-state hy-
identical at all three f1uorines. Therefore the coupling between equivalent nuclei perfine coupling. Since the metal is not diamagnetic, we should concern ourselves
does not affect the resonance frequency. with the possibility of 11 first-order effect - related, therefore, to the Knight shift.
Ramsey and Purcell [4.18] proposed another mechanism utilizing the elec- This mechanism of coupling was originally proposed by Frohlich and Nabarro
tron spins, which was substantially larger because, as we shall explain, it allowed [4.21]. As Yosida [4.22] has explained, however, the Frohlich-Nabarro effect is
the two nuclei to interact with nearby electrons, in contrast with the orbital mecha- actually included in the second-order calculation. We shall discuss the physical
nism where only one nucleus is on the same atom as the electron that is polarized. reason shortly, but for the present, we shall simply ignore any first-order effect.
We may schematize their mechanism as shown in Fig. 4. 12. Van Vleck [4.23] also discussed this effect.
132 133
1be effect of the magnetic moment of a nucleus at any lattice site is to We now tum to the actual calculmion of these effects. For simplicity we
make that site a region favorable for an electron of parallel magnetic moment shall calculate the interaction between the two nuclei directly rather than compute
but unfavorable for an electron of anliparallel moment. In order to take advanlage the changes in the spatial distribution of electron spin. However, the oscillatory
of the magnetic interaction, an electron of parallel moment will distort its wave natun; of the charge will be apparent from the answer.
function to be larger in the vicinity of the nucleus. The distortion is brought We consider, therefore, an electron-nuclear coupling 7ten involving only
about by mixing in other states k of the same spin orientation. As we shall see, two nuclei of spins II and 12 and for simplicity, treat only the effect of the
the result is as though only states above the Fermi surface were added. TIle wave s-state coupling. We have, then,
functions of the Bloch states are added so as to be in phase with the unperturbed
function (Fig. 4.13) at the nucleus in order to interfere constnlctively at that point, 7ten = 81\" ['Yl'Yeh211 • L Slb(rl - R I ) + "I2'Yeh2 12 . L: Blb(r/ - R2)]
but because of the spread in wavelengths, they rapidly get out of step as one 3 I I
= 'Ii, + 'Ii, (4.205)
moves away from the nucleus.
where we have allowed the nuclei to be different by using two values of gyro-
magnetic ratio "II and /2 and spins II and 12.
,,-,, We must lake into account the exclusion principle for the electrons. Two
(a) , , methods are available. We could use (4.205) to find penurbed one-elecrron func-
~
tions and then fill these functions in accordance with the exclusion principle. Or,
we could do penurbation theory in which we used the many electron functions
as the unperturbed states. It is this laller procedure that we shall utilize, since it
(b) •x emphasizes that bas.ical1y we are dealing with a many-electron problem and that
"'-----./ x-o "'-----./ the states we use are but one approximation.
Fig.4.13a,b. The UnllerturbM fun<;lion and two of the higher states mixed in. The nucleull Let us therefore consider a many-electron Slate 10) with energy Eo and
is at z = 0, which gUIlrAntees th"t the WIIYes be in ph"se lit z = O. Nole thl\llhc IIdmixed excited states In) with energy En. and compute the second-order energy shift
Waycs beal wilh one Allother. (a) Two of lhe WllVes mixed in by lhe perturbation. (b) due to ?ten. We shall as usual assume the total wave function of the system to
Unperturbed function
be a product of the electron and nuc1e:lr functions. Denoting the latter by 1/;0
with energy Ecn where 0 represents the nuclear spin quantum numbers, we shall
As a result of the beats between the unperturbed and perturbed functions, the wish to compute the second-order energy shift of states 10)"'0'.
original uniform distribution of spin-up charge density (we neglect the variations
Therefore, for the second-order shift, we have .6E~~ of the slate 10)"'0':
due to the lanice charge) is changed to have an oscillatory behavior, which dies
out as one goes away from the nucleus. The characteristic length describing the .6£(2) = L (Oal'h'enlna')(na'I'HenIOa) (4.206)
atlenuation is the wavelength of electrons at the Fermi surface. The resulting 00' ,(Eo + EO') - (En + Eo-')
",0
charge density of electrons whose momenl is parallel to the nucleus is shown in
Fig. 4.14. Since the electronic energy differences are generally much greater than the dif-
134 135
The first two terms in the brackets represent the changes in energy we should have
were onc or the other of the nuclei the only one. The last two terms represent the 10) = _1_ L(-I)P P[A(I)B(2)C(3) ... ]
.jjiif p
extra energy when bOlh are simuhaneously present, and are thus an interaction (4.213)
energy. Since it is only the interaction thai we wish 10 calculate, we shall consider jn) = _1_ L(-I)P P[A'(l)B'(2)C'(3) ... )
the last two terms only. We have then .Jiif P
Of course the pennutation causes any function to vanish if any two functions such
.1E(2) "" ' " (Oal1tdno')(no'j1i21 0o ) +complex conjugate (4.208)
"'" L..- Eo - £ n as A and B are identical. Consider the matrix element of a perturbation V, which
n,oJ
is symmetric among all the electrOns; that is, it is unchanged by interchanging
Now. fTOm the form of 'HI and 112. we can write them as the elecrron numbering:
HI '" I) . Gl;:; L I\fJG11J
(nWIO) = ;.,. L (-I)P+I"
P=Z,II,Z
(4.209) vN! pp'
,
L
'Ii, = I, . G, =
(J=z,l/,z
I,pG,p
x J P'[A'(I)B'(2) .. .j'VP[A(l)B(2)C(3) ... ]dT (4.214)
where Gj and G2 do not involve the nuclear spin coordinales. By using these Since V is symmetric, it is only the relative ordering of the states that counts.
relations, We can express this fact by defining the pennutation pll. pll is the pennutation
that, following p, gives the same ordering of electrons as pi alone. That is,
A£(') ",,(OIGjpln)(nIG,p,IO) ,,( II I ')( 'II I)
.u 00' "" L L E _ J;' L 0' 1/3 0' ex 2{j' 0' +C.C. pO p = p'. By making this substitution into (4.214), we get
fJ,P' II 0 LJJj
J
<)',
"
= L..- L..-
,,(OIG,pln)(nIG,p' 10)
E ~ (alhpI,p'la)+c.c.. (4.210)
(nWIO) = -2,
N. P
L (-I)('P+P") P" P[A'(I)B'(2) . . .j'
,pit
/J.P' n O.un
X VP(A(I)B(2)C(3) ... )dT
To evaluate the energy of (4.210), we should need to specify the nuclear
states 10), lust what they would be would depend on the total nuclear Hamil- = -2,
N. P,P"
L (_1)1'" JP" P[A'(1)11'(2) ... j' V P[A(I)B(2) ... )dT
tonian, which would include such things as the coupling of the nuclei to the
external static field Ho, the dipolar coupling between nuclei, and so forth. It is
convenient to note that whatever the states 10) may be, the energy ~E:;;! is just
= L(-I)I'" J P"(A'(I)B'(2) ... j'V(A(I)B(2) ... jdT (4.215)
what we should find as the first-order penurbation contribution of an extra tenn '"'
where the last step follows because V is unchanged by relabelling the electrons.
in the nuclear Hamiltonian 1felf given by lei us now consider V to be a sum of one-electron operators:
" [ " (OIG'P1n)(nIG,p,IO)] V = LV<Q (4.216)
1felf = L.J
(1,p'
11(112(1' L..
..
E _ J;'
0
+ C.c.
LJJJ
. (4.211)
,
By expressing the couplings G] and G2 explicitly, we obtain where V(f) depends only on the coordinates of the lth particle. Consider, for
example, the contribution from l = I:
L 5,6(r, - L 5,6(r, -
1felf = C"1 ]. ,
L..
(01 Rj )Io)(nl
(E
,
_ E,)
R,)I0)
·h +C.c.
(nWilO) = L(-I)I'" J P"[.4'(I)B'(2) .. .j' V(1)[A(l)B(2)C(3) . . ·ldT .
pit (4.217)
fI 0 I (4.212)
where This will vanish unless the state In) contains B(2)C(3). . Let us therefore
64.'
C = -9-;11"21';11'·
write such an excited state as
136 137
(nlVtlO) , I:<-I)P"
p"
J
P"IA'(I)D(2)C(3) ... ]' Now. the Fermi funclions as well as the energies E kll depend on the energy
associated with the electron spin quantum number. For example. the electron
x V(I)IA(I)B(2)C(3) ... ]dT spin Zeeman energy changes with 3. However, the Fermi levels of the spin-up
=0 irA' is idenlicalto any of the funclions n,e,D. etc. and spin-down distributions coincide. Thus. at absolute zero. there is a continuous
range of Ekll up to the Fenni energy and a cominuous mnge of E k ,.. from the
, (A'IV(I)IA) othe.,.;,. . (4.219)
Fermi energy on up. The matrix elements of the 6·functions vary slowly with
It is clear thai the various values of I in (4.216) will simply pick out the various energy. Therefore the variation of 'Herr with the electron spin energy is very small,
states A. B. C in 10), and the sum over ex:cited stales In) will pick out the states and we may forget about the energy of the electron spin coordinate, writing to a
A', n', and so on, which are not occupied in 10). We can therefore write (4.212) good approximation (4.223) in the fonn it would have in zero magnetic field:
as
'Ii," ' C L II
'Ii,"' C L II' ('15Ii)(,'151') ·1,
k,.s occupied kk'; 11,11'
k', s' unoccupied (klo(r - RIl!k')(k'lo(r - R,llk)f(k)[1 - f(e)]
(ksIS6(r - R,)lk's')(k'iIS6(r - R2)ks) x E E +C.c.
X ·12 + c.c. (4.220) k - k' (4.224)
Ek. - £1.:'"
We can now perform the sums over" and s':
where we have replaced Eo - En by Ek. - £1.:'/," since lhe stales
and En Eo
differ in energy solely by Ihe rransfer of one electron from the stale lk,,) 10 LII·(,151,')(iI5I,)·I" L IIp(,15pliXi I5p'I')[w
Ik', S'), The tenns "k, s occupied", "k'. i unoccupied", refer to whether or nOi
these siales are occupied by eleclrons in the wave funClion 10).
.,11'
.,.'
11"":'""
(4.225)
If we define the functions P(k.s) by
,L ltp[,p' T, {5p5p'}
PoP'
p(k, s) = I if k, s is occupied in State 10)
= ° if k,3 is unoccupied in state 10) (4.221)
But, as we saw in Chap. 3,
1
, }5(5 + 1)(25 + I)opp'
we can easily remove the restrictions on k,3 from the summation:
(4.226)
0IJIJ' 1
,-- since S=-
'Ii,"' C L 2 2
k,lI;k',.' which gives, finally.
P(k,,)(1 - p(k',il]
x E E ·/z+c.c. (4.222) ,C {" (kl<5(r - Rtllk')(k'I<5(r - R,)lk)f(k)(1 - f(e)]
k. - k'" 'Heff = I t· .1'2- L.
2 kk' E k - Ek ,
In order to express the variation of 'Herr with the tempemture of the electrons, we
must average 'Herr over an ensemble. This will simply replace IJ(k, 3) by f(k. 3), + c.c.} (4.227)
the Fermi function. We have then,
= A I 2II'/z ,
" II' (kslS6(r - R J )J'/s')(k's'IS6(r - R:z)lks)
'Herr = C L. where AI2 is a constant independent of spin. We now evaluate the matrix ele-
k,II;k',If' E kll - E k 'II' ments in terms of the Bloch functions:
f(k,,)[1 - f(e,i)] ,
X ·..I2+c.c. !f;k = 11k ( r ) e ik· r (4.228)
Eklf - Ek'If'
where. as before, lIk(r) has the periodicity of the lanice. We have. then,
, C L II' ('15Ii)(iI5I') ·1,
(k'16(1' - R2)lk) = U~,(R2)Uk(R2)ei(k-k')' R] (4.229)
(kl<5(r - Rtlk')(k'I<5(r - R,)lk)f(k")11 - f(k',i)]
X E E +c.c. so that.
kif - k'.' (4.223)
138 139
(klo(r - RI)Ik')(k'lo(r - R,)lk) I: co, I(k - k')· RI,]f(k)[1 - f(k')]
, (R)'(R) (R)i(k-k')'(R 2 -R I ) (4230) k2 k,2
= Uk'(R2)Uk' I uk 1 "k 2 e . k,k' -
If we assume that R, and R 2 are equivalent sites (as, for example, in a simple = (-,-)6 j'J [COS (kR COS 0) cos (k'R COS 0')
mctal), and define 2'11' k2 _ k,2
(4.231) + ~'i:::n-'.(kc:'R::..:c::o,,-';:;0,,):::,':::n.,(;;.k..:'R.:..:c:::o,:.:.O'.!..!)]
k2 _ k,2
we have
= C1112
'. '" 1".,(0)I'I".(0)I'2co, [(k - k') . R12J f(k)[1 _ f(k')] X f(k)[1 - f(k')jk'k"dfldfl' dkdk' (4.236)
HeW 2 L- E -E, The integrals over dn = -2'11'd(cos 0) and dn' = -2'11'd(cos ( 1 ) are readily
k,k' k k
performed to give, for the summution of (4.236),
64 2 2 1
= I"[r9'Tr 'Yc/1'Y2 h 4 1 j'J sin kR sin k'R kk'f(k)[1 -
-- - f(k')Jdk dk' (4.237)
X I: 1".,(O)l'I".(O)I'co, [(k - k')· RI,]f(k)[1 - f(k')] (2'11')4 R2 k2 _ k,2
I.:,k' E1,; Ekt (4.232) This integral may be evaluated at absolute zero by noting that the limits on k'
can go from 0 to 00, nO( just kr to 00, since if k' < kp for each k = k J , k' = 1.:2,
It is not possible 10 evaluate the summation without either some further approx- there is a k = k2, k' = k J which has the opposite sign for the integrand. The
imations or some explicit infonnation on the k dependence of wave functions range of k l is then extended 10 cover -00 to +00, and the integral is evaluated
and energy. If one assumes spherical energy surfaces, an effective mass m*, and by a contour integral. The infinity at k = k' is avoided by taking a principal parI.
that !uk/(O)12 and lu/.:(0)1 2 may be replaced by a v;\lue appropri:ltc to k and k' The final result is
near the Fenni energy, one can evaluate the sums. In these terms, 2
'Heff = - 9 '11' ,;,1I2 1i2m *!Uk F(0)j4
Ek = 2m*k
", 2 (4.233)
(sin 2k p R - 2krR cos 2kFR) I I
x I ' 2 (4.238)
This gives us R'
We note that this expression indeed corresponds to an oscillatory behavior as one
'Herr = I, . h ~ 1r 21';'YI1'2 h1 Iu kF(O)]" varies R. For large distances the coupling goes as
X 2m' I: co, [(k - k')· R12] f(k)[1 ~ f(k')] (4.234) (4.239)
fl 2 I }.;2 _ }.;/2
k,k
The number dN of states in k-space within a solid angle ef[} and between two We see that the dependence on lu(0)1 4 will cause the coupling to be large for
spherical shells of radii k and dk is large-Z atoms. In fact the coupling for the heavier elements is substantially larger
than the direct dipolar coupling.
dn k 2dk Had we used the conventional dipolar fonn of coupling between the nuclear
dN = 41f 21f 2 (4.235)
and electron spins appropriate to "non" s-states, we should have obtained
Denoting the angle between k and R l 2 as 0, that between e and R12 as 0' (see ~ = [r • r _ 3(l1 . R J2)(l2' R I2)] B (4.240)
Fig.4.15), and using R for IR 121, we have neff 1 2 R2 12
12
where B12 is a complicated function. It vanishes if there is no "non" s-character
to the wave function as seen by a single nucleus; for large distances, B 12 typically
falls off as 1/R12' For a discussion of B 12 the reader is referred to the paper
by B/oembergen and Row/and [4.19]. In the case of molecules such as PF3,
Gutowsky et al. [4.17] show that the assumption of p-type wave functions on
Fig.4.IS. Relalive orienlations of k, k', and R I 1 both P and F together with the "non" s-state coupling gives a coupling of the
lonn
140 141
(4.241) analysis of the electron-nuclear interaction. Denoting the wave vector of a Fourier
when averaged over all orientations of the molecule in the external field. The component by q, Yosida points out Ihut the Frohlich-Nabarro effect is the q = 0
coupling does nol vanish when averaged over molecular orientations, since the leon (infinite wavelength). In the second-order perturbation, a component q '" 0
induced spin magnetization itself depends on the orientation of the molecule with joins--two electron slates k and k' that satisfy the relation
respect to the nuclear spins. k' = k + q (4.242)
Couplings such as (4.240) have the same spin dependence as thai of the
direct dipolar coupling. To emphasize this similarity, the coupling is often referred Since in a proper second-order calculation the excited and ground states mllst
to as the "pseudo-dipolar" coupling. On the other hand, a coupling such as differ, we see from (4.242) that we must exclude q = O. Yosida treats q = 0 by
AI2II • h has the same fonn as the electrostatic exchange coupling. Since in first-order theory, but q f. 0 by second order, adding the results. The answer he
our case the physical origin is not an exchange integral, the tenn is referred to obtains in this way is identical to that of Bloembergen-Rowland and Ruderman·
as the "pseudo-exchange" coupling. Kittel, taking a principal part as mentioned on p. 141. The simplest way to see
The effect of pseudo-exchange or pseUdo-dipolar coupling on the width and that the answers will be the same is to consider instead the Knight shifts.
shape of resonance lines can be analyzed by simply adding these tenns to the There are two ways one can calculate the Knight shift. The first method is
dipolar tenns. The number of situations that arise are very npmerous. In liquids. to assume a unifonn static field andJO compute the first-order nuclear-electron
the pseudo-dipolar coupling averages 10 zero, but the pseudo-exchange does not, coupling of the polarized electron state. The second method is more complicated.
giving rise to resolved splittings. We assume a static field but one that is oscillating spatially with wave vector
For solids, both tenns have effects. The pseudo-exchange lenn, since it com- q. For such a spatial oscillatory field there is no net spin polarization, since the
mutes with [~ = [1% + h%, has 110 effect on the second moment of the resonance field points up in some regions of space as much as it points down in others.
but does increase the fourth moment. As Van Vleck discusses, this must mean The usual first-order interaction vanishes. If one now goes to second-order
that the central portion of the resonance is narrowed, but the wings are enhanced perturbation in which one matrix element is the electron-nuclear coupling, the
(see Fig. 4.16), since the fourth moment is more affected by the wings than is other, the electron-applied field interaction, a nonzero result is found. That is.
the second moment. The fact that the central portion appears sharper gives rise the static field induces a spatially varying spin polarization. Let us choose the
to the tenns exchange narrowing or pseudo-exchange narrowing. Where a real maximum of the static field 10 be at the position of the nucleus. When q gets
exchange interaction exists. as in electron resonances, the exchange narrowing very small (long wavelength), we expect that the result must be the same as if
may be very dramatic. the field were strictly unifonn. Therefore the limit of the second-order answer as
q ...... 0 must be the usual Knight shift. This result can in fact be verified.
In evaluating (4.238) using principal parts. one takes the limit as k ' --+ k.
Fig.4.16. Solid curve, the res- This procedure, as with the Knight shift, includes the first-order perturbation
\ onance shape wilh vanishing repopulation contribution. or the Frohlich-Nabarro effect.
pseudo-exchange coupling. The
dashed curve shows the shape
Before concluding this chapter, we should consider the role of electron spin
142 143
so that all matrix elements of 'H.sz to excited states In) vanish: 5. Spin-Lattice Relaxation and Motional Narrowing
(nl1iszIO) = 7.hHo(nIS, 10) = °. (4.246) of Resonance Lines
The ground state is therefore strictly decoupled from all other states as far as
the spin Zeeman coupling is concerned. 1be applied field is therefore unable to
induce any net spin, and there is no phenomenon analogous to the unquenching
of the orbital angular momentum.
The fact that the spins actually couple to a magnetic field makes this result
seem strange. Intuitively we except that, given a strong cnough magnetic field,
the spins must be polarized. The parndox is resolved by considering an example, 5.1 Introduction
the hydrogen molecule. The ground state is the singlet bonding state, but there
is a triplet antibonding state. In the presence of an applied field, the states split, We turn now to a discussion of how the nuclei arrive at their thermal equilib-
as shown in Fig. 4. 17. rium magnetization via the process of spin-lanice relaxation. We shall find it
convenient to cliscuss two techniques for computing TI·
The first method is appropriate when the coupling of the nuclei with one
Energy
another is much stronger than with the lanice. In this case an allempt to com-
pute the population changes of an individual nucleus due to th~ coupling to th~
lattice is complicated by the presence of a much stronger coupling of the nuclei
S-l,M._O among themselves. The first method makes the assumptio.n that the strong c~u.
Triplet
piing simply establishes a common temperature for t.he SpillS and that the .laHlce
Ag. 4.17. Erred of lhe applied field /10 coupling causes this temperature to change. There IS a close analogy wllh the
on the singld. And triplet-spin stales of process of heat transfer between a gas and the walls of its contain~~, i~ which the
S.O,M.-O a hydrogen mole<:ule. If 110 were large role of the collisions within the gas is to maintain a thermal eqmhbnum among
Singlet enough, a triplet slate would be lowesl
Ilnd the ground state would pOSSCliS & the gas molecules. In the collision of molecules with the wall, we con~i~e~ the
magnetic moment molecules to have the velocity distribution appropriatc to themlal eqUlhbnum.
H, As we shall see the first method leads to a fomlUla for TI that is particularly
convenient when the lattice is readily described in quantum mechanical tenus.
As we can see, for large enough Ho the S = I, /tI/s = -I state crosses For example, relaxation in a metal involves the transfer of energy to.the conduc-
the S = 0, Ms = 0 state. The ground state is then a triplet state, corresponding tion electrons, which are readily thought of in terms of Bloch funcuons and the
to a spin polarization. However, since the singlet-triplet splitting in zero field exclusion principle. . '
is several electron volts, the crossing of levels could never be produced by an The second method is that of the so-ealled density matnx. Although thiS
auainable laboratory field. is a completely general method, it finds its greatest utility for systems in w~ich
Further insight is obtained by considering the effect of a hypothetical mixing the lallice is nalurally described classically and in which the resonance WIdth
of a triplet state into the ground state. If the result is to induce a net spin is substantially narrowed by the motion of the nuclei. Moreover, when motio~
polarization in the positive z-direction on one atom, it induces an equal and takes place, the relaxalion time T2, which describes coupling between the nUcl~I,
opposite spin polarization in the negmive z-direction on the other atom. Clearly becomes long, and it may be a very poor approximation to assume that a SplO
there is zero net spin Zeeman interaction with an applied field in the z-direction. temperature is achieved rapidly as compared to T l . lllUS the second ~ethOO
Since such a spin polarization gives no net lowering of energy, it is not in fact is useful when the first one fails. Because of their large mass, the motion of
induced. the nuclei is often given very well in classical tenns. In fact an attempt to
Note, however, that if the two atoms are dissimilar, there may be a differ- describe the motion of molecules in a liquid quantum mechllOically would be
ent induced orbital moment which, through the spin-orbit coupling, could then quite cumbersome. Consequently the density matrix. method is well suited t.o
induce such a spin polarization. Thus, in a molecule such as H I, the iodine or- discussing cases in which motional narrowing takes place. An added feature IS
bital magnetization could induce a spin polarization into the bond, giving a spin that both TI and T2 processes (relaxation of I z and I z or III) can be treated by
contribution to the chemical shift on the hydrogen as well as on Ihe iodine. lhe density matrix method, provided there is motional narrowing.
144 145
The density matrix method is very closely related to the conventional time- would enter only when we tried to express the wave function of the total system
dependent perturbation theory. Actually the two are entirely equivalent. The in terms of the wave functions of individual spins.
density matrix method, however, gives results in a particularly useful form. It We shall say that any system whose population obeys (5.1) is described by
is ideal for treating problems in which phase coherence is important, and in a temperature T, even when the system is nOt in equilibrium with a reservoir.
fact the density matrix or a mathematical equivalent is necessary to treat such Equation (5.3) enables us to make a simple plot to schematize the populations.
problems. In any event, much of the fonnalism of this chapter applies equally We illustrate in Fig.5.1 a case based on our first interpretation: we consider the
well to systems other than spins. For example, dielectric relaxation can be treated population of the various energy states of a single spin of I = acted on by a t
by these methods. static magnetic field.
As we see, the two approaches complement one another, one applying to
the broad resonances of a rigid lattice and the other being most useful when the
resonance has been narrowed by nuclear motion. I
,
I
I
I
5.2 Relaxation of a System Described by a Spin Temperature I I
Exponential I
-'h envelopo ".i=-~
A system with a set of energies E u , Ell' and so on, which is in thermal equilibrium
with a reservoir of temperature T, occupies the levels with probabilities p(Ea ), +%
,
p(E/I), and so forth, which arc given by (a) (b)
+% - - -
P(Ea ) e-E~/kT .'ig. 5.1. (a) Energy levels ar II spin 3/2 nucleus. (b) Bllr grlll>h af POpullllion vel'$US energy.
- so that, since (5.1) The lengths or the bal'S are determined by the exponential em'e1ope
P(EI1) - e E.lkT
146 147
We shall assume that the Po's obey simple linear rate equations. Introducing
~) ._0 = P_I/2P_I/2W,;_O (5.5) Wmn as the probability per second that the lattice induces a transition of the
system from m to n if the system is in state m, the rate equation is
The inverse reaction from the circle 10 the cross will have a rale dNldOo .... ~,
given by dpni ::::: ""
d LJ(Pm W mn - pu W nm ) (5.13)
dN)
dt P_J/2PI!2 WO_J: (5.6) '"
0-. =:
'48 149
where
L Wmn(E m - E n )2
1 m,1I
= (5.28)
2 L:E~
n
I " 2 (5.19) Equation (5.28) was first derived by Gorter on the assumption that 1(3 - (31, 1« (3
-;::t -ZLJE" [5.1]. As we can see, this restriction is not necessary.
00 n
The great advantage of (5.28) is that, by postulating a temperature, it has
again in the high-temperature limit. Thus
taken into account the spin-spin couplings. The rate equations, (5.14), would by
" E' themselves imply that there are multiple time constanlS that describe the spin-
dE dj3~n lattice relaxation, but the assumption of a temperature forces the whole system
(5.20)
dt<::-dt~ to relax with a single exponential.
We now tum to evaluation of (5.14). Since the system is always describable by We may get added insight by viewing our states n as being nearly exact
a temperature, we have solutions of the nuclear spin Hamiltonian, between which (since they are not
_ P e(Em~En){J
exact states) transitions take place rapidly to guarantee a thermal equilibrium, but
(5.21)
PII - m between which the lanice also makes much slower transitions. After each lattice
We shall furthennore assume that when the system is in thennal equilibrium transition, which disturbs the nuclear distriblllion, the nuclei readjust among
with the lattice, the transitions between every pair of levels are in equilibrium. their approximate levels so that the lattice once again finds the spins distribllled
This is the so-called principle of detailed balance. Denoting by p~ the value of according to a temperature when it induces the next spin transition. Our formalism
Pn when the spins are in thennal equilibrium with the lattice. the principle of implies that treating the states n as being exact makes a negligible difference in
detailed balance says thai the answer.
LWmn = PIIlOWnm
Pm or that (5.22)
I.
5.3 Relaxation of Nuclei in a Metal
P
w:mn-- W nfllL
---!:!... = W nm e(Em-En).B" (5.23) We now turn to an example of the application of (5.28). We shall consider the
Pm relaxation of nuclei in a metal by their coupling to the spin magnetic moments
where PI. = I/(kTL)' of the conduction electrons. This is the dominant relaxation mechanism.
By subsliluting (5.21) and (5.23) into (5.14), we find In a T, process, the nucleus undergoes a transition in which it either absorbs
or gives up energy. In order to conserve energy, the lanice must undergo a
dE =.!. L: pm W mn [1 - e(Em-En)({J-fJ!.l](E" - Em} (5.24) compensating change. For coupling to the conduction electrons, we may think of
dt 2 m,n
the nuclear transition as involving a simultaneous electron transition from some
Now, expanding the exponential, we find state of wave vector k and spin orientation .5, to a state k l , l. We may think
of this as a scattering problem. Denoting the initial and final nuclear quanlum
dE (5.25) numbers as m and n, respectively, we have that the number of cransitions per
dt second from the initial state of nucleus and electron Imks) to the final state
Now Ink'i), W mkB,lIk' B" is
e- f3Em 1 - (3E m + (/32 E~/2!) .....,
pm=
Zoo
~
Zoo
=
Zoo
(5.26)
W mkB,nk'B' = 2; 2
l(mkslVlnk' i)1 o(Em + E kB - En - Ek'B') (5.29)
Thus, combining (5.26) with (5.25) and equating the resultant dEldt to that of where V is the interaction that provides the scattering, and where (5.29) assumes
(5.20), we find that there is an electron in Iks) and there is /lone in Ikll). The total probability
L: Wmn(Em - En)'] per second of nuclear transitions is obtained by adding up the IVmks nk'B' 's for
dP = (P _ P) ~ ,"m""n_~::;;;- _ _ (5.27) all initial and final electron states. We have '
dt L [2 LE~
n
150 151
W mn '" L
ksoccupicd
Wmks,Ilk's' (5.30)
x Jl"k(O)I'I"k,(O)I' f(k, ')11 - f(k', ,')Jg(Ek • A)
1,,'6' unoccupied x g(Ek " A')o(Em - En + Eks - Ek'6,)dEJ.,dAdEk'dA'. (5.37)
The sum over "ks occupied" is, of course, equivalent to summing over electrons.
We can remove the restrictions on ks and k's' by introducing the quantity Pks' We integrate first over dA and dA ' , utilizing the relations of (4.153, 154) to per-
which is defined to be unity if ks is occupied; zero, otherwise. This gives us form the integrals, and introducing again the average of [uk(0)1 2 over the energy
surface, Ek' (IUk(0)[2) EJ:' We also assume that the energy Eks appearing in the
W mu '" L vVmb ,llk's'Pks(l- Pk'6') (5.31) Fenni functions remains constant on a surface of constant Ek' an assumption
ks;k's' that would be fulfilled unless the spin energy depended on the location on the
By averaging (5.31) over an ensemble of electron systems, we simply replace surface Ek . The integralion over dEI,,' is then easy because of the delta function.
Pb by the Fermi function f(Eb), which we abbreviate as f(k,s): We have that
where we have chosen the nucleus I to be at the onglO. For the electron ~,~~+~-~,+~-~ (5~
wave function, we shall take a product of a spin function and a Bloch func-
tion uk(r)exp(ik· r). Therefore the initial wave function is Since E", - En. the nuclear energy change, is very small compared with kT,
the Fenni function f(Ek6 + Em - Ell) may be replaced by f(Eb)' Actually,
Imks) '" Im)ls)uk(r)e ik • T (5.34) doing so makes W mn = W nm . It is, in fact, at this point that the slight difference
between W mn and Wmn arises. It is the slight difference, we recall, that gives
It is a simple matter to compute the matrix element of (5.29):
rise to the establishment of the thennal equilibrium nuclear population. We are
(mkslVlnk's') '" 81r Iclnh2(mII[n). (sISls')uk(O)uk'(O) (5.35)
allowed to neglect the difference here in computing W"'>I' since we have already
3 included the effect in (5.23).
which gives us Moreover, since both u(Ek ,) and ([uk,(0)1 2 ) Ell are slowly varying func-
tions of E k" we may set them equal to their values when Ek '" Ek" In fact we
2: (mII.ln)(nllo,lm) shall evaluate U(Ek ,), and so on at E ks ' This gives us for the integral in (5.39):
OI,OI'=Z,Y,z =
x (,15.1,')(/150,1,)I"k(O)I'I".,(O)I' J<luk(O)I')~e'(E)f(E)[1- f(E)JdE (5.41)
x o(E", + Eb - En - E k '6') (5.36) o
where we have set the lower limits as zero, since the only contributions to the
We can substitute this expression into (5.32) CO compute W mn . We are once again integral come from the region near the Fenni surface. E'" E F .
faced with a summation over k and k ' of a slowly varying function. As before, Since (5.41) is independent of the spin quantum numbers sand S', we may
we replace the summation by an integral, using the density of stales g(Ek , A) now evaluate the spin sum of (5.39):
introduced in Sect. 4.7. This gives us
2:('15.1,')(/150,1,) = 2:(,150 5.,1,)
S,6' 6
W"'>I '" r; -9-'C Inti
27l' 647l'2 2 2 4 '"'
L.
I I
(mIIOIln)(nIIOI,[m)(s[SOIls )(s 1501 , Is) = Tr {SOISOI' } = 00101' !S(S + 1)(2S + I)
0I,OI'6,S'
'" 00101,/2 since 5 '" ! (5.42)
152 153
This gives us
W mn :::: ~4 1l'31l3'Y;'Y~([uJ.:(0)12)h,£l(EF)kT L [(m[Io [n)[2 (5.46)
W mn :::: ~ 1l'3h3'Y~'Y~ L(mIIQC[n)(n[Iolm) •
We note that Wmn is proportional to the temperature T. This fact has a
•
x J(IUk(O)I')~.'(E)f(E)[1 - f(E)JdE (5.43)
simple-physical interpretation. When the nuclei undergo a transition, they give an
energy to the electrons that is very small compared with kT. Most of the electrons
Now are unable to take pan in the relaxation because they have no empty states nearby
f(E)[1 - f(E)} = -kT :~ (5.44)
in energy into which they can make a transition. 11 is only those electrons in the
tail of the distribution that are important. Their number is proportional to kT.
which follows directly from the fact that We can write (5.46) as
1
W mn = L.;j L(mll;.ln)(nllj.lm) (5.48)
iJ 0
where the coefficient i = j is aoo, and where aij for i 1: j falls off rapidly with
the distance apart of the nuclei i and j. 11lese terms arise because the elecrron
wave function extends over many nuclei, so that more than one nucleus may
Fig. 5.3. f'mdions j(E), t -j(E), and scatter the electron from a given initial 10 a given final state.
j(E){I - j(E)I. The $Olid line .how.
I(E)(I -/(E)I By returning to (5.47) and employing our fonnula for T I , we have
L (mll.lnXnll.lmXEm - En)'
I •• ~,n~,~.
'00 ~=------
Since /(0) = 1 and /(00) :::: 0, and since /(£)[1 - /(E)] peaks up only TI = aoo Z- EE;n
00
within a width kT, it is also closely related 10 a 5-function when in an inlegral
of other functions that vary slowly over kT. L (ml[?t.I.J1n)(nl[?t. l.Jlm)
= _ a~ cm~,~n~,.,--- __=-=::- _
f(E)[1 - feE)] = kT6(E - E.) (5.45) EE;n
(see footnole 7 ). By utilizing this fact, we have, finally, 00
aoo o=::e,y,::
(5.49)
=--
7 Equation (5:4.5) can be derived simply. Let C(E) be a slowly varying function of 2 Tr {?t'}
energy. Then, utlh1.lIIg Lhe fact that j(E)[l-j(E)J is nonvanishing within only kT about
Ep, we c~n expand C(E) in a power series "bolll EF: A similar expression is found by using (5.48). The important point is that we do
2 2 not need to solve for the explicit eigenstales and eigenvalues, but only evaluale
G(E)=GCEFl+CE-EddC[ +(E-Ed d CI + the traces in a convenient represent:ltion.
dE 2! dE2 .
Thus 00 SF _ Ep
J G(E)j(E)[1 -j(E)]dE = G(E~') J j(E)/1 - jCE)ldE The first term, using C5.44), gives G(EF )kT. The second term vanishes, since the integrand
is an odd function of E - EF, and the third term gives a contribution proportional to
" dG 00 ,
J(f.,-E~,)j(1-J)dE+dE7
",pC I (kT)3, as seen by changing the integrand from E lo ::e := ElkT. If we neglect the third
I and highcr terms, the answer is just what we should have if we replaced j(1 - n by
00
+dE j(E-Ep )2jCt-ndE ... kT6(E - Ep). The correction!! are generally of order (kTIEd 7 smaller, the exact form
SF 0 S 0 depending on the functional dependence of C on the energy E.
154
155
For our problem of a single spin, the quantum numbers m and 11 would
label the 2I + 1 eigenstates of I r • Then, using the fact that (5.59)
(5.50)
The T appearing in the Korringa relation represents only one contribution to the
[Ir , I z ] ::: iIy etc. (5.51) relaxation time - that due to the coupling of nuclei to the magnetic moment of
,!I-state electrons. One expects that the experimental TI should, if anything, be
we find shoner. It is therefore interesting to examine a table given by Pines {SA]. In it
L Tr {['H, IO']2} ::: _1'~h2 H6 Tr {(I; + I;)} (5.52) he lists e;Jlperimental Tj's, those computed from (5.58) (the Korringa relation)
• and those computed from (5.59), using Pines' theoretical values of relx& and
I!O(E.)/e(E.).
E Tr {'H 2 } ::: 1~'12 H6 Tr {I;}
• Thbl~ 5.1. Experimental and Lheoretical T I '. (all tillW$ in ms)
so that, since Tr{I;} ::: Tr {I;} :::Tr{I;J,
T 1 (Experimental) T. (Korrillga) T. (Pines)
L Tr {[?i.E.I' t5O±5 88 232
a :::-2 (5.53) 15.9±0.3 10.3 18.1
Tr (11') 2.;5±0.2 '.1 2.94
3.0±0.6 '.3 4.'
For a Fenni gas of noninteracting spins, one can show that X~ given by
1
2h2 5.4 Density Matrix - General Equations
Xo:::: Teo(Er) (5.57)
As we have remarked, the concept of a spin temperature is nOl always valid.
where we have put subscripts "0" on XS and e(E F ) 10 label them as appropriate
We lum now to discussion of a mel hod of attack that is very useful when the
to noninteracting electrons. In this approximation one has
spin temperature concept breaks down - the technique of the density matrix. An
C>H)2 h "(2 exceptionally good discussion of lhe densily matrix is that of Tolman [5.5}.
TI ( - - • - - --" (5.58) The method has the further advantage of giving one a discussion of both
H 41fkT "(~
T l and T2 processes in a natural way. It is ideally suited 10 treating problems in
Equation (5.58) is commonly called the "Korringa relation", after Dr. J. Korringa which the resonance is narrowed by lhe bodily motion of the nuclei. It is also
who first published it [5.3]. It provides a very convenient way to use measured applicable to broad line spectra, whcre it can in fact be used for an alternate
Knight shifts to predict spin-lattice relaxation times. A more accurate expression derivation of our equation for TI of Scct.5.3. As we shall see, the method is
is obtained from (5.56) and (5.57): simply a variant of the usual time-dependent perturbation theory, but one that is
in a panicularly useful fonn.
156 157
We begin by considering a system described by a wave function t/J, at some Therefore
instant of time. and ask for the expectation value (M~) of some operator such PMztln = L PUm(mIM~ln)
as the x-component of magnetization. M~. We have. then, m
158 159
For the future, we shall omit the bar indicating an ensemble average to simplify The density matrix is the quantum mechanical equivalent of the classical
the notation, but of course we realize lhat whenever the symbol (J is used, an density (J of points in phase space, and (5.82) is the quantum mechanical fonn
ensemble average is intended. of Liouville's theorem describing the time rate of change of density at a fixed
Of course the wave function t/J, describing whatever system we are consider- point in phase space.
ing, will develop in time. Since the 11 11 's are independent of time, the coefficients In the event that 1i is independent of time, we may obtain a formal solution
of (5.82),
C
n mUSI carry the lime dependence. It is straightforward to find the differential
equation they obey in terms of Hamiltonian 1i of the system, since (](t):: e-(i/l)'ltI{](O)e(i!l)UI (5.83)
_'!. a~ = 'H~ (5.18) In terms of functions U'I> which are eigenjimcrions of Ihe Hamiltonian 'H, we
, at have, for example,
which gives, using (5.61),
11 dCn
(k1{](t)lm) :: J uke-(i/')'H'{](O)e(i/l)'ltfumdr
--;-1 Ln -dUn::
t
Lcn'Hu ..
.. :: J (e(i/l)'HI Uk ) •{](O)eWA)'HIUmdr (5.84)
We can pick out the equation for one panicular coefficient, CI;, by multiplying
By utilizing the fact that HUm:: EmU"" and using the power series ex.pansion
both sides by uk and integrating:
of the exponential operator, we get
h dCk
-7 - = 2:Cn(kl1tln) (5.19) (5.85)
I dt n
for the time-dependent matrix element in lerms of the matrix element of t! at
This equation is the well-known starting point for time-dependent perturbation
t :: O.
theory. We can use (5.79) to find a differential equation for the matrix elements
So far we have talked about the density matrix without ever exhibiting
of the operator P, since
explicitly an operator for e. For the sake of concreteness, we shall do so now.
We shall take an example of a spin system in thermal equilibrium at a temperalUre
T. We shall take as our basis states, Uri, the eigenstates of lhe Hamiltonian of the
problem, 'Ho. The populations of the eigenstates are then given by the Boltzmann
factors, giving for the diagonal elements of (!:
(5.80)
."
= -,i
i
E [ckc;.(nIHlm) - (kl'Hln)cnc:"l
e- E ... / kT
CmC~ :: =--'0--
where, as usual,
Z
(5.86)
160 161
for a group of noninteracting spins, as we did in (2.88), the vanishing of the In the high-temperature approximation we can expand the exponenlial, keeping
off-diagonal elements of !! would make the transverse components of magneti- only the first tenns. By utilizing the fact that Tr {M z } = 0, we have
zation vanish, as they must, of course, for a system to be in thennal equilibrium.
More generally, we see from (5.85) Ihal the off-diagon..l elements of fI oscillate (M,) = ~ T'{M,(,- ~+)}
harmonically in time. If they do not vanish, we expect that there will be some ob-
servable property of the system which will oscillate in time according to (5.75).
BUI we should then not have a true thennal equilibrium, since for thennal equi-
~ ~ Tr{ (~~h::oI;) } (5.93)
librium we mean that all properties are independent of time. Therefore we must
assume that all the off-diagonal elements vanish. Note. however, from (5.85)
Now, in the high-lemperanlre limit, Z = 2I + I. Since Tr {4} = j I(I + 1)(21 + 1),
we get
(which applies to the situation in which the basis functions are eigenfunctions of
the Hamiltonian) that if the off-diagonal elements vanish at anyone lime, they (M z ) = ~~fI2 I(l + 1) H (5.94)
vanish for all time. 3kT 0
We have, therefore, which we recognize as Curie's law for the magnetization. The density matrix
gives, therefore, a convenient and compaCI way of computing thennal equilibrium
(nlelm) = (omll/ Z )e-En /k'l" (5.88)
properties of a syslem.
It is worth noting that the operator for f! is on a different fOOling from most One situation commonly encountered is that of a Hamiltonian consisling of
other operators such as that for momentum. In the absence of a magnetic field, a large time-independent interaction 'Ho, and a much smaller but time-dependent
the latter is always li.\J/i. For a given representation the density matrix may, tem H,(t). The equation of motion of the density matrix is then
however, be specified quite arbitrarily, subject only to the conditions that it be de i
Hennitian, thai its diagonal elements be greater than or equal to zero. and lhat dt = !i[e, 'Ho + Hd (5.95)
they sum to unity. There is therefore no operator known a priori. However, in
certain instances the matrix elements (nJelm) can be obtained very simply from If HI were zero, the SolUlion of (5.95) would be
a specific operator for fl. When this is possible, we can use operator methods to e(t) = e -(i/" )1io/ e(O)e(i/h)1i oi (5.96)
calculate properties of the system. We now ask what operator will give the matrix
elements of (5.88) bearing in mind that the u,,'s, and so 011, are eigenfunctions Let us then define a quantity e* (the siar does not mean complex conjugate)
of 'HO. by the equation
Using the fact that e(t) = e -(i/")1iot e* (t)e(ijh )"Ho/. (5.97)
-"Ho/kT -Em/kT
e um=e Urn (5.89) If, in fact, HI were zero, comparison of (5.96) and (5.97) would show that e*
(which can be proved from the expansion of the exponentials), we can see readily would be a constant. (Note, moreover, Ihat at t = 0, e* and e are identical.)
that the explicit fonn of e is For small?tl, then, we should expect e* to change slowly in time. Substituting
(5.97) into the left side of (5.95) gives us the differential equation obeyed by e*:
e = ~e-"Ho/kT (5.90)
Z -~[?to, el + e- O/")1iotd: * e(ifh)1iot = ~(e, Ho + ?til (5.98)
We can use this expression now to compute the average value of any physical t
property. Thus suppose we have an ensemble of single spins with spin I, acted We note thai the commutator of e with ?to can now be removed from both
on by a static external field. Then ?to is the Hamiltonian of a single spin: sides. Then, multiplying from the left by exp (il1l)'Hot) and from the right by
exp (-(i/h)Hot), and defining
(5.91)
Hi = ei1iot/h?tle-(i/h)"Ho/ (5.99)
We shall illustrate th~of the density matrix 10 compute the average value of
the z-magnetization (1I1 z ). It is we get, from (5.98),
162 163
Equation (5.100) shows us, as we have already remarked, thai Ihe operator r/ We can make a closer approximation by an iteralion procedure, using (5.109) to
would be constant in time if the perturbation ?i1 were sel equal to zero. get a bener value of U·(t l ), to put into the integrand of (5.108). Thus we find
The transformation of the operalor?il given by (5.99) is a canonicallTans-
formation, and Ihe new representation is termed the interaction representation.
The relationship of U and (J. is illustrated by considering the expansion of the
·e(l) = ,'(0) + *J[{, *J"
o
,'(0) +
0
(e(O), "H;(,")Jd," }. "H;(")] d"
,
wave function tP in a form
tP = L:ane-(i/l)E.l un
n
(5.101) = ,'(0) + *J
.
o
(,(0). "H;(")]d"
instead of 2 t tl
tP=L:Cnu n (5.102) + (D JJ
o 0
((,'(0). "H;(''')J. "Hi(")]d" d'" (5.110)
n
where the un's and En's are the eigenfunctions and eigenvalues of the Hamil- We could continue this iteration procedure. Since each iteration adds a term
tonian ?io. In the absence of ?iI, Ihe an's would then be constant in time. We one power higher in the perturbation 'Hi, Ihe successive iterations are seen to
shall show Ihat Ihe matrix ana~ is simply (nlu·lm). We note first that replacing consist of higher and higher perturbalion expansions in the interaction ?i I, For
I?(O) by U· in (5.83-85) gives that our purposes, we shall not go higher than the second. Actually, we shall find
it most convenient to calculate the derivative of (J*. Taking Ihe derivative of
(nll?·lm) = e{i/Ii)(En -E".)t(nll?lm) (5.103)
(5.110) gives
Since (5.101) and (5.102) give the same tP, we must have
(5.104)
d,;,(1) = *["(0). "H;(I)] + US J o
((,'(0). "Hi(I')]. "Hj(l)]dl' . (5.111)
166
167
'H~ff '" - ,MHo - w/-y)I, - 1hH I I r
(5.139)
I
+ "E1 IiH0t1 IJ,k+"2
Ir;
L'H2 j
k,j
(5.132)
and a new zero of chemical shifts
This new effective Hamiltonian reminds us by the presence of the tenns '171, '" Uk - (j (5.140)
'Hqlr; that if spins interact with one another in the laboratory frame, they do also which makes
i/the rotating frame. It also reminds us from the tenns involving UIr; that when
there is more than one chemical shift, one cannot be simultaneously exactly on 1(z = --I hHoI, (5.141,)
resonance for all nuclei.
Since 'H~tr is independent of time, we can solve formally for en(t) in terms 1tp = L 7 'IHOUkI,1r;
1
(5.14Ib)
k
of its value at an earlier time (t '" 0) by
where in getting the second equation we drop terms such as 7iult.
eR(t) '" e -i'H~ffL/. e R(O)ei'H~ffl/lI (5.133) Clearly the fonn of (5.141) is the same as that of (5.127). We have simply
It is imponam to keep in mind that in general the interaction representation redefined 7 to COlTeSpond to a different zero of chemical shifts.
differs from the rotating coordinate transformation. Both may be used with a While (5.138) guarantees that the Zeeman term falls in the midst of the
large time-independent Hamiltonian 'Ho and a time-dependem ?it(t). resonance lines, it seems to imply through (5.137) that we know the individual
e- (e expressed in the interaction representation) is related to e by uk's, and go through some evaluation to get 7i. What is imponant to realize is
that there is some arbitrariness in the division between 'Hz and 1tp which really
e- (t) '" ei'Hol/h efJ)e -i11{Jtlh (5.134) corresponds to the arbitrariness in the choice of a zero of chemical shift.
whereas eR is related to e by
eR(t) '" e-K..>tI. e<t)eiloIll. (5.135) 5.6 Spin Echoes Using the Density Matrix
If we have the Hamiltonian of (5.114),
Now that we have obtained the differential equation for the density matrix in
l (t) '" e -K..>oll. e<O)eiwoti. (5.136) the laboratory frame, in the interaction representation, and in the rotating frame,
These do become the same if we are exactly at resonance (wo '" w), but othelWise we tum to methods of solution. The rest of this chapter is devoted to that task.
would differ. If the only time dependence in the problem is from HI, the rotating We will first discuss several problems involving pulse techniques, then we tum
coordinate transformation eliminates it. to use of the density matrix to discuss relaxation. As topics involving pulse
The division of the Hamiltonian of (5.127) into ?i z and ?ip is nOt the techniques we shall first examine how to use the density matrix to derive the
only division one might think to make. For example, if one defines an average spin echo, giving us a chance to use this formalism to examine a problem whose
chemical shift for the system of N nuclei, 7i, by the equation physics we have now seen both classically and quantum mechanically (using
1 wave functions). We then turn to some new physics which underlies one of the
7i'" NLt1k (5.137) major techniques employed in NMR today, Fourier transform medlOds. To do this
k we first set the stage by deriving a result important to linear response theory, the
then one might prefer to have the Zeeman energy be the average value including theory underlying Fourier transform NMR. This result is the response of a system
the chemical shifts. We can do this by making a division into an 'HZ and ?il> to a 8-function excitation. Anned with that result, we next go on to explain what
defined as is meant by Fourier transfonn NMR, and then prove the fundamental theorem
which specifies what it measures. We then further develop a concrete appreciation
(5.138')
of the density matrix by exploring what happens to the actual matrix elements
under the action of an alternating magnetic field.
'H'
p''" 'LJ
' 1 hH o(Uk - -U)I,k +"2I ",--?tij
, (5.138b) The most general approach to solving the equation is to pick an appropriate
k j,k family of wave funclions as a basis set, then conven the operator equation to a
Note, however, that (5.138a) can be converted to (5.127) if we define a new 1, set of coupled differential equations between the matrix elements of e and ?i.
called i, Thus if we pick a basis set 10), where 0 stands for the set of eigenvalues which
distinguish the states, we convert the equation
168
169
d(! i H,
- = -«(!H - ?i(!) to (5.142)
dt h
d, • p
j;
d (aleld) = D(aleIP)(P1Hla') - (aIHIP>(Plela'» (5.143)
I
r------ T ----l
I iI
In many cases there are then straightforward melhods of solution. Once one has
I I "
obtained the (alula'rs, one can then calculate the expectation values of observ~
abies using (5.77), o t1 t2 t]
Fig. 5.4. III vs time, giving the definition of the times 0, tl, t2, and 13 which spc<:ify the
(M.> =T,{M.e) = L:(a'IM.la)(alela') (5.144) beginning and end of the tr/2 pulse (I = O,td and of the tr pulse (t = t2,13)
0',0"
The weakness of Ihis appro,\lch is that it may fail to reveal important physical
insights available from other methods. As in Chap. 2, we approximate this as
The "other methods" are based on manipulating operators, and are analogous 1i"err = -1'hH I T;I; (5.146)
to the quantum mechanical derivation of the echo carried out in Chap. 2. We now
tum to several illustrations of the use of the operator approach. Later we return Then, utilizing (5.83), we write
to use of (5.143). I? n(t I) = e iWI tl f" I? n(O+)e -iWl 11 I"
We begin with a derivation of the spin echo of a group of non interact- = eiW1 tl f z I?n(O-)e-iW1 tIl" (5.147)
ing spins using a (rr!2,7r) pulse sequence. We deal with this problem by first where
transfonning to the rotating frame so that our basic equation is given by (5.123)
WI = 1'HI (5.148)
d(!R i
--;It = "h«(! llHeff - 'Herru n) Likewise, during the interval tl to t2, where HI is zero, we have
As in Chap. 2, we assume that the static field is inhomogeneous with a distribu- fm(t2) = e i ,.Il o(t2 -( 1) I, I? n(t l)e -i,.lt o(t2 -11 )1. (5.149)
tion function p(ho)dho for the N spins, where dN, the number of spins whose
resonance lies between ho and ho + dho, is given by (2.119): Defining [as in (2.122)]
(5.145)
A similar argument holds of the other discontinuities. During the time in- I?n(t, ho) = X(7r)T- I (t - T, ho)T(T, ho)X(rr/2)l?n(O-)
terval from t = 0 to t = t I or from t2 to tJ, the Hamiltonian is given by (2.117) x X- 1(7r/2)T- 1(T, ho)T(t - T, ho)X-I(rr) (5.153)
1teff = -1'1i(h oI z + H IIz) Thus, when t - T = 7, or t = 27,
170 171
(l/l(2r, 110) '" X(7r)X( 7r(l.)e n(O-)X- 1(7r(l.)X- 1(7r) (5.154<1) where
Tr{G} ",Tr{I+X(-7r!2)(1 + -Yli:'~lz)X-I(_7r/2)} (5.160b)
= X(3./2)eR(0-)X-'(3./2l (5.154b)
'" ..!.-e-x....II. e -'Ho/kTeilollf• (5.155) Now X(-1(/2) '" X-I (1f!2). Hence, using (2.124) (X- t ('lf!2)lzX(Jr!2) '" -llIl
Z we get
Since 'Ho '" -'YhHolz, we get
,hN J
en(tz) '" TZTy ... TBT"{;l(O-)T; I Tjj I ... TylT I Z (5.167)
'" (21 + I) P(ho)d11o {G} Tr (5.160<1)
172 173
We chose to neglect all tenns in the Hamihonian but the tenn involving ,+
HI when HI is on. That approximation is often useful (as illustrated above) but
would not be allowable if one were trying to understand what happens when HI
JC'(t)d'
,-
= C (5.169)
is comparable to the line width. The general fonnula (5.167) does nOI necessarily
include this approximation, but the approximation is readily introduced if desired. But, 'by definition, m(t) is the response when this integral is unity, i.e. when
C = I. or. alternatively, if M(r) is the response when C 4- I,
meT) = M(T)/C . (5.170)
5.7 The Response to a 6-Function
While in principle one can set C = I, in the process one conceals some units.
Now lhat we have looked at a familiar problem with the density matrix fonnalism. and thus complicates the use of dimensional arguments to check the correctness
we tum to a problem we have not treated quantum mechanically before. the of fonnulas. We will therefore keep C, realizing that eventually we will utilize
response of the system to a o-function. This topic is imponam in pan because, (5.170). Indeed. to keep the units explicit. we will replace C by a product H TfJ:
as we saw in Chap. 2. if we know this response we can use it to calculate both
Hro == C (5.171)
I and I'. But it is also important because it underlies the theory of Fourier
transfonn NMR which we take up in the next section. We therefore stan by where one can think of H as the height and 1"0 the duration of a rectangular
recalling (2.156): pulse which. in the limit that rO goes to zero, represents the 6-function.
~
To find mu(r) we need to solve the density matrix with the Hamiltonian
J,
~
e(0-) = _1_
Z(oo)
(I + ,/oHoI,)
kT
(5.184)
contains only spin-spin terms and 7 is chosen such that the Uk'S are all zero,
the first trace always vanishes. The effect of the operators Tp is to modulate the
coefficients of the oscillations at Wo over times of the order of the inverse of the
Following the argument after (5.158), we drop the first tenn in the parenthesis spectral frequency width, L1w.
since it contributes zero magnetization, getting
176 177
Thus, for times short compared to the inverse of the spectral frequency
width, ~, we may write mzz(t) = X{w~} [cos (woOTr {Io;Trl yT p1}
Tr l y
721i2 Tr {12} +sin (woOTr {IyTplyTj;I}] (5.195)
mzz(t) = 7 Z(oo)kT Ho sin wot with t <: l/L1w (5.190)
~
While (5.195) looks relatively simple, involving just a few traces, its actual
• evaluation may be enonnously difficult since the equation contains all the content
It is straightforward to show that the term a for a many-spin system is the static of the theory of line shapes! Nevenheless, the equation gives a compact statement
susceptibility Xo. Hence for attack by the various fonnal methods which have been developed to deal with
the line-shape problem.
mu(t)=xowo sinwot for t<t:l/L1w (5.191)
In principle, this mu(t) will yield both the absorption and dispersion spectra
This result has a simple classical significance (Fig. 5.5). utilizing (2.156)
=
z x~z(w) = J mu(T) cos (wT)dT
o
(2.156)
=
y
X~:I:"(w) = J mu(T) sin (wT)dT
o
....ig.55. A magnetization, Alo = xoJ1" initially
along the :--direction is rotated through II.
small angle dB = "'I J1 TO by a field J1, of 5.8 The Response to a ,,/2 Pulse: Fourier Transform NMR
duration TO, along the laboralory :I:"-direction.
The rC:$Ulting magnetization has .. component
Mod6 along the laboratory ",-direction. At later We tum now to the topic of Fourier transfonn NMR, a method of doing magnetic
times, it precesses about z, producing .. nonzero
resonance which is today all pervasive in the field of high resolution liquid spec-
, component along the :I:"-axis which oscillates all
sin Io'(Jt tra, and which fonns the basis for one of the most important other developments
in NMR, so-called two-dimensional Fourier transfonn NMR (which we take up
in Chap. 7).
Considering the 6-function to be a field in the x-direction of strength H>- Ho,
In 1957, Lowe and Norberg [5.6] discovered an imponant theoretical result
on for a time 7"0 < I/wo, we see that the field tilts the thermal equilibrium mag-
which was confinned experimentally by them in conjunction with Bruce [5.7].
netization xoHo to have a component in the positive laboratory y-direction.
Lowe and Norberg showed theoretically that in solids with dipolar broadening,
Since we are seeking a response (Le. an x-y magnetization plane) linear in
the Fourier transfonn of the free induction deCa'yfoliowing a -rr/2 pulsegave
the driving tenn, H, we keep the angle of rotation, L18, small
the shape of the absorption line. They verified their result experimentally with
(5.192) a single crystal of CaF2 by comparing their experimenlal free induction decays
with the Fourier transfonn of Bruce's steady-state absorption data taken on the
giving
same crystal.
My = XoJIO(7HTo) = HToxowo (5.193) In 1966, Ernst and Amfer.fon [5.8] pointed out that there were substantial
This magnetization precesses at Wo, in the left-handed sense, giving a sin wot experimental advantages to the use of pulses over steady-state meth<X1s when
dependence to M:I:"' lllUs, utilizing (5.170), mo;o;(t) will be given as dealing with complex spectra. (We discuss advllntagcs later in this section.) Their
method, which consists of first recording the free-induction signal, then laking ils
mu(O = xowo sin wot (5.194) Fourier transfonn, is referred to as Fourier transfonn NMR. The method allows
for times short compared to the dephasing effects arising from chemical and on~ to obtain both the absorption spectrum and the dispe-;;ion spectrum'"":", TheY
Knight shift differences and spin-spin couplings. ~how ttlatthe cosine transfonn of the free induction decay gives one whereas the
We see, therefore, that (5.188) and (5.189) have simple physical significance. sine transfonn gives the other. (See also the book by Ernst et al. [5.9].)
In fact, we can usefully rewrite them as
178 179
[
These remarks are reminiscenl of linear response theory discussed in Sect. with WI = "(HI. Taking UR(O-) to represenl thermal equilibrium, we have,
2.11 where we found thut X' and i l are given by cosine and sine transforms of I utilizing (5.155-157) generalized for the Nobody Hamiltonian as in (5.184),
the function m(T), (2.156):
~n
(0-)= (0-)=
U Z(oo)
_1_(1 +7
hH 1
kT
o ,) (5.204)
J
~ ~
x' = m(T) cos (wi)dT X" = J m(T) sin (wT)dT (5.196) whence, as in (5.150a) and (5.151)
o •
where m(i) is the response of Ihe syslem 10 a S-function which we treated in 1
un(t,) = Z(oo) (lhHO
1+ kT X(wltt>I~X (Wltl) -I) (5.205)
Sect. 5.6. where
We therefore demonstrale that in the linear response regime al high temper-
atures, a 1f(l pulse (in faCI, any pulse) is equivalent to a S-funclion excitation, X(D) == e+ i8Jz (5.206)
hence Ihat the transforms of the free induction signal give ;( and X". Let us con-
Now, ulilizing (2.55),
sider, then, Ihe Hamiltonian of (5.124) describing a group of N coupled spins,
allowing there 10 be both chemical and Knight shifts. We then have a Hamiltonian X(8)I z X-'(8) = I y sin 8+ I~ cos 8 (5.207)
'HI given by keeping only Ihat part of the spin-spin coupling which commutes
so that
with the Zeeman coupling as in (5.172)
1C = 'Hz +'Hp +1-{1 with (5.197) UR(tl) = Z(~) (I + 1::0 [Iv sin Witt + I z cos Wtttl) (5.208)
The first term on the right will nOt contribute 10 the magnetization, following the
argument of (5.160) and (5.161) (i.e. it is allihal remains as T-+oo).
Moreover, the term involving I z is, apart from a constant of proponionality,
(5.198) the same as the temperature-dependent contribution to a (! R corresponding to
thermal equilibrium. Thus, it contributes nolhing to transverse magnetization.
(The validity of neglecting the tenns involving 1 and I z can be verified by direct
calculation of (Mz(t)) and (M!I(t)).] We are left, then, with
(5.199) I 1liHo.
UR(tl) = Z(oo) ---;;r
Sill (wltdIv (5.209)
18. 181
Thus, since 1tz + 7-ip arc independent of time, we get from (5.83) Utilizing (5.196) and (5.219) with Hro = I, we have then
e(l) = T(t - tt)e(tl)T-I(t - til =
= -yIiHo sin (Wltl)T(t _ tl)R(t\)lyR-I(tl)T-IU - t\) ..
i(w) = ."1 j
Sin 0 0
(Mz(r»o cos (wr)dr (5.223.)
Z(oo)kT (5.213) =
Now
T(t - tl)R(tl)
Xl/(w) = .'0 J(Mz(r»o sin (wr)dr
sm
(5.223b)
o
= exp [iwo(t - t 1)1:] exp (-i7-ip(t - t I )/1i.) exp (iwt I [:)
Equation (5.223b) includes the result of Lowe and Norberg (they did not
= exp (iwotI:) exp (-i1tpt/fi) exp (iHptdh)exp [i(w - WO)t1 1:] include chemical or Knight shifts). Equation (5.223) includes the result of Ernst
p
= T(t)T I(tl) exp (j(w - WO)tl 1:) (5.214) and Anderson that the cosine and sine transfonns yield respectively the dispersion
and the absorption. [It is useful to keep in mind that the tenns "cosine" and "sine"
If, now, we neglect the effect of Hp and (w-wo)t I during t\ we can approximate
imply a particular phase convcntion. Ours is defined by (5.221).]
Trl(tl)ei(""-~)tl/. ~ I (5.215) The validity of (5.218) rests on the approximation that during the pulse we
can neglect the effect of Hp and (w - woH I. Since HI' leads to a spectral width,
getting Llw, this approximation is equivalent to
(5.216)
and (w - WO)tl «I or (5.224.)
Therefore, the expectation value of the transverse magnetization (Mz ), following
"YHI »Llw and "'fHI» Iw - wol (5.224b)
(5.158), is
The classical derivation of the b-function response given in (5.192-194)
(M.(t) = ,1m {g(t)I.} involved assuming that a magnetic field H satisfy
2h2H
= -y . (Wltl )Tr (I z TIy T- 1)
O Sin (5.217) H»Ho (5.225)
Z(oo)kT
where the field H lasts for a time ro. This condition signifies that the magneti-
This result is identical to the result of (5.187) except that sin WI tl replaces
zation is rotated during a time less than a Lannor period in H o. This appears to
-yHro. We introduce the notation
be a much stricter condition than (5.224b). The reason one can use an HI «Ho
2,.2H is that HI is rolating at a frequency w which is nearly the same frequency as the
(Mz(t»o = -y t 0 sin (O)Tr{IzTlyT-I} (5.218)
Z(oo)kT free precession frequency WOo Thus, while HI is on, the total phase angle about
the Ho axis which the spin advances is wtl. If a l5-function had kicked the spin
the magnetization following a pulse with
at t = 0, the subsequent precession angle in the same time tl would have been
wltl == 0 (5.219) wotl' But since I(wo - w)tll and Llwtl «I we can take wOtl = wtl' Thus, the
HI builds up the rransverse magnetization over the time it> but does not in the
Then we can write
process inrroduce a phase error relative to a "zero time", or 6-function, creation
sinO of transverse magnetization followed by free precession for t I.
(Mz(t))e = 7n zz (t)-H (5.220)
, TO We note also that this argument explains why there is nothing magical about
Now, (Mz(t»e is just the signal we would observe from a pulse in which exciting the free induction decay wilh a 0 = 7T/2 pulse. Equation (5.221) shows
-yHltl = 0, produced by a rotating field which at t = 0 lay along Ihe x-axis that any angle of 0 will do. [The fonnula (5.221) appears to blow up for sin 0 = 0,
according to the convention of (5.114). This is also the rotating field generated which simply reflects the vanishing of (Mz(t»o for that case].
by a linearly polarized field along the x-axis given by There are several very imponant advantages to the Fourier [Tansfonn ap-
proach. The first is that since one is recording the NMR signal when the HI is
Hz(t) = Hzo cos wt (5.221)
zero, one eliminates noise from the oscillator which drives the HI. For steady-
where, following (2.83), state apparatus, the oscillator noise is reduced by bridge balance, but typically
such an approach reduces noise which arises from amplitude modulations of HI
Hzo = 2H I (5.222)
but does not reduce noise which arises from frequency modulation.
182 183
If one is dealing with a complex spectrum which has many absorption If
lines well separated from one another, as in many high resolution spectra of
liquids, there is an additional advantage to the Fourier transfonn method. A
Vsig(t) = Vs cos (wst +,p)
steady-state search necessarily spends much of its time looking between the = Vs(cos ¢ cos wst - sin r/J sin wst) (5.226,)
resonance lines of such a spectrum. The search procedure is thus inefficient.
By pulse excitation, all the spins are induced to broadcast their presence at Vrcr<t) = Va cos (witt + 0) (5.226b)
their characteristic frequencies, thus one immediately gains in dam collection Then,
by eliminating the dead time of a steady-state search. Of course, one must use
a broader band system, since the steady-state system need only have the band Vout = Vs cos [(ws - wn)t + ,p - 0] (5.227)
width necessary to examine a single line whereas the pulse system needs a band In expressing the reference, we have intnxluced a phase angle 0 to represent the
width large enough to include the whole spectrum. However, since the noise fact that the phase of the reference can be set by the experimenter.
voltage goes as the square root of the band width, one still is better off with the Suppose 0 = 0, then
pulsed approach.
An important point is that the Fourier transfonn method is the basis for the (5.228)
enonnously powerful two-dimensional Fourier transfonn techniques which have At firs! sight, this expression appears to be equivalem 10 (5.226a). However, there
revolutionized NMR. is an ambiguity since the cosine is unchanged when the sign of its argument is
II is useful at this point to make several remarks about instrumentalion. The changed. Thus, one does not know whether wn < Ws or ws < Wit from recording
signal M:c(r) consists in general of a sum of oscillating signals (corresponding to (5.227) alone. To resolve the ambiguity, we record a second signal with (} = -1r/2.
each resonance frequency which has been excited) which decay as a result of line If 0 = -1r/2
broadening and relaxation phenomena. In general, all of the spread in oscillations
is small compared [Q the Larmor frequency. To record these frequencies takes Vout = Vs cos [(ws - wn)t + (<f; + 1l"12)] (5.229')
very high frequency response. It is customary to use a technique called "mixing",
which in essence translates the fTequency of oscillation while preserving phase = Vs sin [(WS - wll)t +,pl (5.229b)
relationships. We will not go into the details of how mixers work, but essentially Together, (5.228) and (5.229) enable us to determine both the sign of (wS - wn),
they work as shown in Fig. 5.6. The mixer has three pairs of terminals, one at and the phase angle ,p.
which the signal is applied, one to which a reference voltage is applied, and one Recording both such signals is called quadrature detection. It requires two
which is the output, the "mixed" signal. mixers, whose reference signals differ by rr/2 in phase, and the sels of recording
apparatus to record the output signals from both mixers.
In the example above we took the reference phases to be "0" and "-1rI2". To
Mixer apply (5.196) we must recognize that there is an absolme meaning to the phase
since we took the rf driving magnetic field to vary as cos wt in our definition
Signal Mixer of Xl and XII [see, for example, (2.92), (2.95) and (5.198)]. I{ is best, then, to
voltage signal
input o---+~ output consider thal the two reference phases are
(5.230)
where 0 1 is unknown until one goes through some procedure to set it. Using 01>
Reference voltage we then have from the first mixer
input
Voul = Vs cos [(ws - wn)t + ,p - Oil (5.231')
"'ig.5.6. Schematic diagram of a mixer. The radiofrequency magnetic resonance signal
is fed inlo the signal voltage input terminats; a referenee voltage whose phase can be and from the second mixer
adjusted is fed into the reference volLage input. The voltage at the mixer signal output
goes to recor<Hng equipment (oscilloscopes or multichannel digital recorders) Vout = Vs sin [(WS - wn)t +,p - Oil (5.23Ib)
184 185
These two outputs together enable one to detennine Ws and ; - 8 t . If one has a E p = (pl1tzIJI) = --yhHoJt = -1""'OIJ . (5.241)
procedure to detennine 81 in absolute tenns so that one can set it to zero, the cos
output gives the dispersion signal and the sin output gives the absorption signal. We use these as basis states to express e, (5.235), in matrix fonn:
A(I
dt pen I') L;:«P Ien III
P = '"i .... I
Ii )(/1 /I j'Hefflp)
5.9 The Density Matrix of a Two·Level System
"
- (/tl'Herrlpll)(Jt"lenll/» .
We have seen examples of the use of operator methods in conjunction with
A somewhat more compact natalion is useful, defined as
the density matrix to calculate the results of a sequence of pulses. These meth-
ods are extremely powerful and help us see what is happening in physical tenns. (1Ignl- ,) = g+- (5.242)
However, there is also great insight to be found in actually following what is
and so forth for the other matrix elements.
happening to the individual matrix elements of the density matrix for an example
Recognizing that
which we already understand thoroughly, the effect of an altemating field on a
spin! system. [The spirit is that of (5.143).] The Hamiltonian in the laboratory (pll,I"') = 0 (5.243)
system for this problem is then
unless p 1= p', we then get four equations such as
'H = -"(hHoI: - "'(1tH I (Iz cos wt - I,I sin wt) (5.232)
d"++ . [
where we assume a positive "( and a rotating field.
--:it = -lw\ g+_<-Il,I+) - (+ll,l-lg-+] (5.244)
As in (5.115), we write But (+IIzl-) = t, so that
'H. = -"(h(HoI: + e'loItl z Ize-iloltlz) (5.233) de++ iWI
Defining, as in (5.116),
dt = -T(e+- - e-+) (5.245.)
'86 '87
If we take the differences instead of the sums of the two pairs of equations,
(5.255b)
we get
Expressing these in matrix form, we get
~(E?++ - l?--) = -iwt(l?+_ - l?~+)
d' (5.249)
(M+(t)) = 1" L:(,.len(')I,/)(P'Ir+I,.) and
/,,/,'
(5.256)
(M,(I)} = 1/' L:("len(I)I,I)(/II,I") •
/,,/,'
These equations state that the off-diagonal marrix elements are "fed" by the action
of HI on the population difference l?++ - l?-_. By taking the lime derivatives which give, on utilizing the explicit matrix elements of 1+ and I:.
of (5.249), we get
(M+('» = 1"e-+(t) (5.257a)
Jl 2
dtZ(£,++-l?--)+wl(l?++-l?--)=O and
(5.250) (5.257b)
Jl 2
dt 2 (l?+- - e-+) + wI (e+- - l?-+) = 0 These equations show immedialely that the existence of transverse mag-
netization depends on the exislence of nonzero off-diagonal elements of l?R'
Therefore,
Moreover, the z-magnetization is proportional to the difference in the diagonal
l?++(t) - l?--(t) = A cos wIt + B sin wIt elements of l?R. hence it will vanish unless £,++(0 j. l?--(t).
(5.251) The significance of (5.252) and (5.253) is clarified by considering e(O) to
l?+-(O - l?-+(O = C cos wit + D sin wit correspond 10 thermal equilibrium, in which case the off-diagonal terms l?+-(O)
and l?_+(0) vanish. Note., incidentally that (5.73) requires that£,+_(t) and £,-+(t)
A and C are obtained by evaluating the left-hand sides at t = O. Band D
are complex conjugates of one another.
are found by evaluating the derivatives of the left-hand sides at t = O. utilizing
TIlen we get
(5.249). The result is
e++(') - e--(') = ["++(0) - e--(O)] cos WI' (5.258a)
"++(') - e--(I) = ["++(0) - e--(O)J cos WI'
- i[e+_(O) - l?-+(O») sin WI t
and
(5.252) (5.258b)
e+-(I) - e-+(I) = ["+-(0) - e-+(O)J cos WI'
If we let the HI stay on for a time tl. we gel, then, for all later limes t that
- i[e++(O) - E?-- (0» sin WI t
Combining (5.252) with (5.247) and (5.248) gives t = I.,,,[ l!++( 0) - l!-_(O) ] Sill
(M +(» . wlt l
Z
I 1 = i(M:(O» sin Wltl (5.259a)
E?++(t): '2 + '2 [e++(O) - e __ (O)] cos wit and
Suppose. now, we wish to calculate the expectation values of the x-. y-,
(My ('»= (M,(O» (5.260)
(M,('» = 0 •
and z-components of magnetization in the rotating frame. We utilize (5.159)
which agrees with the classical picture of rotation about HI by rr/2 in the left-
(5.255')
handed sense.
I ••
'.9
Following a "II" pulse (Wit) ="11") (nle'(O)lm) = (nle<O)lm)
(M.('» = (My(l)) = 0
(5.261) =0 unless n=m=k (5.265)
(M,(<) = -(M,(O» .
Again, the result agrees wilh Ihe classical result. (kle'(O)lk) = (kle<O)lk) = I
Equation (5.257a) Iclts us useful facts about Ihe off-diagonal elemenls of
ell: Then, taking the mm matrix elemen! of (5.111), and using (5.103) also, we have
a) fH- must be nonzero for Ihere to be transverse magnetization. d (mI'
d' e (t)lm) = d'd (mlel m )
b) u+- must be pure real for the transvcrse magnetizations to lie along the
c)
x-axis.
u+- must be pure imaginary for the transversc magnetitation to lie along = *~ !(ml"(O)I,:)(nll£jlm~ - !mll£iln)(nle'(O)lm)!
B
Ihe y-axis. A
Since use of an HI which lies along Ihe x-axis will produce components of
magnetization perpendicular to Ihe x-axis (assuming thennal equilibrium al t =
+ G)' j !mll"(O)1ti(t')1ti('~
0'"
- !ii(")e'(O)1ti('~
.....
0), it will necessarily produce a purely imaginary u+- no matter what angle, C D
Wltl, is chosen. To produce magnetization along the x-axis, HI mllst have a - l£j(<)e'(O)l£i(l') + l£i(<)l£W)e'(O)llm)d,'. (5.266)
component nonnal to the x-axis. Such a magnetic field will produce a u+- with , '" " ..... '
+ (mll£;('JIkXk/l£j(,')lm)jdt' (5.267)
S.lO Density Matrix - An Introductory Example
We convert Ihe malrix elements in the integrand by (5.107);
Since the fonnal equations of the density matrix get rather involved, it is a
good idea al this point to consider an example that will make ~he notation more (mll£; (')1 n) = c(;!')(Em - E. )'(mll£ I (<)In) . (5.268)
concrete. We shall calculate the probability per second of transitions from a slate
k to a state m. We consider thaI only state J.: is occupied at t = O. This assumption Wc shall also adopt a convenient abbreviation. We shall define the quantum
nU~bers m, n, and so on, in tenns of the corresponding energies, measured in
is not necessary, but it has the advantage, Ihat, with it, (dldtXmlulm) is directly
radians per second:
Ihe probability per second of a lransition.
Thus inilially all en's are zero except Ct;. As a result, at t = 0, the only Em Eft
nonvanishing element of the density matrix is (klulk), which is equal to I. By T sm T sk ~. B.~
using (5.103), we see that Thcn we have, using (5.268) and (5.269),
190 191
t
:'(mlelm)" ~2 J (ml1i,(t')lk)(kl1i,(t)lm)e'(m-k)(t'-t)
The funclion Gmdr) is called the "correlation function" of "HI(t), since it
tells how "HI at one time is correlated to its value at a later time. For a typical
o perturbation we have
+ (ml1i, (t)lk)(kl1i, (t') Im)e'(m-k)(t-t')Jdt' (5.270)
(5.277)
So far. our equations have been quile general as 10 the nature of the perturbation
1i 1(f). For example. it could vary sinusoidally in time. We shall assume for OUT Thus, if "HI (t) and "HI (t + r) were unrelated, we could average the two teons in
example, however,that1f\(t) varies randomly in lime. By this we mean thai we the product separately, gelling
shall consider a number of ensembles of systems of identical 110 and identical
(5.278)
e(O). However, we shaU consider that1iI(t) varies from ensemble 10 ensemble.
with properties described below. We shall therefore perfonn an average of the However, we have, for r =- 0,
ensembles. We denote Ihis average by a bar. Then (5.279)
For typical physical systems, the penurbation "HI (t) varies in time, owing to some
physical movement. For times less than some critical lime rc;, called the "correla-
tion time", the motion may be considered negligible, so that hl(t) ';::' "HI(t +1').
+ (ml1i, (t) Ik)(k 11i, (t') Im)e'(m-k)( t-t')ldt' (5.271) For r > Te , the values of "Hl(t+r) become progressively less correlated (Q "HI(t)
as l' is lengthened, so that Gkm goes to zero. Thus Gmk(r) has a maximum at
As an example, suppose 'Ji\(t) were the dipolar coupling betwccn nuclear mo- r = 0, and falls off for 11'1 > TC;, as in Fig. 5.7. A function "HI(t) with the above
ments in a liquid. It might vary in lime. owing to the thennal motion of nuclei propenies will be called a "stationary random function of time".
in the liquid. These motions would in general be different among various parts
of the fluid, all al Ihe same temperature. We shall assume, moreover, that the
ensemble average, such as
depends on t and e only through their difference 1', defined by Fig.S.7. Function G... t(T) (or
lypical physie,,1 s)'slem
II
t-t'=T (5.273) T
192 193
+=
~ (mlelm) =
1
h
2
J GlIlk(T)e-i(m-k)r = Wkm (5.281)
but the set of values covered remains unchanged. As a physical example, the
dipole-dipole coupling of a pair of nuclei depends on their relative positions. If the
-= nuclei diffuse relative 10 each other. the coupling takes on different values. The
where Wkm is the probability per second of a lransition from stale k to m. possiple values are independent of the rate of diffusion, since they depend only
Equation (5.281) is closely related to the well-known resuh from time- on the radius vector from one nucleus to the other and on the spatial orientation
dependent perturbation theory: of the moments. However. the duration of any magnitude of interaction will
-=
..-T,1ong
Jmk(W) may be thought of as the spectral density of the interaction matrix G mk .
We expect that Jmk will therefore contain frequencies up to the order of IIT c I
(see Fig. 5.8). In tenns of Jmk' we have
,
~ T" short
J J \
W :: Jmk(m - k) (5.28la) o Wl W
kill h2
Fig.S.9. Jml(l<oI) for three values of correlation lime, iIIustr"ting the variations that keep
Now. in a typical case, the interaction matrix element (ml1t 1(Olk) runs over the area under the curves const"nl as r c changes. Jml(l<oId is a maximum for lhc curve
of medium r c
a set of values as time goes on. As one vanes something such as the temperature.
the rate at which (ml1t\(t)!k) changes may speed upor slow down (Tc changes),
In the event thai Tc «::: I/(m - k). J",k(w) extends far above the frequency
of the transilion. It is then often a good approximation to say that
Jmk(m - k) '::! Jmk(O) (5.285)
lei us define a quantity 0' such that
+=
"'it:. 5.8. Typical spectral density plOl
2(~)Jm'(O) = J Jm,(w)dw (5.286)
-=
194 195
where 2(a/Tel is lhe width of a rectangular Jmk(W) whose height is Jmk(O) and Then we have
whose area is the same as that of the real Jml:(w), From our remarks, a ~ I.
Then we have, combining (5.284), (5.284a), and (5.286), W km = ;2 I:i:a,.,(m
, - k) (5.293)
196 197
We shall assume h z = h y = hz, so that 5.11 Bloch-Wangsness-Rcdfield Theory
h; = !hij (5.300)
We tum now to a more general treatment of the density matrix, following the
where hij = hi + h~ + h;. Then, since m - k = wo' the Larmor frequency ideas of Redfield [5.10], which are closely related to a treatment of relaxation due
to Wangsness and Bloch [5.11]. All of the basic ideas are anticipated physically
...!- = 2')'2 hij TO (5.301) in the basic work of Bloembergen et al. [5.12]. The development of the basic
Tl n 3 1 +wfiT6 equation of Redfield's theory is a generalization of the treatment given in the
This function is ploned in Fig. 5.10. We see that it indeed has the sort of variation previous section for computing transition probabilities. Redfield shows that the
we had predicted earlier from our general arguments based on the "constant area" elements of the density matrix obey a set of linear differential equations of the
under the curve of Jmk(W). We note thal the minimum comes when WOTO = 1. following fonn: 4
where the prime on the summation indicates that we keep only those tenns for
whieh 0' - 0" = {J - {JI. The diagonal pan of this equation (that is, the part we
should have if we kept 0' = ai, {J = {JI only) has the same fonn as the "master"
By utilizing this fact, we can calculate the minimum value of T 1 from
equation, (5.13). The conditions under which (5.303) or (5.304) hold are given
(5.301) and obtain, in agreement with (5.289)
in tenns of R OtOt '(j{3" T c , and a time interval .1t. .1t defines a sort of "coarse
_I) __I ')'~h~ (5.302) graining". That is, we shall assume that we Ilever try to follow the details of
T} min - 3 WO e· for time intervals less than .1t. We must be able to choose such a time,
Now, when TO gets very long, the field Hz simply gives rise to a static line subject to the simultaneous conditions that
broadening (in this case, since only two discrete values can occur, the stalic line .1t:» T c and (5.305)
consists of a pair of lines at ± h z ) and "&
is simply related to Ihis static width.
1
Of course, wo is known for a typical experiment. Equation (5.302) therefore -;;--'---- :» LIt (5.306)
tells us the remarkable fact that knowledge of the rigid lattice line breadth, R aa ',{3{3J
plus the resonant frequency, enables us to compute the most effective relaxation Equation (5.305) will permit us to set the limits of certain integrals as ± 00, as
possible from allowing the line broadening interaction to fluctuate. In general the we did in (5.281). Equation (5.306) will guarantee that, during .1t, the density
correlation time is changed by changing the temperature of the sample. Although matrix does not change too drastically, so thai our perturbation expansion has
(5.302) enables us to predict the minimum value of Tl' it cannot tell us at what validity. Since the R aa ',{3{3"s are comparable to the inverse of the relaxation
temperature the minimum will occur unless, of course, we know the dependence limes, these conditions are equivalent to saying that Tj and T2 are much longer
of TO on temperature. We note in addition that the measurement of T} as a than T e . These are also the conditions of motional narrowing. The reader will
function of temperature can provide infonnation about the temperature variation pehaps note that these are the very conditions stated in Section 2.11, which must
of whatever physical process produces the fluctuations. hold if simple time-independent transition probabilities are to hold. What we are
doing here is to generalize Ihe usual time-dependent perturbation theory so that it
includes the coherence effects associated with phase factors in the wave function.
4 For compactllCS5 in writing, we use the notation eo,," for the matrix clement
(aiel a').
198 199
The beauty of (5.304) is that it provides us with a simple set of linear differ- these equations are fewer in number than are the equations of the elements of
ential equations among the elements of the density matrix, which in principle we the density matrix, their solution may be considerably simpler than the original
can always solve. They wiJIlead to a set of "nonnal modes". We note that there set. This trick will work when the relaxation mechanism and operators are such
is a great deal of similarity to the rate equation describing population changes. thai the expectation values of the operators pick out only a small number of
Moreover, expressions for the Raot'(J(J"s are given by Redfield's theory, so that the possible nonnal modes. We shall illustrate this use of (5.312) shortly. First,
we have the relaxation times given in tenns of the atomic properties. however, we tum to a description of the derivation of Redfield's fundamental
Before outlining the derivation of (5.303) and (5.304), we remark that there equation.
are twO ways of utilizing (5.303) or (5.304). In the first method, one solves for the Our starring point is the basic equation for the time derivative of fl·, (5.111):
behavior of each separate element of the density matrix and then computes the
time dependence of the physical variables of interest (such as the x-components
of magnetization M z ) by the fundamental equation
d:" = *[,(0), :>lj(t)] + G)' Ja [[,(0), :>lj(t')], :>lj(')]dt' (5.313)
We then use (5.303) to express the time derivative (dldt){laQ'" In fact, since
matrix coincides at t = 0 but whose penurbations 1i1(t) are different. (We are
therefore flot allowing an applied alternating field to be present. That is, we
are describing relaxation in the absence of an alternating field. One could add
the effect of the alternating field readily. We discuss it in the next section.) We
(5.309) shall assume that the ensemble average of 1i 1(t) vanishes. This amounts to our
assuming that1iI(t) does not produce an average frequency shift.~
we have Let us discuss this point. In general we expect
d{l~a' = i[a _ 0"]. + ei(a-a')! d{l(.ta' (5.310) (5.315)
dt fl aa' dt
This relation enables us to transfoml (5.303). By substituting into (5.303) and where J(q is a function of the spin coordinates, and Hq(t) is independent of spin.
utilizing (5.309), we find For example, we saw that hi (t) had this fonn if it consisted of the coupling of a
fluctuating magnetic field with the X-, y-, or z-components of spin. In that case q
d{lQ'a'
= i(a' - a){laa' + L RaQ" ,{3{3' {I{3(J' took on three values corresponding to the three components x, y, and z. If hI (t)
d' P,P' represented a dipole-dipole interaction of two spins. there would be six values
, of q corresponding to the tenns, A, B, ... , F into which we broke the dipolar
= -,. [I?, 7-lo]Q'Q" + L Roo',pp'I?PP' (5.311) coupling in Chapter 3.
l P,P' Since we are dealing with stationary penurbations, the ensemble average
We then substitute this expression into (5.308), obtaining of 1ij(t) is equivalent to a time average. In general we assume that the time
average of Hq(t) vanishes, causing 1ii(t) to have a vanishing ensemble average.
d(Mz )
---=
d'
L:
a,Q",pjJ'
(*[e, :>lo] ••' + n••"pp,gpw)(a'IMz1a) (5.312) As a consequence we shall set
(5.316)
Although it is not obvious from (5.312), under some circumstances the right-
hand side is proponional to a linear combination of (Mz(t», (My(t», and ~ If there is a shifl, we <;ao include the lwerage shifl in ho, redefining 1-£1 (t) lo give a
(AdAt», giving us a set of differential equations similar to those of Bloch. If zero shifl.
200 201
=~"...."...,. . . .
where the bar indicates an ensemble average. This means, we have remarked,
that we cannot let 1-£1 (t) be a time-dependent driving field such as that applied Lqq,(w):= J Hq(t)Hq,(t + T)e-i... r dT (5.324)
to observe a resonance. o
On the basis of (5.316) the first teon on the right of (5.313) vanishes when By utilizing the fact that Hq(t)Hq,(t + r) is real' and is an even function of
averaged over an ensemble. T, it is convenient 10 define the real and imaginary parts Lqq,(w):
We proceed in a similar way to compute the aa' matrix element of the ,+=
second teon on the right of (5.313). By utilizing the fact that Re {Lqq,(w)} = 2" J·H·,"(t~)H~,-,(~t-+-T') cos (wT)dT := kqq'
<P1?i; (tll/l') = e'(ft-ft')'<PI?i, (tlIP') (5.317) -=
(5.325)
and defining
T=t-t'
1m {Lqq,(w)} = -
=~~~~~
J
Hq(t)Hq,{t + T) sin (wT)dT
(5.318) o
Since the only imponant contributions to (5.319) come from teons that satisfy
the condition a - a' = fJ - If, we can combine the first twO teons on the right
of (5.319) as
202
203
(5.334)
Introducing spin quantum numbers sand !l, and lanice quantum numbers / and
(5.330) /', we replace 0' by sf, and so on. Then we assume that the lattice remains in
thermal equilibrium despite the spin relaxation:
Equation (5.329) relates de"ldt at time t > T( to eO al t ::: O. II is the first
leon in a power-series expansion. In order for the convergence to be good. it e-"/IkT
(5.335)
must imply that f!ppl at time t nOI be vastly.different from this value al t = ~. ll}1' = Off' Ee "/"/kT
This implies thai we can find a range of tImes t such that t:> Tel but sull I"
1!~1J,(t) ':l!' Upp,(O). This laner condition implies that
We then find the differential equation for
1 d L• • I..d.
~--'--- :» , (5.331) r
d/ ll /I'uu ') = o/f'llfl' dtan' (5.336)
R".. ,~W
The important trick now is 10 nOle thai if (5.331) holds true, we can replace and sum over f. The result, in the high temperature limit, is simply to give a
modified version of Redfield's equation, with density matrix for spin a replaced
l'flIJ'(O) by i!P13,(t) on the right side of (5.329). By this step, we convert (5.329)
by the difference between u and its value for thennal equilibrium at the lattice
into a differential equation for [/, which will enable us 10 find [/ by "integra-
temperature q(T).
tion" at times so much later than t <:: 0 that l?PIJ,(t) will no longer be nearly its
We therefore simply assert that for an intcraction in which the lattice couples
value at t = O. The resulting equation is (5.303).
to the spins via an intcraction 'HI (which, to the spins, is time dependent),7 the
The physical significance of our conditions is now seen to be that we never
role of the lanice is to modify the Redfield equation to be
ask for infonnation over time intervals comparable 10 Te, and that in this time
interval the density matrix must not change too much. In practice. this implies du;o-' ~ +i(Ol-Ol'-.B+.8')t.. (5.337)
- d - = L. Roo-',fJ.B,e [ufJ.B' - u.BfJ,(T)]
that
t .oft'
TI. T2 >Tc (5.332) where a,a',fJ,' stand for spin quantum numbers, and where u.8.8,(T) is the
As we shall see in greater detail, the condition Tc < T2 is just that for which thennal equilibrium value of ufJ.8':
the resonance lines are "narrowed" by the "mOlion" that produces the fluctuations
C"/J/ kT
in 'HI (t).
"~~,(T) = 6~W Le WIlT (5.338)
Because
~"
~oJJfJ = R/J/J,OOl (5.333)
That (5.337) should hold true is not surprising in view of our remarks in Chapter
(that is, the transition probability from 0' to fJ is equal to that from fJ to 0'), I concerning the approach to thennal equilibrium. We note here. however, that
the solution of the Redfield equation is an equal distribution among all states. our remarks apply not only to the level populalions (the diagonal elements of u)
This situation corresponds to an infinite temperature. Clearly the equalions do but also to the off-diagonal elements.
not describe the approach to an equilibrium at a finite temperature. The reason is
immediately apparent - our equation involves the spin variables only, making no
mention of a thennal bath. The bath coordinates are needed to enable the spins
to "know" the temperature.
A rigorous method of correcting for the bath is to consider that the density
matrix of (5.313) is for the total system of bath and spins. Since in the absence
of HI the spins and lattice are decoupled, we may take the density matrix to
7 "HI involves both spin Rnd lattice coordinRlcs. If we treat the lattice quantum me-
consist of a product of that for the spins, a, and that for the lattice, ~L. We chanically. the lattice variables are operators, and 'HI dOClS not involve the lillle explicitly.
take for our basic Hamiltonian '110 the sum of the lattice and the spin Hamil- If we treal the lattice classically, 'H\ ;n\'oln'$ the time c:o:plicity. Thill this must be so is
tonians (which, of course, commute). 'HI commutes with neither and induces evidcnt, since the coupling must be time dependent to induce spin transitions between
spin stales of different energy. However, it is time indcll·cndent when the lattice ml'lkcs a
simultaneous transitions in the lattice and the spin system. Then we have simultaneous transitiOIl that just absorbs the spin energy.
204 205
5.12 Example of Redfield Theory where q = x, y, z, and
1to = -"(n/tHol: = -liwolz (5.340)
We tum now to an example to illustrate both the method of Redfield and some
simple physical consequences. The example we choose is Ihal of an ensemble where. wo is the Larmor frequency. We characterize the eigenstates by a, the
of spins which do nOI couple to one another but which couple to an external eigenvalues of (5.340). These are wo times the usual m-values of the operator
fluctuating field, different al each spin. The external field possesses X-, y-, and I z . (Herem = I, I-I, ... -I). We shall continue to use thenotalion a,d,fJ,fJ',
z-componenls. This example possesses many of the features of a system of spins however, rather than m, in order to keep the equations similar to those we have
with dipolar coupling. However, it is subslanlially simpler to lTeal; moreover, it just developed. The matrix elements (al1tI(t)la') are
can be solved exactly in the limit of very short correlation time. For the case
of dipolar coupling, Ihe nUClUations of the dipole field arise from bodily motion (al'H,(')la') = -,"h L:Hq(')(allqla') . (5.341)
q
of the nuclei, as when self-diffusion occurs. The correlation time corresponds 10
the mean time a given pair of nuclei are near each other before diffusing away_ Then the specrral density functions JOtpo'P'(w) are
OUT simple model gives the main qualitative features of the dipolar coupling
if we simply consider the correlation time 10 correspond to that for diffusion.
In panicular, then. our model will exhibit the imponant phenomenon of mo-
tional narrowing, which has been so beautifully explained in the original work
of BJoembug~n et al.
Before plunging into the analysis. we can remark on cenain simple features
that will emerge. At the end of this section we develop these simple arguments
funher. showing how to use them for more quantitative results. We now use the symbol kqq.(w) introduced in the preceding section as
We may distinguish between the effects of the Z·, y-. and z-eomponents
+~
of the fluctuating field. A component Hz causes the precession rate to be faster
or slower. It. so to speak. causes a spread in precessions. It will evidently nor kq¢(w) = 4J 'H'q"(")H"q-,T.('--:+-:T">e- iw.. d, . (5.343)
contribute to the spin-Ianice relaxation because that requires changes in the com· -~
ponent of magnetization parallel to Ho. but it will contribute to the decay of Clearly the fluctuation effects, correlation time, and so on are all associated
the transverse magnetization even if the fluctuations are so slow as to be effec- with the kqq"s. For simplicity let us assume that the fluctuations of the three
tively static. In fact, as we shall see, it is Hz that contributes to the rigid-lattice components of field are independent. That is, we shall assume
Line breadth. The phenomenon of motional narrowing corresponds to a son of
averaging out of the Hz effect when the fluctuations become sufficiently rapid. Hq(t)Hq,(t + r) = 0 if q "" q' (5.344)
The z- and y-components of fluctuating field are most simply viewed from For example, (5.344) will hold true if, for any value of the component H q ,
the reference frame rotating with the precession. Components fluctuating at the the values of H q• occur with equal probability as IRq'1 or -IHq,l. We note
precession frequency in the laboratory frame can proouce quasi-static components that kqq(w) gives the spectral density at frequency w of the q-component of the
in the rotating frame perpendicular to the static field. They can cause changes in fluctuating field. With the assumption of (5.344) we have, then,
componenls of magnetization, either parallel or perpendicular to the static field.
The former is a T 1 process; the latter, a T2 process. Clearly the two processes are (5.345) "
intimately related, since the magnetization vector of an individual spin is of fixed
length. The transverse componentS of fluctuating magnetic fields will be most We now seek to find the effeci of relaxation on the X-, y-, and z-components of
effective when their Fourier spectmm is rich at the Larmor frequency. For either the spins. To do this, we utilize the second technique described in the preceding
very slow or very rapid motion, the spectral density at the Lannor frequency is section, that of finding a differential equation for the expectation value of the
low, but for motions whose correlations time, is of order I/wo, the density is at spin components. Let us therefore ask for (dldt)(Ir ), r = x,y, or z. By using
a maximum. The contribution of H~ and H y to the longitudinal and transverse (5.312), we find
relaxation rates therefore has a maximum as , is changed.
Let us consider, then, an interaction 1t1(t) given by d(I,)
- - = "
L. i h[ ll,1to Joo,,(a' I.. 0'l l+) "L. Roa',f3fJ'llf3J3'(U ' IIrla)
'H\(t) = -,"h L:Hq(t)Iq (5.339) dt 0,0' o,o',P,P' (5.346)
q
206 207
The first tenn on the right, involving 'Ha, can be handled readily: +1'~ :L ({3IIq]a)(aleUqlr - Irlq}IP)kqq(fJ - a)
cr,fJ,q
L , 2.r.Q.
h
'H'o)Cf(f,(a'IIr 10) = _,i Tr {(g1iO
,
-1toe)Ir}
=.,~ I: (Wqlo)(ol[[I" IqJ, eIIP)kqq(P - 0) (5.352)
0,.
O',fJ,q
j
= Ii Tr {eHoI, - e I , Ho} where, in the last step, we have utilized the faci that kqq(w) is an even function
j of w.
= Ii Tr {e[Ho,!,]} To proceed further, we must now specify whether r is x, y, or z. First let
= -i'YnlIo Tr {e[I~, I r ]} (5.347) us consider r = z. Then, since lr will cornmUie with [z. we get nothing from
q = z in the last line of (5.352). Since matrix elements of ['" vanish except
If r = z, this term vanishes. If l' = x, we have for L1m = ± I, the only states (0' and (3) that are joined by I q for q = x have
-i1'nHO Tr felIz, I:t'J} = -i1'nHO Tr {ily!?} 1,6 - al = wo, the Larmor frequency. Since [Iz.I z ] = iIy and [Iz,I y] = ilz , we
have, then,
= +1'n Ho(Iy ) (5.348)
,~ DW, 10)(011[/" I,lellPlk,,(o - P)
If r = y, we get --YnHo(l:t'). Thus we have
',p
L
0',0"
*[e, 'h'o]O'O',(c/llrlo) = 1'n«1) x HO)r (5.349) = ,,~[ I:<PII, 10)(0IiI" - ie1yIPl]k,,(wo)
',p
which is the driving tenn of [he Bloch equations describing the torque due to the = i1'~ku(wo)Tr {I:t'Iye - I:t'ely}
external field. The second tenn on the right of (5.346) involves the relaxation = i-y;ku(wo)Tr {U:t'I y - IyI:t')e}
terms;
= --y;k:t':t'(wo)Tr{IzU}
L RO'O",(3(3'!?(3/3,(O"IIr la) . (5.350)
= -1'~k:t':t'(wo)(lz) (5.352a)
0',0" ,/3,(3'
The tenn q = y gives, in a similar manner,
As we have seen, RCl'O" ,(3(3' is itself the sum of four tenns [see (5.330)]. We shall
discuss the first teon. JO'(3O"(3'(O" ~ 13'). By using (5.345), we find -1';kyy (wo)(Iz) . (5.353)
All told, then,
~ L JO'(3O"/3'(o/ - ,O')e(3(3' (0"1 Ir 10')
2h 0',0",(3,/3' L RO'O', '/3/3'!?/3/3'('/ 1I z 10) = --y;( k:t':t'(wo) + kyy(wo) J(I z ) . (5.354)
= 1; I: (oIIqIP)(P'IIqlo')(P1dP')(o'II,lo)kqq(o' - P') 0',0" ,(3,(3'
O',O",/3,(3',q By combining (5.346), (5.349), and (5.354), we have
"" -y; L ({3'jlq]O")(o'jlrlq!?I,O')kqq(o' - {1) (5.351)
O",(J' ,q
z
d(I ) "" -y«1) x Ho): -1'~[k:t':t'(wo)+ kyy(wo)](Iz) (5.355)
dt
where the last step follows from the basic properties of orthogonality and com- This equation relaxes toward (Iz ) "" 0 rather than the thennal equilibrium value
pleteness of the wave functions 10), and so on. We are able to "collapse" the 10· To remedy the silUation, we should replace U by e- f!{T), as discussed in
indices 0 and .0. but we cannot do the same trick for 0' and p' because they rhe preceding section. This result simply makes (Iz) relax toward the thennal
occur not only in the matrix elements but also in the kqq's. equilibrium value 10:
In a similar way one can obtain expressions for the other three [enns in
RO!O!',(3(3" getting finally d~,) = 1((1) x Flo), - 1~[k,,(wo) + kyy(wo)I«I,) - 10) (5.356)
L: RO'O', ,(3(3' e(3(3' (ollIr 10) This is clearly one of the Bloch equations, with T 1 given by the expression
0',0" ,(3,/3'
= 1; I: (PIIqlo)(oII,Iq - IqI,)eIPlkqq(o - P) (5.357)
O!,(3,q
208 209
We can proceed in a similar way to find the relaxation of the x-component. dMx Mz dM y My
For it, the value q = x contributes nOlhing, but q = y or z does. The situation & = woM:l: - T2x ili = -woM:l: - T2y (5.363) '1
for q = y is similar to the one we have just discussed, connecting states 0' and
{J which differ by woo On the other hand, when q = z, the states 0' and {J are Using (5.363) and (5.362b), we get
the same (Iz is diagonal), so that 0' ~ (J = O. The spectral density of Hz at zero dM1.. M2 M2
frequency enters. Therefore we find M ~ -dt = lvfxwoMy - _z - MywoM:l: - - "
T2z T2y
L R ua' ,PP'(JPfJ,(O"II:l: 10') = -'Y~[kyy(wo) + ku(O)](I:l:) (5.358) M; MJ
u,u',fJ,P' (5.364)
= -T2x - T2y
which gives for the derivative of (I:l:):
We now introduce the precessional motion by replacing M:l: and My
d(l.) ,
----;u- = 1'n«1) x NO):l: - 'Yn[kyy(wo) + ku(O)](I:l:) (5.359) M x = M1.. cos (wot + 4»
(5.365)
My = -lvh sin (wot + 1')
For this equation there is no effect of replacing (J by (J - (J(T), since in
thennal equilibrium (I:l:) = O. Equmion (5.359) and a similar one for (Iy ) look in (5.364), getting
very much like the transverse Bloch equation describing a T2 process, except
that the T2 for (I:l:) differs from that for (Iy ). Labeling these as T2:l: and T2y lvh dM1.. = -Ml (COS 2 (wot + r/J) + sin 2 (wot + r/J») (5.366)
we get ~ ~x ~y
1 Averaging this equation over one period, we replace cos 2 (wot+r/J) and sin 2 (wot+
= 'Y~[kyy(wo)+ k:z(O)]
1',.
- (5.360a) !
r/J) by getting
1
-1' = 'Y~[k:r::l:(wo) + kzz(O)] (5.360b)
dM1..
dt
= -M1.. ~(_I_ +
2 T2z T2y
_1_) = _ M1..
T2
(5.367)
"
If we did not have the driving tenn from Ho, these equations would cause -!... _ ~ (_1_ + _1_) (5.368)
(I:l:) and (I y ) to decay with the rates T2z and T2y respectively. However, in T2 - 2 T2z T2y
general the precession rate wo is much faster than I/T2 :r: or I/T2y. Thus, we
must average the T2 effects over the precessional motion. That is, the amplitude Combining (5.360) with (5.368), we see that
of the transverse magnetization will decay slowly compared to the precession 1 1
frequency. T2 = 2Tl + 'Y;kzz(O) (5.369)
h is straightforward to detennine this rate. We define the transverse magne- Our relaxation mechanism therefore leads to the Bloch equations. Of course
tization M1.. by the equation one cannot expect that in general the Bloch equations follow from an arbitrary
M1.. = iMx + jM y so that (5.361) 7-tl(t), and one would have to study each case to see whether or not the Bloch
equations resulted.
To proceed further, we need to know something about the spectral densities
of the x-, y-, and z-components of the fluctuating field. We shall once again
Theo assume a simple exponential correlation function, with the same correlation time
7"0, for q = x, y, and z:
dMl=2M dM1.. (5.362.)
dt 1.. dt Hq(t)Hq(t +7") = H~exp(-I7"I/To) (5.370)
which gives
=2M dM:l:+2M dM y (5.362b) - TO
Zdt Ydt kqq(w) = Hi 2 2 (5.371)
I +W 7"0
From (5.359) and (5.360), we get
In tenns of (5.371) we have, then,
210 211
f
relative to its neighbors by diffusion. In Ihe time T, a spin will precess an extra
phase angle 6¢ over its nonnal precession:
6¢ = ±l'nlH.:IT (5.315)
(5.372)
After n such intervals, Ihe mean square dephasing tJ.¢2 will be
one rotating at the Larmor frequency), these results are reasonable. since the T, We note that the shorter T (Ihal is, the more rapid the mOlio~), the narrower Ihe
corresponds to a change in the z·magnetization. Such a change is brought about resonance. This phenomenon is therefore called motional narrowing. We see that
by "static" fields in either the z- or y-directions in the rotating frame. since in Ihe motion narrows Ihe resonance because it allows a given spin 10 sample many
the rotating frame the effective field is zero (HI. of course, is zero). But "static" fields H~, some of which cause ilia advance in phase; olllers, 10 be retarded. The
x- or y-fields in the meating frame oscillate at WQ in the laboratory frame. dephasing takes place, lIlen, by a random walk of small sleps, each one much
On the orner hand, lhe decay of the x-magnetization must arise from the less Ihan a radian.
"static" y- or z-fields in the rOiating frame. Since the z-axes of the laboralory and In conttast, when there is no motion, a given spin experiences a constant
rotaling frames coincide, for the z-field it is the static laboratory component that local field. II precesses either faster or slower than the average, and Ihe dephasing
counts, but for the y-field, il is the laboratory component at the Lannor frequency of a group of spins arises from Ihe inexorable accumulation of positive or negative
Ihal is important. We note that in the limit of very rapid motion (wOrD <:: 1), and phase.
assuming an isotropic fluctuating field The contraSI with "collision broadening" of spectral lines is greal. In thai
H'• • H'y = H'• (5.314) case Ihe phase of the oscillation is changed by each collision. Since the frequency
is unperturbed belween collisions, Ihere is no loss in phase memory except during
Tl and T2 are equal. Physically, for our model, this result signifies that for a very a collision. Since each collision gives a loss in phase memory, a more rapid
short correlation time, the spectral density of the fluctuating field is "white" to collision rate produces a shorter phase memory and a broader line. With motional
frequencies far above the L1.rmor frequency, so that the x', yo, and z-directions narrowing, there is no phase change when H r is changing from one value to
in the rotating reference frame see equivalent fluctuating fields. another because the change is very rapid, but lIlere is a phase change during
The two terms in the expression for T z have simple physical meanings. One the time a given value of H r persists. More rapid motion diminishes the loss in
tenn depends on H~. It represenls the dephasing of the spins due to the spread phase memory in each inlerval.
in precession rates arising from the fact that H ~ can aid or oppose Ho. This term We have considered just one (enn of our expression for T2. The other term,
can be readily derived by a simple argument, which we give below. The second
term, as we shall explain, results from broadening pf the energy levels due to
which involves H;,
clearly has the same dependence on TO as does the spin-
lattice relaxation. We interprel it as the bro,1.dening of the line due to the finite life
the finite Iifelime a spin is in a given energy state. of a spin in any eigenstale as a result of the spin-Ialtice relaxation. The lifetime
Let us now tum to a simple derivation of the first term of the equation for
T2. We assume the field as a value lH~1 for a time r; then il jumps randomly 8 Of course the word ~motioll~ here refers to tn.nstation of the position of the nudeWl,
to ± IH~I. Such a change in field in practice arises because a nucleus moves not to the change in spin orientation.
'I' 213
is finite because a field in the V-direction can change the .z-magnetization. We comes in. For example, when the relaxation arises from the coupling of one
should estimate the order of magnilUde of the lifetime broadening to be nucleus with another by means of their magnetic dipole moments, the E and F
terms of Chapter 3, which involve the product of two raising or IWO lowering
.cl.E =!!... or (5.380) operators, connect stales differing in energy by 2hwo.
T, Our formulas show us that the measurement of T 1 and T2 will enable
c,E I us to detennine TO. When the fluctuations in the interaction 'Ht(t) arise from
L!w=-=-
h T, bodily motion that varies with temperature, we can use resonance to study the
Assuming isotropic fluctuating fields, we see that our example actually gives temperature variation of TO. Often one has a "barrier" to motion and an activation
energy E such that
_I = _I + _ 1 (5.381)
T2 T 2, 2T1 (5.384)
where T2, is the broadening due to the spread in the .z-field. The quantity 1fl'2', where TOO is the value of TO for infinite temperature. The temperature variation
is often called the sUlllar broadening; the term lilT) is called the nonsecular of T t or T2 gives one a convenient measure of E and TOO. The narrowing
or lifetime broadening. More generally, we replaced lilT) by IIT I" where T I, smdies of Andrew and Eades [5.13) performed on molecular crystals, provide
(which gives the nonsecular broadening) is related to T). one such example. Another one is the work of Holcomb and Norberg [5.14) on
As we have remarked, if we consider the secular broadening, we nOte that self-diffusion in the alkali metals, and subsequently, Seymour [5.15] and SpoJcas
as TO decreases, T2 increases, oc the line narrows. Of course, we have seen that (5.16) on aluminium. Here one has the interesting fact that, using resonance,
as one increases TO (slows the motion), the validity of the Redfield equations these workers could measure the self-diffusion rate of both Li and AI (for which
ceases when T2 ";:! TO. For longer TO'S, we cannot apply the Redfield equation. there is no radioactive isotope for use in the conventional tracer technique).
The longest TO for which the Redfield theory can apply, then, is TO = T2 , or
_1 = _1 , 82TfJ
= "'( - or (5. 382) 5.13 Effect of Applied Alternating Fields
TO T2 nz
7n!H,lrO = 1 (5.383) So far we have excluded applied alternating fields from the time-dependent cou-
As we can see from our simple model, Ihis is just the value of TO at which a pling 'H)(t). Let us now assume that such fields are present, giving an extra tenn,
'H2(t), in the Hamiltonian. We can include its effect in a straightforward way,
Iypical spin gets one radian out of phase before there is a jump. For longer TO'S,
the spins are dephased before there is a chance for a jump. That is, they do not as has been shown by Bloch. Introduction of 1t2(t) simply replaces 1tj(t) by
dephase by a random walk. The line breadth is then independent of the jump Hj(t) + Hi(t) in (5.313):
rate, giving one the temperature-independent, rigid-lattice line breadth. d' .
The twO contributions to the line breadth (secular and nonsecular) are plotted :, = *le'(O),Hj(t) + H;<t)]
. 2'
in Fig. 5.11.
If one analyzes the relaxation via other mechanisms, the same general fea- + (*) Jo 11.(0), (Hj(t') + H;<t'))], (Hj(') + Hi(t))]d"
(5.385)
tures are found. The fact that more than one transition may be induced will make
important the spectral densities at frequencies other than 0 and wOo Often, 2w0 The effect of Hi in the first tenn on the right vanished when averaged over
an ensemble. For that reason we were forced to consider the second tem. In
general the contribution of 'H 2(t) to the first term does 1I0t vanish, since 'H 2(t) is
/11 ..!.... T', identical for all members of the ensemble. If 1t2(t) is not too strong, we expect
T',
0' Ihat the first-order tenn is all that is necessary, and therefore we neglect the role
t
1" - of 'H 2 in the integral. Physically, this approximation amounts to our saying thllt
T',
Secular (1~) and non$OCu-
Fig. 5.11.
dt/ de·) de')
lar (Tn broadening versus TO. For (5.386)
the eXilmple in the led, T[ =2Tl dt = dt 1f. 1
+ ill n'!lax
Toe!l/wo I where de*/dt)1f.2 is the rate of change of e· due solely to 'H2, and de*/dt)relax is
Toe!I/'r"IH.1
214 215
the rale of change of r/ we should have if 1t2 were zero. What we are neglecting dM M x H()
are, therefore, nonlinear effects in the interactions. --=/ t+ Mo-M where (5.390)
dt Tj
Under what circumstances do we expect this approximation to work? The
answer is that neither perturbation must change 'l too much during the time Mo = xoH(t) (5.391)
t (which is the upper limit of lhe integral), for the 'Hi [enns in the integral
and H(t) is Ihe instantaneous applied field. Explicit solution shows that the
represent the fact that 'Hi
is acting on a (/. which is not in fact e-(O) bUI must
solution of (5.390) differs significantly from the nonnal Bloch equations (assum+
be corrected for the driving by 'Hi. Since we wish to choose t as about Te. this
ing T 1 = T2, but Mo = xoMo) only when the line width is comparable to the
requirement means that
resonance frequency.
1("I?t,I,,')IT< <0:' (5.387) We can combine (5.386) with (5.311) to obtain the complete differential
equation for the density matrix, including an applied alternating field:
"
If 1t2 is too large to salisfy (5.387), we should then attempt 10 solve first for the
combined effect of 'Ho and 'H2. using perturbation theory for 'HI- For example, dUoot'
- i
d - = -,,(Eo-' - Erx)l!rxtY L.. (l!rxtY'(o "I 112(t) I0)
+ -,i '" '
t l 0"
we may use a rotating coordinate transfonnation rather than a transfonnation 10
the intemcrion representation, thereby converting 11i(t) to a static interaction, - (,,(?t,(')I,,")eo""'] + L Ro"'JJ~'[e~p' - e~p'(T)J (5.392)
which could then be removed by a second transformation to the interaction ~JJ'
representation of the effective field. where, for l!fJIP(T), we use either (5.388) or (5.389), depending on the circum-
It is interesting to note that there is in fact a very close similarity between stances.
the interaction representation and the usual rotating coordinate transformation In order to make (5.392) more concrete, let us suppose we have a two-level
that renders H I a static field. Both are transformations to rotating coordinate system. Thus we may have a spin !
particle quantized by a static field in the
systems. The interaction representation is a transformation to a system rotating z-direcrion. We label the states 1 and 2 and find that the density matrix has four
at the Lannor frequency, whereas the usual transformation goes to a coordinate elements Ull> 022, un, U'2t·
system rotating with H t . In the relaxation tenns Ra~ ,/Jll', the only tenns of imponance (as we have
For simplicity we shall assume that (5.387) is satisfied. Note that since T, seen) involve cr - crl = fJ - If. The only tenns Ihal count are therefore
and T2 are much longer than T c for all our equations to be valid, (5.387) can
still be easily satisfied evcn under conditions of saturation. R Il ,22 = R22,1I == I/TI
(5.393)
There is one further consequel'lCC of the addition of a tenn 112(t). We have R12,12 = R21,21 == - 1/1"2
remarked that when we treat the latticc classically, the density matrix relaxes
By assuming 1i2 joins states 1 and 2 only, and denoting (l11i2(t)12) by 1i12(t),
to its value at infinite temperature rather than to the thennal equilibrium value
we find
e!:T)o
dUl1 dU22
f!{T) = _'_e-1{,/"· (5.388) d. =-dt"
Z(T)
=
t!22 - ",1 - [m(T) - ell (T)] + ii; ( UI2 ~
1L21
(t) ~ (t)n )
- Iq2 c:21
where Z(T) is the sum of states. When the change of 1i2 is small during the
time T e , which characterizes the "lattice" motion, we expect that 1i2 looks like a TI l (5.394)
"static" coupling to the lanice and that we should assume that the system relaxes
del2 Ul2 i i
,
at each instant toward the instantaneous density matrix:
216 '17
If we write 6. Spin Temperatnre in Magnetism and
'H2(t) = V cos wt (5.397) in Magnetic Resonance
where V is an operator, and define Wo by the relation E 2 - El == hwo, we have
in the limit of small V (that is, no saturation)
e-El/ kT 1 -El/kT
1'" = ell(T) = e EI/kT + e E21kT ';::' -e
2
IT2
)
WOT2 ell
V I 2W O
e22
218 219
6.2 A Prediction from the Bloch Equations solids but not for liquids? Redfield has given an explanation. We will discuss
the conditions for validity of the Bloch equations later in this chapter, after we
Let us consider a simple resonance experiment with a rotating magnelic field of have discussed some important background material on spin temperature in cases
angular frequency w. lI"ansverse to the stalic field B o• tuned exactly 10 resonance where there is no alternating field HI present.
w: ,Ho (6.2)
We discuss it by means of the Bloch equations. 6.3 The Concept of Spin Temperature in the Laboratory
It is convenient 10 transfonn to a reference frame rotating at w with HI Frame in the Absence of Alternating Magnetic Fields
along the x-axis as is done in Sect 2.8. Exactly at resonance. the Bloch equations
become Let us now tum to a discussion of the application of the concept of spin tem-
dM~ Mo - M~ (6.3a) perature to magnetic experiments not involving resonance. A typical system we
----;It "" -..,MyH I + Tl might consider is a group of N spins of spin I, gyromagnetic ratio '1, acted on
by an external field Ho, and coupled together by a magnetic dipolar interaction
dM~ Jkrz (6.3b)
-;U""-T2 represented by a dipolar Hamihonian Hd' We denote the Zeeman Hamiltonian
by Hz. The solutions of the Schmdinger equation are then wave functions tPn
dMy "" -vM HI _ ~ (6.3c) of energy Ell of the total system.
dt ,: T2
Suppose we now orient M along HI sO that at t "" 0 Mz "" Mo. Jvly "" M: "" O.
'li",,: ('liz + 'lid)"" : E""" (6.6)
From (6.3b) we see that M z will decay 10 zero in a time T2 • The low HI steady- Unfonunately, (6.6) is exceedingly difficult to solve, depending as it does on the
state solution of the Bloch equations shows Ihat they describe a Lorenlzian line coordinates of 10'22 spins.
with a frequency widlh One can assume, however, that if the spin system is in thennal equilibrium
with a reservoir of temperature 8, the various states n of the total system would
I
&:- .~ be occupied with fractional probabilities PII given by the Boltzmann factor
T,
For solids typically , I'll "" .!..e-
Z
E,,/k8 (6.7)
Llt..,) ';!' '1Hnei,hbor ~ '11l ~ '1.-../1(/
J
+ I)/Z (6.5) where Z is the panition function
(tJ a
where H ~gl . I bo r is the nuclear magnetic dipole .field due to neighbors, and a is (6.8)
. .
the distance to the Z nearest neighbor. For typical solids Llt..,) IS of order of a "
few tens of kilocycles (e.g. .t1w S! 211'" X lO kc for AI metal). Quantities such as the average energy E and magnetization M ~ would then
Funher examination of (6.3a) and (6.3c) shows that they do not involve M~ be given by
and that if M and M z are initially zero they would remain zero were it not for
the tenn invofving T t . If T} ::> T2, therefore, and for times up to about T I after E:E~& ~~
M has been orienled along H I, we still have My "" M: "" O. "
Therefore these eqlw.tions predict that wilen M is aligned along HI, it wUl (6.9b)
decay to zero in a time T2, typically of order 10- 4 to lO-5 sec. "
Experimentally this prediction (decay in T2 when M is along H) is found As we have remarked, Van Vleck recognized thai expressions such as (6.9)
to be correct for liquids, but it is not correct for solids. Rather, for solids it is could be evaluated wilhout solving the Schmdinger equation because they could
found that as long as HI is turned on and is sufficiently strong, the decay rate be expressed as traces. For example we can express the panition function as a
of M~ is much more like T, than a time T2 which characterizes the line width. lrace as follows.
Redfield first stated this fact on the basis of his steady-state experiments, but (6.10)
he did not actually perfonn lhe experiment we have described. That experiment
was first performed by Holton et a!' [6.6]. Why do the Bloch equations fail for " "
Since the trace is independent of lhe panicular representation used 10 evaluate it.
220 221
we can use a convenient representation. For example. we could evaluate (6.10) Another useful property of (6.16) to note is that when Ho :: 0, M :: O.
in principle by using as basis functions the eigenfunctions of the z-component Thus, suppose Ho were tumed 10 zero so suddenly that M does not have time to
of spin Izk of all the individual nuclei. However, to do so, we need to lake one change. Immediately after Ho :: 0 we have a case where M f. 0, Ho :: O. But if
more step, expansion of Z in a power series. there..were a temperature, (6.16) shows M must be zero. Therefore we can use
It is often valid for nuclei and for eleclTons to use the high temperature (6.16) to conclude that at this instant of time the system is not describable by a
approximation. We expand the exponential in a power series, keeping only the temperature.
leading lenns. Then the sums are easy to do. Another quantity of great utility is the entropy ~. We know from statisti-
cal mechanics that Ihe entropy measures the degree of order in a system. In a
2
Hl1i } N', reversible process in which there is no heat flow into or out of a system, the
Z=Tr { 1-/.;8+2/.;28 2 -'" =(21+1) +2k26 2Tr {'Jt}+··(6.11)
entropy of that system remains constant.
where we have used the faci that Tr {1t} = 0, as can be readily verified for both E+kBlnZ
q= (6.17)
Hz and 'Hd- 8
Using these methods one finds Evaluating Z and E we get
C(H2 + lJ2) C (H 2 +H 2)
E= 0 L where (6.12) ~::Nkln(2I+I)-- 0 L (6.18)
8 2 8'
N"f2fI 2I(I + 1) (6.13)
C= 3k 6.4 Adiabatic and Sudden Changes
is the Curie conslant, and HL is a quantity we call the local field, which is of
the order of the field one nucleus produces at a neighbor (several Gauss) and is The significance of these results is more fully realized by considering the behav-
defined by ior of the spin system when the applied field Ho is made a function of time. For
simplicity, let us assume the spin system is thennally isolated from the outside
CH' - 1 T< {H'} (6.14)
L- k(2I+l)N d world and that it mayor may not be in [hennal equilibrium. The first of these
assumptions is satisfied if the experiments we conduct are perfonned on a rime
Since the trace in (6.14) can be computed, HL may be considered to be precisely
scale short compared to the spin-lauice relaxation time. The assumption implies
known. One finds that the Hamiltonian of the system includes only variables intemal to the system
H[:: -y 2r!2[(I + I) ~)1/rjk)6 (6.15) since spin-lanice relaxation results from tenns involving variables both intemal
j and external 10 the system as illustrated by (5.33). If the system is in internal
One can compute the magnetization M and finds it obeys Curie's law themlal equilibrium, a spin temperature applies, so that (6.18) holds. We can then
consider three cases.
M= CHo (6.16)
8 a) The Hamiltonian is independent of time (the applied field is static). In
Note that this is a vector equation, so that M and Ho are parallel. this case, the average energy is constant in time, whether or not the system is
A moment's reflection shows that (6.16) is truly remarkable. It states that describable by a spin temperalure. If the system has many parts which are strongly
the vectors M and Ho are parallel no matter what the size of Ho as long as the coupled together, but is not initially in a state of internal equilibrium, we expect
high temperawre approximation is valid. Suppose Ho is small, comparable to that irreversible processes within the system will eventually bring the system into
the local field a nucleus experiences from its neighbors. One might then suppose an internal equilibrium describable by a spin temperature B. During that process,
that nuclei would tend to line up along the direction of the local field, not along the energy is conserved since a time-independent Hamiltonian corresponds 10 a
Ho. It would seem reasonable to suppose that the degree of polarization one system on which Ihere are no applied forces. Proof of the constancy of the energy
could achieve per unit of applied field would be less when Ho «..H L than when is left as a homework problem (Problem 6.5).
Ho:» H L . Equation (6.16) shows Ihat this intuitive argument is incorrect - the b) The Hamiltonjan changes slowly jn time. The criterion of slowness is that
degree of polarization per unit of applied field is independent of the size of Ho at all times internal transfers of energy shall be fast enough so that the system
relative to the local field. Not only is this true of the magnitude of M, but also is always describable by a single temperature B. Under this circumstance, Ihe
of the direction as well. changes are reversible, and the entropy of the system remains constant.
223
222
c) The Hamiltonian changes discontinuously in time. Such a change could system. The condition will be satisfied if we make the changes on a rime scale
occur if Ho can be changed quickly. The term "discontinuous" means that the rapid compared to the spin-lattice relaxation time T!. Frequently one has T1's
change is so fast that the various spins making up the system do not change of seconds, and by cooling may achieve Tl'S of hours. Such long limes are
practically infinite.
direction during the change.
The second condition we must satisfy is that after each small change in Ho
Let us now investigate these various cases more fully. we must allow a new temperature to be reached before making another small
I) Time-Independent Hamiltonian
change. This condition is typically that we must change Ho slowly on a time scale
defined by the precession period of nuclei in the local field of neighbors (I/-yHL)'
Consider a system not in thermal equilibrium initially. Let us suppose that the This time scale is a few tenths of a millisecond. Between these time intervals
parts of the system are coupled together. We then expect that eventually the there is a readily achievable range for which Ho can be changed adiabatically.
system will achieve an internal equilibrium described by a "final" temperature For an adiabatic change we have a constant entropy. Thus from (6.18) we get
o. If we know the energy of the system at t = 0 (call it Eo), we can compute that (HJ+HVI02 remains constant. If we stan in an initial field Hi at temperature
I~e temperature Or as follows, making use of the fact that for a time-independent 0i and change the field adiabatically to a final value Hr, the temperature Or is
Hamiltonian the energy is conserved. _ . given by the relation
Utilizing (6.12) which relates the energy E to the temperature, and applymg H.I2 +HI.2 _ H2r +H2).
conservation of energy, we get that
E=Eo or (6.\9)
8r - Of
(6.23)
Sf = -
C(H6 +Hl)
Eo (6.20)
by taking H;:> Hl, and Hi Ht
<: corresponding 10 what happens when a
sample initially in a strong magnetic field has that field turned to zero. Then we
find
We can compute the magnetization Mr which finally results after internal thermal
8r H L
equilibrium is reached by means of Curie's law -=-<:\ (6.24)
8i Hi
eRo (6.2\) Thus, the final temperature of the spins is much colder than the initial tempcrnture.
Mr=--
Sf If initially the spins were in thennal equilibrium with a thermal bath (such as the
where we take Of from (6.20). lauice cooled to liquid helium temperatures), the final temperature would be a
During the process of establishing internal thermal equilibrium, the entropy good deal less than the bath temperature. Note that the bigger Hi and the smaller
is not constant since irreversible processes are taking place. But eventually the H L, the greater the cooling. To reduce HL' it is common to dilute the magnetic
entropy O'r is given by the thermal equilibrium value for a temperature Or atoms.
Figure 6.1 shows Or versus Hr for an initial field Hi much larger than the
C(H 2 +H2)
O'r= Nkln(2I+ l)- ~02 I. (6.22) local field, and for an arbitrary initial temperature 0i' Note that as one lowers the
f
If the Hamiltonian can be divided into two pans which commute, the energy
0,
of each pan is separately conserved. The system then cannot be expected to reach
an equilibrium described by a single temperature, but rather is expected to reach
an equilibrium in which each commuting pan is described by a temperature. We
encounter this situation when we apply these ideas to the rotating frame later in
the chapter. In some cases the subsystem may not have enough complexity to
make us confident it will eventually be describable by a temperature, but even Fig.6.1. The final temperalure 6 f
so we ohen blithely proceed to assume that a temperature is achieved. reached al a final magnetic field
I1f ror adiabatic changes in the
2) Slow or Adiabatic Clwnges applied field rrom an initial value
ilL at tcml>erature 8;. Noll' lhat 8r
We keep in mind that to be adiabatic, a change in Ho must satisfy two con- does not chllnge much ...·ith IIr for
ditions. The first is that there should be no heat flow into or out of the spin IIr < II..
",
225
224
field from its initial value. the temperature drops until the applied field becomes thennal equilibrium between the spins and lattice in the field Bo. If Yo:» H L•
comparable to the local field. Funher decreases in applied field do not then the magnetization is small, but nevenheless it is not zero.
produce much lowering of the temperature. One of the remarkable features of (6.25) or Fig. 6.2 is that when H f = 0,
The magnetization can be computed from Curie's law if the temperature is Mr = O. As we have remarked, this is a very general consequence of Curie's law
obtained from (6.23). For the case that there is an initial magnetizmion Mj in an no matter what the temperature. (Note, however, that the derivation of Curie's
initial field Hi much larger than HL. the final magnetization M r is then easily law utilizes the high temperature approximation. At low enough temperatures
shown 10 be this approximation breaks down, allowing spontaneous magnetization to be a
Hr possible zero field state). What is remarkable is that the degree of order of the
Mr = Mi (6.25)
spin system is just the same when M = 0 as when M = Mi. Although it is
V/H r +H2L
2
clear that a system with net magnetization is ordered, how can there be order
This result is shown in Fig. 6.2. NOIe that during me cooling process, M r remains when AI = 01 The answer to this paradox is that even when H 0 = 0, spins still
at the initial value Mj until Hr gets comparable [0 the local field HL. because experience magnetic fields owing to the presence of their neighbors. A typical
for large values of H. 9r is propor1ional to Hr. spin will point either with or against the local field it experiences. For a highly
ordered system, there will be a substantial excess pointing with the local field
rather than against it Since the local fields at different nuclear sites have ran·
dom orientation (we here rule out highly ordered spin arrangements such as in
a ferromagnet), there is no resultant muaoscopic magnetization resulting from
the alignment along the microscopic local fields.
3) Sudden Switching
We have considered a process which is reversible. Now we tum to one where
things happCn suddenly, resulting in irreversibility. Suppose we describe the sys-
tem by a wave function t/J. TIle time-dependent Schrtxl.inger equation is
Fig. 6.2. MlIgno:ti:tation AI, versus
"pplicd fidd IIf for adiabatic h a~
dl"ngcs in IIf. It is &lI5lImed that -i {It = 7i(l)~
III > II~. Noto: thllt M, is indepen-
dent of 'II{ for II~ >11£, and goes = -M·H(l)~+7id'" (6.27)
to zero when IIr goes to :tetO
II, where M is the tOlal magnetic moment operator and 1id the dipolar coupling.
The time dependence of 1i arises because the applied field H is time dependent.
For the case of sudden switching, we take H as independent of time except
This equation is ploued in Fig.6.2. Note that if we lower the field until it
for t = 0, at which time it jumps discontinuously from one value to another.
is zero, M f = O. However, since the field changes are reversible (and in fact
Denoting by 0- and 0+ times just before and just after t = 0, we then have
at all times keep the entropy constant), we can recover Mi by raising Hr from
zero back up to its initial value Hi. The recovery of Mr in such a process d~s 0+
1I0t involve spin-lattice relaxation since, as we have postulated, everything is
done on a time scale shon compared to Tt! This curve is often a surprise to
~(O+) - "'(0-) = - J*7i(l)~(l)dl =0 or (6.28)
0-
those of us who first learned magnetism by studying magnetic resonance, since
(6.29)
we learned that one needs TJ to produce magnetization from an unmagnetized
sample. Actually, if one uses Curie's law and (6.23), one sees that if one stans since 1i(t), though discontinuous, is nevenheless never infinite. Thus we have
in zero field with spins which are in thennal equilibrium with the lauice (8 = ()/), that the wave function just before the switch is identical to its value juS! after.
the mere act of adiabatically turning on the static field to H o ,. HI. will produce We can utilize (6.29) to see that a sudden change in H produces a change
a magnetization. The size is in the expectation value of the energy E. The expectation value of energy at any
time t, E(t), is
(6.26)
E(l) = (~(l). 7i(l)~(l)) (6.30)
where Mo is the magnetization one gets when the spin-lattice relaxation produces
226 227
Thus whence we gel
E(O-) = (,,(0-). 'II(O-),,(O-» E(O+) = -C H(O-) . H(O+) _ C H~ (6.40)
= -(M(O-» . H(O-) + ('IId(O-» (6.31) 8j 8.
As discussed earlier, immediately after switching the field. the spin system
where (M(O-» and (1{d(O-» are the expectation values of magnetization and is in general not in internal thermal equilibrium even though it was in thennal
dipolar energy at t = O~. defined as equilibrium at t = 0-. If we wait a long enough time, we expect a temperature
(M(O-» " (,,(0-). M,,(O-» to be achieved. Call that time if and the temperature 8r- We can compute 8 r by
recognizing that fOT t > 0 the Hamiltonian is lime independent, hence the energy
('IId(O-» " (,,(0-). 'IId"(O-» (6.32)
is conserved. 1Oerefore.
Ukewise
(6.41)
(6.33) But
The fact that "'(0-) ::: "'(0+), however, means that the expectation values of E(tr) = _CIH'(O+) + HlJ
both magnelization and dipolar energy are the same at t ::: 0+ as at t ::: 0-. This 0,
resuh expresses Ihe physical fact Ihal all spins point the same direction al t ::: 0+
and E(O+) is given by (6.39). Thus
as they did at t ::: 0-.
Thus we can write H(O+)' + H'
8r - (J. " (6 42)
(6.34) - 'H(O ).H(O+)+Hl . .
This equation is very useful since in general the system just before Ihe The significance of (6.42) can be seen by considering a particular example in
sudden change is assumed 10 be in internal thennal equilibrium with temperature which the applied magnetic field is turned suddenly from its initial value. Ho,
OJ. Thus we can compute the expectation values by the methods of Sect. 6.3. to zero at t = O. Then H(O-) ::: Ho. H(O+) ::: O. giving
So far we have nor employed the density matrix notation, simply to make Or = 0, . (6.43)
Ihis chapler accessible to those readers who have not yet become familiar with
it. We can write our results compactly by recognizing that (6.29) is equivalent We contrast this with the result of turning the field slowly 10 zero
to saying that lhe density matrix f! obeys the relation HL
Or::: OJ Ho (6.44)
(6.35)
The sudden tum-off leaves lhe temperatu~ unchanged, the slow tum-off produces
Thcn
cooling.
(6.36) The results of sudden changes are summarized in Fig. 6.3. for Ihe case
Assuming thermal equilibrium at t ::: 0- at a lemperalUre Bi we get described above in which Ho is lurned suddenly to zero at t ::: O.
Adiabatic and sudden changes in Ho have been very useful in magnetic
_ cxp(-'II(O-YkO,) resonance. One of the first uses of adiabalic changes was to measure the magnetic
riCO ) = Z(O ) (6.37)
field dependence of the spin-lattice relaxation time. In 1948, Turner el al. [6.7]
Therefore were studying [he spin-lattice relaxation time TI of protons in insulators. The
relaxation times were, in some inslances, many minutes long. To observe Ihe
('IId(O-» = Tr {e(O-)'/{d}
dependence of T 1 on stalic field, they used a field cycling lechnique in which
Tr {'lid cxp (-'II(O-)lkO,») they observed the resonance at a high field. bUI allowed the spins to relax. in lower
(6.38)
- Z(O-) fields to which they cycled belween their observations. Pourui [6.8] discovered
Manipulation, using the high temperature approximalion plus the definition (6.14), that the nuclear relaxation times in a crystal of LiF were so long thai he could take
gives his sample out of the apparatus into Ihe earth's magnetic field and then return it
CH'L with only a small loss in magnetization. With Ramsty [6.9] he demonstrated that
('IId(O-» = (6.39) if this sample were removed from the strong field 10 a small static field (40 Gauss
0,
228 '29
Fig. 6.3. Magnetic field (11), wave function exclude the magnetic field, it is hard to observe a magnetic resonance in a metal
(l{!), energy (E), spin temperature (0), and in the superconducting state. One starts the field cycle with a magnetic field of
III i
/
rnagnc~ization (M) as functions of lime
for the case of a magnelic field turned sufficient strength to suppress the superconductivity. When one turns the field
suddenly to zero to zero, one achieves two effects: (1) the nuclear spins are cooled and (2) the
~
P
sample becomes superconducting. With the field zero, the nuclear spin tempera-
'V!v ture relaxed towards the lattice temperature so that when the magnet was turned
on again after a variable time T, the metal returned to the nOTlllal state and the
spins warmed to a holler temperature than they had before the cycle began. The
increase in temperature is measured by a swift pass though resonance. By vary-
£1 i - - - - - !-
ing T, one could deduce the zero-field spin-lattice relaxation time. The NMR
results showed that as one lowered the sample temperature below the supercon-
,, ducting transition temperature, Te , the nuclear relaxation rate was at first faster
,, than it would have been if the metal were normal, then at lower temperature
, ~N;;O~ _ _ it became slower. The temperature dependence of the nuclear relaxation rate is
{%<!~f!~1,~~~/1f ~
I dramatically different from the temperature dependence of ultrasonic absorption,
, as was shown by Morse and Boltm [6.15]. They found that ultrasonic absorption
dropped rapidly relative to its value in the normal metal on cooling below Te ·
"I
The contrast between the behavior of two low energy scattering processes, nu-
clear relaxation and sound absorption, is difficult to understand in a one-electron
-----'--"----- theory of metals, but finds a natural explanation in the Bardeen, Cooper, and
Schrieffer (BCS) theory of superconductivity. These experiments constitute a di-
rect verification of the concept of eleclTon pairs, which is the basis of the BCS
or less), subjected to an audio-frequency magnetic field, and then returned to the theory. (See, for example, Leon Cooper's Nobel Prize lecture [6.16].)
original magnetic field, he could detect aI the strong field the resonant absorption Anderson [6.17] working with Redfield, combined field cycling with the
of energy by the spin system when in the weak field. application of an audio-frequency magnetic field applied while the static field
The experiment we have just described is closely related to the famous ex- was zero to heat the spins to plot Ollt the zero-field absorption characteristics
periment on negative temperarures by Pound and Purcell [6.10]. They 100 started of a spin system. Thus they used field cycling to give them the sensitivity of
with the LiF sample in a strong field, but removed {he sample to a solenoid whose resonance in a strong field to monitor the effects they produced in zero field.
field they suddenly reversed, producing a situation in which the magnetization Abragam and Proctor [6.18] did further studies establishing the validity
is anliparallel to the static field. They then returned the sample to the strong of the spin temperature concept, again using LiF. An important result of their
field where they observed the relaxation of the magnetiz.1tion component along experiments was their observation of the transfer of energy between the twO spin
the static field. They point out that such a circumstance, with the upper en- systems (Li 7 and F19) even in a static field which greatly exceeded the dipolar
ergy Zeeman levels more populated than the lower, corresponds to a negative fields exerted by the nuclei on their neighbors. We discuss these experiments
Zeeman temperature. Any system with an upper bound on its energy levels can further in Sect. 7.10.
in principle have a negative temperature. In addition to the original paper by
Pound and Purcell, we call the reader's attention to the wonderful account of
this experiment and discussion of the negative temperature concept to be found 6,5 Magnetic Resonance and Saturation
in Van Vleck's [6.11] lecture on the concept of temperature in magnetism. (Pro-
fessor Van Vleck cautions that there are some incorrect numbers in his paper.) The analysis of magnetic resonance by Bloembergen et lIl. using standard per-
This imponant experiment in which a population inversion was produced is the turbation theory is given rather compactly in (1.32) - the differential equation
forerunner of the maser and the laser. for the population difference, n, between the two energy levels of a system of
Demagnetizing to zero field is the basis for the experimenls of Hebel and 4
spin particles,
Slichter [6.12] and Redfield and Anderson [6.13,14] to measure the nuclear
dn no-n
relaxation in superconductors. The problem here is that since superconduclOrs - '" -2W(w)n + - - (6.45)
dt T1
230 231
W(w) is the probability/second that a spin will be flipped by the radio-frequency will eventually produce large effects. Whenever a perturbation of small size can
field H t . Standard perturbation theory shows produce a big effect, it is dangerous to treat it lightly. He therefore eliminates
W(w) = i-y2 Hrg(w) (6.46) the time dependence by transfonning to a reference system in which the Hamil-
tonian..is essentially time independent. The residual time dependence is not of
where g(w) is a funclion nonnalized to unit area having the same dependence on the dangerous variety. For such a transfonned system, the energy is conserved.
frequency as the absorption line -that is, it expresses the fact that the frequency Moreover, the system is highly complex, consisting of many interacting spins.
of HI must be close to resonance for H, to induce transitions. no is the thennal One can thus predict that after a sufficiently long time the system will be found
equilibrium population difference, and TI the spin-lattice relaxation time. in a state of internal equilibrium. That is, it will be in one of its most probable
It is always JX)ssible, at least conceptually, to consider TI infinite, in which states. Phrased alternatively, the energy states will be occupied according to a
case (6.45) is especially simple to solve. Boltzmann distribution at some temperature 8.
n = Ae- 2Wf (6.47) We consider the Hamiltonian 1t, given by
To satisfy condition (ii) we expect that we must consider times less than 1/1V. 'It = -~h(Ho-wh)I.+HII.J+1f.:
Thus, though it is always easy to meet condition (i) by making H, small, no + term oscillating at ± w, ± 2w (6.49)
matter how weak HI, if we wait long enough we violate (ii). [Note we are
In deriving (6.49) we have used relations such as
requiring here also that I/W < TI, otherwise the Tl tenn would rescue condition
(ii) even for times long compared to I/Wj. -Iz sin? + I y cos; = e- i/ .? Iyei !.? and (6.50)
We have the interesting problem, therefore, that we do not know how to
integrate the equations of motion beyond a time for which n is almost its value (6.51)
at t = O. (See Fig. 6.4).
to transform products such as
The solution of this problem was found by Redfield [6.19] in a truly re-
markable paper, the more so when one recognizes that it was his first work on ei!,w.l IzIye -iI .w. t (6.52)
magnetic resonance. In it he shows that the Bloch equations, when applied to a
to expressions involving sin Wit and cos Wit. The term 'H~ is that part of the
solid, violate the second law of thennodynamics. The essence of his approach
dipolar coupling which commutes with Ii' Physically, it is the part of 1id which
is to note that a resonant time-dependent perturbation, no matter how weak,
is unchanged by rotation about the z-axis. (The two statements are of course
equivalent since one can consider Ii as generating rotations). II is the sum of
the tenns A and B discussed in Sect. 3.2.
". ~Rrgiolt com'elJliOIl(lI/H:1(llrlJo/io/f
of l"lliitiity of
/lleor)'
The fonn of 1t~ is
1f. = -'"'(hI· He + 1f.3 = 1f. z + 1f.~ (6.56) where {IJH 2 } is the second moment of the resonance line. Following our earlier
convemion of omitting primes. we shall now use H L for H~, using the prime only
Now, in the absence of HJ, 'Hz and 'H3 commute, since the fact that when we wish to distinguish between the local field in the laboratory reference
H3
[Iz , 11:3] = 0 was the definition of1f.3. Under this circumstance 'Hz and would frame and the rotating reference frame.
separately be constants of the motion. However, if HI f. 0, ['Hz, H~] f. 0, and the It is important to notice that the Redfield assumption leads to Curie's law and
Zeeman and dipolar systems can then exchange energy. Since H is independent that the vector nature of the law shows that the nuclear magnetization is always
of time, the total energy is conserved. Moreover, the system is very complex. parallel to the effective field when the system is describable by a temperature in
Redfield therefore postulates that no matter what the state of the system at t = 0, the rotating frame. Thus, if one is exactly at resonance, where He = iHI' the
a long time later it will be in a state of internal equilibriulll described by a magnetization is perpendicular to the static field.
Boltzmann distribution. In other words, there will eventually be a temperature 6 Since the fonn of (6.58), (6.59) and (6.60) is identical to that of the cor-
which can be assigned to the spins. We can thus say that the density matrix g is responding equations in the laboratory frame, most of the equations of Sect. 6.4
can be immediately applied to the rotating frame by repl:lcing Ho with He and
e-'H/k9
e=
Z
(6.57) HE with Ht[.
with Z = Tr {exp(-H/k8)}, where 'H is the effective Hamiltonian of (6.56). Of 6.6.1 Adiabatic Demagnetization in the Rotating Frame
course, we expect that after a long enough time Redfield's hypothesis would be
fulfilled (unless there were some hidden selection rule which we have overlooked, An adiabatic demagnetization in the rotating frame can be perfonned readily,
such as the fact that 1f.z and H?t
are perfectly isolated from one llnother if HI as was demonstrated by Holron [6.20]. Suppose that initially HI = 0, and the
sample is magnetized 10 Its thermal equilibrium value kMo. Let us shift Ho far
is zero). But the really important question becomes, how long does it take to
reach equilibrium? The answer to this question clearly depends on the size of above the resonance value wI'"'(, and tum on HI' C'Ne assume Ho - wI'"'( to be
HI, since HI is needed to prevent isolation of the dipolar and Zeeman systems. much bigger than HL and HI.) We now have an effective field which is virtually
We return to the question later, but for the present consider that the time is short parallel to M. Next we change Ho to approach resonance at 11 rate sufficiently
enough to make the establishment of a temperature practical. slow to satisfy the criterion for a reversible change. C'Ne are here assuming T l
to be infinite, which we achieve in practice by perfonning all experiments in a
time shorter than T j ).
We then have that
6.6 Redfield Theory Neglecting Lattice Coupling
M = Mo He (6.63)
The significance of (6.56) can be appreciated by calculating again the energy E, JHi+ H [
entropy a, and magnetization M. We find easily that
Notice that M is parallel to He' as is required by Curie's law. Thus, as
E = _C(H; + H'I> (6.58) one approaches resonance, M changes direction, always pointing along He. In
8 general, the magnitude of M also changes, unless Hi ~ HE. We can experimen-
tally measure M by suddenly turning off HI, leaving M}o precess freely about
234 235
Ho. The induced voltage immediately after turn-off is proportional 10 M',t.. (Use
of a phase sensitive detector enables one to measure Al'1: and 1L1y , but for these
experiments My = 0). One can measure M z by noting Ihm though M'I: decays
MQ
to zero within a time of order I/-yH L, after tuming off HI, M l does not change. ------------ ----- -----------------_.-
One can thus wait till M',t. has decayed, and then apply a 7r/2 pulse which rotates
M z into the x-y plane for inspection. The theoretical values of M'I: and M l are
HI HI (6.64)
M. = M - = MOr~==;;
He ./H2 +H2
vel,
ho ho
M, = M - = MO'f~~=,;' (6.65)
He I H 2+H2
V el.
where ho is the component of the effective field in the z-direction: fl i . fI,
)'ig.6.S. The transverse nmgnetization At" produccJ in a demllgnetization experiment in
110 == (Ho - w/-y) (6.66) the rotating frame, as a runction of the strength or alternating field If, employed. Ih is
assumed to be turned on when If a is we11 off resonance, with the slImple initially at its
Notice that if one does an adiabatic demagnetization exactly to resonance, the thermal equilibrium magnetization {\'fa
value of magnetization
HI achieve the full magnetization at resonance. Were one to reduce HI slowly to
~=~ 0~
/H2 +H2 zero after arriving at resonance, M would shrink to zero. In this manner all of
V I L
the order represented by Mo prior to demagnetization would have been put into
will persist indefinitely as long as HI remains on (actually, when relaxation to the dipolar system, thnt is, into alignment of spins along their local fields. One
the lattice is included, it decays, but on a time scale typically of order T I ). We could, at a later time, slowly turn H t back on again, thereby recovering the
contrast this prediction with our earlier conclusion from the Bloch equations that magnetization.
M:z; decays to zero in T2 , where T2 ~ I/-yH1•. The factlhat M does not shrink
as long as He is kept constant, and that M precesses in step with HI is often
6.6.2 Sudden Pulsing
described by the graphic tenn "spin locking". If lIf is comparable 10 El, M:z;
will be less than Mo. However, the "loss" of magnetization is recoverable. Were A situation frequently encountered in experiment and interesting to contrast with
one 10 go back off resonance, M would grow back to M o when H;:» Hl. adiabatic demagnetization is the effect of a sudden change in He. We treat an
Figure 6.5 shows (6.67). especially simple case, that of suddenly turning on HI at time t = O. We assume
When Redfield proposed his theory. the fact Ihat the spins were locked to that before we tum on HI, we are off resonance by an amount ho, with the
He was one of the most surprising results. It is, of course, nOlhing but the rotating system initially magnetized to Mo along Ho. We utilize the ideas of sudden
frame equivalent of the statement that the magnetization in the usual laboratory changes discussed in Sect. 6.4.
frame adiabatic demagnelization has a one-to-one correspondence with Ho, as The sudden tum-on of HI is so fast that the system has the same wave
expressed in (6.25). function or density matrix just after turn-on as it had before. The dipolar energy
Note that if one pulses on Ho when (Ho -w/-y):»HI and H t • and then Ed which depends on the relative orientation of spins is thus the same at t = 0+
changes H o so thai one passes through resonance, continuing until one is far on as at t = 0-. Moreover, since the dipolar Hamiltonian 1t~ is the same in both
the other side of the resonance (Ho - w/-y negative), one will have turned M an- laboratory and rotating frame.
tiparallel to Ho. Moreover, although near to resonance, one might have M < Mo _ C1I2 H2
(if HI < Hd; by the time one is far from resonance one would have M '";;£ Mo· Ed = -~ = -Mo.::.L. (6.68)
This experiment provides a simple means of lllrning over the mngnetization. 6] Ho
One can see from (6.64) that if one demagnetizes exactly to resonance, the where 61 is the lattice temperature. The Zeeman energy is
magnetization will be the full Mo provided 1I I :» HI.. TIle resonance signal is
then as big as can be achieved using a 1r/2 pulse. If HI ~ HI,. one will not
Ez = -M·He = -Moho (6.69)
236 237
The tolal energy. E = Ez + Ed, is then 6.7 The Approach to Equilibrium for Weak HI
E = -Mo{ho + HVHo) ~ - Moho (6.70)
We saw in Sect. 6.5 that standard perturbation theory predicted that following
A long lime later a spin temperature will be established together with a magne- the turn-on of a weak HI the population difference 11 would go to zero for long
tizalion M parallel 10 He giving times, although we recognized that we could not rigorously apply perturbation
theory to times greater than l/W. The requirement of a weak HI was necessary
E=_C(Hi;Hl) =_M(HJ;eHl ) (6.71) in order that perturbation theory be valid for at least short times. Of course, since
M z is proportional to n. this implies M z would go to zero. In Sect. 6.6, however,
But the lolal energy is conSClVed once HI has been turned on, so that, equating we saw that M z would, under thesc conditions, go to an equilibrium valuc
(6.70) and (6.71), we get
il 2
Mz)equil = Mo , 2 °H 2 (6.74)
M = Mo H~O:~2 (6.72a) 10 + L
o
Hillo
L
where we have assumed Hr« 116 and fIr «: HE (although we note that the
equilibrium expressions in Sect. 6.6 were not limited to weak H J). It is therefore
M z = NIo H2 H2 (6.72b)
0+ L clear, as we suspected, that for fong limes, perturbation theory does flat give cor-
,,2 rect predictions. For short time intervals, however, it must be correct. Recalling
the proportionality between M. and n, we know that
M: = Mo H2 : H2 (6.72c)
o L
This equation shows that exactly at resonance, M would vanish. The null is vel)'
d~~, _ -2W(w)M, (6.75)
sharp, M:t varying linearly with ho, so that observation of the null provides a for times short compared to I/W(w). How can we describe M z for longer times?
precise method of observing exact resonance. The solution to this problem was worked out by Provotorov [6.21] in an
II is interesting to contrast these results with conventional saturation theory. elegant paper utilizing powerful techniques. Rather than outlining his analysis,
For a spin! system, M: is related to the population difference n by the equation we will give an altemative derivation of his result.
/'fm We note that in the absence of HI, thc Zeeman interaction in the rotating
~-- ~~ frame is just
2
Saturation theory says that the equilibrium population difference, assuming infi- 1iz = -Iflhol z (6.76)
nite Tl' is n = O. hence M z = O. Equation (6.72c), on the other hand, says M z Let us make the assumption that we can assign a tempcrature 0z to this Zeeman
will be zero only when saturation is perfonned exactly at resonance (ho = 0). Hamiltonian, and 0d to the dipolar 'Hl This assumption may not be rigorously
Conventional saturation theory assumes the transverse magnetization vanishes as correct, bUI it is al least simple, and corresponds to the facts at the time HJ is
well as M z . Equation (6.72b) shows that in general, M:I; is large, and may in turned on. Immediatcly before turning on H] the dipolar system is at the lauice
fact be larger than M z . temperature 01, since 1t3 is the samc in thc rotating or laboratory frame. In
If Hi:» Hl. (6.72) has a simple geomelTical meaning. After H] is pulsed rhe laboratory frame we have
on. M precesses about He. The component of M parallel to He cannot decay
CHo
without energy exchange to the lattice. but Ihe componenl perpendicular can Mo=-- (6.77)
decay since Ihe local field gives a spread in precession frequencies. Thus, after 8,
several times l/-yHL' M will be parallel to lIe. and will have a magnitude given But in the rotating frame
by the projeclion of the il'lilial Mo on He. Clio
Mo=-- (6.78)
8z
Inasmuch as Ho ~ h o , Oz «0,. Thus the Zeeman temperature in the rotating
frame, 0z, is very cold compared to the dipolar temperature tJ d . Turning on II]
°
couples the two reservoirs and they approach the final equilibrium value given
by the analysis of Sect. 6.6. The coupling of HI produces lTansitions between the
238 239
energy levels of the Zeeman and dipolar systems which we assume are governed
by simple rate equations for the population of the various slates. (This assumption
is similar to our postulates of Sect. 5.2. It is quite common in all cross-relaxation ". Slope b,OW" I,om f'tnuroutioll tlltory
calculations. Provotorov makes it implicit in his work when he evaluates the
relaxation times.) Since there are many states, a large number of coupled rate
equations similar to (5.13) result.
As has been shown by Schumacher [6.22], when the two systems are char· ".
acterized by temperatures, the many equations reduce to twO coupled linear rate
equations, one for (1/8z), the other for (I/8d) much as the many equations rep- Filial rolue gi~fm by
resented by (5.13) reduce to a single rate equation (5.27). But the conservation Uedfield O,eo,y
of energy gives a relationship between 8z and 8J :
Ch 2 CH2
,
0 _ ~ ::::: const
___ (6.79) Fig.6.6. M. "eUIl! t during saturation ~ given by the argument in ~he te,,~. The equlltion
8z 8d ror M. 115 Il runction or lime is round by joining lhe initilll region, where dM./dt is
known rrom perturba~ion theory, to ~he equilibrium "alue, give" by Redfield theory, using
Equation (6.79) is a first integral of the coupled equations, so that one of Sc1wlllochu's observation lhat ~he approa<:h lO equilibrium involves a single exponential
the resultant time constants is infinite, and a single exponential results. Since
M, oc 118z, this means M~ relaxes according to a single exponential towards its
equilibrium value. Using the fact that M, ::::: Mo initially, we get an equation for
M, as a function of time, 6.8 Conditions for Validity of the Redfield Hypothesis
M" - M,,)equil::::: (Mo - M~)equil)e-t/T (6.80)
We have noted that lhe concep! of spin temperalure in the rOlating frame has as
The only unknown in this equation is 1". We can, however, easily calculate il as a basic requirement neglect of certain lime-dependent !erms in the Hamil!onian
follows. Taking the derivative of (6.80), evaluating it at t : : : 0, and comparing
with (6.75) (which must be valid initially where perturbation theory is correcl),
J
transformed to the rotating frame. This means, in essence H a,. Hi + Ht. On
the basis of the previous seclion we can now add a second requirement. We
we get
saw that in the absence of HI> the equality of dipolar and Zeeman temperalures
~ :::::2W(w) Mo (6.81)
in the laboratory reference frame implied inequality in the rotating frame. Spin-
T Ala Mz)equil lattice relaxation will allempl 10 bring the two temperalUres together at the lanice
temperature in laboratory frame in a rime TI. On the other hand, the presence of
Using (6.72c) for M')equil we get
H I attempts to equalize them in rotating frame. The two tendencies are in conflict.
2 2 2 The slrength of the tendency is indicated by the corresponding relaxation times (a
_I ::::: 2W(w) h 0 +HL::::: 7(72 H2 (h 0 +H2)
II 9(W)
(6.82)
1" H2 IH2 short relaxation lime means a correspondingly strong tendency for the associated
L L
equilibrium). Thus a spin temperature will be established in the rotating frame
This result is the same as that firs! found by Provotorov, as indeed it must be since only if 1" is much less Ihan Tt. Thus we find that a spin temperature is established
we have made the same approximations as he. The complete time development in the rotating frame and that we muSt employ Redfield's approach provided
of the magnetization is therefore
T,
-:» 1 or
M,::::: 2Mo 2 ho+HLexp -1r"f HI2(h~+Hf)
ho+H L
[2 2 (
HL
2 2 y(w)t )] I (6.83)
1r"'(2 Hr (h~ ;r
l )9(W)T1 ,. I (6.84)
This is almost exactly the conventional condition for saturation. Note that the
is usually valid (see Fig. 6.6). longer T I , the smaller the HI which will satisfy (6.84). In particular, frequently
(6.84) is satisfied even when HI <. HI..
240
241
6.9 Spin-Lattice Effects statistical mechanics for which there are a number of collision terms, each of
which changes the distribution function f and for each one of which one could
So far we have considered the case of magnetic resonance on a time scale shon
compared 10 the spin-lattice relaxation time. In many instances one performs
l compute af/Dt. Thus. if the Zeeman energy -M· He and the dipolar energy
(1t3) ..w ere not in thennal equilibrium with one another. the magnetization would
transient experiments which satisfy this condition. However, an equally important I grow (or shrink) unlil equilibrium was reached. ThaI process clearly produces a
case arises when one does experiments on a lime scale long compared (0 T. as
when one employs steady-state apparatus. One can still salisfy the criterion for ·l contribution to the three time derivatives (6.85-87) which we have not included
in the equations. The lattice coupling. represented by Ta • T/>, and Te , must be
validity of the Redfield theory given by (6.84), but the spin temperature 8 in the expected to push the spin system out of thermal equilibrium in the rotating frame,
rotating frame is now determined by the coupling to the lattice. As we shall see, but the internal couplings of the spin system counteract that trend by producing
this slatement by no means implies thai () = 8,. bUI only that it is a function energy exchanges within the spin system. Changes within the spin system must
of 8/_ Fortunmely it is very simple to generalize oue previous treatment 10 in- conserve the total energy of the spins. If then we consider the rate of change of
clude the lattice. As a maHer of fact. in Redfield's famous paper he considered the expectation value of energy, E, we can ignore any changes which simply
just this case, but we have put off consideration of the lattice relaxation until this redistribute energy within the spin system. and keep only changes associated
Stage to simplify the discussion. with the lattice.
In general, when the spin system exchanges energy with the lattice, the Using the fact that
internal equilibrium of the spin systcm in the rotaling frame is momentarily
disturbed. The basic assumption we make is that the cross·relax;lIion bctween
-E = -M,ho - °
M.H, + (1id ) (6.88)
the dipolar and Zeeman systems is so rapid compared to T I that, following we can find the rate at which the lattice coupling changes E by taking Ihe time
an exchange of energy of the spins with the lattice, a new spin temperature is derivative of (6.88):
rapidly established. Thus, the lattice always finds the spin system described by a
I dE DMz: DM~ a 0
temperature in the rotating frame. We considcr that thcre are three basic relaxation . dt = -h0 /5t - H'/5t + at (1id )· (6.89)
equations for the classical magnetizations lI{zl and Mz: and for the expectation I
value of the dipolar energy (1t~), which we write down phenomenologically in 1be only conttibutions to the time derivatives we need consider are those given
a fonn chosen to assure that in the absence of HI the spin system would reach by (6.85). (6.86), and (6.87) since other contributions do not change the tOlal
thennal equilibrim with the lattice. energy E. Employing (6.85-87). assuming that M always lies along He' and
using (6.58) and (6.59). we readily find an equation for the magnitude of M
DMz: 1110 -Mz:
(6.85) dM
7ft = Tg -
I
= -(M", - M) (6.9Oa)
dt Ttl
(6.86) ••
and for the spin temperature 8
244 245
where p is a quantity expressing the fact that the local field after a jump is not 7. Double Resonance
completely random relative to its value before the jump. It can be calculated. Its
value depends on the jump mechanism (vacancy, interstitial, etc.). The 2 arises
because the local field is a coupling between pairs of nuclei, either of which can
jump.
Thus, simple measurements of T1(J enable one to measure a10mic motion or
molecular reorientation when T m is comparable to T1(J arising from other causes.
One of the most important developments beyond the original concept of mag-
netic resonance is so-called double resonance in which, as the name suggests,
one excites one resonant transition of a system while simultaneously monitor-
ing a different transition. There are many reasons for doing double resonance.
The goals include polarizing nuclei, enhancing sensitivity, simplifying spectra,
unravelling complex spectra, and generating coherent radiation (e.g. masers and
lasers). The field has been characterized by a rich inventiveness which seems to
continue unabated to the present day, and which makes the task of understanding
all that has been done rather overwhelming. It is useful, therefore. to find some
means of classifying the work into broad areas which involve related concepts.
We shall employ three broad categories.
The first category of double resonance makes use of spin-Iauice relaxation
mechanisms. We call it the Pound-Overhauser double resonance after two of
the important pioneers. The method involves a family of energy levels whose
populations are ordinarily held in thennal equilibrium by thennal relaxation pro-
cesses. If one saturates one of the transitions (1.34) one so-to-speak clamps their
populations together (Le. forces them to be equal). The thennal relaxation pro-
cesses then repopulate all the levels. producing unusual population differences
which may possess useful properties (for example, the upper of two energy levels
may have a larger population than the lower, or a population difference which
is nonnally small may become large). Included in this category are methods of
dynamic polarizaliotl of nuclei (the Overhauser effect and solid effec!), electron-
nuclear double resonance (ENDOR). and masers and lasers.
A second category depends on cross-relaxation phenomena, hence we call it
cross-relaxation double resonance. The fundamental concept is that if two spin
systems can exchange energy (i.e. cross-relax), one can detect the absorption of
energy by a resonant altemating field tllned to one spin system by lise of a second
alternating field to monitor the temperature of a second spin system. Various
experiments involving cycling the "static" magnetic fall in this category, as does
the Hartmann-Hahn method which is the basis for the technique referred to today
as "CP" (meaning "cross polarization") introduced by Pines, Gibby, and Waugh
for enhancing the sensitivity of CI3 spectroscopy. [n general these techniques
246 247
require that some appropriate cross-relaxation process exist and that it be much 4
thermal equilibrium. Thus when he salurated Ihe j to transition, the intensity
faster than certain spin-Ianice relaxation processes which would otherwise destroy 4 -4
of the to transition increased by a factor of ~ times its normal intensity.
the effect. We will discuss below why such intensity changes occur in connection with the
The third category depends in general on the existence of spin-spin couplings Over.IJauser effeci. By means of a series of experiments saturating the various
which in many cases must not be unduly obscured by either spin-lattice relaxation transitions while observing the OIhers, he verified that the spin-Iallice relaxation
or cross-relaxation. We shall therefore call it spin coherence double resonance was via electric quadrupole coupling.
because it depends on the ability of spins to precess coherently for a sufficient The second double resonance experiment perfomled was by Carver (7.2).
time to reveal the spin-spin splittings. Typically, one here makes use of the faCI Ollerhauser, while still a graduate student at Berkeley, had predicted (7.3J that
Ihat when two nuclei are coupled, changing the spin orientation of one nucleus if one salurated the condUClion electron spin resonance in a metal, the nuclear
changes the precession frequency of the nuclei to which it is coupled, so that the spins would be polarized 10000foid more strongly than their normal polarization
second nucleus can reveal in this way when Ihe first nucleus is being subjected in the absence of electron saturation. (For an excellent summary of dynamic nu-
to a resonant alternating magnetic field. clear polarization by one of the important pioneers, see [7.4].) Crudely speaking,
Among the examples of this category are topics known as spin decoupling, Overhauser predicted a polarization of the nuclei which they would have if the
spin tickling, spin echo double resonance (SEOOR), coherence transfer, and electron spin BollZmann factor were used in place of the nuclear spin Boltz-
two-dimensional Fourier transform NMR (2D-Fr NMR). 1be last of these is mann factor. It may be hard for readers today to appreciate the deep scepticism
fundamental 10 many important areas of resonance ranging from NMR imaging with which Ollerha"ser' s proposal was greeted by the resonance community, be-
to delennining the struClure of complex biomolecules. cause today there are many schemes for dynamic polarization, and the principle
Our purpose in introducing three classifications is pedagogical. Other scien- forms the basis for many other important techniques. However, in 1953 Ihere
tists might pick different classifications. Of course, most schemes of classification was a widespread (Ihough short-lived) belief Ihal Ollerhauser's scheme must
are not perfect. For example, the nuclear Overhauser effect leads 10 effecls in 20- violate the second law of thennodynamics. Carver's experiment was not only
FT NMR which are exceedingly useful and important. Moreover, 2D-Fr NMR Ihe first demonstration of the dynamic polarization of nuclei, but perhaps more
involving only one nuclear species is not a double resonance experimenl (only importantly showed that Ollerhauser's revolutionary Ihoughl was correct. His
one oscillator is used) but it can be conceptually viewed as one in which lhe abil- concept reoriented the thinking of resonators. While his concept stimulated the
ity of a large HI 10 excite all the nuclei in a spectrum obviates Ihe necessity of subsequent invention of other schemes of dynamic polarization, perhaps more
having a separale oscillator for each NMR line. It is not our intention in picking imponant was the stimulus it provided to exploration of other novel effects of
three classifications to imply that there is only one idea involved in each category. pumping transitions and doing double resonance.
Indeed, within the groupings we employ there are numerous important innova- Figure 7.1 shows the first dynamic polarizalion of nuclei, the Li nuclei in
tions. For example, it was many ye~ after lhe firsl spin decoupling experiments lithium metal. To perform this experiment it is necessary to have an HI of 510
were perfonned that 20 Fourier transfonn spectroscopy was invented. 10 Gauss to salurate the condUClion electron spin resonance. One also wants a size
of metal particle small compared to the skin depth at the electron frequency so Ihat
aU the nuclei are polarized by the electron saturation. Accordingly, Carver used
a solenoid 10 generate a static field of about 30Gauss, which put Ihe electron spin
7.2 Basic Elements of the Overhauser-Pound Family
resonance at 84 MHz, and the lithium nuclear resonance at a frequency of 50kHz.
of Double Resonance At that, low frequency, the nuclear resonance was too weak to be observed
directly (Fig.7.1a), but popped up from the noise instantly when the electron
The first double resonance experiment was carried out by Pound p.l] on the saturating oscillator was turned on.. Carver calibrated the degree of polarization
Na 23 nuclear resonance in NaN03. His aim was to prove that the spin lattice by observing the proton resonance in mineral oil (Fig. 7.lc). He went on to show
relax:ltion mechanism was via time-dependent electric field gradients. He com- Ihal dynamic pol:uization did not require a melal by working with the proton
puted the thermally induced transition probabilities between the various spin resonance of liquid ammonia in which sodium had been dissolved. The sodium
states based on this mechanism. In the NaN03 crystal, the Na resonance is split atoms in such solutions ionize, giving free electrons whose resonance is readily
by an axially symmetrical electric field gradient into three distinct absorption saturated.
lines corresponding to transitions between the m values of ~ to !' ! -4,
to and The principles underlying the Overhauser effect are in fact identical to those
-4 -lto He predicted and observed that saturating any Olle transition (e.g., on which POUlld based his experiment. The Overhauser effect provided a strong
4)
the ~ to would produce level populations of all the states which differ from impetus to the development of double resonance melhods, in part because Over-
248 249
Fig.7.1. Demonstration of the Overhauscr Hence we take m,.
the eigenvalues of I z.as another good. quantum number. The
........."p.p................. ~
nudear polarization effed on Li 7 nuclei in
metallic lithium by Carver. The osdlloscope
energy eigenvalues are then
a picture shows nuclear absorption plotted
vertically versus magnetic field. The mag-
E = "fefi.Homs +Amlms - ",(n!lHom,
nelic field excursion is about 0.2 Gauss. The
top line shows the normal Li 7 nuclear res-
ms =±~ m, = ±~ (7.3)
OnanCe (lost in noi~ at the 50kHz fre- It is convenient to label states and wave functions by a convention
quency of the NMR apparatus). The mid-
dle line shows the Li 7 nuclear resonance (7.4)
enhanced by saturating the electron spin
resonance. The experimental conditions of where c = 2mS, p. = 2m,. A state with ms = +~, m, = -~ is written as
(3) and (b) differ only in turning on the
electron spin saturating oscillator. The bot-
1+ -). The selection rule for transitions induced by an applied alternating field
tom line shows the proton resonance from a is Llms = ± 1, Llm, = 0, or LlmS = 0, Llm, = ± 1. The first corresponds to
glycerine sample containing eight times l'lS an electron spin resonance, the second to nuclear resonance. (See Sect. 11.3 for
many protons under the same expt'rimenlal a detailed discussion).
conditions, from which one concludes the
Their resonance frequencies, We and Wn, are
c . J.i7 nuclear polarization was increased by a
~:, factor of 100
(7.5.)
(7.5b)
hauser had made an ingenious and daring prediction which many talented physi-
cists judged to be wrong, and in pan because he addressed a topic of interest to a There are thus four allowed transitions. They are shown in Fig. 7.2. The relative
community of scienlists much broader than resonators. Although Pound did not size of IAI and ("fnhH) influences the appearance of the energy level diagram
recognize that if his concept were applied to other systems it could lead to large slightly as is shown in Fig. 7.3. If the nucleus and electron under consideration
nuclear polarizations, there can be lillie doubt that even without Overhauser's are far apart, typically IAI < l"fnfi.Hol (Fig.7.3a). rf the nucleus and electron are
contribution, Pound's invention of double resonance would have caused a bur- close together, usually IAI > hu rlHo I (Fig. 7.3b). The fonner case is encountered
geoning of the field. for a typical nucleus in a solid which has a low concentration of paramagnetic
centers. The latter case is encountered for nuclei of paramagnetic atoms, or for
nuclei which are very close neighbors of a pararragnetic center.
For the rest of our discussion we adopt the convention of drawing figures
7.3 Energy Levels and Transitions of a Model System
which look like Fig. 7.2a, though we do not mean to imply thereby anything
about the relative size of IAI and hn!lHol.
To understand the principles of the many applications, one can consider a very
!
simple system consisting of a nucleus with spin 1= coupled to an electron of
spin S = !' acted on by an external static magnetic field Ho. The Hamiltonian
I
w .. '" 'Y.. 11 0 + .!!.I
++1
for this system is 211
,e» l"fnl. These assumptions make Sz nearly commute with 11.; hence mS,
(iI) Nudl'lIr r<!!mlllttlC,. IrIlIlSiliOlI$ (b) I;·/,.,.trlill V!ill "'SOllllllce 1rl1!l)'ilirJIIs
the eigenvalue of Sz, is a good. quantum number. Only the term AlzSz of the
electron-nuclear coupling gives diagonal terms. so the Hamiltonian is effectively Fig.7.2a,b. The energy level diagram and allowed transitions of a system consisting of an
electron of spin S = !
coupled to 3 nuclcus of spin I =t
acted ou by an externa.t static
(7.2) magnetic field flo. The figure assumes 'Yn is negative
250 251
++-- ++--
The resul/ of using the more general form offunction of (7.8), whether as a
--+- --+- result of solving the simpler HamUlonian of (7.1) more exac/ly, or as a result of
solving the more general (7.7), is thai, on application of an applied alternating
magnetic field, transi/iolls other t!lan those shown in Fig.7.2 become possible.
We adopt the convention of calling transitions other than the four in Fig. 7.2
"forbidden transitions" .
-+--
In the absence of applied alternating fields, populations of the energy lev-
-+-- els of the combined spin system are given by the Boltzmann factors when the
(a) DisrulI/ IIlIdt'('$ (h) N"urb)' 'IIICIt-,ts system is in thennal equilibrium. As discussed in Chapter I, the achievement
Fig. 7 Ja b. The cITed of lhe size of the electron-nucle;o.f coupling II/h relative to the of thermal equilibrium can be thought of as resulting from lransitions induced
nucle'at ~nallce frequency Col on Lhe appearance of the energy level dillgn"us. For nuclei by the coupling to a thennal reservoir in which the thermally induced transition
far from the dectron, lJ$ually iAI < h'.h 110 For nuclei ne~r. to the electron, .rrcquen~ly
I·
IAI > b.h/fol. The figu~ ll$IIumes a n~"tive 1'n and A positIVe. What would It look hke probability We'l,c,y from state 1!"'1) to state Ie'l') is related to the rate Wc"'"e.,
for positive "fa? by
in general
252
, 253
Fig. 7.5. Some thermally induced mp'
ping rates which invotve ~he transi-
lion of the electron spin. The direc- (7. 13c)
tion of the lIrrow indicates the direc-
tion of the trallsition corresponding
to the rates 1V21 , 1V12 , ete.
(7.13<1)
PI = P2 (7.19)
Whence
Fig. 7.6. The Overhauscr effcet. Thermally induced transitions lVi; shown attempt to
Illaintain thermal equilibrium. An applied al~ernaling field induces elcetron spin transi- (7.20)
tions at 8 ra~e IV., between the 8ta~es I and 2
254 255
We have stated that the Overhauser effect produces a nuclear polarization. Let This expression is the same as would be found for the nucleus if acted upon by
us compute the nuclear polarization. the external field only.1
In a state le11), the nuclear spin expectation value is '//2 (7]:< +1 or -t). Comparing (7.23b) with (7.26), we see that electron saturation has increased
Therefore. the average expectation value of nuclear spin I z • (Iz) is (Iz)·.by the ratio
256 257
the ammonia. A similar effect should arise with lWO nuclei. as was recognized 4-- FIg. 7.7. The energy levels ImsmJ> of II
by Nolcomb (7.5]. Bloch [7.6J, and Solomon [7.7], and was demonstrated by p"ir of nuclei (posilive .,'5). An IIpplied
allernalins field induces lhe transition 1
Solomon in a pioneering double resonance experiment in HF utilizing the HI to 2 and 3 to 4 in which rnJ remains fixed,
and FI9 resonances. w but ms chanses. The lransition rllte IV ill
In Sect. 3.4 we saw that the dipolar coupling between nuclei in a solid could defined in lhe text
be used to obtain infonnalion about struclUre of molecules or solids. In liquids
the dipolar coupling is averaged out owing to the rapid translational and rota- w l---.·
tional motion. Nevertheless. the dipolar coupling still plays an important role
in detennining the relaxation times. It has tumed out that the nuclear Over-
hauser effeci (both stalic and transient) gives important information about which
resonance lines conespond to nuclei that are physically close to one another
in a molecule. It is thus playing an important role in struclUre detenninalions
of complex molecules. especially when combined with two-dimensional Fourier numbered 1. 2. 3. and 4. We assume that we are irradiating the S transition.
transfoml methods. We therefore seek to understand the Overhauser effect in producing a transition probability per second of W between states 1 and 2 and
liquids in greater detail. between 3 and 4. To find what this does 10 the [-spin polarization, we calculate
In order to see more clearly how the liquid degrees of freedom make an (It) given by
Overhauser effect possible, we examine a concrete example, the nuclear Over- ,
hauser effect of a molecule containing two spin! nuclei. An example is HF, the (It) = L (il[:li)')i = ~(Pl + P2 - P3 - P4) (7.3Ia)
system studied by Solomon. We shall demonstrate that saturating the resonance .=1
of one nucleus (the S species) produces a polarization of the Other species (the In a similar way
[ spins) and then show that if one disturbs one species from thennal equilibrium
with a pulse, the effect is revealed in the transient behavior of the other species. (S:) = !(PI + P3 - P2 - P4) (7.3lb)
It is, then. convenient to write the Hamiltonian as To find the Pi'S. we must solve the four rate equations for the populations
(7.29) Pi (i = I to 4)
where 'H 12(t) is the dipolar coupling between the two nuclei, augmented. perhaps, dp,
dt = PI WH + P2 W 24 + P3W34 - P4(W41 + W,n + W 43)
by any pseudoexchange or pseudodipolar coupling (Sect. 4.9). and where the
explicit inclusion of t emphasizes that in liquids the dipolar coupling is in general + <Po - p,)1V (7.32<1)
time dependent since the radius vector from one nucleus to its neighbor within
dp3
the molecule is continually changing direction. There is also coupling of the dt = PI W 13 + P2 W23 + 1'4 W43 - P3(W3t + W32 + W34)
nuclei in one molecule with those in another. We omit these effects since we are
concerned with demonstrating principles. Our calculation would therefore apply +(P'1-P3)W • (7.32b)
rigorously only to a case in which the molecules are present in low concentration etc. Adding these two equations. gives
in a liquid which does not otherwise contain nuclear spins. Since the Hamiltonian
d(P3 + p<j)
commutes with both It and St, we can label its eigenslates by mJ and ms, the dt = PI (W14 + W13) + P2(W24 + W23) - P3(W31 + W32)
eigenvalues of [z and St.
-J,,,(W<l1 + W42) (7.33)
The energy levels then become
In the steady state (where all dp./dt's vanish). (7.32) shows that as W
E = -"YJ'iHomJ -,sliHoms + Amrms (7.30)
is increased. 1'3 and Pot become progressively more nearly equal. In a similar
where A is the pseudoexchange coefficient. the only tenn of 'H'12 which does manner. PI and 1'2 become progressively more nearly equal. We therefore set
not average to zero under the molecular tumbling. If A .,; O. the resonances of PI = 1>2 and P3 = 1'4, obtaining from (7.33)
the I-nuclei or of the S-nuclei are doublets. We shall for simplicity assume that
A is zero, SO that both the I resonance and the S resonance consist of single W14 + W13 + WZ4 + W23) F
~=~ ( =1'1- (7.34)
resonance lines. Figure 7.7 shows the energy levels. labeled by mrms, and also W.U + W31 + W"2 + W32 G
258 259
which defines F and G. To calculate these transition rates we need several equations. From (5.281)
Now. utilizing (7.12) and (7.18)
1 +~
(7.35) Wkm
_ = 2"
h J Gmk(T)eXp[ - i(m - k)TdT) willi
-~
and so on for W13. W2". and W23. Since EPi = I. PI = P2. P3 = p". and
Gm.(T) " (ml1i l (l)!k)(kl1i , (t + T)lm)
P" :::: PI (FIG), we get
Taking
hence (7.300)
Gmk(T) = Gmk(O)eXP(-jTI/Td
Utilizing (7.34) and (7.35) we can get an exact expression for F and G. However,
B = -t(S- I+ + S+ r)(1 - 3 cos 2 8)
in the case of a liquid. the high temperature approximation is valid even for
electrons (perhaps liquid He is an exception!) so that
Bij = I + (Ej - Ej)/kT (7.39)
E = -i(S+ J+) sin 2 ee- 2i6
giving
we get that
(I,)
= _I_{ W"I(E4 - E
I ) + W31(EJ - E I )
Uo = ~~ (/6) ((1 - 3 cos' 8)'),.(+ -1[+5-1 - +X- + Ir5+1 + -)
2kT 2(W41 + W31 + W 42 + W32)
(7.44)
+ W 42(£4 - £2) + W32(£3 - E2)}
(7.40)
2(W"1 + W31 + W 4 2 + W 3 2) where «(1 - 3 cos 2 8)2)" .. means the average over the 411" solid angle of (I -
3 cos 2 8)2. Thus
Examination of Fig.7.7 shows that W41 is llie transition from 1- -) to «1_3cos2 8)2)" =~J(I-3COS28)2sin8d8d¢=4/5 (7.45)
1+ +) in which both spins flip up. W31 is the uansition from 1+ -) to 1+ +) • 4.
in which the I-spin flips up, W 42 is the transition from 1- -) to 1- +). also a Likewise
transition in which the I*spin flips up. Finally W32 is the transition from 1+ -)
(sin 2 8 cos 2 (J),,'/I" = 2/15 (7.46)
to 1 - +) in which the S-spin flips down and the I-spin flips up.
We shall define a new notation which makes these processes more explicit
by defining U/l1 where M is the totnl chnnge in mr + ms in the tmnsition: (sin" (J)o\?r:::: 8/15.
In this manner we get
(7.41)
Suppose Iwsl »wl (as for S being an electron spin, in which case ws is negative) dP4
dt = W ll {PI - Pol - [PI (T) - 1'4(T)j}
and thai W~7"; «: I. Then
+ W",{)" - P4 - [p,(T) - p4(T)j}
(I ) = ''"'5 (3/5 - 1/10) = ~ "W5
z 41.:1' (3/5 + 1110) 7 41.:1'
(7.49) + W",{p3 -Pol - U'3(T) - 1',1 (T)]} (7.57)
which is negative for electrons. with similar equations for dpddt for i = I, 2, and 3.
On the other hand, if there were a strong Fenni contact tenn, so that Uo »UI We saw in Chap. I that the fact that Wij oF Wj; is important in producing
or U2, as with the conventional Overhauscr effect, a thermal equilibrium population. That is the physical significance of (7.56).
Clearly the family of equations represented by (7.54) describes a system which
( I ) = _ lIws (7.50) will relax to the thermal equilibrium populations p/T). We can then replace the
z 4k1' Wij'S by the Um's, taking
which for negative Ws is posltlve. Therefore, dipolar relaxation produces an
Overhauser effect of the opposite sign to the conventional Overhauser effect W32 s:' W23 = Uo
since for dipolar coupling U2 »Uo. W42 s:' W2.1 = VI = W3l S! W I 3
We have carried out the solution for the steady-state polarization of the 1· W31 s:' W l 3 = U2 (7.58a)
spins. If one is doing pulsed experiments, as is frequently the case for double
resonance experiments or for two-dimensional Fourier transform experiments, the and introducing Ui by Jhe relations
pulses disturb the 1-S spin system from thermal equilibrium, following which the WI2 s:' W21 == ui = W34 s:' W"3 (7.58b)
thennal processes bring the system back to thermal equilibrium. One then speaks
of a transient nuclear Overhauser effect. Thus, in general, one wishes to find the U: differs from UI (7.47) by the subslitution of Ws for WI. Making use of these
time dependence of the observables (I~(t) or (S~(t) for some sort of initial relationships, one can now show Ihal
conditions. Solomon, in [7.7J, calculates the transient response and demonstrates d
dt (Pl +!J2 -1'3 -1',,) = 2UI [(pol -]>2) + (P3 - PI)]
what happens in a set of classic experiments. To do the calculation, one needs to
solve for the time dependence of the Pi'S, starting with equations such as (7.32) + 2UO(p3 - 1") + 2U,(P4 - P'> - 2U I [PI (T) - P2(1') + P3(T) - PI (T)]
with W = O. Thus we have - 2Uo[P,(T) - p,(T)]- 2U I [P,,(T) - p,(T)] (7.59)
dp4
dt =PJWI4 +P2 W 24 +]J3 W 34 -P'I(W41 + W"2+ W"3) (7.51) From (7.31) one can show that
BI4 = I +cH where (7.52) Therefore (7.59) can be rewritten utilizing 10 and So to represent the thermal
equilibrium values of (I~) and (S~) as
EI ~E4
cl" = kT (7.53)
d(I,) = [10 _ (1,)](U0 + 2UI + U,) + [So - (5,)](U, - Uo) (7,61,)
we get dt
dp4 In a similar manner, we get
dt = W 41 (PI - P4) + W 42U>2 - p,,) + W 43 (1'3 - P4)
+ W41PIC14 + W42P2C24 + W43P3C34 (7,54) (7.61b)
,,, '63
These two coupled linear differential equations have solutions consisling ~, =1++)
of real exponentials. They may be viewed as a nonnal modes problem with I,
imaginary frequencies, i.e. real exponentials. There are thus two time constants.
,, ~J=t+-)
, , ,,,
These equations show that as long as (5,) = 50 and (I,) = 10, both d(5,)/dt
WI':, Woo ,,, ,
and d(I,)/dt remain zero. However, if one disturbs either spin system from
thennal equilibrium, both spin systems will respond. Thus, if one makes ($,:) = 0 ,,, w,. ,, ,,,
by applying a 7fn pulse to the 5-spins, not only will d(S,)/dt be differem from
, IV4.J: : 11'.,.
~1=1-+) ,, ,,
zero, but also d(l,)/dt will be nonzero. ,, t
In general, then, when one applies pulses to a system of coupled spins, all ~.=r--)
spin populations will be found to respond as a result of the relaxation. We pose
Fig. 7.3. The use or • rorbidden tnonsilion which flips oolh lile dN:tron and nucleus si-
the detailed solutions of these rate equations as a homework problem. multaneously, !Vea , to produce nuclear polilfiZlltioll. h is I181SUl'll«I thallransilions which
ilwolve an e1edfOn sl>in-nil> only are lhe only sip;nilicallt lhermlll PI"OCes6e$
In order for the Overhauser effect to work, the nuclear relaxation process (such
as W13, W24) cannot be allowed to short circuit relaxation in which both an
electron and a nucleus flip, such as W23 or W!1. It is not always possible to
meet those conditions.
The condition which one can, in general, be quite sure will hold is that pure
electron spin-flip processes (such as W 12 or W34) are much the fastest W;j's
because electrons couple more strongly to tile lanice than do nuclei. Jeffries [7.8]
and independently Abragam et al. (7.9) recognized that the so-called forbidden FIg. 7.'). A rorbidden transition 1Iltemalive to that or Fig. 7.8, which producl'3 Iloclear
polarization or the opposite sign
transitions were not strictly forbidden in many useful cases, and that one could
use them to good advantage in achieving polarization. In fact, Erb et al. (7.10)
independently discovered the effect experimemally. Using this scheme, Jeffries Even though the transition probability Wen may be small, it is frequently
and his colleagues at Berkeley, and experimentalists at Saclay collaborating with possible to achieve strong enough alternating magnetic fields to make it larger
Abragam. have obtained proton polarizations of over 70%, and have made a than the thennal transition rates Wij in which a nucleus flips, so that the transition
rich variety of applications. Invention of this technique was another major step connected by Wen produces effective population equalization.
forward. The phenomenon is often referred to as the solid effect. Lei us analyze the case of Fig. 7.9 for which the transition between states
There are two possible forbidden transitions. They are shown in Figs. 7.8 I (-IPt = 1+ +)] and 4(IP4 = 1- -)) is saturated. Since we assume Ihe only thermal
and 7.9. The transition of Fig. 7.8 can be induced by an alternating field parallel transitions of consequence are those shown, we immediately can write down
to the static field when the simple isotrOpic electron-nuclear coupling of 0.1) is
solved to the next higher order to include the effect of the tenn AUz 5 z + IySy ) 1'1 = P4 (7.62,)
in admixing 1+ -) with 1- +).
The transitions of both Figs. 7.8 and 7.9 can be induced by alternating fields (7.62b)
perpendicular to ]fo when the more general Hamiltonian, (7.6), is solved to
adequate precision. For example, the dipole-dipole coupling between a nucleus (7.62c)
a distrance r from the electron makes the transition matrix elements of Fig. 7.8
and 7.9 of order 'Yefilr 3 Ho times the matrix element of Fig.7.6 in which only (7.63,)
an electron is flipped. The ratio also depends on the angle the static field makes
with the axis connecting the nucleus and the electron. (The effect of the dipolar
coupling is expressible more precisely in terms of the contribution of the dipole- (7.63b)
dipole coupling components AZ'Zl, Ay'y" and A,,'z')'
264
265
relative to the donor atom. The solution is a straightforward generaljzation of that
P3 : ~~B::,4",-J~_ (7.630)
of Sect. 7.3, using the N quantum numbers mi, the eigenvalues of fZi'
2+8 12+ B 43
Using (7.2Ib), ..E = '1ehHOmS + I: AjmjmS - I: "fnhHOmj (7.68)
) 1 812 - B<l3 (7 64)
(1z) = '2(P1 +1'2 -P3 -P4)= 2" 2+ IJ 12+ B 43 . Electron resonance occurs when Ams "" ± I, Ami"" 0 for all i. 1'he frequency
We IS
To appreciate the meaning of this expression we look .at l~e high temperature A;
limit (evaluation of the expression for all temperatures IS given as a homework We = "feHo + ~ -;;m j (7.69)
problem). •
I 7enHO (765)
Since there are many possible values of Ai, and since each mi can be +!
(1,) : 2 2kT . or -i, (7.69) describes many frequencies, in fact virtually a continuum, rather
than a set of resolved lines such as in the case of an electron interacting with
so that the enhancement over the normal polarization is only a single nucleus. For a sample consisting of many donors, there are 2 N
various ways of assigning the mj's.
(7.66)
The problem of resolving the hyperfine lines may be likened to that of a
man with several telephones on his desk, all of which ring at the same time. If
the full Overhauser effect. he tries to answer them all, he hears a jumble of conversations as all the callers
speak to him at once. Of course his callers have no problem - they hear only
one voice, though he hears several.
Feher recognized lhal each Si29 nucleus experiences the hyperfine field of
7.7 Electron-Nuclear Double Resonance (ENDOR) only one electron. Thus the nuclear resonances are sharp. Each Si 29 site gives
rise to two resonance transitions, corresponding to whether the electron hyperfine
A double resonance experiment of great historical importance was performed by field aids or opposes the applied magnetic field:
George Feher [7.11]. He named it electron-nuclear double resonance (ENDOR).
The purpose of this technique is to resolve otherwise unresolvable resonance Wnj "" "fnHo - AjmS (7.70)
lines. The concept involves in essence observing a nuclear resonance through The nuclear transition frequencies of Si 29 nuclei located at each distinct crystal-
its effect on an electron spin resonance. Fehu was studying the electron spin lographic location relative to the donor atom give rise to a distinct pair of lines.
resonance of electrons bound to donor atoms in silicon. 5 % of the silicon nuclei, For many sites, the Aj is sufficiently large that these lines are well resolved from
the isotope Si 29 with spin ~, possess a nuclear magnetic mome~1. As a result the other transitions corresponding to other sites.
of the large radius of the orbit, the electron spin couples magneucally to many Recognition that nuclear resonance would give resolved lines even though
Si29 nuclei. Since the Si 29 nucleus can have two spin orientations, and since the electron spin resonance did not was a great insight on Feher's part. But, how-
Si29 hyperfine fields depend on where in the donor orbit the nucleus sits, there ever, the nuclear resonance signal would be weak since the donor concentration
are many different local fields Si 29 nuclei can produce at the electron. (_10 17 per cm 3) is typically so low. Feher conceived of the clever method of
The essential situation is shown by considering a Hamiltonian of the one detecting the nuclear resonance 1;ly its effect on the electron spin resonance.
electron interacting with N Si 29 spins. (For simplicity, to avoid the needed Consider the situation in Fig. 7.10. We show 4 levels only, which represent
discussion of the statistical effect of an isotopic abundance, let us suppose all Si the four energies we get if we hold all mi's fixed except for those of the jth
sites had Si 29 atoms.) We take it to be of the foml nucleus:
N E '" CrehHo + L Ajmj)ms + Apnpns - "fn1iJIOmj (7.71)
11. "" "fchHoSz + S· L A;Ii - "fnhHolz (7.67) i • j
i=1
Clearly we can define an effective static field acting on the electron of
where Ai is the hyperfine coupling of the ith nucleus (for simplicity we assume
A-
a simple Fermi contact form of coupling rather than the more general tensor Heff = Ho + L -kmj (7.72)
fonn). The Ai'S are determined by the crystallographic location of the ith nucleus ji-j"fe
for this case.
266 267
'It,-I++l Tn, ,
'
,I· I', I"
7.8 Bloembergen's Three-Level Maser
:,,'/',
.,1'
I... I '
..~
.;... I A maser (microwave amplification by stimulated emission of radiation) works
ill ,
I
I ,"" , I on me principle that irradiation of an absorption line associated with transilion
I ","" I t
'¥l-l-+) , I" I p, p, between two energy levels will lead to a net emission of energy when the upper
' ~evel is more highly populated than the lower one. Townes (7.13]. and [7.14J, and
..Ll.. Independently Prokhorov and Baso'll [7.15J, recognized that if such a situation
(,) 'v, 'e) could be achieved, it could be used as the basis of a new type of oscillmor.
Townes achieved population inversion with ammonia gas by physically separating
1J>l '" (x/! E' -E')
(-;;:r- lhe molecules in the upper energy level from Ihe lower energy level by means
of properly shaped electric fields, in essence a molecular beam apparatus. The
frequency of the oscillator fell in the microwave region.
J'1g. 7.10a-c. Ft~,'s scheme for observing nuclear rcsomlll(e by nlOniloring il.$ effcct on
the electron spin resonance. flg\lrc 7.IOa shows lhe stales of (7.71) for the electron and Shonly Ihereafler, Bloembergen recognized Ihat one could produce the en-
the jlh nucleus, with the eleclron transition being shown between stales I llnd 2, and ergy level inversion by use of the Overhauser-Pound principle of level repopu-
thermal relaxation pro«sscs such as in the Overhauser efrect. Figure 7.IOb shows the lation through saturation. It required a system with three or more energy levels
populations produced by saturating the electron transition. Figure 7.1Oe: shoWlii the effcct (Fig. 7.11).
of an Ildiah2.tie pIlIJ!:IlI.ge through the resonant frequency of states 1 and 3, with lV~ turned
off. When 'V~ is turned back on, it finds the electron transition bct"'een stales I and 2
momentarily ul1!laturated
268 269
we have that I h, going to 4 h is a hanl way to improve signal·to-noise! These facts illustrate
that when one is using averaging there is a big premium on setting variables
">1 a~
P3 such as HI> modulation amplitude, sweep excursion, etc., to the optimum. A
making the upper level more highly populated than the lower, the condition for corollary is when looking for an unknown, be willing to set the equipment to
maser operation at the frequency (E 1 - ~)lf1. give ~ maximum signal even though you may thereby distort the resonance line
B/oembergen pointed out that the theoretical situation he described could (for example, from using too large a modulation, or from panially saturating).
be achieved with paramagnetic ions, as was shonly demonstrated [7.16]. After the resonance has becn found you can focus attention on it with parameters
Bfoembergetl's concept has been of enonnous value. Level inversion by adjusted to avoid distonion.
pumping another transilion is the principle on which all lasers operate. The
pumping may be done optically, as when a light source is used, or by the equiv-
alent of pumping (which generates nonequilibrium populations between a pair 7.10 Cross-Relaxation Double Resonance
of levels) in chemical reactions (chemical lasers) or electric discharges (electric
discharge laser).
Prokhorov and Basov [7.15] independently developed many of the same No mailer what one does to improve one's apparatus, one eventually reaches
ideas. the limit of the current state of experimental an. What then if the signal is still
too weak to see? We tum now to the use of double resonance, assuming there
are two resonances which can be excited, one the "weak" one too difficult to
see directly, the other a "strong" one observable by conventional means. For
7.9 The Problem of Sensitivity example, the weak one might result from a low abundance species with spin S,
whereas the strong one might result from an abundant species with spin I. (For
Sooner or later all resonators want to observe a resonance wh.ich is too weak simplicity, we use the term "rare" to imply the species whose resonance is weak.
to be seen. Resonances may be weak because the number of spins is low, for Since a low ""( could also make the resonance weak, "low ""(" can be substituted
example since the isotopic species is rare, or the nuclei occupy special positions for "rare" in most cases.)
such as being surface atoms on a crystal or neighbors of an imperfection. The If an H I is applied at resonance to the rare species, those nuclei absorb
resonances may be weak because the nuclear ""( is small. energy and their spin temperature rises. If the rare and abundant spins could
The development of superconducting solenoids with their high magnetic exchange energy, the abundant spins would thereby get holter, and their signal
fields has helped enormously in increasing sensitivity of NMR equipment be- amplitude would diminish. If the abundant species is thermally isolated from the
cause the quanta absorbed in resonance are thereby increased. Reduction of outside world (i.e., has a long TI), this temperature rise can be made quite large
noise by narrow banding also helps. The practical problem used to be that a and thus readily observable merely by silling on the weak resonance long enough.
narrow bandwidth was achieved by using long integrating time constants, and By going to low temperatures, TI of many systems can be made exceedingly
those in tum meant it took a long time to sweep across the resonance line. By long. Thus the crucial question becomes how well can the two species exchange
the time one was near the end of the sweep, the apparatus was likely to have energy.
drifted, giving distorted line shapes or spurious signals from drift in the apparatus The mixing of the Iwo spin systems was sludied by Abragam and Proctor
baseline. The advent of the multichannel digital signal averager has dramatically [7.17]. They employed the Li 7 and FI9 resonances in LiF. Their experiments were
changed all that by permitting cumulative addition of a number of rapid sweeps. part of their studies of the fundamentals of spin temperature and of adiabatic
If the appamtus drifts, one stops sweeping, readjusts the equipment, and contin- demagnetization. Working in a field of several thousand gauss to observe the
ues sweeping. Moreover one sweeps only as long as is needed to achicve the resonances, they prepared the system in some nonequilibrium state in the strong
desired signal-to-noise ratio. TIle signal accumulated is proportional to the num~ field (for example, they inverted the F I9 magnctiz..'ltion), removed the sample to a
ber of sweeps N. Since the noise is random and both positive and negative, its lower static field, allowed the Lj7 and FI9 spins to mix, then returned the sample
amplitude increascs as .JFi, giving an overall increase in signal-to-noise ratio to the strong field for inspection of the Li 7 and F19 resonances. In this way they
proportional to NI.JFi = .JFi. found that in a field of 75 Gauss the spins come to a common spin temperature
As a practical maller, this means one must increase the averaging time in a mixing time of 6s, that the time was unobservably shon at 30Gauss, and
fourfold to gain a mere factor of 2 in signal-to-noise. This increase may not longer than the TI 's (several minutes) above lOOGauss.
matter if it means going from I min to 4 min, but if one is already averaging for
270 271
A detailed study of cross-relaxation was made by Bloembergen et al. [7.18] W = 0). One therefore looks for a matrix element which connects the initial state
and by Pershall [7.19] for both electron spin systems and nuclear spin systems. (two Li 7 spins up, one FI9 spin down) to the final state (two Li7 spins down,
They show that the crucial problem is the failure of the Zeeman energy to match one FI9 spin up). If we denote the Li 7 by S and the FI9 by I, and use subscripts
when two different nuclear species undergo mutual spin Aips. The mismatch in i, j, k to distinguish individual nuclei, a matrix element of the dipolar !enn B.
Zeeman energy must be made up by the dipolar coupling between the spins. t:
(aIS; Ib), between states a and b flips only olle U 7 spin. To describe two spin
The simplest sort of process arises when two nuclei have nearly the same 'Y. For flips we must go to second-order perturbation to get the effective matrix element
example, HI and F I9 have 'Y's which differ by 5%. Owing to the existence of (al?ierrlb) joining the initial and final states:
tenns such as S+ I- or S- 1+ in the dipolar coupling [the B tenn of the dipolar
Hamiltonian of (3.7)], the dipolar coupling couples states in which the proton (al1iofflb) = L: (al1i~I'~'~1idlb) (7.77)
spin flips up (down) and the fluorine spin flips down (up). If we could consider ° b °
the 'Y's as being identical, we would then have a situation such as shown in where one matrix element might involve st IZj and the other st Ii. There
Fig. 7.12. Because the individual energy levels match, the initial state of the two
spins, indicated by the two x's, has the same energy as the final state indicated
are other possibilities, for example st I":
combined with SziSt The energy
difference Eb - E c depends on what the initial and intennediate states are, but
by the two o's. A system started in the x state will undergo a transition to the
is of the order of magnitude of the Zeeman energies involved.
o state by means of the S+ I- part of the B term in the dipolar coupling. For
In general, if one considers a process Wilh I1.F fluorine spin flips down and
HI and F t9 , the fact that the 'Y's differ by 5 % means that in a strong magnetic
n Li Li spin flips up, with !lwLi or flWF energy change per Li or per F spin, the
field such as 10 kGauss, the Zeeman energy difference between the x and the
energy mismatch L!E between initial and final states is
o states would correspond to the energy of FI9 or HI in a field of 500Gauss.
Such an energy mismatch would prevent the transition unless there were some (7.78)
other energy reservoir whose energy could change to make up the difference.
A possible candidate is the dipolar energy reservoir. For a pair of spins, it has (7.79)
a typical value of 'Yl'YSh2/r3, which, expressed in units of magnetic field, is
only a few gauss for reasonable values of r. Thus, it cannot make up for an
energy mismatch of 500Gauss. If, however, the static field were much lower, dE
nLi = 2 nwF =0.17
the mismatch would be correspondingly reduced, and the mutual flips might
become possible! but for
Returning to LiF in a field of IOkGauss, the Li 7 resonance occurs at
16.547 MHz and the Ft9 resonance occurs at 40.055 MHz, a ratio of 2.420. llr = 2 Hl.i = 5 11E = 0.066
Clearly the mismatch here is even worse than for H t and F t 9. What Bloember- flWF
gen and his colleagues recognized [7.18,19] was thai 2.4 is close to 2.0, hence Thus, the more spins we allow to flip simultaneously, the closer we can make the
a process in which two U 7 nuclei flip up and one F I9 nucleus flips down comes energies of initial and final states match. However, more spins flipping requires
much closer to satisfying energy conservation than a process in which only one more dipolar terms acting simultaneously and thus requires third- or higher-order
Li7 flips. Now, Li 7 has a spin of ~' but for the sake of argument we are going perturbation expressions. The first-order matrix elements are of order HI., where
to pretend it has a spin !'
since the explanation is then simpler. To estimate the HI. is some sort of dipolar field of one nucleus at a neighbor. The second-
rate, one must utilize a fonnula for transition probabilities such as (2.182) (with order expression (7.77) is then of order Hl/Ho, where Ho is the static field.
Each increase in order of perturbation contributes a factor HLlHo. Thus, since
H L <Ho, one expects the perturbation order which works to be the lowest one
Ms = -1/2 M l = -1/2 for which 11E ~ y'nnwL' where WL is the local field in frequency units, and n
]<'ig.7.12. A mulual spin-IIiI' from thc the total number of spins which flip.
(x) lo lhe (D) slale cxchangcs energy
belween the [-spins and the S-spins.
Since the energy mismatch 11E is proponionalto Ho, and the cross-relaxation
Shown verlically are lhe allowed matrix elements are inversely proponional to a power of H o, one can see that the
energy levels. For lhis cxample thc cross-relaxation rate should be a steep function of Ho, as is illuslfated in Fig. 7.13
M s = +1/2 cnergy spacing is assumed to l>c lhe
same for lhe two spc<.:ies, and bolh showing Pershan's measurements of the Ho dependence of the cross-relaxation
Spccies 5 Specic" J nudei have spin t rate in LiF.
272 273
.'ig.7.0. Dah. or Pershiln for the cr0S5- When one has a single crystal in which some nuclei have a quadrupole
T1\ YS. magnetic rield in LiF relaxation Lime 121 ror l,i7_1~19 in Li(~
splilling and others do not, there are magnetic fields other than zero in which
versus static magnetic field for three
the energy level splittings of the two systems become equaL Cross-relaxation is
10
l / crystal orientations (sec 17.191)
then r.apid. Edmonds {7.26] has reviewed nuclear quadrupole double resonance.
/ J
7.11 The Bloembergen-Sorokin Experiment
••< An important advance in cross-relaxation double resonance was discovered by
B
•• Bloembergen aod Sorokin in their studies of the cesium halides (7.27J, especially
CsBr. In fact, they found vinually all the essential elements of what has come to
be known as the Hartmann-Hahn method which we discuss later, though it seems
l,, / evident that the full significance of the results of Bloembergen and Sorokin was
not appreciated owing to the fact that they discussed the problem entirely in the
0.1 I ,t Csl33 rotating frame instead of using the doubly JOtating frame which we take
up in the next seclion.
,I / We give a brief account here of what they found, then return in the next
section to discuss their results a bit further. using the concepts of Hannumn and
o.Oll-~-l('-7:-~--~
o SO '"
Hahn [7.28] and of Lurie and S/ie/lter [7.29] concerning spin temperature in the
doubly rotating frame.
We shall discuss two important discoveries made by Bloembergen and
Ma9"etil: field IGallssl
Sorokin. 1be first has to do with spin-locking (Sect.6.6.1). We saw in Chap.6
that if the Ht is sufficiently strong (6.84) one must use Redfield's ideas of spin
Clearly the use of cross-relaxation as part of double resonance leads in a temperature in the rotating reference frame. Then, if one puts the magnetization
natural way to cycling the static field between low values for cross-relaxation along the HI in the rotating frame, it will persist for rimes as long as those
and high values for observation. for which the spin system may be considered to be thermally isolated from the
Anderson [7.20). working with Redfield, combined field cycling with Ihe surroundings. In their sample of CsBr, Bloembergen. and Sorokin found that the
application of an audio-frequency magnetic field applied while the static field Cs t33 spin-Ianice relaxation time was lo-2Dmin at room temperature, ten thou-
was zero to heal the spins to plot out the zero-field absorption characteristics sand times longer than the Dr Tt 's. Thus they expected to be able to maintain
of a spin system. Thus they used field cycling to give them the sensitivity of (i.e. "lock") the Cs lJJ magnetization lined up along the HI)es for many minutes.
resonance in a strong field to monitor the effects they produced in zero field. Yet, when they auempted to observe the Cs l33 while sweeping through the line
Redfield [7.21], Fernelius [7.22], Slusher and Hahn [7.23], Minier [7.24], using an HI for which Redfield theory should apply. they found no signal even
and others utilized field cycling to observe quadrupole splittings of nuclei near to when they took only 15s to go through the line. They found that the explana-
foreign atoms. If one has a sample which is a powder (e.g. a metal sample where tion was that the short Br T t causes the Cs·Sr dipolar coupling to fluctuate in
one employs a powder to overcome problems of the skin effect), an advantage time, thereby providing a relaxation mechanism for Tie Eq. (6.9Ib). Since TIe
of demagnetizing to zero field is that the quadrupole splittings are then iden- specifies how long the spins can be locked. the Br T I thereby limits the Cs 133
tical in all the individual crystallites. There is therefore no longer any powder spin-locking. A simple view of their discovery is that the Br T} contributes to
broadening of the NMR transitions [7.21-24]. To observe very large quadrupole the relaxation rates Tb and Tc of (6.87) and (6.9tb). The short relaxation time
splittings in powders, there is a problem of coupling the energy absorbed by the of the Br results from the modulation of the Br electric quadrupole interaction
quadruple transition into the rest of the spin system. It is necessary then to adjust by means of lallice vibrations.
the amplitude of the "audio" frequency alternating magnetic field according to Thus. we see that if one nuclear species has a dipole coupling to a second
principles of Hahn and Hartmann discussed later in this chapter. species with a fast spin-lattice relaxation time, Tie of the first species will be
Weitekamp et al. [7.25} have employed field cycling to zero field together shortened, and one can detect the existence of the second species by comparing,
with Fourier transfom NMR to simplify dipolar spectra of powdered samples. for the first species, Tl with T( e'
274 275
Their second discovery was thaI if they utilized two oscillators. one tuned to
the es l33 resonance, the other tuned near the 8r79 or BrS l resonance. they could
7.12 Hahn's Ingenious Concept
utilize the rapid spin-lattice relaxation time of the bromine to polarize the es 133 .
This effect they labeled the transverse Overhauser effect. In their experiment, The problem of matching the Zeeman splittings of two different species could be
they applied a strong rf field H de. tuned exactly to the Cs resonance. and a solvro if by magic one could apply one magnetic field to the I-spins, a second
rather weak rf field HI )Br tuned to a frequency'; which differed from the Br magnetic field to the S-spins. How can one do this to spins which are neighbors
resonant frequency VBr by an amount given by on the atomic scale?
The magical solution was found by the Wizard of Resonance. Erwin Hahn
(7.80)
and demonstrated by the Wizard and his Sorcerer's Apprentice Sven HartmiJnll
As we explain rigorously in the next section, this condition can be understood by [7.28]. Hahn recognized that alternating magnetic fields have a negligible effect
considering the effective fields Heff)Br and Heff)es in their respective rolating on nuclei unless the frequency of altemntion is close to the precession frequency.
frames. In the bromine frame. Thus if he applied two allernating fields HdJ and Hds at frequencies WJ and
ws, respectively, the frequencies being chosen to satisfy the respective resonance
21r(v - VBr) 'H)
(7.81) conditions,
HeffBr='
) k +~ IBr
11l,
Since HdBr is small, ~=n~ ~=U~ (7~)
(7.82) HIlJ would have negligible effect on the S-spins and Ht)s would have negligi-
Heff)Br ';;t k2'IT(v - vBr)/rBr
ble effect on the I-spins. Each species could then be viewed in its own rotating
Likewise. reference frame. Figure 7.14 shows the two rotating reference frames. Note that
the z-axis is the same in both frames. It is in fact the direction of Ho in the
laboratory. In Fig. 7.14 the magnetization vectors MJ and M s are also shown.
since the Cs is tuned exactly to resonance. The Bloembergen-Sorokin condition Note that in a general situation they do not lie along either the z-axis or the
is a special case of the equation respective HI'S, though they might if special ways of preparing the syslem were
used. For the case where an M is not parallel 10 its HI, it will precess about the
(7.83)
HI in the rotating frame with precession frequency [}J or [}s given by
which has come to be known as the Hartmann-Hahn condition. As we shall see,
~=nHI)J %=~HI)s (7.8~
when this condition is satisfied, the two spin systems can come to a common
spin temperature in the doubly rotaling frame. Equation (7.83) is a condition for These equations suggesl immediately that we can make the precession frequencies
rapid cross-relaxation in the doubly I'OIating frame. match by adjusting the ratio of the H)'s to satisfy the relation
Since Bloembergen and Sorokin analyzed their experiment by means of the HI), is
single rotating frame of the es 133 , they did not have spin systems which came ~~=W n~
to a common temperature in that frame of reference. They discussed the problem
in tenos of an Ovemauser effect (indeed they called it the transverse Ovemauser
effect). They showed that when (7.80) was satisfied. the es magnelizarion Me.
and the Br magnetization M Br satisfied the relation
(7.84)
In Sect. 7.16 we shall see how we can get their result in a simple manner y,
by combining Redfield's results on the calculation of Meq, (6.9Ia), with the idea
of spin temperature in the doubly rotating frame. ,/' --",n/
,
, ),"/
M, \
... __/ I /l,p
flg. 7.14. The I'Olaling reference frames ahow lhe axes r, and rs along lhe respeelive rf
fields Ill), and III )$' Note lhill the z-axis is in common
276
277
This condition is known as the Hahn condition. It corresponds to making the as energy is exchanged, the sizes of M r and Ms change. In equilibrium 8r = 85 ,
Zeeman splittings of the I-spins quantized along Hdr in the I-spin rotating Causing 85 to increase produces a heating of 8r.
frame equal to the Zeeman splitting of the S-spins quantized along H1)s in the We see, therefore, that there is a way of coupling two different spin systems
S-spin rotating frame. so that they can exchange energy. We need now 10 do three things:
Satisfying the Hahn condilion matches the Zeeman spJiuings, but in refer- l. set the qualitative arguments on a firm quantum mechanical basis;
ence systems which are exotic to say the least Our real objective in producing 2. describe the experimental steps involved in putting Hahn's idea to use;
the matching is to permit the two spin systems 10 couple. How can we use the ex- 3. analyze the expected results.
otic matching 10 couple the systems? We know that the spin systems are coupled
via the dipolar interaction. For two spins of unlike species that portion which
broadens the line is the A term of the dipolar coupling 1idA
7.13 The Quantum Description
1idA = 'Yr,}S h2IzSz(l - 3 cos 2 8rs ) (7.88)
T
rs We star! by writing the Hamiltonian of the system in the laboratory frame.
'J-i = 'J-izr(t) + 'HZS(t) + (1{d)r / + ('Hd)SS + ('J-iJ)rs (7.89)
Since the z-axis In the rotating frame is the same as in the laboratory, this
coupling is unaffected by the transformation of spin variables to the rotating where 1{zr(t) is the Zeeman energy of the I-spins. It includes both a static inter-
frame. (Note: Sometimes confusion arises as to what variables are transformed. action with field Hok and a time·dependent interaction with the two alternating
In quantum mechanics one can formally introduce operators which transform fields. It is most convenient to consider that rotating fields have been applied,
only I, only 5, or only r, or any combination. Since r rs and 8r s are unaffected instead of linearly polarized alternating fields. We have, then,
by rotations of spatial coordinates about the z-axis. (7.88) is unaffected whether
1{zr(t) = - "fr'il· [kHo + i(HI)r cos w:rt + j(Hlh sin W:1t
or not r is transformed. As we show below, the usual approach transforms only
I and 5 in order to remove the time dependence of the HI couplings to the + i(Hl)S cos w:st + j(Hlls sin w:st] (7.90)
respective spins.) where W:1 and Wzs may be positive or negative, 10 represent either sense of
Taking the classical view, if Ms is precessing around HI)S in the S·spin rotation, and where
rotating frame, M~s will be oscillating sinusoidally at ils = 'YsHt)s. This
produces, via 'J-idA (7.88), a time-dependent coupling to M:r. But M:r is the 1= I:Ij (7.90,)
j
component of M r transverse to HI)r, hence the coupling produces the same
effect as applying an alternating field along the z-direction of frequency ils. is the total spin vector of the I-spins.
When the driving frequency as matches the I-spin resonance frequency il r, the The terms ('J-id)rs, etc., represent the magnetic dipolar coupling of the I-
S-spins cause the I-spins to absorb energy, and Mr to nutate away from the spins with the S-spins, and so forth.
direction of HI) [. We now wish to transform to a rotating reference system. In so doing,
We have described two spin systems which are coupled. We know with we are following Redfield as discussed in Chapter 6. However, our problem is
coupled systems that when the natural frequencies coincide, resonant transfer somewhat different from his since we have two rotating fields. We therefore wish
of energy results and we need to worry about the back reaction. Thus for two to transfonn in such a way that we view the I-spins and S-spins in Iheir respective
coupled pendulums, one al rest initially, the other set in motion, after a while the reference frames. The transformation is readily accomplished by introducing the
first reaches a maximum amplitude, with the second one at rest. Then the energy unitary operator T defined as
exchange reverses, the pendulum which was driven now drives, the pendulum
T = exp (iw~rIzt)exp (iw zs5zt) where (7.91)
which drove now is driven.
If we had only a pair of spins, a similar situation would occur. However,
typically there are many coupled I-spins, many S-spins, perhaps coupled, per-
I, = I:I,j S, = I:S,k (7.92)
;' k
haps not, but in different neighbor environments in any case. For such a case it
are the total z-components of angular momentum of the two spin species.
becomes useful to assign a temperature 8/ to the I -spins in their rotating frame
We define a new wave function 1/;/ by the equation
and a temperature 85 to the S-spins in their rotating frame. Under these circum-
stances, M r always poinls along HI)r and Ms always points along Ht)s, but (7.93,)
278 279
then substituting T-1ljJ' for ,po SchrOdillger's equation becomes and similarly for 'H~SS' To these may be added the pseudo-dipolar and pseudo-
exchange couplings when their size is large enough to be imporlanl.
_!!. atjJ' ::; 'H'.// (7.93b)
i at or II is the tenn 1£3,5 which gives rise to the effects observed by Bloembergen
and S.orokin [7.27] in their studies of CsBr. They found that the rapid bromine
where H' is a lransfonned Hamihonian. spin-lattice relaxation could communicate itself to the Cs nuclei through this lenn
Explicit evaluation of 'H' using the techniques of Chapter 2 gives when the Cs nuclei were quantized along their own HI.
'H! = - 1'11/[(Ho + W:lhl)lz + (H ')1 IzJ We can view the various tenns of (7.96) as energy reservoirs of Zeeman
or dipolar energy. Sioce the various tenns do not commute, they can exchange
-1's1J[(Ho +w,shs)Sz + (H,)sSzl + ~Il + ~lS +~ss energy. Such processes may be tenned cross-relaxation in lhe double-rolating
+ lime-dependem terms we ignore . (7.94) reference frame. The rate of cross-relaxation will depend on how the energy
The lenns 1t~11' elC., represent thai pan of the dipolar coupling 1tdl1 that com- levels of the different tenns match, on the hcat capacities, and on the strength
mutes with the Zeeman inleraclion between the spins and the static laboratory of coupling as measured by the failure of tenns to commute with one another.
field Ho. These terms are usually called the "secular part" of the dipolar inter- Thus, we note that the Zeeman tenn of the I-spins 1£Z! commutes with the
aclion. We write them OUI explicitly below. Zeeman energy of the S-spins 'H.zs. However, as long as (Hlh /: 0, 1iZ! does
The lime-dependent terms are of two sons. One variety arises from the not commU{e with either ~11 or ~,s' and we can transfer energy between
nonsecular parts of the dipolar coupling. They oscillate at frequencies Wl1. wzS. 1izl and 1i311 or 1i3,s' Moreover, 1t~,S provides a coupling mechanism to
or w::/ ±wzs, The second sort arises from couplings of the [-spins to (HI)s transfer energy between 1tz! and 1tzs [provided (H tls /: OJ.
and the S-spins to (Bd,- These oscillate al a frequency (wz/-wzS). Since wz/' All these remarks lead one, following RedM/d, to assume that if one waits
WzS. and (WzI ±wzS) are all far from any of the energy level spacings in the long enough, the various pans of (7.96) will come 10 an equilibrium in which
rotating frame, they can be neglected. One must remember, however, that it is the system can be described by a common temperature 8. For some purposes, it
conceivable that should there be a quadrupolar interaction added, and should the may also be possible and convenient to assume that various parts may comc to
two nuclei have similar ,'s (as do, for example, Cu63 and CuM), the frequency common temperatures faster than the whole system achieves a single temperature.
(Wzl - wzS) might, in fact, be close to a possible transition. This is the viewpoint Lurie (7.29] adopts to calculate some cross-relaxation times.
If one chooses We therefore, make the assumption that when the system has achieved a
common temperature it is described by a density matrix I! given as
(7.95)
exp(-l£'lk8)
where Hoo is a particular value of Ihe stalic field Ho around which we laler
wish to make variations, then the two nuclei are exactly al resonance. Note that
,= T, {e,p( 'Ii'lk9)} (7.99)
the two w's are negative if the 1"S are positive, representing the fact that nuclei where 1£' is given in (7.96). In lenns of I! we can calculate the average energy
of positive, rotate in the "negative" sense about Ho. Making use of (7.95), E and the average magnetization vector (M r) in Ihe high temperature approxi-
defining ho by Ho - Hoo '" ho, and neglecting the rime-dependent lenns of 1t', mation
we have
E = T, {,'Ii'}
1£' '" -1'rh[hoI: + (Ht),I:z] -1's1i[hoS: + (HdsS:z]
= _ C,[(H j )7 + h5 + HCl + Cs[{Hj)~ + hij]
+'H?t11 +'H~'S+1-l?tSS (7.96) 9
(7.100a)
It is convenient to define the Zeeman energies l£zl and 'H.zs by the equations
(M r ) = T, {e1r H } = Cr(~"'>r (7.100b)
(7.97)
where C, and Cs are the Curie constants given in tenns of the number of J. or
S-spins per unit volume N J or Ns, and Boltzmann's constant k by
C _ 1'yh2I(l + I)N ,
(7.98) J - 3k etc. (7.101)
202 283
the S-spins among themselves. All these effects can be rigorously and simply
included. In the process, Lurie demonstrated that it is nOl necessary for the HI's £= (1.112)
to be large compared 10 the local fields and further demonstrated the coupling in
cases where (H»/ has been turned to zero. In the process the magnitude of (M,) drops from its initial value (M/)i to
We now tum to an analysis of the Lurie experiment. We shall assume a final value (M/lr which, in view of Curie's law, is
throughout that spin-lattice relaxation can be neglected during the times of the
experiment. Spin-lattice processes can be included readily, but one must ~ care- (M/)f/(M/); = 1/(1 + e) (1.113)
ful in so doing to include the sort of transverse Overhauser effects descnbed by We now suddenly turn off (Ht)s. Once again the system immediately after the
Bloembergen and Sorokjn [7.27]. change has the same wave funclion as it did jusl before. The expectation value
We begin by the demagnetization process. This brings (M/) down along of 'Hz, and of the dipolar energies is thus unchanged, but that of 'Hzs is zero
the x/-axis, ils magnitude being given by (1.105). Let us call this magnetization since (Heff)S = O. The tOlal energy E} is therefore
(M/}j. During this process, since (Ht)S is zero, 11.zs commutes exactly w.ith
the rest of the Hamiltonian. So, likewise, does 'YshSz. Therefore, (Ms) remains CJ!(H,)j + HlJ
(1.114)
unaffected, and points along the static laboratory field Ho. The rest of 11./ is at Of
a common temperature 8j which we can compute from (7.l00b) and (7.105): Immediately afler the tum-off of (H»s, (Ms) is nonzero. Therefore, we see
0; = C/(H,)//(M/); (1.106) that we do not have thennal equilibrium. If we wait for a sufficiently long time,
the entire system will come to a common temperature 8n" In the process, (Ms)
This temperature is, of course, very much lower than the lattice temperature 8/. decays 10 zero. This is an irreversible decay. We shall in fact calculate the entropy
We now tum on (HI)s suddenly. In such a rapid process, the state of the increase below.
system does not change. The dipolar energy and the Zeeman energy of the I-spins When the system has reached the final temperature 8rr, the energy is Eff.
are therefore unchanged. The S-spin Zeeman energy E s Using (7.100a). we have
(1.101) Ejr=-CJ!(H,)j+HlvOff . (1.115)
since (MS) and (Heff)S are perpendicular. The total energy of the system Ei But Eff = E} since the spin system is isolated from the outside world and has a
Hamiltonian independent of time. Therefore, using (7.114) and (7.115)
is therefore
8r =8ff . (1.116)
Eo = _ CJ!(H,)j + Hl1 (1.108)
I 8j Using Curie's law we see thai following the lurn-off of (Hils, (M/) does nOl
After a sufficiently long time, the S-spins Zeeman energy comes into thenna! change.
equilibrium with the rest of the system at a common final temperature 8r· The For one complete on-off cycle, therefore, we can argue that (M,) is reduced
final energy E r is then by the factor 1/(1 + E).
We can repeat the argument for another on-off cycle. The magnetization
_ CJ!(H,)j+H1J+Cs(H,)~ (1.109) M/(N) after N cycles is thus given in tenns of its value M/(O) prior to the first
Er- - 8r cycle by
But since the total system is isolated and its Hamiltonian independent of lime, M/(N)/M/(O) = [1/(1 +e)IN (1.111)
its energy cannot change. Therefore,
When £« I, as in our experimenls, we can write this as
Ej=Er (1.110)
N!J(NVM(O) = e- N <
This gives us that
where £ is given in (7.112).
8j _ C,(HI)l + Hl] (1.111) Double resonance is possible even when (H 1)/ «HL. Experimentally we
-- 2 =--
8r C,(H1)1+HlJ+Cs(Ht)s 1+£ accomplish this by perfonning adiabatic reduction of (HI)' after M, has been
brought along (H I)' in the rotating frame. After (HI), ::::::0, /11, ::::::0 from Curie's
where law, however, we have retained the order in the I-spin system, the order now
284 285
being with respect to the local field [7.30]. (Hj)S is now cyc~ed o~ ~nd off N 60
.. '.... = '",," I f1/StC F1g.7.J6. In(Ah) versus (lldl
times. After the Nth cycle, (Hlh is adiabatically returned to Its onglOal val.ue. II.l 1.5K (or N ..... 22, ,011. =
The resulling M[ is then observed by rapidly turning off (Ht)/ and observmg 50 o '.,.. - '"," 2mst>c
o 1. . - t",= 1001/Stc
low = 1,2, and IOms, (lId7 =
the free induction decay. 2.30G. For this value or (11.)7,
.~ (111)' =: S.tG salisfies the
\
'1.~.
TIle analysis for the case when (Ht h «. H L is essentially the same as • IIlI.hn condilion. The solid line
given above; however, now, the tenn C[(H1)iI8 no longer appears in (7.108) is cakulated (rom (7.112) and
and (7.109), and E: reduces to
• (7.117)
•
E: =C5(HI)~/C/Hl (7.118) • • • •
a
7
Lurje studied Li metal for which the 93 % abundant isotope Li gives the
strong or I-spin resonance; the 7 % abundant isotope Li 6 is the weak or S-spin
system. To observe the effect of the double resonance, he measured the amplitude \. a .'
of the Li 7 free induction decay afler N mixing cycles.
Figure 7.15 displays Lurie's data demonstrating the effect of N on the
destruction of the Li 7 signal. The solid line has no adjustable parameters apan
\"
from the N = 0 intercept.
10023456 7 II ') 10
U·1I 1 (gUlI$J)
Figure 7.16 shows how the spin temperature idea works even when the
Hahn condition is nOl perfectly satisfied. Note that the solid line has no adjustable
parameters. The deviations for larger H I )6 and shoner ton and tolr arise because
there is not enough time for a common spin temperature to be established.
In this example a very large destruction is produced. Alternatively, one can
conclude that a much smaller fraction of Li 6 nuclei could produce an observable
destruction. In this case one begins to worry about the long distance that energy
must diffuse from the hot Li 6 spins to couple to the distant Li 7 's. This problem
has been investigated experimentally by several authors [7.23.31.32].
It is interesting to follow the changes in energy and entropy of the spins that
take place during the double resonance. The essence of the experiment is the
heating of the I-spins brought about by the contact with the hot S-spins. There
is a net flow of energy into the system as a result of work done on the S.spins.
The destruction of the I magnetization corresponds to an irreversible loss of
10 0 40 SO 120 160 200 order, that is, an entropy increase. The energy of the system is, of course, the
Numbtf Qll'lIlst$ (N) expectation value of the Hamiltonian of (7.96). Basic theorems of quantum (and
t·lg. 7.15. Lurit's experimentsl dll.ta showing In(M7) versus N at 1.5K. (l1d7 =:= 2..14~, classical) mechanics tell us that the total energy remains constant as long as 1t
(1I 1 )e = 504 G, nearly satisfying lhe lIahn tondillon. 4. =: tow =: 4 ms. The _lid IlIIe IS does not explicitly depend on time. Rea.mngements of energy within the total
calculll.l«l usirl& (7.117) II.nd (7.118) system, even when 1£ is independent of time, we identify as a heat flow within
286
287
the system. Changes of the total energy due to varialion of an external parameter (1 = k[N/ In (2J + I) + Ns In (25 + 1)]
we call work on or by the spin system. C/[(Ht)~ + HlI + Cs(H I )1 (7.123)
We can follow the cycle by considering work and heatlfansfer in the rotating 20'
frame. Consider one complete cycle of (HI)S on and off. We stan by turning on
(Note that (7.105), describing the adiabatic demagnelization, follows from use
(HI)S suddenly. Bearing in mind that the 5 Zeeman energy is
of (7.IOOb), Curie's law, together with the requirement that the entropy given
(HZS) = -(Ms)· (H,)s (7.119) by (7.123) remain constant.] Since we have worked out the temperature at each
part of the cycle, we can use (7.123), together with the approximation that E: is
and that during the sudden tum-on the dipolar energy does not have time to
small, to find that in one complete cycle starting at temperature 8 there is a total
change, we see that it takes no work to tum on (HI)s since (Ms) is initially
increase in entropy (1r - O"j of
zero. The establishment of the 5 magnetization causes (1tzs) to go from zero to
a negative value. That is, there is a heat flow from the 5·spin Zeeman reservoir CS(Ht)~
to the rest of the spin system. That this is the direction of heat flow is reasonable ur-O"i= 82 (7.124)
since the initial zero (MS) in the presence of a nonzero (Ht)s can be viewed Half of the increase occurs following the tum-on, the other half following the
as saying that 1tzs has an infinite temperature. Associated with this heal flow
between systems at different temperatures there musl be an entropy increase.
tum-off at (HI)s. We note that the larger (H tls, the larger the change of enlTOpy
per cycle. Since the existence of M, is a sign of order, we see that a large (H t)s •
Following establishment of (MS)' we tum off (Ht)S' Using (7.119) we
can see that we must do positive work on 'H.zs in the process. [Note that during
leads 10 a large destroction of M" as was, in facl, already expressed by (l.118). I
the tum-on or tum-off, which takes place suddenly, there is no time for heat
flow, so that we can compute the work done solely from the changes in (1tZS)
given by (7.119).) Following tum-off, (Ms) decays irreversibly to zero. Again 7.16 The Effects of Spin-Lattice Relaxation
there must be an entropy increase associated with the irreversibility. We are now
ready to repeat the cycle. Note that we have done a net amount of work on the In Sect. 7.11, we saw that Bloembergen and Sorokin observed IWO striking ef·
spin system, and thai there has been an irreversible loss in order. feclS in their pioneering double resonance experiment. As we have mentioned,
Had we turned on (Ht)S in the next cycle before (Ms) had been able they gave a full explanation of their resullS. We tum now to an ahemative de-
to decay, the S·spins would have done positive work back on us. In fact, had scription of their experiment to show how one can think of their experiments in
(Ms) not decayed at all, we would have gotten back as much work as we put teons of two concepts: (I) spin temperature in the doubly rotating frame and (2)
in when we turned off (H ,)s. We would not then have done any net work in a Redfield theory of the equilibrium spin temperature reached in the rotating frame
cycle. Moreover, apart from the effects at the original tum-on, there would have when spin-Iauice relaxation is included. Thus, we show that their experiment
been no irreversible loss in magnetization of either spin system. We must allow can be viewed as a combination of Redfield's experiments on saturation with the
a sufficiently long time for the irreversible features to occur. Hartmann-Hahn and Lurie·Slichter experiments.
The entropy 0" of the system can be calculated starting from the basic equa- As we explained in Sect. 7.11, Bloembergen and Sorokin observed a species,
tion [, (Cs 133 in their case) which had a long T\. The (H I )/ they used was such that
E+k81nZ one should consider the I·spins to be described by a spin temperature in their
u = 8 (7.120)
rotating frame. They observed that for the [-spins, TIl? was deteonined by the
where Z is the partition function. Making use of the high-tempemture approxi- T1 of the 5-spins, leading to a circumstance that (T'e)/ «(TI )/·
mation, we can evaluate (7.120) to get They observed further that if they were on resonance for the I-spins
288 289
MI = ±M/(O"Ho)1s , (7.127) we can set
1/ Ta ! = (T1 )! (7.133)
where M/(ot. Ho) is the I magnetization when the I-spins are in thermal equi-
librium in a static field Ho with a lauice al temperature (J/. for the S-spins we will have. in analogy to (7.132),
To understand how (Tt)s can determine T 1g for a case in which (Hils = aMzS Mos - M"s
0, we follow the Bloembergen-Sorokin analysis. We consider a case in which --- with (7.134)
/)t - TIIS
(HI)j::> HE
where HL is some sort of local field. Then the I-spins. if exactly
al resonance, are quantized along the I-spin x-axis, the direction of (H I )/. and and (7.135)
the S-spins are quantized along the z-axis, the direclion of Bo_ The term 'Hd/S •
given by (3.55), involves products such as I;k5zp. which have matrix elements oMzs Mzs
(7.136)
which are diagonal in Szp. bUI are off-diagonal in J;J;p- However, owing to the ~=-TbS
spin-Ianice relaxation time of the S-spins, the diagonal matrix elemenlS of Szp where TbS ...... (TI)S.
are functions of time with a correlation lime of (Tl)S' Defining We will have three dipolar relaxation equations similar to (6.87):
Ckp=
"'fnSh2
3
rkp
(I-3cos 6kp )
2
(7.128) ~(7i )_
fJt dfJ -
( 7idll)
T,11 (7.137a) •
BlQembergefi and Sorokjfl show that
_'_=.2..l: C 2
(TI,,)1 1i 2 k,p kp
5(5+1)
3
(T,)s
I +wrl(TI)~
(7.129)
(7.137b) I
( 7idlS)
The reader should compare this expression for the relaxalion of (Iz ) along (HI)z (7. 137e)
TeiS
with the expression for TI of (5.372) which applies to the relaxation of (I,,)
along Ho. Physically, the fact that the z-component of the S-spins is fluctuating where we have neglected the tenns like (~) of (6.87).
produces a time-dependent magnetic field on the I-spins along the z-direction. If we assume that (Tt)s <: (Tt ), the lattice will flip individual S-spins much
A field of this polarization is ttansverse to the I-spin quantization direction, more rapidly than individual I-spins, so we can neglect any rates dependent on
and can therefore relax the I-spin magnetization along its "static" field in the I-spin flips.
rotating reference frame. The factor TI/(I +wr,Tls ) gives the spectral density at Therefore
the frequency WII which can produce flips of the I-spins quanlized along (HI)!. (7.138)
If (HI)s is turned on to a level sufficiently strong that the $-spins obey
the Redfield spin-Iemperature condition in their rotating frame (6.84), then the and both will be much shorter than Tell - Likewise
magnitude of the S-spin magnetization, Nls, and of the I-spin magnetization,
M" will be given by an argument such as is given in 5ecI.6.9. The calculation TIl& <:Tb/ (7.139)
we described above for (Tle)l can be viewed as a calculation of Tb' defined by In analogy with (6.91) we will then find for the Bloembergen-Sorokin case
a modified (6.86): - 0
E = - (ho)M: s - (H,)/M:r./ + (lldfJ )
z / = - M:z:d
oM-
- w'lh
, (7.130) + (7i?'/s) + (~SS) . (7.140)
at 'I Then
(7.131)
aE I oM:s aMzl
at = - (IO)S---al - (Hd,---at
The z-component of M, will not be relaxed directly by the time variation of a"~ a"~ a 0
1-$ dipolar coupling since that coupling commutes with I:. Thus in the equation + at ("dll) + a,!"d/s) + at (7idSS ) (7.141)
aM:J Keeping only the fast tenns which relax at rates comparable to (Tt)s, we get
- - = MOl - M:J (7.132)
at (Til)!
290 291
BE = -(ho)S (Mos - M.s) _ (1i~15) _ (1i~SS) (7.142) Ml = " 1(1'0)51 MOl (7.150)
at T as TclS TeSS iJ (ho)S
We now assume the system has a common spin temperature B so that The ratio l(ho)sl/(ho)S is either +1 or -1 depending on whether ws lies below
or above resonance respectively. The two possible signs reflect the fact that 8 is
M - Cs(ho)s (7.143a) positive in the fonner case [Ms parallel to (Hcff)s] and negative in the laHer
zS - 8
[Ms antiparallal to (Heff)s].
The size of MJ is larger than its thermal equilibrium value in the ratio
M - ClCHill (7. 143b) ishJ' This result is just the result of the Overhauser effect, hence the name
:d - ()
"transverse Overhauser effect".
In considering this result, we see that it follows very directly from the
o
(1i dIS ) = -
CsHJs
e (7.144.) realization that if the S-spins are far from resonance
Ms =Mos (7.151)
(1t~ss) = - Cs:~s (7.144b) and that the I-spins and S-spins are at a common spin temperature in the doubly
rotating frame.
where Hys and H~s are defined by (7.144). We could achieve the same result by an appropriate adiabatic demagneti-
To find the equilibrium spin temperature, we set dE/dt = 0 and substitute zation experiment. We tum on (HI)J and we tum on an (HI)S which satisfies
(7.143) and (7.144) into (7.142), gelting the Hanmann-Hahn condition at a frequency Ws far from resonance. We then
sweep ws onto resonance. The two spin systems will then mix. However, since
Cs = Mos(ho)s/(ho)~ + Hl ls + Hl ss ) (7.145) they have comparable heat capacity, the common spin temperature they reach
() Tas TaB Tcls TeSS will be holter than the temperature the S-spins would reach if the spins did not
If one is off resonance (lto)s by a sufficiently large amount compared to mix. So we tum off (H.)s, go off resonance. magnetize the S-spins again, and
the local fields, one can keep only the [enn involving (ho)S in (7.145), getting repeat cooling the I-spins several times more. In this way, we can bring the
I-spins asymptotically 10 the value given by (7.150). However. we can only hold
Cs = Mos (7.146) this result for a time (TI/?)! given by (7.129). The Bloembergen-Sorokin method
(J (lto)s
gives one a steady-state MJ along (HI)J'
BUI, since Note that if one sits with the S-spins exactly at resonance [(ho)s = 0],
(7.145) and (7.146a) show that the equilibrium values of Ms and MJ would
Ms = Cs(ho)s (7.147.) vanish even if one satisfied the Hartmann-Hahn condition. Thus, to use the fast
8 relaxation of the S-spins to generate a strong I-spin magnetization, the S-spins
(7.147b) must be far from resonance (above or below).
MS=MoS
Equation (7.145) can be viewed as giving the spin temperature of the S-
Recognizing that spins. If the Hartmann-Hahn condition is satisfied, the I-spins will rapidly come
to this spin temperature. However. if one is far from the Hartmann-Hahn condi-
CsHo (7.148) tion, the I-spins will not reach this temperature, as is shown from the experiment
Mos=--
8, of Lurie et al. (Fig.7.16).
we get, combining (7.146) and (7.147),
M = C/(Hl)/ = C/(Hlh CsHo = (Hdl MOl (7.149)
r 7.17 The Pines-Gibby-Waugh Method of Cross Polarization
8 Cs(ho)s 8, (ho)s
But from (7.145)
The methods of Hartmann and Hahn or Lurie et al. involve detection of a rare
'"a (HilI = 'Ysl(1to)sl
species by observing its effect on an abundant species. Pines, Gibby, and Waugh
so thai we gel the Bloembergen-Sorokin result that [7.33] pointed out that under some circumstances it is preferable to perform
292 293
the experiment in a manner to observe the rare species. Specifically, suppose
We now lum on (HI)c again, repolarizing the C13· S along their HI. We
C 13 is the rare species (its natural abundance is 1.1 %), and HI is the abundant
again tum off (Hde and record the C l3 free induction decay. Each repetition
species. as when one is working with solid hydrocarbons. Each nonequivaicni of this cycle will heal the HI nuclei in accord wilh (7.112) and (7.117), so
C 13 site will have its own chemical shift. One would like to record each C l3
that the C13 signal gets progressively smaller until it becomes necessary to tum
chemical shift. We have seen (Sect. 5.8) thaI the Fourier transfonn method is
off (HI)}! to allow Ihe proton spin-lattice relaxation time to replenish the HI
then a very efficient way of collecting data. since a single C l3 7r/2 pulse will
magnetization.
excite all the C 13 lines simultaneously. If onc observes the HI resonance, one can
We turn now to a few remarks about the nature of the rare spin spectra.
only measure the individual C13 resonance lines using a point-by-point search.
If we recall that in general the spin-temperature approach requires that we be
However, if one brings the HI magnetization along (Hdll. the hydrogen HI,
working with solids. we should distinguish single crystals from powders. In a
one can rapidly polarize the C 13 nuclei by turning on their H I> (HI )e. to the
single crystal a given type of C I3 chemical site might have several different bond
Hartmann~Hahn value. Within a cross-relaxation time, Ihe C I3 will be polarized.
orientations with respect to Ho. For example, in a benzene molecule (C6H6), Ho
Since there are many more hydrogen atoms than C 13 nuclei, in this process the
might lie along one CH bond, in which case il would also lie along one other in
HI spin temperature will hardly change. Therefore
the molecule, but would make a 60" angle with the other four bonds. Thus we
-- GdHlk would have two C13 lines, with one twice the intensity of the other. If the sample
Me o (7.152)
is a powder, the C 13 spectrum of this bond would be a smear. By magic-angle
But spinning (discussed in Sect. 8.9) one could narrow this pauem to a single narrow
8 = Gu(H1h, and line at the average chemical shifl position of C I3 'S in Ihal molecular position.
(7.153.)
Mil
1'dHde = 111(Hd" so thai (7.153b) 7.18 Spin-Coherence Double Resonance -Introduction
Me = M u Ge
Gn
(I'll)
1'e
(7.154) The Overhauser-Pound and the cross-relaxation methods of double resonance are
both dependent on the existence of relaxation processes of sufficient vigor. In the
Suppose. now, Ihal initially MI-I has ils {hennal equilibrium value in Ho case of Overhauser-Pound schemes it is necessary that appropriate spin-lattice
mechanisms are Strong. In the case of cross-relaxation double resonance. cross-
GIIHo relaxation times must be sufficiently rapid (relative to spin-lattice relaxation)
MH = Mil (0/) = - - (7.155)
0, for the method to work. We now tum to the third family of double-resonance
The corresponding MC<8/) is given by schemes. For these schemes to work, Ihere must be a coupling between the
Ho two spin systems which manifests ilself in splillings of the spectral lines which
MdO,) ~ Ce9; (7.156) would exist in the absence of the coupling. We observe the double resonance
by producing some fonn of modulation of this coupling. Crudely speaking, we
Ulilizing these equations in (7.154), we find observe the effect on splittings of the I-spin resonance produced by exciting the
Me ~ MdO,) 711 (7.157) S-spin resonance. If 8w is a typical such spin-spin splitting. the essential condition
7e for viability of Ihis third family is that the effects of Ihe splitting not be obscured
This is the Bloembergen-Sorokin enhancement, characteristic of an Overhauser by relaxation processes. If one thinks of 8w as some sort of beat frequency, Ihe
effecl. period of the bealS is 1/8w. The requirement is then physically that the coherence
Having polarized Ihe C13·s, one turns off Iheir H" and records the free of this beat nOI be interrupted by relaxation. If T stands for the relaxation time
induction decay. Meanwhile, one leaves on the proton HI for two reasons: (1) (spin-spin, or spin-Ianice) which could interrupt the coherence, observation of
to maintain the prolon magnetization [spin-locked by (Hl)lI] and (2) to provide the coherence requires that T:p 1/6w. The requirement of long relaxation times
so-called H 1_C13 spin decoupling. This last topic we take up later in the chapter, shows that spin-coherence double resonance tends to work under circumsr.1nces
where we see that the presence of the strong H I acting on the protons effectively where the other methods do not.
wipes out the HI splillings of the C I3 resonance. By this means the C I3 resonance
consists of a family of lines each of which goes with a given C13 chemical shift,
but without splittings arising from Ihe C I3_H I spin-spin coupling.
294
295
7.19 A Model System - An Elementary Experiment: Letters next to each other in the alphabet have chemical shifts comparable to
The S·Flip-Only Echo their spin-spin couplings. Letters not adjacent have big differences in resonance
frequency compared to their spin-spin coupling, either because the chemical
shifts·are quite different, or because the nuclei are different species (HI versus
To begin the discussion of spin-coherence double resonance we are going to C I3 ). An A2 system has two nuclei with the same chemical shift. An AB2 or an
imagine ourselves to be working in resonance before the invention of all the AX2 system has three nuclei. In the fonner case the chemical shift differences
varied double resonance methods which exist today. We suppose that we have between the identical pair are not large compared to the A-B spin-spin coupling,
observed a HI resonance which is split into a doublet by coupling to another in the laller case (AX2) they are large. A notation AMX means three nuclei
nucleus (for example a C 13 ). Since the C 13 affects the HI spectrum, we ask with big chemical shift differences. Note that if two nuclei are different species,
ourselves is there not some way we can use the HI resonance to detect when the they would be called an AX system.
C l3 is brought to resonance? We shall treat an AX system, hence take
We first describe the model system, a pair of coupled spin ~ nuclei I and
1t "" -1IliHo[, -1sliHoS, + AI,S, (7.162)
S with a Hamiltonian similar to that of (7.29)
The energy levels are
1{ "" -"nhHo1, -1shHoS, + 1-£12(t) .
E"" -hwO/ml - hwosms + AmimS (7.163)
Initially we shall ignore relaxation effects, considering 1-£12 to be indepen-
dent of time. For simplicity we take it to be where
(7.158) WOI -= 11HO (7.164.)
This is the fonn of the indirect spin-spin coupling in liquids. If I and S repre-
(7. 164b)
sent different nuclear species, the tenns I~5~ and I1I 5 11 are nonsecular. Then we
approximate If we apply alternating fields at frequencies WI or ws (near to the resonance
frequencies of the I-spins or the S-spins, respectively), the resonance condition
(7.159)
i,
This is also the fonn which the dipolar coupling between two nonequivalent spins
WI =wO/-amS (7.165.)
takes if we keep only the secular pan, see (3.55). For indirect spin coupling in
liquids, it is conventional to substitute J for A, and to speak of the J coupling.
ws "" wos - ami where (7. 165b)
Since we wish to include dipolar coupling between unlike nuclei, we utilize A
as the coupling constant.
If I and S are the same species but have different chemical shifts, we can a -= Alll (7.166)
296 297
I-spin Fig. 7.17. I-spin absorption spedrum before ~fter
~bslH"ption for S-spins 100% l>olarizcd with (a) (,) Ibl
(,) "'5 = +t = -t
and (b) "'s vcrsua
I-spin scar<:h frcqucncy 1<1/
n
1.111 .1/2 w'I·~/2
", 1.1 11 - a/2
n ",
wtI· a/2
I (J Idl
I_spin
~bsorplion
(bl
-1/2
n
1.1'1' ~/2
", n
wOI-~12 1.1 01. a/2
",
,
1.1 01
,,
ms: 1/2
,
I
(e) If! l
Simple and beautiful as this scheme may seem, It ~Ias cannot be ~sed exactly the total of alii-spina (~,f). Note that though the s.-::trum of ~ilhu group (I) or group
(2) alone i$ changed by Ripping the S-spins, the loIal intensity pdtem (~.f) is unslTec:l.ed
as described without first gelting the S-spins polanzed, a very difficult .~k
indeed. In fact, at nonnal applied field Ho and temperatures 8, Ihe condillon
"Ys hHo/ k8 <. I holds, so that to a good approximation one has an equal number of their prior history. An experiment which displays a spin's memory is easy to
of spins ms = +l
as ms =-l· . . concoct. At t = 0 we apply a trn.
pulse to the I·spins with a sufficiently strong
We can reanalyze the experiment very Simply by saYing tha: half of the (HI)! to flip the spins in both the I absorption lines. We view the I-spins in the
I-spins have neighbors with ms = +l, Ihe ot~er half have mS = -'f.' Let u,~ call reference frame rotating at wO/. In this frame, we will have two magnetization
these two groups (1) and (2) respectively. Figure 7.18 shows the before and vectors of equal length, one precessing at angular frequency af2, the other at
"after" spectra for the two groups produced by the 11" pulse on the S-spins. Group -0./2. If we assume that immediately after the 7((2 pulse the I magnetization lies
(I) will have a resonance at WI = WOI - 0/2 before the S-spins are flipped, and along the y·axis [Ihe situtation if (Hj}1 were along the x-axis], the magnetization
a resonance at WI = WOI + all after. Group (2) will have a "before" re~nance will obey
at WI -- w01 + a/2 ' which will be shifted to WI = WOI - all. after the
.•
S·spinS are
flipped. We observe of course the spectrum of all the [-SPinS, which IS the sum (Mi('» = C;i (cos (at/2) + i sin (a'l2)]
of the spectra of the two groups (1) and (2). As we see from Fig.7.18e,f, the
total spectrum is the same before and after. It looks as though our scheme has
+ ~oi [cos (ain) _ i sin (ai(2)] (7. 167a)
failed. .
Fortunately, though the simplest version of our concept does not work, It
can be modified very simply in any of several methods to be made to work. The = COi(eiat/2 +e- iat / 2) where (7. 167b)
2
essential point is that anyone [-spin belongs to either group {l~ or group (2).
Since eirher group (I) or group (2) is affected by flipping the S-spms, both. group Co = (MI(O,» (7.168)
(I) and group (2) can tell the difference between whether or not the S-~PInS are
flipped. We must do an experiment which utilizes the memory that SpinS have gives the thennal equilibrium magnetization of the I-spins for the lattice tern·
perature 8,.
298 299
Then, at a time t1f we apply a 7r pulse to the S-spins. The group (1) spins Fig.7.20a-<:. p31 NMR signal versus time
for diethyl phosphite I(C2IhOhPHO). (a)
will instantly change their precession frequency from WOf - a/2 to WOf + a/2 and The precession of the p31 nuclei due to
the group (2) spins from WOf + a!2 to WOf - a/2. Thus, both groups will reverse lheir coupling to the IP nuclei; (b) and
their precession directions in the rotating frame. (Readers who are not content (c) the precession of the p31 nuclei with
the HI nuclei inverted al the indicated
with the present informal treatment can look ahead at Sect. 7.24 for a fonnal times. The dashed lines in (b) and (c) show
justification of the relationships we use here.) whal the precession would have been had
We can utilize the complex notation to express this rotation reversal. If a lhere been no II! inversion. (Data taken
by Charles Pennington wilh assislance from
vector is represented by a complex number A with initial components At: and
Je3n-Philipl>C Anscrmel, Dale Durand, and
Ayobeying David Zax.)
Thus
(Mly(l» = Co cos (at.f2 - a(t - t.)f2) (7.172)
At time t - t 1r = t 1r (t = 2tll")' (Mfy(t» will once again be Co. In other
words, we have produced an echo of the I-spins by flipping the S-spins. The S- A useful way of viewing this experiment is to plot the phases 4>1 (t) and
spins so-to-speak cause the I-spin "field inhomogeneity" (a two-valued function). 4>z(t) of the two groups of I-spins where we define their phases at t = 0 as being
By flipping the S-spins we reverse this field inhomogeneity. given by (7.169). From (7.171), they are
The theoretical time development of (Mfy(t» is shown in Fig.7.19. An
experimental demonstration of this phenomenon by Pennington et al. from the
author's laboratory for the p31 resonance of liquid (CzHsOhPHO is shown in 4>,(1)= C'f2J+ atf2 fOI t <tll"
(Tr/2) + atll"!2 - a(t - t n )!2 fOI t> t1f
Fig. 7.20.
(7.173)
4>,(t) =
C<(2) - «tf2 fOI t < t 1f
(7r!2) - at 1r / 2 + aCt - t n )/2 fOI t > t>r
These are shown in Fig.7.21.
If there were more than one paicof spins I-S, each pair with its characteristic
chemical shift, the situation would be more complicated, but the same principles
would apply. Let us label each I-S pair by k. (k = I to N if there are N distinct
pairs.) We shall assume the I-S coupling is only within a given pair. Then we
have N Hamiltonians of the fonn
o It. Fig. 7.19. The y·componenl of
the I-spin lllflgnetizalion versus 1ik = -liWOfkIzk - flWOSkSzk + AkIzkSzk (7.174)
S_spins time, t. Al t = t", the S-spins
inverted are in\'erled by " 1f pulsc so that, defining ak = Akl!l,
300 301
PhilSe iogln (7.176)
4'lltl iod 4' l ll)
,
{Mj(O}""" = C;i l)exp [- i(!7 lk - (lk!2)t]
The ideas we have just developed lead in a natural way to the concept of spin
decoupling, one of the earliest goals of double resonance experiments. We will
follow a pedagogical rather than a historical order.
The goal of decoupling is to simplify spectra. A typical NMR spectrum
consists of many lines arising from the combined effect of chemical shifts and
spin-spin couplings. Decoupling is a process which effectively eliminates the
spin-spin couplings. It is very important both for high resolution spectra of liquids
and spectra of solids, for example to eliminate the effect of proton spins on
C I3 spectra. We shall discuss removing the spin-spin coupling between different
nuclear species such as HI and CU. The basic idea on which all decoupling
schemes work can be understood physically as follows. The existence of two
lines in the I-spin spectrum corresponds to the fact that the S-spins have two
Fig. 7.22. Thc phll$C IIngles ~1 k(t) and ~H(I) versus time, t, for a ca.sc with lWO 1-5 pllit'll orientations, up and down. If we can cause the S-spins to flip back and forth
(k = 1 and 2). At 1 = I., the S·spins lire nipped by 11'. Al t = 'It:r, ~u(t) =,pH(I) for between the up and down orientations sufficiently rapidly, we should achieve an
bosh spin groups, indicating thllt bolh components of I magnelization for II given pair
sre collinear. Nole, however, lhal since lhe two pairs have differenl chcmical shifls, lhe effect much like the motional narrowing of resonance lines. That is, an I-spin
I magneli."aliOllS of lhe different pain are out of phase wilh each other will precess at a time-averaged frequency rather than at one or the other of two
discrete frequencies.
To study decoupling, we are going to pick a scheme which is easy to analyze.
= Coi L(exp [ _ i(WOlk '- ak!2)t]
2 , We will look at the effect on the I-spins of applying a sequence of 7r pulses to
the S·spins. We shall see that as the time between S-spin 7r pulses gets shoner
+ exp { - i(WOlk + ak/2)t] (7.175)
and shorter, the I-spin spectrum goes from two distinct lines separated by Q in
in the laboratory. Relative to some frame TOtating at WI we have magnetization angular frequency, to a single line at the average frequency.
(Mj(t)"",s' Defining There are two general approaches we could lake to showing this result. One
would be to try to calculate what the I-spin absorption spectrum Xl/(w) looks
302 303
like as we flip the S-spins al a progressively more rapid rate. Or, we could look (aJ F1g.. 7.23
at Ihe Fourier transform of the absorption line, which we do by looking at the (a) The magoetiution (M,,(t» ver-
sus I for the CAlle of a ,.. pulse applied
time development of the I-spin magnetization following a rr{l pulse applied to to lhe S-spins at I = /'" producing a
the I-spins. In the absence of S-spin pulses, the I-spin magnetization will consist refocusing al t = 21 .. , followed byan-
of two components, one oscillating at WOI + 0/2, the other at WOI - 0/2. So, if other,," pulse at t = I~, with a sc<:ond
refocusing at time I". (b) The effect on
we choose Wf = WOf. in the Wf reference frame the components will precess (M 'J'(t)} of applying a pair of ... pulses
at angular frequencies +a{l and -0/2. If the system were perfectly decoupled, to the S-spins at times I.. and I = 31 .. ,
a would become effectively zero, so both components are at rest in the WOf >t. with nsulting echoes at t = 2t.. and
<t.
reference frame. and the transverse I-spin magnetization will be a constant in
time. (We are not including relaxation effects in our Hamiltonian. If we did, the
(b)
I-spin transverse magnetization would decay rother than remaining a conSlllnt in
time.)
We shall follow the Fourier transfonn approach for whi<.:h, in facl, we have
already set the srage in Secl 7.19. There we analyzed what happens if we apply
a single 'lr pulse to the S·spins al a time t = t ....
Let us then look back at Fig. 7.19 and at (7.171), which show that at time k
t = 2t,.. the magnetization (Mt(t)} has returned to its value at time t = 0+, II
immediately after the 1r{l pulse. lllOugh Ihe magnelization al the time of the 11" '. 21.
pulse is smaller by the factor cos (at ...{l), the 1( pulse has produced an echo al
t = 2t.... If we now think of ourselves as having a fresh start at t = 2t ... , we
realize we can repeat the echo by applying a second 'K pulse to the S-spins. If
lal
we apply il at a lime t~, hence t~ - 2t... after the first echo, we witt again fonn
an echo. It will occur at a time 2(t~ - 2t ... ) after the first echo, or at a time til
given by
til = t~ + (t~ - 2t... ) (7.179)
see Fig.7.23a. If we chose t~ to be given by
t~ = 3t... then (7.180)
(7.181) Ibl
so Ihat the time delay of the second echo after the first is identical to the time
delay of the first echo after the initial signal. We can repeat this process again
•
and again. The situation is shown in Fig. 7.23. ./2 i'E~-::"':;:'--:7<E:~-::"':;:'--:7i;<E:---
,
The phase angles ;,(t) and ¢2(t) of Fig.7.21 will then appear as shown
in Fig.7.24a. If we choose a shoner time for t,.., as shown in Fig.7.24b, the
".
maximum excursions of ¢l (t) and ¢2(t) away from tr{l will be less, so the
magnetization versus time will appear like the solid curves in Fig. 7.25 mther Jo'ig.7.24. (a) The phl\.Se angles .pI (t) and .p,(I) of the two components of I-spin magneti-
than the dashed curves. Clearly, (Mfy(t)} will be a periodic function as long as =
zation. The,," pulsoes: applied to the S-spins at I t .. and 31 .. cause.pl and .p, to become
there is no relaxation, with period, T, equal 10 2t,... We can think of (Mfy(Y)} as ~ua.l at I = 21 .. and 'It,.., correspo~ding to the maxima in the curves of (M,.(I)} shown
m Flg.7.23b. (b) The effect on <h(t) and .p,(t) of shortening t .. to half its value in (a)
consisting of a constant with a superimposed periodic ripple. Figure 7.25 shows above. The echoes (.pl(t) = .p,(t)J occur more frequently, and the maximum difference
that the shoner t ... , the smaller the ripple, and the more (Mfy(t» approaches a between .pl(t) and th(I), which occurs lotI = t .. , 31 .. , 51 .. , dc., is reduc:ed
constant in the reference frame rotating at wOf, hence behaves as though a were
zero. Thus, we have "decoupled" the S-spins from the I-spins.
304 305
Fig. 7.25. The effect on (Mfr(l» of halving the We can summarize for the case thai we flip the S-spins sufficiently fre-
Lime I ... The solid-dashed curve corresponds to
I .. of Fig. 7.24a, the solid-only curve corresponds quently: following a 7r/l pulse applied to the I·spins. they precess as though the
/ Lo Fig. 7.24b. Note that ('""ffr(t» (!Ill be Vlewed coupling to the S-spins were turned off. This method of decoopling was in fact
as a steady value with a superposed ripple, that proposed by Freeman et al. in 1979 [7.37]. They call it "spin-flip" decoupling. In
the ripple is periodic with period 21 .. , and that
the amplitude of the ripple diminishes as I.. is their paper, they discuss its advantages over OIher methods, as well as discussing
shortened interesting variants.
Another method of producing such a zero time-average z-component is to
apply a constant HI at the S·spin resonance which will cause the S-spin to
precess about the HI' As a result. the z-component of 5 oscillates sinusoidally
We can eltpress the eltistence of the ripple mathematically if we wish. by at angular frequency 'Ys(HI)s, This method is historically the first method em-
recognizing from Fig. 7.25 that since (M/y(t)) is a periodic function with period ployed by Royden (7.34] and by Bloom and Shoo/ery (7.35].
2t r it can be eltpressed as a Fourier series To analyze it. we talee the Hamiltonian to be
=
(MJy(t» = L All cos (21fftt(J') with (7.182) 1(. "" - "tJItHoI: - 'YshHoS: + AI;S;
n=O - ')'sMH ds(Sz cos wst - Sy sin WSt) (7.186)
T = 2t.". (7.183) We have included the Zeeman interactions of the two spins, their spin-spin cou-
pling, and the interaction of a rotating magnetic field (Ht)s with the S-spins.
The coefficient of the constant tenn, Ao. then turns out to be We have nof included (he interaction of (H ds with the I-spins since we assume
A = (M (9 » sin (at 7l"/2) that Ws is close to the S-spin resonance, but far from the I-spin resonance.
(7.184.)
o J I at"./2 We now view the problem in a double reference frame which is the lab
frame for the I-spins and one rotating at ws for the S-spins. That is. if the
which shows that as atrn shrinks. Ao approaches (M/(UI»'
original Hamiltonian ?i and wave function 1/1 obey
The higher coefficients are
300 '<)7
This Hamiltonian can in fact be solved exactly, but we will leave the exact (m/mslIIz:lm~ms') = (m/IIz:lm'/)om m' (7.198)
s' S'
solution as a homework problem. Instead, to solve this equation, we assume thai
so that the energy changes aE are
we have applied a strong decoupling field so that
(7.193) A.E(ms') = "UltHo - Ams' cos (J
(ho)s
and treat the tenn AlzS z as a perturbation. It is then convenient to define axes = "'1lhHo - Ams',=~~~~ (7.199)
(x', y', Zl) such that Zl lies along (Heff)S' and the y- and y-axes are coincident ";(ho)~ + (Hl)~
(Fig. 7.26).
Corresponding to the two values of ms' = ±!. there are thus two I·spin res-
onance tines whose spacing is a function of how far ws is tuned away from
, the resonance frequency wos. At exact S-spin resonance, (ho)s = O. the I-spin
resonance consists of a single line. Figure 7.'2:7 shows how aE(ms') - 'nhHo
goes with (ho)s/(HI)S, demonstrating the collapse of the splitting as one tunes
the S-spins to resonance.
• ,
ting, i1E - '"f,U/ o , seal in the I-spin res0-
nance as a fundion of (ho)s the amount, the
S-llpin ill off resonance, showing that the two
transitions collllp6e to a single transition as
the S-spins approach their resonance condi·
A tion (ho)s = 0
FIg.1.26. The effective field acting on the S-spil"llll
in their rotating frame as a result of (lids nd
of being some..·h...t off-resonartee for the S-.pins
((ho)s ct OJ An
This gives
Sz = Sz' cos 8 - Sz:! sin 8 with (7.194.)
-A/2
cos 8 = r~(;;;ho:;)g,S="" (7. 194b)
J(HI)~ + (ho)~
-A
The unperturbed Hamiltonian ?to' is then
Ko = --YJhHOlz - "YS(Heff)SSz' (7.195) The exact solution of (7.193) introduces another feature, the fact that the
direction of the effective magnetic field acting on the S-spins depends on mI. the
with eigenstates lm/ms,) which are eigenfunctions of I z and Sz" The pertur- orientation of the I-spins. This same problem arises in electron spin resonance
bation AlzS z then becomes and is treated in Chap. 11, (11.83-88). There it is shown that as a result the
AlzSz = Alz(Sz' cos (J - SZI sin 8) . (7.196) frequency of transition for .1ml = ± I is modified, and in addition transitions
nonnally forbidden in which ms' changes become possible.
But Ihe tenn in SZI has zero diagonal element in the Im/mS') representation, Decoupling by applying a steady altemaling field was. as remarked previ-
giving for the energy eigenvalues ously, demonstrated by Royden [7.34J and by B/oom and Shoo/cry [7.35] based
E'mJm S' = -""f11IHoml - "Ysh(Heff)sms' + Am/mS' cos 8 . (7.197) on ideas of Bloch {7.36J presented in an invited talk at an American Physical So-
ciety meeting. In Fig. 7.28 we show data from Bloom and Shoolcry, demonstrat-
A weak rf field tuned close to the I-spin resonance can then induce transitions ing experimentally the collapse of the splitting as the S-spin oscillator approaches
with a selection rule given by resonance.
308 309
.'lg. 7.28. Dcmonstntion by Bloom and Noise Modulalion
Jll1111ul1J Jl
Slwolery of the effect on the F19 resonance
of varying the p31 resonance frequency for
an (11 dp of 800 Hz. L}v is the amount (in
Hz) that the p31 is off-resonance. As L}v
"" "' "" ",
zoo Ht
"" "' 400 Ht
No'M;;;;; lar
No double ,esona~
Fig. 7.Z9. Comparison of the dceoupling with random noise and with a coherent rr fre-
quency. The FIB spedrum of CHFCl z is observe<! while the proton frequency region is
strongly irradiate<!. The spectra are taken for several different frequency offsets .1", =
"'s - '"'2, where "'5 is the proton resonance frequency and "'2 is the carrier frequency
of the strong rf field. The rr amplitude is identical throughout all the spectra and is
"Ys1h/2'K = 1370Hz. The random noise is obtained by a binary phase modulation of
the carrier with a 1023-bit shift register sequence with a shift frequency of 3 KHz. (Prom
[7.401)
310 311
Fret/lIt'IIC)' of Clf 1800 pl,lse mell$lfred _ 50~~-~~--'~~~-~~
• 45
,e/atil'e 10 Clfi) mllill line (k1I~)
• r-'.~12~.,----_----,~~.,---__~4~.,---- .,:-__---,4~.~ , 4.
'2
>-
o Spin Echo
,, .~ 35
.... ....
, • SEDOR
,
• • • • •
......• : ~ 30 , ,
:
~ 20
~
.. '•
.; "";;25
~20
~ .~
: • • •
1.'3
'."§
40
60
•
•
•
•
•
§
z
g 15
~ 10
5
\
..
~ ••
• " ' ,•
.L--'-£-,--L-c--J~~',---'-.-,+c__--'-~J.
o 0 0 0
~ • PoSitiol' of 0,·) 1.07 1.08 U)9 1.10 1.11 1.12 1.13 1.14 1.15
j 80
•
.' mallllllle
1-10'*'0 (kG / MH%)
Fig. 7.31. Measurement ofMaki:Jwla et at (7.521 of the spin·echo (0) and SEDOR (.) line
shapes of Pt lU for a sample of small particles of Pt metal supported 011 J\12~ whose
'ooL ---'--- ---.J surface is coal.ed with C 13 0 molecules. The metal particles have diameters of a few tens
of A. The Pt lU spin echo gives the totlll line shape of all the Pt ltS in the particle. The
Fig. 7.30. Boyce's observation of SEDOR in • Cu powder containing O.~ at. % Co near SEDOR data, involving pt 1U _C 13 0 double resonance, and lin add-subtract method, give
the Cuu resonance fT«luence of pure Cu at 9901 Gauss and 1.5 K, showing the satellite the NMR line shape of the PllU nuclei in clo5e proximity to the CO molecules, i.e. the
due to the 1 .... _ 1 transition of lhe first neighbor. The parameters UBed are T = 250"" surfll.Ce layer of Pt atoms
1I is :::::5Ga:" for ~"If pulse of the Cu t .... - t
transition.. IOO echoes 'llere averaged for
each point and the dots are larger th.n the S(:a-ller and dnft. Note that t.he ?"uch more
abundant Cut3 nuclei a long distance from the Co, the SOotalled Cu malll hne, do nol the surface Pt atoms only, they employed an add-subtract technique in which on
show up in the Co SEDOR
alternate spin echo cycles they applied a C 13 'If pulse coincident wilh the Ptt95 'If
pulse. For those PI nuclei far from the C t3 , the echo was unaffected by the Ct3
would continue to dephase during the second interval, producing a smaller echo 'If pulse, whereas for the surface atoms, the Ptt95 echo was diminished by the
at 2T. C I3 pulse. Thus, subtracting the Ptl95 echoes when the C I3 pulse was applied
An example of a resonance delected in this manner is shown in Fig. 7.30. from those when il was nOl, the signal from the Pt l95 nuclei not bonded to C t3
Boyce [7.49-51) studied a dilute alloy of Co in Cu. He wished to observe the nuclei vanishes.
nuclear resonance of Cu nuclei which are near neighbors of the Co nuclei. For To analyze the SEDOR signal, we nOte that the precession angle. 8, of the
such a dilute alloy, the neighbor resonances are weak in amplitude. frequently I-spins off resonance by (1I o)J during an interval T is
hidden under the tail of the resonance of Cu nuclei which are distant from
impurities, the so-called "Cu main lines". Boyce was able to reveal the "hidden" 8 = (")'(ho)1 - ams)T (7.200)
resonance of the first neighbor by doing a spin echo double resonance in which where (ho)1 represents lhe extent to which the particular I-spin is off resonance
he observed the effect on the Co spin-echo amplitude of a 1r pulse applied over due either to magnetic field inhomogeneities or chemical shifts. Suppose, then,
a sequence of frequencies close to the Cu main line. Since distant nuclei do not that at t = 0 we apply a 1r(2 pulse with (H')I along the I-spin x-axis in their
produce much field at the Co, the "main line" has negligible effect on the Co rotating frame. This puts the corresponding magnetization M«ho)J) along the
echo height. But the first neighbor has a big effect. In this instance the use of +y-axis (Fig. 7.32a). At time T- just before the I and S 1r pulses, 8 has the value
double resonance can be seen as a way of selecting pairs of nuclei which are 9, (F;g.7.32b):
close in space.
Makowka et al. [7.52] utilized pt195_C13 SEDOR to detect the surface layer 8, = 80 =f a-r(2 where (7.201)
of Pt nuclei for small metal panicles of Pt on whose surface they had adsorbed a
80"" ,/(ho)iT
monolayer of CO enriched to 90 % in C 13 . The metal particles. which were tens
of angstroms in diameter, were supported on A1203, a typical supported catalyst. The I-spin 1r pulse reflects M«ho)J) about the x-axis, producing the situa-
Makowka et al. observed Pt'95 spin echoes. Figure 7.31 shows their data. The tion shown in Fig.7.32c.
straighl Pt 195 spin echo gives lhe line shape of the Pt nuclei in the small metal During the next time interval, T, 8 advances an angle 82 (Fig.7.32d) given
particles. This line is over 3kGauss wide! To observe the Pt195 resonance of by
312 313
(a) t: O' (b) t :T-
7i(I,5) = -'I"[(l>o),I, + (HI),I.)
, , ~ l's rl[(h o )sS: + (HI)SS~J (7.205,)
M
M
e, " 7i(I) + 7i(S) (7.205b)
,, , ,
t/J(t2) = ex p ( -~1i(l)(t2 - til)exp ( -* 1i(S)(t2 - til) t/J(tl)
. (7.207)
c
, :~, =ex p ( -~1i(S)(t2 - til)ex p ( -*1i(l)(t2 - til)tJ>(t1) I
,,I a
l
M " : e" h
, ' The significance of (7.207) is that in analyzing the effect of the rf pulses, at
a given time, we can treat the effects of (HI)I independently of those of (HI)S'
II
M
Thus, we can treat first the I-spins, then the S-spins, or vice versa. We do not
FIg. 7.31a-d. The I-spin magnetization for spin echo double resonance (SEDOR) seen in
the I-spin rotating frame, M is the magnetization component off-resonance by (h o), for
have to treat both spins at once. Of course, the conclusion of this theorem is
a particular valu!) of"'s. At ! = 0, (lIJ), is applied along the ~-axis, producing a 11:/2 only valid during the time we can neglect the spin-spin coupling.
rotation of M onto the y-axis (a). During the time r, M precesses through a net angle 91 We now employ this theorem to analyze SEDOR using the density matrix.
given by (7.201) (b), At t =
r, 11" pulses are applied along ~I and ~s to both the I-silins We consider a sequence in which the I-spin and S-spin pulses are defined in
and the S-spins, so that at ! = T+, M is at the position shown in (e). During the next
interval, T, M precesses through the angle 92 given by (7.202), producing the situation the doubly rotating frame. At t = 0- the spins are in thennal equilibrium. At
of (d) t "" 0 we apply a 7r/2 pulse with (H I )/ along the I-spin x-axis. We denote it
as an X/(7r/2) pulse. We wait a time T, then apply simultaneously X I (7r) and
(7.202) XS(7r) pulses. We then follow the density matrix e(t) and the magnetization at
later time, t, such that t > T, looking for an echo at t - T = T or t = 2T. We
where the upper signs of (7.201) and (7.202) go logether, as do the lower signs. denote this sequence with the notation
Thus M«ho)/) makes an angle .c18 (Fig. 7.32d) given by
XJ(.fl) ... T ... [XJ(')'XS(')]'" (' - T)
.c18 = 81 - 82 = ~aT . (7.203)
We shall assume the I-spins are off resonance by (hO)1 and the S-spins by
Since Ll8 is independent of 80 or (Ito) I, it is independent of magnetic (lIo)s, to allow for chemical shifts, and use the notations of (7.176)
field inhomogeneity or of the existence of multiple chemical shifts. Since ms is
equally likely to be either +~ or -~, two resultant magnetization vectors, equal fh "'" l'J(h o ll - WI ils = Is(ho), - WS (7.208)
in length to MOI/2 form, making angles d8 of aT or -aT with the negative Then,
y-axis. Thus
(I+('» =Tc{I+,(t)} (7.209)
(My (2T» = -MOl cos aT (7.204)
To find U(t), we must first know its value at U(O-), corresponding to thennal
It is useful for future reference to derive this result using the density matrix .I equilibrium. We thus take
formalism. But first we demonstrate a useful theorem.
While the (HI )'s are on, we customarily neglect the spin-spin coupling, 0 - =..!..(l
<><)Z+
1iWOIIz+1iwosSz)
kT
(7.210)
giving in the doubly rotating frame a Hamiltonian
314 315
with Z the partition function. In calculating (1+(0) we need keep only the teon
,,0 )_ '!WOt
= ZkT1, (7.211)
( 1+(2T) = hwOI
ZkT
Tr{I+e- iat ,S,2T(_1 )eiat ,S,2T}
y
(7.218)
The ql,lantity in the curly bracket is just what we would have if (ho)t and (ho)s
We thus have
vanished, and the system developed solely under the influence of the spin-spin
+ fIWO! + . coupling, a1zS z . Thus. chemical shifts and field inhomogeneities are refocused,
(I (t» = ZkT Tr{1 exp(l(flt1~ + flsSz - a1~Sz)(t - T»Xt(7I")Xs(7I")
but not spin-spin couplings.
x exp (i(fl!1~ + flsS~ - a1~S~)T)X t(1r(2)1z X J 1(ir/2) l'sing the mjmS representation to evaluate the trace, and recognizing that
x exp(-i(flt1~ + flsS~ - al~S~)T)Xsl(1r)XJI(1r) (mjmsII+lm~m~)
x exp(-i(fltl~ + flsS z - a1zSz)(t - T»)} (7.212) = (~II+I-1)6msm~6ml,I/26m~,_1/2 (7.219)
Utilizing that we get
1
Xt(1r!2)I z X/ (1r!2) = 1y (7.213) Tr {I+e -illt.S.2T Iyeiat.S.2T}
and insening = L: (~II+I ~ !)eiamST (- !msllyl!ms)eiamST (7.220)
1
X s I (11' )X J (If)X t(lf )Xs(7I")
m, I
(7.214) But
I
before Xt(7I"/2) and its inverse after X J 1(1r/2) we then transfoon I y = ~(r - 1+) and (7.221)
I
X t (1r)Xs(1r)exp (i(fl[ 1~ + flsS z - alz S z )T)Xs (lf)X/ J (71")
l
(~Ir-I-~) = ( - W-I~)
= exp (-i(fl t 1z + flsSz)T)exp (-ialzSzT) (7.215)
= I so that (7.222)
and
Xt(1r)XS(lf)1yXSI(1r)XJJ(1r) = Xt(lf)1y X r (lf) = -Iy
l
. (7.216) (7.223)
Substituting, we then get
In the high temperature approximation, Z '" 4 so that
+ hWOj +.
(I (t) = ZkT Tr{l exp(l(flt1z + DsS~ - alzSz)(t - T))
(1+(27) = - hWkOt i cos aT . (7.224)
x exp(-i(flt1z + DsSz)T)exp(-ia1~SzT)(-Iy) 4·T
Using for thennal equilibrium
x exp(ia1z S z T)exp(i(fl t 1z + flSSz)T)
x exp(-i(flt1z +flsS~ - a1zS~)(t - T»} . (7.217a) MOt = xoHo = 'Yfl(Iz}
This expression can be simplified immediately since the teons exp (WsS~(t -T)) ')'2h 2 I(I + I)
and exp( ~WSSzT) on the left of (-1y ) commute with everything between them XO = 3kT (7.225)
and the teons exp(iflsSzT) and exp(-WsSzU - T» to the right. Hence, they
we get
can be combined, giving a factor of unilY. We then get
liwOj
(I+(t» = ~~~ Tr {1+ exp (i(Dt I z - a1zSz)(t - T)) (I,) = 4kT " 10 so that (7.226)
SEDOR, X s '(ll')X/'(ll')XI(ll')XS(7r)
7.22 Two-Dimensional IT Speelra - The Basic Coneepl
so that at time 21' we will have had for the transformation of the teon exp [i(ilII:+
nss: - aI:S:)1'} We now arrive at two-dimensional Fourier transfonn magnetic resonance. one
of the most exciting and powerful developments since the original discovery of
Basic S-flip only: exp [i(ill I: - ilsS: + aI:S:)1'] magnetic resonance. In this section we introduce the concepts using a very simple
Conventional I echo: example, the basic S-flip-only experiment discussed in Sect. 7.19. We shall find
exp (i(-ilrI: + ilsS: + aI:S:)1')
that it already contains all the essential ingredients of two-dimensional magnetic
SEDOR, exp [i(-nII: - ilsS: - aI:S:)1') . resonance.
The idea of two-dimensional Fourier transfoonation is due to JeenU. He first
As in the transfoonation from (7.217a) to (7.2l7b). we can combine the teons presented it in lectures at the Ampere International Summer School. Basko Potze.
involving exponentials of ilsS:(t - r) or flSS:1' on the left of (-I,,) with their Yugoslavia, in 1971 {7.55). An excellent review by FreemlJn and Morris [7.56]
inverses on the right of (-I,,). What counts are the tenns involving I: or 1:5:. describes the explosive development of the field, and states and documents that
Thus we see that: "Ernst was the first to appreciate the great potential of the method". Indeed, both
Freeman and Ernst with their collaborators provided many of the fundamental
The basic S-flip-only 1'1" pulse changes the sign of the a teon but leaves applicalions. Aue et al. gave a treatment of basic significance and broad generality
the sign of the il/ alone. Therefore, it refocuses the spin-spin coupling at [7.57] which also gives a detailed treatment of concrete applications. It is a
t = 2r but leaves the I-spin chemical shift alone. fundamental reference for workers in this field. Recent reviews include those by
The conventional echo 11" pulse changes the sign of both spin-spin and 8ax [7.58} and Turner [7.59}.
chemical shift terms. Therefore it refocuses both spin-spin and chemical We shall describe in later sections some of the panicular fonns of double
shifts at t = 21", eliminating both. resonance, but OUf goal in this section is to introduce the concepts using S-f1ip-
The SEDOR 1l' pulses change the sign of the chemical shift tenn only. only double resonance (Sect. 7.19).
Therefore they refocus the chemical shifts, at t = 21', effectively eliminat- We saw in Sect. 7.19, (7.177), that for a spin I off resonance in the frame
ing them. but leaving the spin-spin coupling untouched. rotating at WI by
A useful variant on the SEDOR pulse sequence was introduced by Wang et ill =wOI-wl
a1. [7.53.54). The time of the I-spin 7f pulse is held fixed. The S-spin 7f pulse
the magnetization was
318 319
(Mt(t» '" i~O(e-i(01+0/2)t+e-i(01-0/2)t) (7.231) of strength b. we could label the chemical shifts as fl/a, n,b. and then
(where the -fl , means the sense of rotation is negative) for t < t... and (M/(t» '" i~Oa e-iOJ&t'e-in/&t2(eiot2/2 + e-iaI2/2)
(Mt(t» '" i~Oe-iOlt + i;Obe-iOatle-iOut2(eibI2/2 +e-ibt2/2) (7.237)
x (e+iat./2e-io(t-t. )/2 + e-iat./2e+ia(t-t. )/2) for t > t'lr' (7.232) and we would get
Exactly at t =2t .... we then have
< <, <
(Mt(t» '" i~Oe-iO/(21.) lal / Ibl <,
/'
This, as we remarked at the end of the previous section, means that at t '" 2t.... the
magnetization is just what it would be if the spin-spin coupling were zero. so that <, <,
only the chemical shift fl , '" WO/-WI remains. [We here assume a unifonn static
field, so that (ho), arises because of chemical shifts.] The situation is illustrated
,11 ""'-'_,.J- _
in Fig. 7.33 where we show (Fig. 7.33a) that for limes greater than 2tll"' the phases
4>1 and 4>2 act as though from 0 < t < 2t'lr there were only a chemical shift, but t. 2t.
".
for t > 2t1l" both chemical shifts and spin-spin splittings exist. In Fig. 7.33b we
have redrawn the situation to represent the situation of a '" 0 for 0 < t < 2t'lf'
H we reset t'lf to a new value. r...
longer than t'lf' the new situation is shown <
v-
< <,
in Figs.7.33c and d. We can plot phases for a sequence of values t,.., t~. t~. etc., lei Idl
as in Fig. 7.33e. I
<,
These plots can be summarized by defining two times. tt and t2 • by /
/ " I
I
I
tl '" 2t'lf (2t~, 2t:. elc.)
(7.233) /
/
/ I
,,
I
In terms of t, and t2, we have that for times after t" ,.• 2t~
I
21~
Utilizing Ihe relation 6(x) :: (1/211") L~:: exp (ixt)dt. we get .'~g. 7.33a~ P~1lSC:S of the two I-spin magnelization components Yersus time for the S-
nIp-only ex~rllnent. <a> The two phases ~I(t) and 412(1), showing the time t. aL which
ieo the ... pulse lS applied lo the S-spins, and the formation of the echo at t '" 2t which
g(wj,"")' 26(w, + il,)[6(..., + ill - a!2) + 6(..., + ill + a!2)J. (7.236) refocuses the spin-.pin splilling. H the spin-spin coupling were zero, the ~'s wo:k1 both
follow t~e dashed curve of slope 0 1 rq>resenling a chemica.! shift only. (b) The ~hll.vior
These dala can be represented as points in a two-dimensional plot (WtoW2) of. eqUlVlllent, for! > 21 .. , to that of (a). (e), (d) Replot.1ll olea) and (b) f('$pectively for ..
(Fig.7.34a). If there were a second chemical shift and a second spin-spin spliuing. value t~ >~". (e) A family of phue trajectorir$ colTesponding to a family of experiments
with different times I. (t~, I~, de.)
320 321
la) IbJ lines extrapolate back to an intersection at t = 2t ... The two lines are diverging
0. ,, -IO~-bI21 --~-------""1-
from their intersection since the chemical shift frequency fl/ is split into two
frequencies. n/ ± a/2.. for times after 2t...
I
I
-t QI,·bI21 -----------t-II The example illustrates the general feature of two-dimensional schemes that
I
I iI there is some initial preparation of the system [e.g. the X t (7f!2) pulse], then
I
-!n I -aI21 ---~-- I
the system evolves for a lime interval tl under one effective Hamiltonian (in
_ 0,
-ttl1u12 1
---+---
----t--
iI
,,
our case, one in which the spin-spin couplings are zero), then data are collected
during a second time interval, t2. for which the Hamihonian lakes on some new
I
I , effective Conn (in our case the full Hamillonian). 1lle experiment must then be
repeated for other values of tl_ These three intervals are often referred 10 as the
-Q l ~ -Q h -Q~~
preparation, the evolution, and the acquisition intervals, respectively.
Fig. 7.34. (a) The lwo-dimensional plol of frequencies WI and W2, 'lhowing poinL5 in "'I
corresponding to the chemical shin WI = -(af ±a/2). (b) A plot similar lo (a) corre- The experiment we have just described is in (act essentially the famous
sponding to two I-spins, each coupled lo a single S-spin with corresponding chemiCl\1 experiment invented by Maller et al. [7.60] to resolve the C l3 chemical shifts and
shift!! aid and nf6 l\nd spin-spin couplings a and b
C I3 _H I spin-spin couplings. They demonstrated the technique using n-hexane
(Figs. 7.35 and 36). They utilized broad-band (noise) decoupling of the protons
iCO a instead of applying a "" pulse as in our example, and they decoupled during
9(Wl.W2) = -2-5(wl + n!a){5(W2 + !1-la - 0/2)+ 5(W2 + n la + 0/2)]
the second time inlerval. t2. instead of during the first time inlerval, tl' as we
iCOb I
+ T5(w, + n ,b)(5(W2 + n,b - bIl) did in our hypothetical example_ They point out one could choose to decouple
during either interval, and remark that since each point tl requires a separate 'I
+6(w, + [h, + WZ)J 17.238) experiment, one needs less data if the tl spectra is chosen to have the fewer
which would give {he plot of Fig.7.34b. spectral lines, Le. is the time during which the protons are decoupled. In a later
This plot allows us {o go along the WI-axis {O detennine the values of the paper, they describe the sequence of our example [7.61]' This general method is
I -spin chemical shifts, then at each chemical shift position (o observe in the referred to as J-resolved 20 NMR in the literature.
W2 direction splitting of those panicular I -spins by the S-spins to which the)'
are coupled. That is, there is a correlation of the WI frequencies with the W2 Qoli"CH;Qol;0i;CHrQol3
frequencies. The physical origin of this correlation can be traced back to the abccba
phase diagram Fig. 7.35, which shows that the single phase line over the time
interval 0 10 2t7: splits into two lines for limes after 2t7:' For t > 2t7:' those
'"
H' .,
".
0'.0
Fig. 7.36. DlIla of Muller, KUI1t(Jr, and Ernsf using lhe pulse !lCquence of Fig. 7.35 for the CI3
spectrum of n-hexane (Clb-CII,-CII,-CII,-CII,-Clb). The ""I-axis gives the combined
I, I, chemical shin and spin-spin spliuing, the loo'2-axiS the chemical llhirL5. Thesoe data show
Fig. 7.35. Schematic representation or one ronn of C 13 2D-re;olved Spedroaoopy aner how a 20 display enables one to lIl""atly simplify unraveling a complex spectrom. Twenty-
Miller ct al. 17.601. The CI3 nuclei are the I-spins, Iii thfJ S-spins. For the inlerval II the lW() experimenL5 Vooe"" coadded for each of the 64 values of tl between 0 and 35 IllS. The
fullUamiltonian acts, but during t, the spin-spin interaction i, lUnled ofT by broad-band authors state lhat the resolution is seve~ly limited by the 64)(64 dab. matrill: used to
decoupling. The situation is similar to S-f1ip-only resonance with the role of the times t l represent thfJ 20 Fourier lransform. The absolute values or the Fourier coefficients a""
lind t2 inlerdlllllged ploued. The undecoupled I D spe<;trum is indiCllted along the ""I-llll:is and lhe proton-
decoul)led spectrum is shown along the W2-all:is
322 323
We shall discuss funher examples in later sections. First, however, we need we can select either the real or the imaginary part of 9(Wl) and thus either the
to take a side trip to deal briefly with a topic the reader will quickly encounter absorption signal or the dispersion signaL
in studying the literature, the problem of 20 NMR line shapes. If, however, we take only the real part of g(WI,W2), it involves (for A real)
(7.244)
7.23 Two-Dimensional FT Spectra - Line Shapes
which is a mixture of absorption and dispersion.
Of course, if one goes along the frequency contour W2 - O2 , one has
In the previous section we developed the Fourier transfonn of the theoretical
expression (7.234) by fonnally extending the meaning of t1 and t2 to negative A al
Re{g(wI,W2)}=- 2 (n )2 (7.245)
times. This is easy to do given a theoretical expression. If, however, we have 0'20'1 + HI-WI
experimental data, we cannot in general guess the theoretical expression. One
then encounters a problem which Ihe reader will immediately come across if which is the WI absorption line. However, the practical problem is that since
he reads the literature of 20 NMR. We therefore take it up briefly. A thorough dispersion extends out farther from the line-center Ihan does absorption, the
treatment is given in the basic paper of Aue et al. [7.57]. mixing of dispersion tends to decrease the resolution between adjacent lines.
Let us therefore reexamine how to treat experimental data considering pos- There are various methods of processing data to deal wilh these maners. In
itive times only. Since the problem is one of mathematics, we assume we have addition to the treatment of Aue [7.57], the review articles by Freeman and
the simplest fonn of experimental data given by an exponentially decaying sine Morris [7.56], Bax [7.58], and Turner [7.59] discuss these issues.
(cosine) wave. We put in the decay since such decays are always present, and to
avoid difficulties at infinite times. Thus, suppose we have
(7.239) 7.24 Formal Theoretical Apparatus I -
The Time Development of the Density Matrix
Then
=
1
9(Wt> W2)"" (211-)2 J =J dtl
..
dt2/(ll,t2)e-llU\t\e-I"'2t2 (7.240.) In discussing the S-flip-only experiment (Sect. 7.19), where we were applying 'II'
o 0 pulses to the S-spins, we assumed that the effect of the S-flips was simply to
1 A shift the precession frequency of the I-spins suddenly from one ms multiplet to
(7.240b) the other. An S-spin originally pointing up now points down, and vice versa. If,
"" (211')2 [al i(OI WI)J[a2 i(02 W2)]
however, one applies a '11'/2 pulse to the S-spins, classically they are now per-
O't+i(OI-WI) O'2+i(!t2-w2)
A pendicular to Ho. How then do we treat what happens in a quantum mechanical
(7.24Oc)
= (211')2 OI
+ (0 1 - Wl)2 O'~ + (02 W2)2 system? Classically, if Sz is zero, the S-spin will not shift the I-spin precession.
That would lead to a precession midway between the two I-spin lines at WOI+a12
In the corresponding one-dimensional case and WOl - a/2. But there is no transition al WOl among the eigenstates of the
l(tl) = Ae(in\-aJ)l\ (7.241) Hamiltonian, so this answer cannot be right. The next closest thing is for the
spin to act as though it now has both frequencies. Thus, a spin which precessed
at WOI + a/2 before the S-pulse would have components precessing at WOl + a/2
1=
g(Wt) = 21l' J e-i"'lt\ l(tl)dtl (7.242) and WOl ~ aI2 after the S-pulse. As we shall see, this latter answer turns out to
be correct. We clearly need a systematic and rigorous derivation of the result.
o
we get Let us therefore treat carefully a concrete problem, a pulse sequence
A 01 +i(JtI -WI) X 1(1I'!2)· .. r ... Xs«(J)···t-r ,
( ) =-
gWI (7.243)
211' O'I
+ (01 - wd 2 where we will explore and contrast what happens when (J "" 11' with the situation
We recognize that the real pan is the Lorentzian line shape (the absorption), and for (J := 11'/2. We wish to calculate Ihe expectation values of 17; and I y in the
the imaginary part is the dispersion. If we use quadrature detection (Sect. 5.8) I-spin rotating frame. We shall use the doubly rotating frame Hamiltonian.
324 325
We assume that the X,(1C(2.) pulse is applied at t == O. At t == 0-, just before. and -(fl[ + all). Therefore, no matter wfu:lt rotation angle we give the S-spins,
we assume the spin system is in thennal equilibrium. (We also treat me case of these are the only frequencies we observe. At most we can change the complex
100 % polarizalion of the S·spins.) We employ the general fonnula coefficients
(I+(t)}=Tr{I+en(t)) (7.246) (-mSIe<tl>l + ms)exp (i(fl, - oms)ttl
where 1+ refers to axes in the rotating frame. as does llR· Since we will be of the time-dependent tenns
working almost exclusively in the rOlating frame, we drop the subscript R to
simplify the nouttion. . exp [- i(fl, - ams)t]
Then, assuming 1{ is given by (7.162) except dunng the (Hlh and (HI)s Since the coefficients are complex, we have at our disposal the amplitude and
pulses, we characterize states by m, and ms. getting phase of the two frequency components by manipulaling (-mslll(t\ >l +ms). To
(1+(t» = L (m/msII+e(t)lm/ms) follow what happens in the actual pulse sequence, we use the method outlined
in (5.167) in which we divide time into intervals during each of which the
m"ms
Hamiltonian can be taken as time independent.
== L: (m,mslr+lm',ms)(m~msle<t>lm[ms) (7.247)
In SO doing, we will use the resulls of (5.253) and (5.254) for the time
"","'S
"',''''s dependence of II under the action of rf pulses. We must first, however, generalize
these equations which apply to one spin for the case of two spins. In so doing,
Making use of the propcnies of the raising operator and of the eigenstates
we consider the effect on II of interactions in the rotating frame. Utilizing the
Im,ms). we have
concept that these interactions are large compared to the spin-spin coupling, we
(m,mslr+lm/ms) == DmsmsDml,I/26m~,-1/2(,II+I-~) employ (7.206) and (7.2m) to treat the interactions one at a time:
:: Dms.m'sDm,,1/26m,,-1/2 (7.2530)
Abbreviating +~ as +, -! as - for m[, we thus get
1{s:::: -7S',(HI)SS~ == -1iwISS~ (7.253b)
(r(t)} = L(-msle(t)1 + ms) (7.248)
ms Thus, taking just Hs
Thus, there are only two matrix elements of e<t> needed, one for each value of
~(m,msle<t>lm"ms)::*(m,mslll'HS - 'Hselm/ms)
ms· .
During those times when me rf pulses are not turned on, we have, usmg
(5.133), that
: : r; L (m,msle<Olm'fmsXm'fmsl'Hslm/ms)
mi',llI s
e<t2) == exp (-iH(t2 - t I )/h)I2(t I) exp (i1i(t2 - t I )/tl) (7.249) (7.254)
where Since
11.:: -hfl,l~ -1iflsS~ + hollSl (7.250) (m'lmsl1islmims) == 6m7m,(msl'Hslms)
using the notation of (7.176). As a consequence we get
(-msle<t2)! + ms):: exp [ - i(fl, - ams)(t2 - tl)]
x (-mSlll(till + ms) (7.251)
!!.(m,msle(t)lml,ms)::
dt
*L
I m "s
[(m/msle(t>lml,ms)(msIHslms)
Thus, if we know II at time tt. we can express it for later times as -(msl1islmsXm[msle<t)lmims)] (7.255)
(r(t» = L(-msll?<'I)! + mS)exp [- iUh - amS)(t - 't)J . (7.252) Equation (7.255) shows that if we are driving the S-spins only (Le. 11, == 0,
m, Hs ,;. 0), the I-spin quantum numbers do not change. There is therefore no
This expression, even without specifying (-msle<tl)l+ms) alre~dy allows us to mixing of matrix elements (m/msl/?(Olmims) with matrix elements such as
draw an imponant conclusion: (1+(0) contains just two frequenCIes -(flT-an) (m,msle(t)lm'lms)' The problem is thus transformed to a one-spin problem.
326 327
As a result, we can immediately utilize the results of (5.253) and (5.254). Note, We now let e<t) evolve for a time T until time T- just before the S-spin pulse.
however, that though matrix elements of 7-{s are necessarily diagonal in ml, Utilizing (7.250) we get
matrix elements of e are not. This is a very important point to which we will
return later in discussing coherence transfer and multiple quantum coherence. (-msle(T-)[ + ms) = exp [- i(n , - amS)T](-msle(O+)1 + ms)
Thus, under the action of 7-{s, acting for a time t, we can write (using ms
=+iloexp[_i(n,_amS)T). (7.261)
for -ms) for the matrix element diagonal in ms: 2
(-mslg(tll +ms) = ![(-mslg(O)1 + ms) +(-mslg(O)1 +iiisl] We now consider twO cases for the Xs(U) pulse, 8 ::: 7f and 8 ::: 7f12. Since
(HI)S will not change the ml quantum numbers, and since (I+(t» arises only
+W-mslg(O)1 +ms) - (-mslg(O)1 + ms)] cos wIst from teons ofT-diagonal in m/, we need consider only the effect of (HI)S on
i the matrix elements (-msle<r-)I + ms). We have if 8 ::: 11', sin wiSt::: 0,
+2:[(-mslg(O)1 + msl - (-msle(Oll + ms)J sin wlS' (7.256)
cos wiSt::: -1. For 8 = 'K!2, sin WiSt = 1, cos wISt::: O. For 8::: 7f
and for the matrix element off diagonal in ms (7.262)
(-mslg(tll + ms) = W-mslg(Oll + ms) +(-mslg(O)1 + msl]
For 8 = 7({l.
+!(-msle<O)1 + ms) - (-msle<0)l + ms») cos wist
i (-mslg(T+ll + msl = W-msle(T-ll + msl
+2:[(-mslg(O)1 +msl - (-msle(Oll + mslJ sin WIS" (7.257) (7.263.)
+ (-msle(T-ll + iiisl]
An analogous pair of equations holds for the effect of (HI) I. Therefore, at t = 0- ,
which also implies
just before the pulse X,(7f12) we have
~[(+mslg(O-ll + ms)
displayed in Fig. 7.18a,b, as well as Fig. 7.21. We will discuss the meaning of
(-msle(O+)1 + ms) =
(7.263) below.
- (-mslg(O-)I- ms)] (7.258) Denoting (I+(t» for a 8 pulse as (I-(t»o, we get for t> T,
But
(r+(t)). = ~ L: [(+mslg(O-)1 + msl - (-iiislg(O-ll- iiisl]
ms
isjust the net I-spin z-polarization at thennal equilibrium for those I-spins cou- x exp [ _ i(n, - ams)r] exp [ - i(n , - ams)(t - T») .
pled 10 S-spins which have a particular ms. Since to a very good approximation (7.264l
ms is as likely 10 be +! -!,
as the (Iz(O-» is equally distributed between the
Utilizing (7.260), this becomes
two values of ms. (I+(t»ll"::: i~o "Ee-iI1,teiams(t-27') (7.265)
Defining ms
(7.259l This result is exactly what we deduced in SecL 7.18 if one replaces T by the
notation tll"' For t ::: 2T it displays the refocusing of the spin-spin coupling. ~e
we get
teon involving n, gives the chemical shift. lfwe set il, equal to zero and utlhze
+ iIo (7.260), (7.264) is identical to (7.171).
(-msle(O )1 + msl = 2 (7.260)
32. 329
For the case e=1ffl we get
FIg. 7.33. The phllSe$ ~ of the I-spins wh)ch rault
Now that we have employed fonnal methods to follow in detail the role of w:(1. or
We can get added insight by taking the S-spins to be 100% polarized, so 'Ifpulses, we return to one of the important applications of spin-coherence dou-
=
that ll[ t 0- only mS =+i
is to be found. Then, for t > T ble resonance, the transfer of the magnetization of one set of spins to the other.
The method of doing this transfer was discovered by Feher in 1956 (7.63,64].
(I+(t}}'/2 = ~[(++ Iu<O-ll + +l - (- + lu(O-ll - +)] In 1962, Baker discovered a related fonn of nuclear-nuclear double resonance
which he called INDOR. Baker pointcd out that it was much like Feher's EN-
x exp [ - i(!h - a/2}T] DOR experiment [7.65]. In 1973, Pachter and Wessels [7.66] invented a pulsed
x (",PI-;(J)/-a!2)(t- T l];ex P[-i(J)/+a!2)(' T)]} version of the INDOR experiment, which they named selective population in-
version. leener's proposal {7.55] for two-dimensional NMR, as we shall see
in Sect. 7.27 includes among other things a homonuclear pulsed version of the
(7.268l Feher experiment. Then a heteronuclear pulsed version, with all the benefits of
This shows that up until t = T, we get precession at a single angular frequency, Fourier transfonn, was proposed by MaJuisley and Ernst [7.67], with modifica-
-(n l - an), corresponding to ms = +!'
but that after the 7:(1. pulse at time T, tions by Bodenhausen and Fneman [7.68], Maudsley et al. [7.69J, and Morris
the precession splits into two components of equal amplitude. one precessing at and FnelTliln (7.70).
3JO
331
Feher was studying the electron spin resonance of donors in Si doped with State 4 P(--)=t-(.1+6) (7.271d)
phosphorous. His work was part of the research that led also to his disc:overy
of ENDOR (Sect. 7.7). He showed that he could transfer the electron polariza- The spin polarizations are
tion to the p31 nuclei. Since his experiment is conceptually very simple and
straightforward, we begin with it.
(Iz) = L: (m/msIIzlm/ms)/l(m/ms)
In keeping with the model system of Sect. 7.19, we assume the energy levels
are those of Fig. 7.39a labeled by jm/ms), and also labeled 1,2,3,4 as shown. = Lml [LP(ml.m s)] (7.272)
In thermal equilibrium, the populations of each state are given by the Boltzmann ml ms
factors, which, in the high temperature approximation, may be wriuen as
Now, though we have neglected the spin-spin coupling in computing the thermal
where P(mb ms) is the probability of finding a given 1-$ pair in the state equilibrium populations, we must tIOt neglect them in calculating the spectrum
Imf, m~). In this ex.pression we have neglected the small spin-spin coupling in since they make transition 1-2 have a different energy from that of transilion 3-4.
compUt1ng ~he .ener~es of the levels. The leading term, t,
does not pr<Xiuc:e any Since the" t"
in (7 .27Ia) is the same in all energy levels, it has no effect on
net magnebzabo~ Sl~nals, so it is only the terms involving m/ and ms which the intensity of any transition, and can thus be conveniently dropped in recording
express net polanzallons.
population probabilities, giving one the population excess above i. (A negative
excess is a deficiency.) These are conveniently expressed in a matrix
la' fbI
l_._
t -A·6 mS=l ms=-!
1
1-"'06
4 m/ "'-! - .1+S - Ll-,
m/ "'+~ Ll +' Ll- , (7.274)
Thus, the resonances for both I-spins and S-spins are doublets, as shown in
1.-0.
-i- oA - 6
Fig. 7.40. For the sake of explanation, we assume the I-spins have a much larger
fmjmsJ
---i°.1 0
6 population difference (Fig. 7.403) than the S-spins (.1:> S). So that we see a
larger intensity in the I-spin absorption lines (Fig.7.40b).
Fig. 7.39. The energy levels (a) and therm.al equilibrium population I)robabilities (b) of the Feher's method consists then of two steps. In the first, he makes an adiabatic
model system. States are labeled accordlllg to the convention lm/ms)
passage of frequency w/ across one of the I-spin transitions (but not the other)
Defining
such as the transition 1-2 for ms '" +!.
Since the adiabatic passage invens the
magnetization, it interchanges the populations of the states involved, I and 2.
hwOI This step leads to the population excesses shown in Fig. 7.4la, and listed in the
8kT ;: Ll (7.270.) following matrix:
ms"'! ms'" -!
!twos
8kT ;:, (7.270b) - Ll-'
we have for p(m/,ms) Ll- , (7.275)
State I p(++)::::t + (.1 + 6) (7.271.) The S-spin population difference corresponding 10 a tranSition in which ms
changes, such as fTOm 1- +) to 1- -) (states 2 (04), would be given by
Slate 2 p(-+):::: t - (.1 - 6)
(7.271b) P(-+)-P(--)=(1 -Ll+6)-<1 -Ll-6)=20 (7.276.)
Slate 3 P(+-}=1+(Ll-6) bl!fo~ the adiabatic passage, but
(7.2710)
332
333
Energy Energy F1g..':40: (a) The energy level and lhcrmal (7.276b)
la) eqUlhbrlUm populations (sc.henlalic) for a p(-+) - P(--) =(t + il +6) - (t - il- 6) =2il+26
CIlS(! in ~hich "'II" "'Is !lID that 11" 6. (b)
1- .J The I-spon and 5-spin resonance intensities
sho,,:,in~ the ~pin-spin splitting. The grealer
after the adiabatic passage. Thus, for ml = -!,the S-spins have an enhanced
I-spm lIItell.'llty an!leS bec:aUS1e "'If:> "'Is
polariz,ation.
On the other hand, for the S-spin transition involving ffil = +i
(the transi-
tion between states I and 3), we have a population difference before of
Population
P(++) - P(+-) = -2il- 26 (7.277b)
Population
which, for Ll"~, is an invened population. The population differences for both
I-Spin S-Spin (7.276b) and (7.277b) are much bigger after this first step than before.
Absorption Absorption Since (7.277b) corresponds to a population inversion of the S-spins [the
Ib) "s; 112 .s: -112 upper energy state ImJms = 1+ -) has a larger population than the lower energy
state ImJffis) = 1+ -)], there is no net polarization of the S-spins.
To achieve a net polarization, one may therefore inven the population dif-
ference of eidlCr (7.276b) or (7.277b). Feher achieved this by his second step, an
adiabatic passage of (H 1)5 across either (but not both) of the S-spin doublets.
Energy Energy Thus, if the passage were across the 1-3 transitions, the 1-3 populations would
lal be interchanged leading to
1--1
"'5= , "'5=-'
""1--- il + 6 - il- 6
il- 6 - Ll +0 (7.278)
11,11/21 , /11.1-1121 the I pulses will have produced a situation in which the M I (!) population is
,, t1 1 11/21 1'1[1- 112 1 inverted, bUI not Ihe MI(-!). so the situation will look like Fig. 7.4Ia. Therefore
" " the XS(1f(l.) pulse (Fig.7.43a) will produce a silUation in the rotating frame such
~
lei (fl " "
L
(a) Ibl
[Hll l Mil-liZ!
Mi l-1I21
Fig.7.428-1. The effect of (lid! pulses tuned to resonance (WI = ,",od on the I-spin
magnetization M,(ms) for the two components ms = and ms+t = =
_1. (a) At I 0,
(Ildl is applied IIlong the z-lIxis in the rotRting frame giving an x/('II'/h pulse; (b) I'll (e I ~·Ig. 7.43. (a) (/fils is applied along the :I:-dire<:tion in the
I = 0+, both At I (t) and M l( -!)
lie along the y-axis; (c) the two components M /(mf) S-llpin rotating frame, giving an Xs(rr/2) pulse; (b) the S-
precellS in o[.posite diredions at angular frequencies ±a/2; (d) at a time t =
11:/0, 111'(2) lIpin magnetization Ms(m,) jUlIt after the Xs('7f!2) putlle.
and MI( - ~) point along the -z- and +z-axes respectively; (e) an (HI)I lying along Ms(_l) and Ms(+l) are antiparaUel; (c::) /lS a result of
the y-axis gives a Yl(lr!2) putse; (f) the I-magnetization vcdors after the Yt(,../2), with their dMerent precell6?on frequencies, M s( ~) and M s( -~)
MI(-i) pointing parallel to the ~-dire<:tion, but MICt> pointing antiparatlel
" are parallel after II. time 1t!IJ
336 337
as shown in Fig. 7.43b in which the two components of Ms point opposite to one 101 Ibl .
another initially. At later times, their relative phase will have changed, making
the magnetization components, in fact, parallel to each other at time 1r/a after
the XS(7f/2) pulse (Fig.7.43c).
Indeed, following the XS(1f/2) pulse, there will be an S·spin free induction
signal. In the lab frame, it can be thought of as two signals, one at wos - 012,
the other at wos + 012, or it can be viewed as a signal at wos which is amplitude
Ly,
IHII I "
Ml l1121 M11-1I2)
Idl r,
"
modulated as sin al/2. This signal is of course similar to what would result from 1'1 IOI-al2)1
using Feher's two adiabatic passages to transfer the I polarization to the S-spins, M111I2)
followed by a 1r/2 pulse to the S-spins to generate magnetization perpendicular
to l • anJt
to Ho. If one performed that total sequence using pulses instead of adiabatic
passage, it would take a total of five 1f/l pulses (two of the I-spins, three for the " ............ ~t-V21 "
S-spins). However, the last two "/fa. pulses can be omitted since the first of them
If J y,
rotates transverse S-magnetization to lie along Ihe z-direclion, and Ihe second lei
t-
rolates the magnetization from the z-direction back to a transverse orientation. In
their technique of selective population transfer, Pach/er and Wessels [7.66] apply MI II121 MI III2I
a 1f pulse to one of the I-spin transilions, invening their population difference.
,1 M 1-1/lJ , "
" ,,M I-1I2)
They use an (Hth much smaller than the spin-spin splitting. Then they apply a
strong 7f/l pulse to the S-spins 10 observe the free induClion decay. The Fourier ,1 I I
transform of the S-spin signal shows one hyperfine line invened with respect I /
to Ihe other. and a large signal increase if "'(, »"'(5' Thus their experiment is
equivalent to (7.283). y, Ihl y,
191
For the sequence (7_283) we have required that w, = wo" exaCI resonance
for lhe I-spins. For Ihe S-spins, XS(1fn) does nOI have 10 be exactly at reso- M11-l/l1
nance. It will still produce the situation of Fig. 7.43b. Then, those magnetization MI I-ll2) Ml l1l2l
components will precess freely at (wos - am,), giving a signal at WO/ which is
amplitude modulated at angular frequency o.
---- -
'. M1l1/l1
"
The sequence (7.283) is almost the Maudsley-Emst sequence. What they do
is (i) remove the requirement that WI = WOI and thereby (ii) remove the special Fi&o 7.44a-b. The effect of the. '_spins bei~g off resonance ("" T
1+'0,). (a) At t 0, (lid, =
phase requirement of the second I pulse. Then, taking a sequence is applied along the lI;.axis m the rotating frame; (b) the .)1.,(_/2). pulse puts th\t'"':O
componenlll of MI(ms) along the y-axis at I = 0+. M/(t) IS the solid arrow, M/( - 2) IS
(7.284) the dashed arrow, (c) the componenls precess at 01 -: am~, where.OI = ""01.- "'I, un~er
the combined effect of the chemical shif~ (flt) and IIpm·spm coupling (ams), (dl--:<h) give
in which t2 is the observation time. and for which (7rn), means an I-spin 1f/l M,(1) Md M,(-!) a~ successively la~er times near to _/a under the assumption th~~
pulse without a special phase condition. they make a 20 Fourier transform. The a< hI so that the relati\'e alignment of the two vectors changes slowl>: compared to their
joint rotation rate 01' An XI(_/2) pul!le would have no effect (or sltu~tlOn 00, wo~ld
fact that WI "WOI produces a situation in which. following the initial (7f/l)/> make M I( 1) inverted rdative to 110 for situation (e), and make M I( -!)
IIlverted relative
the I-spin components precess both with respect to one another and also with to 110 (or ~tuation (h)
respect to the rotating frame. Thus, if one is off resonance by an amount large
compared to the spin-spin splitting, the pattern of Fig. 7.42f will rotate rapidly at (H t) J along the x-axis is applied (Fig. 7..44a). ~ere will be comple~e in.version
(WOI -wI) in the w, frame while the relative orientation of M/(!) and M,(
will change slowly. The situation is shown in Fig.7.44. The initial X,(7fn)
-!) of M,(!) for the situation (e), complete IflVerslon of M,(~!).r0r sl~unuon (h),
no inversion of either component for situation (g), and partial II1verslon of ~h
(Fig.7.44a) puts M,C!) and M/(-!) along the y-axis in the frame rotating at M I ( "',) and M (_') for situations (d) and (t). We see therefore that the chemical
w, (Fig.7.44b). Under the combined influence of chemical shift (il/) nnd spin-
1 /"'1 . '
shift frequency flj will determine the urnes at whIch the t':'"o M, com.ponents
are invened, and thus they will determine the times til which a~ S-SPIfl pulse
spin spliltings the M, vectors rotate and separate (Fig.7.44c). A pallem of the
}.th's at successive times is shown in Fig.7.44d-h. If another 1r/2 pulse with will find transferred magnetization. Of course, a will also determine the general
339
338
range of times for which Mj(!) and Mj(~4) point opposite to one anOlher. (mjmsle<r+)lmlms)
Indeed, Bodenhausen and FreelrUJn [7.68J analyze just this model to show that
= 4[(mlmsle<0-)lmlms) + (mpnsle<O-)lmjms»)
if the I-spins are protons and the S-spins are C l3 , the resulting 20 spectrum
has the proton spectrum in the WI direction and the e 13 spectrum in the W2 .. + 4[(mjmsle(0-)lmlms) - (mlmsle<O-)lmjms»
direction. Thus by observing the C I3 spectrum one gets the indirect detection of x cos (f1[ - ams)r (7·290l
the proton resonance frequencies, and the correlation of the proton and carbon
chemical shifts. for the diagonal elements just before the XS(7C!2) pulse.
It is useful to think of this experiment in tenns of the density malTix analysis There are several interesting limiting cases. First, if n j = 0, the result de-
of Sect. 7.24. We have a pulse sequence pends on cos (amsr), hence is the same for ms = 4 and ms = -4. This
result confinns our poinl earlier thai in the absence of a chemical shift an
Xj(1f!2) ... r ... (X j (1f!2), XS(1f!2» . .. observe S (7.285) Xj(7r!2) ... T ••• X j (1f!2) will not inven only one multiplet Instead, one needs
in which we shall treat the X S(1f!2) as having occurred just after the second an Xl(w!2) ... r ... Yj (1f/2).
Xj(7f!2) pulse. In fact, if f1[r = (7r/2), one has the case of Fig.7.44g, and indeed then
Using our theorem of (7.207) we define r- as just before and r+ as just cos (ill - amS)r = sin (amsr) so the result differs between ms = +! and
after the second X j (1f!2) pulse, r++ as just after the XS(1f!2) pulse. Since ms = ~!. Thus one can inven one multiplet only.
eventually we will observe the S-spins, we seek (mj-jg(t)lmj+) during the Note also that as T goes to zero, (mlmsle(r+)!mlms) goes to
observation time t>r++. Now (mj_lg(r++)lml+) is prcx1uced by XS(7C!2), (mlmsle<0-)Imlms), representing the fact that two closely spaced 'I({l. pulses
hence from elements of e{r+) which are diagonal in mi. At t = 0-, before the are equivalent to a single 7f pulse. We can now use (5.254) to get the S-spin
first Xj(wll) pulse, we assume f! is in thennal equilibrium, hence diagonal in magnetization
both ml and ms. Hence, under the action of the two X j (1f!2) pulses, f! remains (S+(t,» = T, (s+ e<t,)}
diagonal in ms· Thus, (ml-Ie(r++)lmj+) arises solely from elements of e{r+)
just before the pulse XS(7C!2) which are diagonal in both ml and ms. These = ~)mf - 1e<',}lmf+) (7.29la)
mamx elements can be traced back to elements (mlmsle(r-)Imlms) as well
m,
as to elements (=F msle{r-)l± ms) using (5.253): = L(ml _le{T++)lml+)e-icns-amt)t2 (7.29Ib)
(± msle<r+ll± ms) = 1(+msle<r-ll + ms)+(-msle<r-)I- ms)J m,
;
± 1:[(-msle<r-)1 +msl
= ~ L: [emf + le<r+)lm,+) - (mf - le<r+)lm,-)J
- (+msle<r-)I- ms)J (7.286)
2 m,
x exp [- i(ils - ams)t21 (7.291c)
These in tum arise from the same elements at t = 0+, just after the first Xl(-rr!2)
pulse, using where the malTix elements of e(r+) are given by (7.290). Equation (7.29Ic)
is the mathematical expression of the fact that the S-spin signal is detennined
(7.287) by the population differences between the states ms = ! and ms = -4, for
and each value of ml' existing just prior to the XS(7r{l.) pulse. These population
differences oscillate as a function of T, the time difference between the Xj(1r/2)
(Tmsl,(O+)I± ms) = ± ~1(+msle<O-ll + ms) pulses, at the nltes [} I - a/2 and ill + a/2.
We can utilize (7.290) together with (7.275) to follow what the two X I (Tr!2)
-(-msle<O-)I- msl) and (7.288) do to the population excesses. Thus we have for t = 0- , the population excesses
-i
m, '" 6 + Ll cos (ll[ - aI2)r - 5 + L1 cos (01 + a!2)r and since
m/ = ! 6 - Ll cos (II{ - 012),) - 5 - L1 cos (n l + a!2)r .. "( h
M./(ms) = T[(+msle<'+lI + ms) - (-msle<'+)!- ms)J . (7.297)
(7.293)
Therefore, from the point of view of the S-spins. the m/ :: ! line has a population Therefore, using (7.295)
difference 31 t '" r+ of
(+msle<'+)1 +ms) - (-msle<t+)I- ms)
(+ + 1e<'+)1 ++) - (+ -1e<'+)1 +-) = [(+msle<O-)1 + ms) - (-msle<O-)I- ms)] cos (n l - ams)r
"" 26 - A.[ cos (0/ - aI2)T - cos Uh + oI2)r] (7.294.) (7.298)
Moreover,
and the m/ :: -i has correspondingly (+msle<O-)1 + ms) +(-msle<O-)1 - ms)
(- + 1e<'+)I- +) - (- -1e<'+)I- -) = (+mslg(r+)1 + ms) +(-msle(r+)I- ms) (7.299)
'" 26 + L1[ cos (ill - a/2)r - cos <fl[ + a/2)r} (7.294b) Combining (7.298) and (7.299) we get (7.290). In summary, the pulses work
Bodenhamer! and FreemtJn derive the resuil of (7.290) by use of the classical by having the X pulses take (m/msle(O-)lm/ms) into (±msle<t)l=fms),
spin picture (Fig. 7.45). The X t (7Cn,> pulse (Fig. 7.45a) puts the [-magnetization letting these elements evolve, then taking them back imo diagonal elements
along the y-direction in the 1 Jl)(ating frame (Fig. 7.45b). This precesses at (0/- (m/mslg(r+)]m/ms) wilh me second I-pulse. These modified populations are
ams) (Fig.7.45c). The second X,(7fn) pulse r()(ates this vector down into the like those of Fig. 7.4la. Thus we have modulated the populations of a given m/
z-z plane, with a .:-component M,/(ms) given by {Fig. 7.45<1) seen by the S-spins, and set up a case where the S-spin pulse can utilize an 1-
spin population excess. Indeed, the XS(1fn) pulse sets in motion magnetization
(7.295) oscillating at angular frequencies (ns - ami) but with amplitudes and phases
This equation can be converted to an equation for elements of the density matrix dependent on (il, - amS)r.
since The situation we have just described shows that if one has an inhomoge-
neous magnetic field smearing out the chemical shifts, there is no time at which
r, the second rf pulse will find all (he spins properly oriented for inversion of
I: O'
('I (bl the magnetization of one multiplet line. Certainly if two lines corresponding to
l
ms = ± are not resolved, the method of adiabatic passage through one line
will not work. But, since we can eliminate magnetic field inhomogeneity ef-
L .,
fects with a spin echo, we are tempted to look for a pulse solution to inversion
of one multiplet. It is easy to find one. Wha! we must do is refocus the field
Idl
" inhomogeneity without refocusing the spin-spin coupling. A
(e) Yl t • y-
(7.300)
r,
would refocus both the field inhomogeneity and the spin-spin splitting. However,
thinking back to spin-echo double resonance, we see that if we also flip the 5-
spin by a 11" pulse simuhaneously with the [-spins, the spin-spin coupling will
not refocus (see discussion at the end of Sect. 7.21). Thus, the sequence
=
angle (nl - (lmS)T at time T-; (d) an X 1(></2) pulse at I T puLs /If1(ms) int(l the .7:-~ will produce a shuationjust after the last pulse in which the M/(!) magnetization
plane, with a eoml>onent along the I-axil! of -Al t cos (n, - oms)T
is inverted, but not the M/(-!) magnetization.
342 343
Then, a 7r/2 pulse applied to the S~spins will generate transverse magneli~ We tum now to such a method. which has been named the "product operator"
z.ation. It, too, will dcphase if there is magnetic field inhomogeneity, but can be fonnalism. This method was developed independently by at least three different
refocused by an X S (7r) pulse accompanied by an X/(;rr) to prevent the spin-spin groups: Sorensell et al. [7.71], Van den Ven and Hi/bers [7.72] and Wang and
spliuing from being refocused. The same condition 2Ta = 7r of (7.287) will guar~ Stichter [7.53.73]. We shall present the method in connection with the two-spin
antee that at 2T following the first Xs(-;rr/2) pulse the magnetiz.alion will look system. The generalization to more spins is straightforward [7.71]. We confine
like Fig.7.43c. i
ourselves to spin nuclei.
We can write this sequence in a slightly different format as The typical situation one encounters is illustrated by (7.212) which describes
spin echo double resonance:
x/(7fn) ... T ••• X/(1r) ... T ••• Y/('Il"I2) X/(7f)
Xs('II") XS(1fI2).··T",Xs ('II")··.T ... S-echo
(7.3OJ)
with +
(I (t» =
11WO, + .
ZkT Tr{I [exp(l(il/I:: + ilsS:: - uI:5::)(t - T»X/(l'l")Xs(7f)
2aT = 7f (7.304)
x exp (i(il/ I, + ilsS:: - aI:I:)T)X,(7rn)I::xil(7rn)
in which we utilized two lines. one for the I-spins, another for the S-spins to x exp(-i(il,I:: + ilsS, - aI:S,)T) (7.212)
help keep track of the two spin systems.
If the "inhomogeneity" comes instead from the existence of a number of s l
x X '(7f)Xi ('Il")exp (-i(il,I: + ilsS: - aI:S:)(t - '7f»)}
chemical shifts of the I-spins, then one can still transfer I-spin magnetization to
The system has an initial density matrix corresponding to thennal equilibrium
enhance S·spin signals:
given by (7.210):
X,(7fI2) ... T ••• X,(7f) ... T
XS(7f)
••• Y,(7fn)
X S (1rI2)···acquire (7.305)
JO-) =
~
'Z!'(t + hwo/I: +IIWOS
IT
S:) a21~
in which we assume the values of a are rather similar for all chemical shifls. The effect of the rf pulses is to produce rotations. We may symbolize these in
This sequence was discovered by Morris and Freeman [7.70] who gave it a name general by unitary operators R. An example of such an R is X/(9):
"INEPT' to stand for "insensitive nuclei enhanced by polarization transfer".
X/(9) = ei 'z8 (7.306a)
Now, the R's plus TJ and Ts are all simple rotation operators of the general cos (aSztl) = cos (atd2)
form exp(iI:I;8), exp(i/yO), exp(i!..,O). Their effect on a density matrix e(O-) For (7.316), we get
such as (7.210) can be easily followed using equations such as d3
sin (as,'I) = S ( , _ (at
z a 1 3! s2z
+ (at5!1)5 S4z + ... )
(7.311)
We now make use of the spin! property. If we choose as a basis set functions
We now see how to evaluate TISIt;rT1S1 for a = x, y, z. An interesling point
lms), (mS = or +! -!),
the most general function is of the form is that it takes single spin operators, It;r, into a sum of a single spin operatOr and
1{;(t) = C+(t)lms =!) + C_(t)lms = -4) (7.317) a product of spins It;rSjJ. We see therefore that if applied several times, it might
develop terms like It;rJjJS.rS6' But such terms can be simplified as we now
Explicit evaluation gives explain.
S;1{;(t) = C+(t)S;I~) + C_(t)S~1 - !) Suppose we think of either the I-spins or the S-spins in terms of the 2 X 2
Pauli matrix representation. We recognize that spin products such as I",Iy are
= iC+(t)I~) + iC-(t)1 -~) products of one 2 X 2 matrix with another. The result is also a 2 X 2 matrix.
= i,p(t) (7.318)
346 347
Now any 2 X 2 unitary matrix can be represented as a linear combination of the For simplicity, we consider a case in which il r ::: ils::: 0 (exact resonance
three Pauli matrices plus the 2 X 2 identify matrix. Thus. it must be possible to for both spins). We apply an X r (7f!2) pulse al t '" 0, and an Xs(Tr/2) pulse at
reduce products such as Izly to linear combinations of the identity operator with time tl, let the system evolve under TIS from 0 to tl, and after t1 for a time t2.
la's. We define B, and fh by
In fact, we have shown one such example of a product Ialp for the case
B1 :::at, • B2:::at2 . (7.330)
that cr ::: p:
1~:::1 . cr:::x.y.z . (7.323)
Then
Evaluation of Ihe portion involving 5, is left as a homework problem. When these are substituted into (7.335), the two tenns in 111 57; cancel, and the
teons in IzS y combine (cos 2 8212 + sin 2 92fl = 1) giving
348 349
TI S(8 2)XS(1r(2)TIS (8 1)X1(7r(2)I z Xii (7r(2)T1S1(8 1)X5 1(7r(2)Ti!] (82) spin coupling [7.74,75] (Sect. 4.9). The bonding might be direct, as in a C l3 H I
= I" cos 82/2 cos 8 1/2 - I z S z 2 sin 82/2 cos 8 1/2 - I z S,,2 sin 8tf2 . fragment, or remote via an electronic framework as in the H-H splittings in ethyl
(7.337) alcohol (CH3CH20H) between the CH 3 protons and the CH2 protons. In solids,
Therefore, the contribution of (7.329) to e is the dipolar coupling proves proximity. In liquids, thc dipolar coupling shows up
through thc nuclear Overhauscr effect. One thcn distinguishes between bonding
e(tl + t2) = hWOI (lz cos 82/2 cos 9 1/2 - IzS z sin 92/2 cos 9 1/2 and proximity, an important distinction in large biomolecules [7.76] in which a
ZkT long molecule may fold back on itself. A variety of methods such as spin echo
- I z S,,2 sin 9d2) (7.338) double resonance, spin tickling [7.71], INDOR [7.65J and selective population
These various terms tell us what elements (mlmsl,q(t)lm~ms) are nonzero transfer [7.66J have been employed, to utilize the fact that if I and S are coupled,
at time tl +t2' For example, the first term involves I z , hence gives nonvanishing perturbing the I-spin resonance will produce an effect on the S-spin resonance.
All of those methods involve sitting on one line and point-by·point exploring the
elements
other lines. leener's discovery of the two-dimensional Fourier transform method
(± mslu(tl + t2)I~ms) (7.339) converts the approach to a Fourier transform method with all its advantages. In
The term IzS z gives similar elements, weighted with mS: this section we analyze the original Jeener proposal, which now is commonly
referred to as the COSY (correlated spectroscopy) method since it reveals which
(mlmsllzS z Im~ms) = ms(±IIzl=f)Oms,m Omr,±Om/,:,: s (7.340) pairs of chemical shifts are correlated by a spin-spin coupling between them. It
involves a single nuclear species (e.g. HI) and a pulse sequence
Lastly IzS" givcs nonvanishing terms
(7.341,) (7.342)
(± ±!e(t, + ',li'!' '1')
However, since two nuclei are involved, with II = IS but £2r i- £25 owing to
(± Tit"', + ',)1'1' ±) (7.341b) their chemical shift difference, it is better to think of this as
whose further significance we discuss in Sect. 9.1. XI(7r/2), XS(7r/2)··· tl ... X / (7r/2), X S (7r/2) ... acquire I(tz) and S(t2) ,
In Table 7.2 we list the other useful relationships employed in dealing with (7.343)
spin operators.
where a single HI of sufficient strength flips bOlh spins.
We have already discussed the basic principles of coherence transfer in
Table 7.2. Useful relations for unitary operators U. (U is a function of I, or 5, or I and 5) Sect. 7.25. There we saw that if we are observing the S-spins in time interval
two, the density matrix elements (m/-Ie(tl +t2)lmr+) are responsible for the
U/J(1,S)h(I,S)U-l U/l(1,S)U-IUh(l,S)U I
Ucxp(i/(I,S)lU- 1 = exp[iU/(I, S)U-I I signal. These elements, however, have passed through both (mlmslfl(t)lmrms)
and (=fmsle(t>!±ms) during time interval tl. Thus, the I-spin resonance fre-
For spin !: cos (1.8) = cos (8/2) quencies modulate the S-spin signal during 12 as a function of tt. We can express
sin (1.8) = 2/. sin (8/2) these facts colloquially by saying Ihe (=f mslu(t)1 ± ms) matrix elemcnts during
tl "feed" the (In, ~ Iu(t)!ml ±) elements during t2.
Such a description almost suffices 10 describe the pulse sequence (7.343). All
one need add is that at t = 0+ both (=fmsle(t)I±ms) and (ml =f lu(t)lml ±)
are excited, so that during 12 (ml-lu(t)lml+) is fed by both (=fmslel±ms)
and (ml =f lu(t)lml ±). [Note that (mrmslu(t)lmlms) during II also "feeds"
(ml -lu(t)lmr+) during 12, as is shown by (7.291) and (7.290), however, since
the diagonal elements of e are independent of time they do not by themselves
7.27 The Jeener Shift Correlation (COSY) Experiment introduce any I I dependence to e.J
We could try to carry out an analysis using U similar to our approach to spin
Since thc early days of high resolution NMR, an important goal has been to tell coherence. However, Ihis method rapidly becomes quite cumbersome. Indeed,
which nuclear resonance lines arise from nuclei which are bonded to one another. one finds that all 16 elements of U are excited by this pulse scquence! We turn,
The existence of bonding manifests itself in liquids through the indirect spin- therefore, to the method of spin operators.
350 351
Since the two nuclei are identical, we consider a Hamiltonian in the singly Expressing T(t" as
rotaling frame, rotating at the angular frequency w of the alternating field. The
T(t" = Tr(l])Ts(t])T/S(t]) (7.351)
Hamiltonian includes a chemical shift difference so that il r and ils are not
equal. Rather, Ihey are given by we get
(7.344) T(t,)[yT-1(t])
with ur and us being the two chemical shifts. The Hamiltonian in the rotating = Tr(tt)Trs(tt)ciflsS~ll lye-iflsSztl T } (tl )Tr-'(t,)
rf
frame is still l
= Tr(t,)Trs(t,)[yTif} (t,)TI- (tl) (7.352)
(7.345)
Utilizing the fact that
Both nuclei experience the same alternating field, H" so Ihat for any pulse the
T,{ABCj=T'{CABj (7.353)
two spins experience the same phase for HI and the same angle of rotation, e.
Thus, for example, if there is an Xr(e) there is also an Xs(e).
Since we observe both I and S spins, we seek
r s
we pull the T '(t2)T '(t2) from the far right hand end of (7.348), giving for
the first pan of the trace
352 353
(7.358) Table 7.3. EffecU! of operators from (1.357) on an initial operator I,
X s (,,/2)
XI(,,/2)
-a1,S,I,
,,
But of these 16, only the eighl operalors
(7.359.)
,
I 1. 1. 1.
I,
1.
I.
-I.
I,
I.
I, 3
3
• 1,5, •
[:nSz. ["Sz. IzS%. IzS" (7.359b) ,
5 1.5, _1.5, 1,5, 1.5,
[~~:5J x [-~:/.J~
1,5,
=
/.1,5, -(i/2)1,5.
5
,,
7 =
-1,5,5, (i/2lf)5.
=
will give matrix elements of the fonn (7.358). In fact. using the symbol 0 10
Sland for these eight funclions, (7.357) will consist of a sum of terms Tr {I+O} ,
8
1.5, 1,5, -1,5. 1,5,
+1,1,5,5. (i/2 11.5, 8
,
or Tr {S+O} which vanish if 0 is bilinear. Thus. only the terms (7.359a) of the 10 +1,5,1, = 5,,/4 10
3S4 355
In writing expressions such as (7.365), Van den Ven and Hi/bers [7.72) time, t2, after the second pulse that the spins reach the condition of Fig. 7.43c.
introduce a useful nClation which greatly cuts down on the number of lellers one That coherence transfer is involved is also shown by the fact that this term also
needs to write down. They point out that using this formalism one encounters involves
trigonometric factors such as e-inl12sin (ilstt)+e-insf2sin (illtl) . (7.370)
cos (att!2), sin (atzf2), sin U2[t2), sin (ilst t ) (7.366,)
The first term corresponds to a spin precessing near ils during the first interval,
They write these as and near il, during the second. The second term represents precession near ill
during the first interval. and near ils during the second.
C~ == cos (attn.) 5~ == sin (at2n.) ,
IT one makes a double Fourier transform of (7.366). one gets two types
51 == sin (nlf2) 51 == sin (nst,) (7.366b) of peaks. From the term involving cos (atl/2) cos (at2/2), one gets peaks near
in a notation which is self-evident. One merely has to be careful to remember WI"" W2 "" ill or ils. which lie close to the diagonal line WI:: W2 in a (WhW2)
the "i"in the arguments involving a which is not present in those involving l n plol. From the term involving sin (aid2) sin (at2f2). one gets peaks near the
or ils. points WI "" ill. W2 "" ils or WI "" nS, W2:: ill. Such peaks immediately show
Returning to our discussion of (7.365). we now add to this the conuibution that the spins with chemical shifts il, and ns are coupled. Since the frequencies
from 5~ to e(O-). which we obtain simply by interchanging ill and ils in are always il, ±a/2. not il" and ns±a/2, not ils. the peaks are actually
(7.365). The final result is clusters of four peaks. Thus, the diagonal peaks occur at
(P"(l) + s+(t» : ~::; {cos (atl/2l cas (at,J2) WI"" ill +a/2 W2 "" il, +a/2
wI""il , +a/2 W2""nl-a/2
x [e-inI12 sin (il/tIl +e-insI2 sin (nstl)] ill - a/2
WI "" W2 "" nl - a/2
+ sin (at In.) sin (atz!2)(e- in,l2 sin (nstj) + e-insl2 sin (n/tl)]) w}:fh-a/2 W2""n/+a/2 (7.371)
(7.367) Likewise, the off-diagonal peaks are actually a cluster of four peaks at
The various terms in this expression have simple physical meanings. We (7.3720)
note fi~t that we have a term involving cos (att/2) cos (atz!2). This term is
present even as tt and t2 approach zero. It has the terms and another cluster of four peaks at
e-inl12 sin (n,tl) and (7.368a) WI : fls ± af}. • w,: fI/ ± af}. (7.372b)
If we think of the factor sin (atl/2) sin (at2n.) as modulating signals at il,
e- inst2 sin (ilstil (7.368b)
and ils, we note that the buildup of the cross terms exp (-iil/t2) sin (nst}) or
which correspond to oscillation near il, during both t I and t2, or oscillation near exp(-inst2) sin (il,tl) requires a time ""TrIa. In a typical molecule. a spin I
ils during both tt and t2. Of course the actual eigenfrequencies are ill ± an. will have coupling constants to both nearby and more distant nuclei. The stronger
and ils±a!2, but if a<: Iill - ilsl the term "near il," or "near ils" is well coupling constants a will produce their cross peaks in a shoner lime than the
defined. Near t, :: t2 "" 0, this term is independent of a. It therefore does not weaker coupling constants. Thus, by limiting the intervals t} and t2, one can
require a coupling to exist, and represents an effect of uncoupled spins. as is select large or small a's. If, indeed, two spins are not coupled at all, so that a is
evident from the fact that the precession frequency is the same (il, or ils) strictly zero, the cross peaks will never arise.
during both time intervals.
On the other hand, the second term is proportional to sin (ati/2) sin «(lt2/2).
It reaches a maximum when atd2 "" at2/2 "" Tr/2 or
7.28 Magnetic Resonance Imaging
(7.369)
This is the condition we encountered previously in (7.356) as the time which The use of NMR to produce two- and three-dimensional images has, since its
optimizes coherence transfer. This is the time, t" after the first 7(/2 pulse at invention in the early 1970s. increased at an explosive rate. We refer the reader to
which the spins have precessed to the condition of Fig. 7.42d. 11 is also the the books by MaflSfield [7.77] and by Ernst et al. [7.78] for extended treatments
356 357
of the many variants. We present a shan treatment of the principles al this point What spatial forms can h: take? If we simply consider it to be a gen~ral
in the text since it enables us to discuss the method of Kumar, Welti, and Emst, function of r, we can expand it in a power series in z, y, z about a conventent
which utilizes the two-dimensional Fourier rransfonn approach. origin. We assume we have shaped the field so that only the lowest terms are
Two pioneering papers, published within a few months of one another, needed. Then
are the proposals of lAuterbur [7.79) and of Mall.fjield and Crannell [7.80].
which independently propose use of NMR 10 ronn images. The basic concept
ll~(r) = h:(O) + x ({)~~: )0 + Y (a~~v )0 + Z(aa~: )0 + ...
is straightforward. We start with a sample which has a narrow NMR line in a == h:(O) + xG;e +yGl/+zG;z (7.378)
uniform slatic magnetic field. To the uniform field we add a second magnetic which defines the components Gz , G", G: of the gradient G of h~(r) at the
field which is nonuniform. ideally one with a constant gradient. The field gradient
broadens the NMR line. For a constaol gradient (e.g. in the z-direction) the origin:
magnetic field component, HI> parallel 10 Ihe uniform field is G = iG z + iG y +kG: (7.379)
The reader may wonder whether we can guarantee that it is possible to generate
H~=Ho+z (OH')
fu (7.373) an 11 with the spatial dependence of (7.378). After all, h must obey the laws
of physics (V. It = 0, V x It = 0). We explore these aspects in a homework
so that planes of constant z correspond to planes of constant precession frequency. problem. The answer is that we can achieve (7.378), but for there to be a term
Thus, a given frequency interval is bounded by two frequencies which correspond such as kxG z for example, there must also be terms in the i and/or i directions.
to two planes (of constant z). The total NMR intensity in that frequency interval However, since as we have seen the transverse components do not affect IHI to
is proportional to the number of nuclei in the sample lying between those planes. first order, we can neglect them.
Thus the NMR absorption spectrum provides a projection of the sample spin Let us then define two functions, a frequency distribution function, !(w),
density integrated over planes perpendicular to the gradient direction. From a which gives the NMR intensity at frequency wand is normalized to satisfy
series of such projections for various gradient directions, one can reconstruct the
object. J f(w)dw = 1 (7.380)
We are considering, then, a uniform field kHo to which there is added a
small additional field h(r) which is static in time: and a spin density function e(r), which gives the number of nuclear spins in a
unit volume at point r, and which is likewise normalized so that
H = kHo + h(r) with (7.3740)
h(r) :::: ihz(r) + ihl/(r) + kh~(r) and (7.374b)
J (,(r)d'r = 1
(7.381)
dw=;Gdz' (7.383)
Since Ho:> Ih(r)l, the effects of ih;e and ih" are merely to rotate H slightly
without the first oroer changing its magnitude, whereas h~ changes H(r) to firs! Therefore
order:
H = V(Ho + h~)2 + hi + h~
!(w)dw = dz'
f'lJ
%':::consl
dx 'd y"",x,y,z
I"J I I I) (7.384)
The integral in (7.384) is the projection of the spin density funclion e(x', yl, z')
= JH~+2Hoh~+h~+I.;+h~ (7.376) on the z'-axis.
Thus, utilizing (7.383),
Neglecting the terms quadratic in the components of h, we get
359
358
Since this projection depends on the direction of G, and since the magnitude of problem then becomes one of finding a suitable numerical approach for getting
f(w) depends on the magnitude of G, we put a subscript on f(w) : the most accurate value of the integral from the data as given. Of course, in the
.."Gfc(w) = JJ dx'dy'e(z', y', z') (7.386)
process of selecting the approach, one gets guidance as to whal data to collect.
The approach of reconstructing e(r) by Fourier transforming the spin den-
:r'=const sity distributions is one approach to finding e(r) from data for fc(w). Lawerbur
Since fc(w) can be measured, we can view the integral on the right as known [7.79] demonstrated a simple graphical means, and pointed out that Ihe solution
experimentally as a function of z'. to the problem of getting e(r) from spin-density projections was well known
We can think of making a variety of measurements 10 delennine the w from Olher areas of science.
dependence of fc(w) for a variety of different G directions, thereby getting the Another experimental approach was proposed by Mansfield, based on 0b-
projecled spin density for any desired projection axis, z'. serving free induction decays. If one applies an X(ll'"{l) pulse to the spins and
It is now easy 10 show that such measurements enable one to reconSmJct observes the free induction decay in Ihe reference frame rotating at wo, the com-
e(z, y, z). Firsl we take the experimental dara for a given z' and fonn the Fourier plex magnelization is
transform of the data:
(M+(tl) = (M.(t)} +i(M,(tl)
ek = J e-ik:r'dz' JJ dx'dy'e(z',y',/) (7.387a)
= iMo J e(z', y', z')e-iw(z' ,JI,z')ldz'dy' dz'
where
k = H' (7.387b)
= iMo J f!{z', y', i)e-i.,.G:r"d:i dy' dz' (7.39Oa)
'. by the direction of the field gradient, and each point in time corresponds to a
magnitude of wave vector given by (7.392). Clearly, the free induction decay is
also producing f:!k along radial lines in k-space as in Fig. 7.46.
9, k,
.'11:.7.46. The paths in k-space (k = ki') eXI>lored In his initial paper, Mansfield [7.80] proposed using the free induction decay,
by applying the gradient C = G i , in a variety or giving (7.393). Since his initial idea related to crystallography, he explored the
directions i' in a rhUle. The various directions 91, case that the spin density is given by a periodic lauice, and demonstrated the
i:;,
92, ... ,9,. correspond to different directions
-, i;, imaging scheme by making a macroscopic layered sample to simulate atomic
" periodicity.
360
36.
In Fig.7.46 we display a set of k trajectories for a planar sample. Even which is the Fourier transform in the plane z of the spin density. Since f z is
for nonplanar samples, it is convenient to collect data from planar slices in me fixed while recording (M+(t», the data are for a fixed kz • but for a continuous
sample, a sequence of slices thus providing the full three-dimensional object k y . They may be represented as in Fig. 7.47. They thus sample (}k for the z-slice
The methods of selecting well-defined slices are pan of the present an along parallel lines in (k;r:. kll) space.
of imaging. We refer the reader to (7.77] for example. A simple approach to In a full three-dimensional version, one applies the initial rl2 pulse in the
generate a slice perpendicular to the z-direction is to start by applying a gradient absence of gradients. Then. one applies
in the z-direction. If one applies a pulse at a panicular frequency WI, the spins
G=iG z fo< I.
at coordinate z such that
G=iC" for Iv (7.400)
(7.394)
G= kG r for I.
are perfectly at resonance. An X , (7r!2) pulse will then rotate them into the x-y
recording the free induction decay during t r • Then
plane. Spins within a Ll.z given by
(7.395) (M+(t» = iMo J e-izGzlze-iIlG,I'c-irG.I. p{x. y. z)dx dy dz (7.401a)
will also be flipped. but spins farther away in z wiJl be less effectively tilted into with
the :r-y plane. By in fact shaping the time dependence of (HI) I one can improve
(7.4Olb)
on the z-selection.
In any event. if one now turns off Gr. one has produced an initial transverse ""d
magnetization for all x. y within Ll.z. One can then apply gradients in the :r-y
t z = t - (t z + ty) (1.40lc)
plane to record the free induction decay for any direction k = G in the x-y
plane. All planar imaging methods must initially carry out some such precedure This approach gives one the Fourier transfonn of u(r) lines parallel to the k r
to define Ihe slice Lb. axes fonned by the intersection of the planes k>; = constant and kll = constant
Kwnar et al. {7.81] proposed a two-dimensional (or three-dimensional) (Fig. 7.48).
Fourier transform approach to collecting data which has been adopted by many If one is interested in knowing the density pattern in a slice of known %. it
workers in the field. is much faster to collect the data using a z·gradient to define the slice excitation
We start by describing the two-dimensional version in which one first exciles and then work with two-dimensional transfonns.
a slice in the z·direction. One then applies a gradient An exceedingly rapid way to sample all the points in one slice within a
single sweep was proposed by Mansfield (7.82J and called by him the planar
G=iG;r: (7.396,)
echo method. In the planar fonn of this method. one first selects a slice (say
for a time called f;r:. then a gradient
G = iCy (7.396b) "
for a time called t y • During t y • one records the free induction decay. Using
(7.170) and (7.390) one has, neglecting relaxation. in Ihe rotating frame FIg. 7.47 " Fig. 7.48
(M+(t» = iMoL\z J
e-i;r:,..Gzlze-iyG,f, e{x, y, z)dx dy. (7.397)
363
t
at fixed z), then applies a weak steady gradient in the z-direction, and a strong
gradient in ~e y-direction which is switched ahemately between two values +GO
k(t) == , J G(tl)dt' (7.406)
o
and -Go (FIg. 7.49). Thus, following slice selection one has
0 0 0
0 0
k.
0
I
Gy
8.1 Introduction
Hahll's discovery of spin echoes [8.1) demonstrated thaI one could defeat the
0 effect of magnet inhomogeneity in obscuring the true width of magnetic reso-
.y(j,t~ nance lines bec;lUse the echo amplitude M(2T) decayed exponentially willi 2r!T2.
where T is the time between pulses. However, if the magnet inhomogeneity is
large, and if the diffusion rale of the nuclei sufficiently great,
I
Fig.7.S0. The {mth in k-spa.<:e explored in the planar echo method of M(lflSfitld
M(2r) = M(O) "p(-2r/T,)exp [- (1 ~:)' ~ Dr'] (8.1)
where D is the diffusion constant and DHIDz the gradient in static field. In some
Fig. 7.51. Projections of a discrete matrix filled to form the letter P. ProjectiolU in G z instances, then, the r 3 leon might obscure the T2- Or, conversely, if DHlfJz is
or 0, alone do not contain sufficient inforllllllion to uniquely determine the original
object. Projection orthog<mal to the special gradient direction G. conlains all information known, the T 3 term gives one a method of measuring D. The derivation of the
necessary to complelely'del.crmine the original objed distribution diffusion tenn is given in Appendix G.
Carr and Purcell [8.2] invented a clever scheme for eliminating the T 3 tenn
when desired. 'Their proposal initated a class of experiments in which a sequence
ing scheme of Fig. 7.50 has parallel scanning lraces ~1'GJl70 apan in k-space, of pulses is applied 10 the resonance. In this chapter we discuss first the Carr·
which corresponds to a spatial wavelength Purcell concept, then an ingenious modification invented by Meiboom [8.3]. and
, __ 2r __ 2. finally tum to some of the most intricate pulse melhods yet invented which have
A (7.410)
k", 21Gyf], the remarkable propeny of enabling one to eliminate the dipolar broadening of
resonance lines in solids. 11lese concepts were introduced by Waugh [8.4J and
This wavelength imposes discreteness (i.e. "rows'') on the spin density function
Mansfield [8.5], and have been extended by Vaughan [8.6J. as well as a group
elr).
of very talented resonators who have spread Qut from these initiators.
Another way of saying this is that the echo train produced by the oscillatory
gradient G y has a period 2Tb. hence corresponds to observations al a spectrum
of frequencies spaced apan by angular frequency Llw of
1 8.2 The Carr-Purcell Sequence
Llw = - (7.411)
2r,
We begin a discussion of the Carr-Purcell sequence by llssuming we call neglect
Bm L1w implies a ..1z of
the T 3 diffusion tenn. To measure a T2 by the conventional spin echo, one must
L\w I 1 make a sequence of measurements, each al a different value of T. The envelope
LIz = - = - - (7.412)
"fGy ""IC" 27"b of the echo amplitude as a funClion Of2T gives one T2. Carr and Purcell pointed
Each frequency corresponds 10 a row of trees. The weak z-gradient produces the out that the entire envelope could be obtained at a shot if one applied the proper
"view al an angle" shown in Fig. 7.51. sequence of pulses.
366 367
Suppose at t::: 0 one applies a 1r/2 pulse in which HI lies along the +x-axis during the interval after Ihe 1r pulse. It is the fact Ihal diffusion makes rephasing
in the rotating reference frame with the rf frequency w exactly at the Larmor imperfect which causes the next echo to be smaller. In Appendix G we show that
frequency woo Such a pulse turns the magnetization Mo to lie along Ihe +y-axis. if a 1r/2 pulse at t "" 0 tilts the magnetization Mo into the x - y plane, a 1r pulse
If one applies a 1r pulse at t ::: r, also with HI along the +x-axis, an echo is r later produces an echo at 27, M(2r), smaller than Mo owing to diffusion by
fanned at 2r with the magnetization along the -y-axis. If now one applies a 2
1r pulse at 3r, another echo will foml al 47, this time along the +-axis. In this
manner, successive 1r pulses at (2n+l)r(n::: 0, 1,2, ... ) foml echoes at (2n+2)r,
M(2r) "" Mo exp [ _ D (7 ~~) ~r3] exp(-2rIT2)
the echoes fanning along the -y-axis for odd 11, and the +y-axis for even n. Since == Moa (8.2)
all components of the magnetiz..1lion in the x-y plane are decaying exponentially
with time constant T2, the echo sequence likewise decays exponentially with T2. When the echo is formed, we have recreated at 2r the situation which existed at
A sequence of such pulses and echoes is shown in Fig. 8.1, with Ihe sign, t ::: 0, except the magnetization is reduced by the factor a, and points along the
positive or negative, showing whether the echo fonns along the positive or neg- -y-axis instead of the +y-axis. Therefore we can consider the magnetization at
ative y~axis. 27 as another initial condition for applying the argument of Appendix G for the
The behavior we describe enables one to obtain the entire echo envelope in a next interval from 27 to 4r with a pulse applied at 3r.
single pulse train, clearly a convenience. If one is studying a weak signal, so that The result is
the signal-to-noise ratio is imponant, the Carr-Purcell sequence is a tremendous M(4r) ::: M(2r)a::: Moa 2 (8.3)
advantage as we explain. The noise depends on the bandwidth of the apparatus.
One limits the bandwidth so as to pass the echo signal without undue attenua- If we have n cycles 2r, we therefore get
tion. Clearly the same bandwidth can be used for either a Carr-Purcell train or M(n2r) ::: Moa" (8.4)
a convelllional echo. After each single echo, one must wait several times TI for
the system to recover before measuring the next echo. However, in the time Therefore
il lakes to get one echo, we can obtain the full envelope with a Carr-Purcell
sequence. If we take N echo signals to define the echo envelope, we could M(n2r) "" Mo exp [ - (Y ~~y D(n27)~72] exp( -n2r1T2) (8.5)
obtain N Carr-Purcell pulse trains, so each Carr-Purcell echo is recorded N
times. If the spin echoes are taken to correspond to the times of the Carr-Purcell Thus, if we change r but hold n2r constant so that we are comparing two
echoes, the Carr-Purcell train will have a better signal-to-noise by Jiii. different echoes of the Carr-Purcell train a fixed time apart, we can vary the
We have remarked that the Carr-Purcell pulse sequence reduces the effect faclor involving diffusion but we do not vary the faclor containing T2. Hence
of diffusion on the echo decay. The reason can be seen as follows. If there is no we can make the effeci of the diffusion tenn negligible.
diffusion, a spin dephases during the interval r following each echo and rephases It is clear thai what is involved are two facls:
A Carr-Purcell pulse sequence may contain many pulses. If the pulse rotations
---------- deviate slighliy from 'IT, there are cumulative effecls which may become large.
------ ------ ---------- --------- Echo CIII'i'lop"
We illustrate in Fig. 8.2. (This figure is drawn for a negative "'/' hence involves
Fig.S.1. The Carr-Purcell sequence and tlte resulting echoes. As des<:ribed in the text, a right hand rotations.) For simplicity we take the initial pulse to be a perfect 1r(l
positive echo forms along the +y-axis in the rotating frame, a negative echo forms on the pulse about the +x-axis, and we consider diffusion effects 10 be negliglible. At
-y-axis
368 369
J
, (,) (b) Meiboom and Gill [8.3] were the first people to find a solution to this
problem. They likewise used a phase coherent method, but introduced a 90"
-/2 ---0- -----
phase shift into the radio frequency field between the initial 7rn. pulse and the
subsequent 'If pulses. Thus if HI for the 1f/2 pulse were along the +y-axis, Hl
M. -----'7---r f---
, _ "\ y for the :IT pulses would be applied along the +x-axis. As a result, the echoes are
'...... "'./
.• -------
r" O' all formed along the +x-axis for negative 'Y' We leave it to the reader to show
x that if the pulse is 1f - 5 instead of 11", the error does not accumulate.
,
(d)
8.4 Refocusing Dipolar Coupling
---.--'-"'-"'-,,---- y 11le dipole-dipole coupling between neighbouring nuclei provides valuable in-
formation for many purposes, but on occasion it is a problem. For example, it
may obscure chemical shifts, and it may cause free induction decays to be soon-
''''(31)' lived and thus difficult to see. We tum now to several interesting approaches to
Fig.&.2a-e. The effect of ,an imperfect If pulse. The initial ../2 pul$e, assunl«! perfect, eliminating or effectively reducing dipolar coupling.
ro~aleti Mo to the - y-axIS at I = 0+. At t = T - • just bo;orore the .$eOOnd pulse, a typical For a group of spins an ordinary spin echo [e.g. an X('IT/2) ... T ... X(:IT)
spm h~ Pl'('(;~ an .~~Ilt 8 in the z - II plane a.....y from the -v-<!ireo:;tion, o....ing sequence] refocuses the dephasing which arises because they are placed in an in-
to static fl(!ld mhomogenc'J1W:-" We no,"" apply a.1I" ~!J ~ rotation about the +;c.u:is. Fig.
8.2<: sho_ the efTec.t on • spm for whICh (J = 0 hes In tho.!: z - 11 plane. It no longer lies in homogeneous magnetic field. Since the magnelic dipolar coupling among neigh-
the z - 10' plane. Just ~ner '" s«ooo 'II" + 6¢J ~Ise at t ::: (3T)+, the same magnetization boring spins is in some way analogous to a magnetic field inhomogeneity, one
makes an angle 26,p wIth the z -" plane. In Fig. 8.2e we see that if the III of the second might ex:pect that the same spin echo pulse sequence would also refocus the
pulse lies along the -x-axis, the rotation direc:tiOll is reversed, so the _(,.. + 6.,> rOlalion
restores the spin to ils orientation at t ::: 0+ dephasing resulting from magnetic dipolar coupling. If the dipolar coupling is
between different nuclear species (e.g. HI and C I3 ). such an echo sequence does
refocus the dipolar coupling. If, however, the coupling is between like nuclei
t ::: 7 - the spins will have fanned oul in the x - y plane. a typical spin making (e.g. HI with HI), the echo does not work. It is easy to see why. The 'IT pulse
an angle 8 with the -y-axis (for negative ")'). At t "" T we apply a ow + 6¢ pulse inverts all the neighbors, so that a nucleus which was precessing more rapidly
about the +x-axis where 6¢ gives the deviation from being a perfect 'If pulse. than average in the first time interval T, finds its neighbors inverted, and thus
For a spin having 8 "" 7rn..
the spin lies along the x-axis. It is not affected by precesses more slowly in the second interval. The usual echo works because the
the pulse at t"" T. A spin for which 8"" 0 would rotate as shown in Fig. 8.2c. It 11" pulse converts a precession phase lead into a precession phase lag, of equal
will lie an angle 6¢ above the +y-axis in the y - z plane where it will sit until size. If at the same time the rate of precession changes, the spins do not come
7r
2T later when the next + b¢ pulse is applied. The result is shown in Fig.8.2d. back into phase.
The spin now makes an angle U¢ with the -y-axis. Each successive imperfect Nevertheless, the fact that a well-defined relationship exists between the
1'1" pulse adds another 6¢ to the deviation. precession frequency when a neighbor points up versus down makes one feel
The cllmulative effect becomes exceedingly serious in praclice because the that there should be some way to undo the dephasing which the dipolar coupling
HI is never uniform throllgh tile sample. COllsequclIlfy thOllgh part o/the sample produces. In the following sections, we explore these ideas.
may have a ow pulse, other part.r do f1ot.
These sons of troubles have major implications for all multipulse sequences
~uch as those described in the later portions of this chapter. One simple approach 8.5 Solid Echoes
IS to apply the H J for the ow pulses alternating along the +x- and -x-axis. Thus
the sense of rotation about the x-axis reverses in alternate pulses. The result is Powles and MallSfield [8.7] discovered that the free induction decay of a coupled
shown in Fig. 8.2e. Instead of getting a cumulative rotation eITor of U¢ as in
Fig.8.2d, the cumulative error is zero!
pair of identical spin ~ nuclei can be refocused by applying a pair of 7rn.
pulses
shifted in phase by ow/2 with respect to each other. Such a sequence might be
denoted by
370 371
X('rr{2) .. . r . .. Y('II"/2) . .. tt (8.6) only follow what happens to the excess, since Ihe remaining nuclei carry no nel
with the echo occurring when tl := , . This sequence has come to be known polarization.
as the solid echo. Powles and Mansfield demonstrated it experimentally for Let us therefore in a thought experiment reach in at random to examine the
CaSO'1 ·2H20, and showed that it followed theoretically. The pulse sequence pairs. We take n/2 pairs in which both nuclei point up, and n{2 pairs in which
perfectly refocuses spin 4
nuclei which interact only in pairs, but does not re- one nucleus points up, and the other points down. This will give us a net n
focus larger groupings of spins perfectly. In this section, we explain the echo nuclei pointing up. The olher N ~ 2n nuclei in the sample must then be equally
for coupled pairs. This same pulse sequence refocuses the first order quadrupole pointing up and down, hence be unpolarized. We neglect them. Now, among the
coupling of a spin I nucleus for reasons explained in Appendix H. It is there- pairs with one nucleus pointing up, Ihe other pointing down, we label the up
fore imponam for deuterium and N t4 NMR. The existence of echoes related nucleus blue, the down nucleus red. Among the pairs with both nuclei up, we
to quadrupole splittings was first demollStrated by Solomon [8.8] for a case of also label one blue, the other red (it does not matter which of the pair we choose
I := ~. It appears to have been Davis et al. [8.9] who recognized that for the for which color). We thus have n{2 red nuclei with spin up, and n/2 red nuclei
1:= I case sequence (8.6) gave perfect refocusing. with spin down. Then the blue nuclei give the magnetization, since the red nuclei
The fact that one needs exactly sequence (8.6) may seem surprising at have no net magnetization.
first glance. Why does one need to phase shift the second pulse by 11:/2? Why We represent the spin-spin coupling by a term 1t'spin-spin of the form
does one use a '11"/2 pulse instead of a 'II" pulse? To help give a feel for the (8.11)
answers to these questions, we shall derive the result Ihree ways. The first method
is a semigraphical one, utilizing a special value of r. The second derivation For pairs of identical particles of spin ! this is exactly equivalent to dipolar
generalizes the first for arbitrary r. The third derivation utilizes the spin operator coupling, as explained in Appendix H. We assume that apart from 1t'spin-spin
method. we are exactly at resonance for any pulses we apply. Then, following an X(1r{2)
We start with the simple physical picture. However, we must present it pulse, the blue spins lie along the +y-axis in the rotating frame.
rather carefully. Consider a sample consisting of pairs of nuclei. That is. each Now, although we have flipped both red and blue spins by 1r{2, the red
nucleus has one and only one neighbor which is sufficiently close thai only the spins at this point have no net polarization and thus their density matrix remains
dipolar coupling to it matters. In the presence of a static field, the nuclei are diagonal with the two diagonal elements equal to each other. Thus, we can
weakly polarized, there being a slight excess with magnetic moments parallel describe the situation at t := 0+, just after the pulse, as in Fig. 8.3b, in which we
to the static field, H o. If there are N nuclei and if we denote by p+ and p_ represenl the blue spins by vectors I and 3, lyir]g along the y·axis, and the red
the probabilities for a given nuclear moment to point either parallel (p+) or spins by vectors 2 and 4, one pointing up, the other down, along the z-axis.
antiparallel (p_) to H o, we have then that there are N+ and N_ nuclei pointing Spins I and 3 will now precess at rates -a!2 and +a/2 (in the left-handed
parallel or antiparallel, respectively, given by sense) respectively about the +z-axis. The explanation of how the echo fonns
becomes very simple for a particular value of " the value which makes arl2 :=
(8.7)
1r12 (Le. a, := 1r). For that time, the spins will be aligned as in Fig.8.3c.
The population difference, n, then obeys At this time, the y('II"!2) pulse will produce the situation of Fig. 8.3d. Now,
n:=N(p+-p_)
during the next time interval" spins 1 and 3 will not precess (they lie along the
(8.8)
z-axis), but spins 2 and 4 will precess at rates +a/2 and -a/2 respectively so that
p+ and p_ are related by the Boltzmann factor at t:= 2" 2 and 4 will lie along the +y-axis. An echo therefore occurs at t "" 2,. It
p+/p_ := e"tldfo/kT
is interesting to note that in this process we have transferred the net polarization
(8.9) which was initially in spins 1 and 3 10 spins 2 and 4. However, we have not
so that, if ",(hHo «kT, made an unpolarized system become polarized. Thus, we can remain confident
that the other N - 2n spins, which we have said are initially unpolarized, do nOI
P+:= ~(1 + "'(fiHo/2kT) p_ := ~(1 - "'(hHo/2kT) (8.10)
become polarized by the echo sequence.
Note that the expressions for p+ or p_ do fU)( depend on the orientation of the If one chooses a , other than Ihe one for which arl2 := '11"12, the explanation
other nucleus in the pair, because we have assumed that the only field acting on becomes more complicated. Then one needs to utilize the ideas of Sect. 7.24 in
a nucleus is Ho. This assumption should be valid if Ho d> "'(1lIr 3 , where r is the which we discuss the time development of one spin when it is coupled to another
distance between neighbors. Now, it is only the excess of nuclei N+ over those spin which is in a mixture of the spin-up and spin-down eigenstates. By choosing
N_ which can give rise to a NMR signal. To follow what happens, we need a,/2 := 'II"!2 we have made this mixture be in fact pure eigenstates.
372 373
However, a true echo means thai the refocusing condition is independent of
the strength of the source of the dephasing. Thus. if spins are dephased by an
inhomogeneous magnet. all the spins get back in phase at the echo. independent
of how far the inhomogeneity has shifled their resonance from the average field_
In our case. that means the existence of an echo should be independent of a,
h: O' hence of the PrOOuC[ aT, hence of T.
Let us therefore reexamine the situation for a more general T, using Fig. 8.4.
, We start (Fig. 8.4a) with the magnetization along the z-axis, and apply an X('K(l)
pulse to prOOuce the situation in Fig. 8.4b. A time T later the vectors I and 3
y have rotated in the % - Y plane through angles (aT(l) and (-aT(l) to produce
t: 0' l Fig.8Ac. Then the Y(1I'/2) pulse produces the situation of Fig.8Ad in which
4 , magnetization vectors 1 and 3 are in the y - z plane, making angles -aT(l and
arl2 respectively with respect to the y-axis. In Fig. 8.4e we show the projections
of the four vectors onto the x - y plane. Spins 2 and 4 lie entirely in the plane.
We define their lengths as Mo. Spins I and 3 have projections in the x - y plane
,IT: " of Mocos(aT(2.). Now, using the concepts of Sect. 7.22, we realize that each of
the four magnetization vectors is coupled to a spin which is in a mixture of up
and down states.
Since the quantum states of spins 2 and 4 are equal mixtures of up and
down. spins 1 and 3 will each break into two counter-rotating components of
Idl , equal amplitude (Fig. 8At). Spins I and 3 have z-components of -Mo sin(ar(2.)
and +Mo sin(aT(2.) respectively. Thus they have an excess of m = -!
(for spin
2 y
1) and m = +i (for spin 3). We need to calculate what fraction of their stales
I: '['
correspond to spin up and spin down.
, Denoting Ihe occupation probabilily of slate m for spin i as (mIUilm) we
have that
, (8.12)
y from nonnalization. The z-component of magnetization of spin I, M~t> is
lei 2 t : 2'[ M" = Mo (~I",I~) - Mo (- ~Ied -~) (8.13)
,
expressing the fact Ihat if the spin is entirely in i or -i. M z1 is Mo or -Mo
respectively. But
I'Ig.S.Ja·e. The fonnation of a solid echo for two pairs of spin!! (I and 2, 3 and 4) for a
particular value of the time T" at which the refocusing pulse is applied. (8) The thermal M~l = -Mosin(ar(l) (8.14a)
=
equilibrium magnetization at 1 0- showing the vector sum of the four magnetization
veclors. Spins 1 and 3 are the blue sl)ina, spins 2 and 4 are the red spins. (b) Just after so we get
the X(Tr/2) pulse (t = 0+), the blue spina (1 and 3) lie along the y-axis while the red
spins (2 and 4) can be taken to point along t.he positive Il.Ild negative ",-axC!! respeclively. (~Ied~) - (-~Ied-~) =-s;n(aT/2) (8.I'b)
(c) The magnetization vectOrs at times L = T- just before the Y(lr/2) pulse, for the
particular T such that (JT/2 = lr/2 (or or = lr). (d) The magneti:u.tion vectors at time
=
1 1+ immediately llfter the Y(lr/2) pulse. (e) The magnetization vectors at time t 2r, = Solving (8.12) and (8.13) we get
showing that there is a net magnetization IIlong the y-axi!! from spin!! 2 llnd 4. Note that
initially spill!! I and 3 carried the mllgnetization. At t = 2,., the magnetiZll.tion arises Well~) = ~[I - s;n(aT!2)]
from spins 2 and 4
(- ~Ied -~) = 1(1 + s;n(aT!2J] (8.15)
In a similar way, we get
374 375
, Fig.8.4 Caption see opposite page
W",I~) = ~(I + si,(arl2»)
(- ~I",I-~) = HI - si,(arl2)] (8.16)
lal
, Thus', as spin 2 precesses, it will break into two counter-rotating components
whose amplilUdes are detennined by the relative amounts of spin 1 in the up
and down states. That is, the component of 2 which rotates clockwise (looking
, down on the x - y plane) which comes from the down component of 1 will have
amplitude
, , ~o (1 + si,(arl2)] (8.17)
Ibl 2 , , while the counterclockwise component of 2 will have amplitude
3
, 4 , ~o [1 - sin(ar!2)] (8.18)
"-k'::..o"
)3 aT 12 M y1 = ~o cos2(aT/2) + ~o cos2(aT/2)
, 4 ,
M y2 = ~o [1 + sin(or/2)] sin(or/2)
Idl ,
_ N~O [1 - sin(ar/2)] sin(ar/2)
1 ,lTn ,
11
= ~O [sin2(or/2) + sin2(ur/2)] (8.19)
/
/
, Therefore
/
L
M y (t+2) = Mo[sin 2(or/2) + cos 2(or/2)] = Mo (8.20')
lei , By a similar argument
t~,' (8.20b)
~--Mo(os(aTl2l--- 3
1
_-'-;:,-...L__ ,
4
, •
Fig.8.4a·g. Formation of a solid ocho for two pairs of spins (I and 2, 3 and 'i) for a general
No Ho
value of the spin-spin coupling conslant u. (a) The thermal equilibrium magnetization
at t = 0- showing the vector sum of the four magnetization vectors. Spins 1 and 3 are
the blue SpillS, spills 2 and 'I are the red spins. (b) The magnetization vectors at 1=0+
If I , Igi
spin 2
immediately following the X(1r/2) pul$(!. (e) The magnetization vedors at time t = r-,
just before the Y(1r/2) pulse. (d) The magnetization vectors at time t = r+, immedi-
ately after the Y(tr/2) pulse. (e) The projections of the magnetization vectors on the
spin :l' - Y plane at time t = r+, immediately after the Y(tr/2) pulse. (f) The projection of
the magnetization components of a spin 1 on the 2: - Y plane at later times, showing
how its Illagnetization breaks into two counter-rotating components whose amplitude is
determined by the orientation of spin 2. (g) The projection of the magnetization com-
("'T---' ponents of a spin 2 on the z - !I plane at later times, showing how its magnetization
M. vector breah into two counter-rotating components whose amplitude is determined by
-f
M
(os(aTl2)
T l1-sin(aT1211
the z-components of spin 1
376 377
This result is independent of a, therefore corresponds 10 a real echo. Indeed, Now, utilizing (8.23) we have
(8.20) shows that in a powder sample (in which a would take on different values e+iolzS... e(O+)e-iaI,S, ..
for each crystal orientation if it were representing magnelic dipolar coupling),
. = (Iy + Sy)cos(ar/2) + ([ZSl + I z S;l:)2sin(ar/2) (8.29)
spins of all the crystallites would refocus at time 2r.
This exercise has been long, laborious and exhausling, but it does serve which, using (8.24), is equal to
to show how coupled precessing veclors behave in quantum mechanics. Now
we lum to a much more elegant approach to analyze the same problem: use of Re(r-)R- 1 (8.30)
the product operator method. It has the advantage of giving us a simple way of and using (8.23) gives
figuring OUI what the second pulse should be in order to produce the echo. (See
R[(ly + Sy}cos(ar/2) - (l;l: + S;l:)2 sin(ar/2)]R- 1 (8.31)
Seci. 7.26. See also [8.10, lIn.
We start with a system in thermal equilibrium and work in the rotating Comparing (8.29) with (8.31) we see that R must satisfy the relations
reference frame. The density matrix JUSt before the first pulse (t = 0-) is
I y + Sy = R(Iy + Sy)R- 1 and (8.32a)
(8.21)
where A is a constant. Following the pulse X(1r(2), the density matrix becomes (l;l:SZ + IzS;l:) = -R(lzSz + IzSz)R- 1 (8.32b)
We automatically satisfy (8.32a) if we make R be a rotation about the y~axis. To
~~=~+~ ~~
satisfy (8.32b), the rotation must be 1r/2:
The density matrix at time t = r- JUSt before the Y(ll"/2) pulse is then
(8.33)
e(r-) = e- ia1• S ... e(O+)e+iol.S.T (8.23a)
a result which follows from the relationships (Table 7.1) such as
which, utilizing Table 7.1, gives
Y(1r!2)I y y- I (7r!2) = I y
(8.23b)
Y(1r/2)I;l:y- 1(71/2) = +Iz
Now, we wish to apply a pulse which produces some rotalion R which will cause (8.34)
Y(1r/2)lzY(7rI2) "" -Iy and
an echo. The effect of R is to change e from its value at t = r- 10 a new value
at t = r+ given by (8.35)
(8.24) Before leaving the solid echo, we should remark on one more matter. In Fig. 8.3c
We want e(r+) to be such that during the next interval r, e returns to its value we show the spins 1, 2, 3, and 4 at a time ar!2 = 1r/2. We want to refocus the
at t = 0+, just after the first pulse. Thus, the condition for an echo is spins to their condition in Fig. 8.3b where 1 and 3 lie along the +y-direction.
Looking at Fig. 8.3c we note that if we simply reverse vectors 2 and 4, vectors 1
e(2r) = g(O+) but (8.25) and 3 will reverse their precession directions, and thus refocus as desired a rime
e(2r) = e- ial• S ... e(r+)e+io1• S , .. (8.26) r later along the +y-direction. All that is needed is a 7r pulse about the x-axis.
So why do we use a Y(7r/2) pulse instead of an X(7r) pulse? Clearly the idea
If the effect of R on e(r-) were equivalent to changing the sign of a during the will work if Gr!2 = 7r/2. But to give a true echo, the scheme must work for other
interval between the X pulse and R, so that values of a, hence of aT. The easiest test is to look at (8.32). An X(1r) pulse
Rg(r-)R-1 = e+ io1,S, .. e(0+)e- ia1 • S... (8.27) will leave I z and S;l: alone, but will reverse I y • S", I l , Sz. It will therefore
satisfy (8.32b), but it willllot satisfy (8.32a). Note, however, in (8.29) that when
then we would satisfy (8.25) and produce an echo, since putting (8.27) in (8.26) (ar/2) = 7r/2, the coefficient of the lenn (Iy + Sy) vanishes, so we will refocus
we would have this particular product ar. We will not, however, refocus for either a general
e(2r) = e- ioJ• S e(r+)e+ ia1 ,S,.. value of a or of r. Thus an X(7r) pulse will not produce a true echo.
= e- iol• S Re(O+)R-1e+io1,S,T
= e-iol.S, .. e+ial. 5... e(O+) e -iol. 5... e+ io1• S ...
= ,,(0+) (8.28)
378 379
The resulting magnetiUltion is shown in Fig. 8.5c. So far, everything we have Fig. 8.Sa-e. The Jeener-Broekaert method of creating dipolar order from pairs of sl)ins (I
done is similar to our first steps in discussing the echo using Fig. 8.3. Examination and 2, 3 and 1). (a) The thermal equilibrium magneti~ation at ! = 0- showing the vector
sum of the four magneti~ation vectors. Spins I and 3 are the blue spins, spins 2 and 4 the
of the four magnetization vectors in Fig. 8.5c shows that their resultant is zero. red Sl)ins. (b) The magnetization vectors at t = 0+ immediately following the X(1r/2)
We appear to have lost the magnetic order. However, that cannot be true since =
pulse. (c) The magnetization vectors at time t T- ,just befoN! the Y(1r /4) pulse, for the
we know that a Y(71/2) pulse will enable us to recover the magnetizalion. The case (JT/2 = 1r/2(aT = 1r). (d) The magnetizatiOll vedors at t = T+, immediately after
the Y(1r/4) pulse. (e) The magnetization vc<:tors a long time after! = T, The transverse
secret of the Jeeller-Broekaen pulse sequence is the recognition that the spins components of mllglletization have decayed to zero from spin-spin couplings to Qlher pairs,
in Fig. 8.5c have spin-spin order. This can be understood by analyzing what leavmg only the resultant componenls along the ~·direction which are 1/V2 shorter than
the vectors of (d)
happens if one applies a Y(1I"/4) pulse, producing the result of Fig. 8.5d. If now
380 381
we have in our ensemble of spins a weak coupling of one pair with another, there Indeed, because I~S~ is. apart from a conSlanl, the dipolar energy, we can see
will be a slight spread to the precession frequencies ± aI2. Thus, at a much later that the part of /? proportional to I~S~ is a measure of the dipolar order.
time. the transverse components of the magnetization will decay, leading to the If the dipolar system now comains some order, one should be able to "read"
situation of Fig. 8.Se. The remarkable poim about this state is thai every spin has the order by convening it from dipolar order to ''Zeeman'' order, i.e. to net
its neighoor pointing in the opposite direction. This is to be contrasted 10 the magnetization. Juner and Broekaert do this by applying yet one more Y(.,../4)
condition al t = 0-. before the first pulse, when half the spins had a neighbor pulse. which produces an echo a time tt later given by il ;; T. It is not surprising
pointing parallel, half had their neighbor pointing anti parallel. Thus, the dipolar that an echo results since the Jeener-BroekaeIt sequence then of
engergy haI~S;, which was initially zero, now takes on a nonzero value as in x(>m ... r ... Y(r/4) ... T ... Y(r/4) ... ', (8.43)
an adiabalic demagnetization experiment.
We employ the operator approach using a derivation due to Wang et aI. can be seen, in the limit T _ 0, to be
[7.73]. The density matrix at time T after the X(1f!l) pulse but before any other X(rm ... r y(r/4)Y(>/4) ',
pulse is given by (8.23b)
;; X(1I"fl) T ••• Y(1I"/2) it (8.44)
(;'(,) ;; A[(Iy + Sy)cos(a,fl) - (1zS~ + I~Sz)2 sin(aTfl)] (8.37)
which is just the Powles-Mansfield dipolar echo.
The pulse Y(1I"/4) then rOtates the spin veClOrs 10 give To show that this happens in the general case, we take /? just before the
second pulse to be given by (8.39)
e{r+) = AY(>/4)(ly + 5 y)Y-' (>/4)
- AY(7r/4)2(IzS~ + I~Sz)sin(ar/2)y-t(7r/4) e{r + T-);; -2AI~S~ sin(ar/2) . (8.45)
<1{(0-» = Tr{rlaI~S~e{O-)}
This signal is clearly maximized if we choose r and tl equal to each other, and
;; Tr{haI~S~A(I~ + S~)} both of such duration thai
= 0 . (8.42)
383
382
arfl ::: attfl = 7C(2 • (8.50) Then
Comparing this signal with that of the free induction decay following an X(7C(2) 'H.::: --yhhoI~ - -yhH) I~ + L Bij(3I~iI~j -1; .lj) (8.54b)
pulse, i>j
<r+ + S+> = T,{(r+ + S+)AUv + Sv» We now define an effective field in the rotating frame
= iA n{(Iv + Sv)2} (8.51) Herr:::iH, +kh o (8.55)
we see the maximum signal is one half that of a pure X(1rfl) pulse. If there is We shall concern ourselves with cases in which Herr is much larger than the local
a distribution of values of a, there is still a signal at tl ::: r since fields. In Ihal c~se, it is appropriate to quantize the spins along the effective field.
Defining this direction to be the Z-direction, we make a coordinate transfonnation
sin 2(ar!2) > 0 (8.52)
(Fig. 8.6) for all spins, getting
Thus, we can create dipolar order by applying X(7C!2) ... T ... Y('ll14) with T
Iz ::: I",cos8 + I",sin8
chosen such that a T ~ 7Cfl. and wecan inspect the dipolar order by later applying
a Y(7C/4) pulse to produce an echo a time T later. The inspection pulse can be used Ix :::I~cos8-Iysin8
to study the decay of the dipolar order resulting from either conventional spin- Iy = Iv • (8.56.)
lauice relaxation mechanisms (e.g. coupli.ng to conduction electrons in a metal),
giving the inverse transfonnations for an individual spin j
or from motional effects which modulale the strength of the dipolar coupling.
We have treated the problem for coupled pairs. JCClicr and Brockacrt treat Izj::: I Zj cos 8 - I Xj sin 9
the general dipolar case in their classic paper [8.12]. IZj::: Ixjcos8+Izjsin8
Iyj::: IYj (8.56b)
Subsriruting into 'H., we gel
8.7 The Magic Angle in the Rotating Frame-
The Lee-Goldburg Experiment +2
'H.::: -"fhHcrrIZ + L A/l1(8)'H/l1 where (8.57)
/1,1:::-2
Another imponant concept concerning dipolar order is connected with the fa-
mous Lee·Goldburg experiment (8.13] 10 which we now turn. Suppose one has 'H.o::: L Bij(3I~iIzj - Ii' Ij) (8.58)
a number of spins all with the same precession frequency, coupled together by i>j
dipolar coupling. We keep just the secular tenos of the dipolar coupling, giving
a Hamiltonian in the rotating frame of
'H.::: - "fhhoI~
- "fllH II~
, z
"f2h 2
+ :L -,-(1 - 3 cos28ij)(3I~i l~j - Ii' Ij) (8.53.)
i>j T ij
----
I
where as usual I
I
,
h,
I
I~::: r:,I~j. etc. HrH !
r I
ho::: Ho -wl-y (8.53b) H, I ,
It is convenient to collect all the radial and angular lenns of (8.53a) in a simple
symbol by defining
x Fig. 8.6. .The erre<:t;,·e Held in the rotAtin« rl1lme, showing the AXes
X,Z WIth respect to ~,z
I "f2h2
B;j=zT(I-3cos2 8ij) (8.54a)
OJ
384 38S
1-£+1 = L 3BijUt IZj + IZirt} of magnetization perpendicular to Herr as usually occurs when there is dipolar
i>j coupling.
1-£+2 = L 3BijUt It} , To make a precise test of this concept, they did pulse experiments in which
i>j they observed the decay of the magnetization with time following a sudden tum-
1LI = (1-£+1)", 1-£-2 = (1-£+2r and (8.59) on of H t at a frequency w somewhat off resonance. They varied the angle 8 and
the strength of the effective field. To measure the effective dipolar strength, they
-'0(9) = !(3cos 28-1) determined the second moment from the transfonn of the decay curves. They
studied the F ID resonance of CaF2. Figure 8.7a shows their measurements of the
-' ± 1(8} = -! sin 9cos 8 (8.60) square root of the second moment versus (3cos 28- l)n.. AI the magic angle, the
-'±2(8) = -i
sin 28 second moment should vanish. It does not quite do so owing to the nonsecular
terms involving 'H± I and H±2. The effect of these tenns should vanish in the
The 1-£/I.1's satisfy a commutation relation limit of infinite Herr. In Fig.8.7b we show their measurements of the faJl-off of
second moment, nonnalized to its value in the lab frame, versus llw; where
lIz, 1iM]- M1iM (8.61)
(8.66)
as is easy to show by explicit calculations of matrix elements. From (8.61), or by
examination of the explicit form of the 'HM's, we see that only 1-£0 commutes The failure of the dashed theoretical curve to intercept the origin arises because
with the Zeeman interactions of the effective field. In the limit of large effective of slight inhomogeneity of H1. BarntUJl and Low [8.14] made an extensive study
field, we can then as a first approximation drop all terms but the term involving of the systems in which HI was exactly at resonance. In this case
1-£0. We get then a truncated Hamiltonian
3cos29 - I
1-£ = -yhHefflz + !(3 cos 28 - I) L
Bij(3IziIZj - Ii' Ij}. (8.62) 2
(8.67)
i>j
They did experiments in which they turned on HI for a time T, then observed
We can compare this with the secular part of the Hamiltonian of the spins in the the free induction decay a time t after H 1 is turned off. They solved the problem
lab frame,1tIl\b' in the absence of an alternating field, H,: of an interacting pair exactly. For an HI along the x-axis they found
'H'l"b = -"'('iHolz + L Bij(3IziI:j - Ii' Ij)
My(t) = M(O) ~ sin(flT)cos [ (3:,: 2 ) (t + T!2)] (8.68)
i> j
:= 1-£Zeemlln + 'H'dipolllr (8.63)
where
Equations (8.62) and (8.63) are identical in form, except in the rotating frame
the dipolar term has been multiplied by the factor n' -_ (3BI2)'
" - - +w,, (8.69)
4~
3,os 29 _ I
(8.64) In the limit of large HI, this expression agrees with the result of the truncated
2 Hamiltonian:
['This is juSt the tenn -'0(8)]. 8 is delennined by the relative size of ho, the
amount one is off resonance, with HI, the strength of the rotating field. Lee and My(t) = M(O) sin(wl T) cos [ (3:,:2 ) (t + T!2)] (8.70)
Goldburg noted that if they chose HI and ho properly, they could make
A striking feature of (8.68) is Ihal the oscillation after the tum-off of HI is
cos 28=! (8.65) identical to what it would be following an X(7r/2) pulse except for (1) a slighl
in which case the dipolar term in (8.62) vanished. In this manner they could amplitude correction and (2) a change of T in the apparent zero of time t. In a
effectively eliminate the dipolar broadening. An effective field for which (8.65) beautiful set of experiments, Barnaaf and Lowe showed that for CaS04 ·2H20,
is salisfied is said to be at the magic allgle. CaF2, and ice the first zero crossing of the free induction decay moves 10 later
We can rephrase their result by saying that if the effective field is at the times as T is increased. The delay in zero crossing is T/2. for values of T which
magic angle, a spin will precess in the rotating frame solely under the influence are up to about half of the nonnal free induction decay zero crossing.
of Neff, without suffering a dephasing and consequent decay of the components
386 387
I .. ~ 0.6
>J
A
(oF,IU doptlll
H.1I111lJ
H, : 5.1!O.2 Gauss
pulse a time T later as suddenly lransfonning the Hamiltonian to be the negative
of the actual Hamiltonian. Then, over the next time interval T, the time develop-
ment of the magnetization unwinds, returning the magnetization to its value just
0.'
." ~
after the initial pulse.
v 0.' /" In discussing spin temperature (Chap. 6), we talked about irreversible pro-
"-
s A 0.' /: cesses connected with the complexity of a system of many coupled dipoles. We
now tum 10 a remarkable discovery which shows how one can run backwards
'"
."
.,
~ 0.2
/'
the dephasing produced by dipolar coupling, thus showing that it is possible,
even after the free induction decay is over, to recover the initial magnetization.
0.1
This experiment shows that the spin temperature approach does not in all cases
accurately describe the evolution of a spin system. In the process an echo is
~OO~5;:;-0::;.';:;-0:::J~-0:::.2~-0:;.1---'~0'--;0:;.1--:0::.2:-:0::J:-:0~.'---'0:':5,---J0.' fanned. This sort of echo has become known as a magic echo. The first intima-
lal ]cos'e~l tion that one could refocus dephasing arising from dipolar coupling was noticed
2 by Rhim and Kessemeier [8.15,16], who discovered the effect experimentally,
and showed theoretically by an approximate method Ihat an echo was fanned in
• .10. 1
/
/ • which the loss of signal owing to dipolar dephasing was recovered. Following
// this work, done at the University of North Carolina, Rhim went to MlT where, in
';
, /
collaboration wilh Pines and Waugh, he extended and perfected the experimental
techniques and the theoretical analysis [8.17].
-v.,
A
~
/
/'// In order to refocus spins which are defocused by dipolar coupling, one needs
10 be able [0 reverse the sign of the dipolar Hamiltonian. Then, as we have seen,
, 2 / we can effectively run the system backwards in time to undo Ihe dephasing. The
-.,
~
A
//
/
•/0 (oF1lU doped I
Ho lll110l
trick is closely related to Ihe Lee-Goldburg experiment. The critical equation is
(8.62), the truncated Hamiltonian in the rotating frame:
v / /3(05
2
&.,/ (3cos 26-1)
/ 2 ~ 0.01 11. = -",lill
,ell wIz + 2~.
,,,
'"' B·1)·(3[z1z·
I J) -1·I - I·)
)
(8.62)
0:'0--;:-;-----;;7'-----:':c--.J In this equation, () is the angle between the static field Ho and the effective field,
0.1 0.2 0.3 0.4
Heff, and the Z-direction is the direction of the effeclive field. If Heff is nearly
Ibl Il&.ll'lwli....: parallel to Ho,cos8 = 1 and the angular factor, (3cos 2() - l)Il, in front of the
F1g.8.7a,b. The ~ Ilnd ?oldburs experimental results for F" relIOnance in CaF, sins1e
dipolar tenn is + I. However, if () = trn, as when H t is exactly at resonance, the
ct)'st.I~. (a~ The normllh~ second moment. of the Fit resonance for flo par.. llel to the angular factor becomes
~lll J dued,.on, as a function of (3 COlI" - 1)/2, where B specifiCil the orientation of H
III the rotatmg rr.~ (see Fig. 8.6). (b) The normalized second moment iLS II function c:r 3cos 2(} _ 1 I
(8.71)
the .second moment III the lab frame < (4....') > LAn divided by",,' where .... = 11 2 =-1:
110 1$ parallel to the crystal (1101 direetion e' e " elf-
Thus, it has a negative sign. The magic echo makes use of Ihis negative sign to
unwind the dipolar dephasing.
Let us then rewrite (8.62), the truncated Hamiltonian, exactly at resonance,
8.8 Magic Echoes using the coordinates x, y, z in which HI lies along the x-axis:
1
1i = -"'thHII~ - 2" L BjJ.:(3J~jI~k - I j .IJ,;) (8.72)
In Secrs. 2.10.and 5.~ we saw thai one way of understanding how a spin echo ;>k
comes ~bo~t IS to view me first pulse as initialing a time development of the In addition to Ihese tenns, there is the nonsecular tenn
magneUZ3tJon under the influence of the Hamiltonian, and to view the second
388 389
1{nonsec = ~ L 3D jk (lt rt + I j- Ii:) , (8.73) This expression is still not quite what we want since it still has two different
j> k dipolar Hamiltonians.
which we are for the moment neglecting.
How can we utilize the negative sign to unwind dipolar dephasing? For the
\ However, examination of (8.76) shows that 1{u; and 1i zz differ simply by
a cOQrdinate rotation of z into x or x into z. Such a transfoooation is produced
moment, consider an experiment in which we produce some transverse magne- by a rotation about the y-axis. In fact
tization [for example, by a Y(nl2) pUlse]. Let it dephase under the action of
the dipolar system, for a time T, then turn on H I to produce a negative dipolar (8.78)
coupling. Will this refocus the dipolar dephasing? If the density matrix (in the Therefore, we add two pulses, a Y(1T/2) and a y-I (1T!2), where the latter is
rotating frame for our whole discussion) is initially (t = 0-) given by simply a Y( -1T/2) before and after the time interval 2r' giving
where A is some constant which we set equal to I for convenience, the pulse
sequence would produce at a time t a density matrix
(8.74)
I
e{T+2T ) = Y(1T!2)exp [_ *(_ 1{;z2T')] y- I (1T/2)
e{r+2T
I
) =ex p [ -* (r!lHII~-~1{XZ)TI] y- 1 (n/2) ... r,/!]
= ex p ( -*(-1t
... T I -
I
»)
Ifl
... Y(1T!2)
Rhim, Pines, and Waugh label such a sequence a "burst". In fact, they argue that
(8.82)
Table 8.1. Listing of the normali:ted spheriul hllrmonks (or f = 0, I, 2 and the uoorlllalized
8.9 Magic Angle Spinning forms
where (J' is the angle the molecular rotation axis makes with the static magnetic Y2,2(B,4» = a sin' Be,l.
field, and "fjk is the angle made by the internuclear vector rij with the rotation Y2,1 (B, 4» = --/* sin 8 COlI Bel. .:(ot + iy) =z.: + iy.:
axis.
This expression suggested independently to Lowe [8.21] and to Andrew et Y2,o(8, 4» = a(3 cos 2 8 - 1)
al. [8.22] that one could produce the rotation artificially by turning the entire Y2._1(B,4» = V*sin B cos Be-I. .:(z - iy) =:u - iy.:
sample. In that case, the angle fJ' would be the same for all pairs of nuclei
throughout the sample. Then if one chose fI to satisfy the condition Y2._ 2(8,q,) = j"ji; sin 2 Be- 21 • (ot _ iy? = :>:2 _ y2 - 2izy
I - 3 co"9' : 0 (8.83)
392 393
Suppose now we consider two coordinate systems x, y, z and x', y, Zl,
with corresponding angles 0, tP and 8', ¢I defined by (8.84). The Yim(O, tP)'s form
a complete set for a given I, which means that we can always express }'i fi (8', tP')
'. " '. "
in tenns of the Yim(O, ¢)'s: '. ". /'
'.
}'jfi(e', tP') = Lafim l'im(O, tP) and (8.86a)
I
/ /
/
Ibl f f
/
m lal //
f
Y,m(B,~) = L:em"Y,,(B',?,) (8.86b) y/ l /
/
, /
"
It is then slTaightforward, utilizing the onhogonality propenies of the l'im's when
integrated over 4lf solid angle, to show that
(8.87)
'. " '. "
Noting the forms of the Yi", 's in rectangular coordinates, we realize that the '. '. -- Yo
y,
coefficients all'" or Cnll' could be found by explicit substitution of the relationship I I
1(1 Idl
between the primed and the unprimed coordinates:
x' =b II X+bI2!J+b 13 Z )//
f
I
f
J../
I
/ '.
'. '."
/
the z-axis. One utilizes (8.86), the fact that the 5-function is axially symmetric
about the z-axis, and equates the expansion in the lI",(D.?)'s to that in the From the figure we see
Yim(8', tP')'s] . °OR=OO, (8.90,)
We now wish to consider a spinning sample. To do this. we wish to define
some axes (Fig. 8.8). We define first the laboratory z-axis. ZL, along the static Oos = nOR = 00 ' ¢os = tPOIt - nt = 7rn - nt (8.9Ob)
field, and an axis zit about which we will eventually rolate the sample. The angle
between ZL and Zit is 00 (Fig.8.8a). We can then, without loss in generality, We now consider the interaction between a pair of spins j, k which we have
define the laboratory axis XL to be perpendicular to both zL and Zit (Fig. 8.8b). previou.sl~ w~tten as proportional to 3cOS 2 0jk - I (3.7). We now, however.
We define XIt to be coincident with x I, (Fig.8.8c). The axes xn.. !JR, and Zit must dlstlOgUish between coordinate systems. hence we replace (Jjk by OLjk>
are fixed in the laboratory frame (Fig. 8.8d). We then define axes xs, YS, and zs writing the coupling
which are fixed in the sample with zs coincident with Zit at all times, and Xs and 2
3COS 0Ljk - 1 = JI67r/5Y20«(JLjk, tPLjk) (8.91)
YS coincidem with Xlt and Yit respectively at time t = 0, but making an angle nt
at" later times (Fig. 8.8e). n is the angular velocity of sample spinning. Utilizing Then, we can utilize (8.89) to get
these definitions, we can specify the orientation of Ho in the coordinate system
Y20(O, O)Y20(OLjk, ¢I,jk) = LY2"m(00s, ¢OS)Y2m(OSjk> tPsjk) (8.92)
Zit' YIt, Zit (at angle 00R, 4toRl and in the system zs' !IS' zs (at angle Dos, ¢os)· m
395
where 0Sjk, ¢Sjk give Ihe orienlation of the internuclear vector Tk in Ihe coor- We turn now to a discussion of the effect of spinning on a pulse experiment
dinate system xS' Ys. zs which is fixed in the sample and where Oos and ¢os in which we observe the free induction decay of a spinning sample following a
are given by (8.90). -;r/2 pulse. Then we might express the dipolllr Hamiltonian, 11.d as
Substituling the explicit expressions for the Ylm'S from Table 8.1, we get
...,2h2
3 COS 20Ljk _ 1 = (3COS 2~OS - 1) (3cOS20Sjk _ I)
'H<J = E (l - 3 COS 20Sjk)-.-3-(3Izjlzk - I j ·Id
j <k 1 jk
+ 6 sin OOS cos Oos sin 0Sjk cos 0Sjk cos (¢Sjk - ¢os) , I: [FOj' + F1j'(') + F'j,(.)jGj,(Ij,I,) (8.98a)
j>k
+ ~ sin200s sin 20Sjk cos 2(¢Sjk - ¢os) (8.93) where 2 ·2
Since ¢os obeys (8.90b),
o"
Gjk(Tj , h)::= -,-(Ii' h - 3Izj l zk ) (8.98b)
r jk
10s ' ,/2 - fit (8.90b) is a function only of Ihe XL, YL, and ZL components of the spin operators Ij
We see that the second and third tenns on the right vary as f}t and 2f}t respec- and Tk' and
tively, whereas the first tenn on the right is independent of time. Thus if one
(3COS200S - 1)(3 '"
can average the time-dependent tenns, we recover the result of (3.61) with the FOjk = 2 COSVSik- l )
notation
Fljdt) = 6 sin Oos cos 80S sin 0Sjk cos 0Sik COS(¢Sjk - ¢os) (8.99)
~k=~k ~~ 2
F 2jk (t) = ~ sin200s sin 0Sjk coS 2(¢Sjk - ¢OS)
For steady-state experiments (i.e. if one talks about the "frequency" do-
main, perhaps with data which are the Fourier transform of data from a pulse Since ¢os = -;r/2 - f}t from (8.90b), we see that FOjk is independent of time, but
experiment), the tenns involving f}t and 2f}t give rise to frequency modulation F1jk and F2j~' oscillate in lime at f} and 2f} respectively. Explicil examination of
at frequencies f} and 2f}. Such a modulation gives rise to multiple sidebands the spin functions G jk shows Ihal two different funClions which have one spin in
spaced in frequency by f} aboul the ncnnal fequency wo. For frequencies f} common, for example the jth, such as Gjk and Gjl (f of- k), do not commute. This
much bigger than the dipolar line width ~Wd' these sidebands will be displaced makes it difficult in general to determine the time development. Such mailers are
by an amount large compared to the width associated with the time-independem discussed in Appendix J dealing with time-dependent Hamiltonians. Here, we are
tenn. If we neglect the sidebands to focus our attention solely on the main tran- dealing with Ihe case of a Hamilitonian composed of a sum of noncom muting
sition, it will appear narrower. Indeed, since the time-independent tenn of the parts each of which has a lime-dependent coefficient. This is Case II of Appendix
dipolar coupling for the spinning case is simply reduced from its value for the 1. However, the lime dependence is periodic, so we can consider what happens
nonspinning case by the faclOr Cs where over a single period.
Suppose that the period T(::= 21f/f}) is sufficently shan that not much hap-
Cs = 3 CO!;200S - I (8.95) pens during one cycle. Then, as we show in Appendix I, we can relate the wave
2 function at the start of a period to its value at the end of Ihat period !>y the
the/orm of the absorption line should be identical in the two cases except for a unitary operator U(T) given by
scaling of the frequency. Thus, if i(w - wo) is the nonnalized intensity function
without spinning, Ihe normalized function with spinning is(w - wo) becomes ,p(T)' U(T),p(O), exp ( - i7tT) ,p(O) where (8.l00)
1 l'
fs(w - wo) , C/(Cs (w - wo))
At the magic angle, the line theoretically becomes a O-function. In practice, it is
(8.96)
-11. T'fo
= 'H(r)dr (8.lOl)
for large 'Y nuclei, since spinning speeds are typically a few kilohertz, comparable o 0
398 399
Recalling (4.198), we get that, in lhe presence of anisotropic shifts, Y2,0(9 0p, ,pop) = L Coms(P, S)Y2ms (90S• tPos)
'H. = - 'Y;,Ho[1 + (/( _ U)+(/(LO _ 0LO) (3COS:O- I) (8.108)
no,
+ (I(TR; UTn) sin 2ocos2 tP] l z . (8.104) Y2 _2(9 0p , tPoP)
, = L C-2Jl1s(P, S)Y2ms(Oos. tPos)
no,
We consider lz to be the sum of the z components l z k of N noninleracting spins At later limes. owing to the rotation, fos will change with time. bUI since zs is
(k = I toN). We shall analyze the effect of rotation on this Hamiltonian.
the rotation axis, 90S does not change, so
The angles 0 and tP of (8.104) specify the orientation of Ho in the principal
axis system of the shift tensors. We shall need to keep track of various reference
frames. Therefore, we shall define !he spherical angles Oop and tPop to designale
80 5(') =80 , <Po5(t) =.f2 - ill . (8.109)
the orientation of fIo in the principal axis system xp, yp, zp. We keep the same Therefore the mS = ±2 lenns of (8.108) will oscillate at ±2n, the ms = ± 1
definitions for the coordinate systems XL, YL, zL (Ho lies along ZL), %R, YR, terms will oscillate al ± n.
and the ms :: 0 tenns will be time independent.
Zn, and %s, YS, Zs (Fig. 8.8). Collecting lerms we gel
We have then that at t = 0, ZL is at Oop, tPop or expressed in R or S at
00R, 4>oR or OOS. 4>oS. which in tum are described by (8.9Oa). Therefore. we can 'H.= -'YhHo1z[I+(J(-U>
wrile at t = 0
+ ( 1(1.0 - 0'1.0) "L...COms(P,S)Y2rns(8o.'ll"f2 - fI)
t
'H.= -'YhHO[1 +(l( -U)+ (J(LO ;ULO) (3 cos 290p -I) 2N2,0 ms
We can now express the angular functions in tenns of the S-system, using (8.86), Let us first consider just the time-independent term (mS = 0). It will give a
contribution to the Hamiltonian of
400 401
1tlimeindcp= --rhHQTZ{ 1 +(J( -01+ [ (J(L~N~,:I.O) CO,O(P,S) erties of the Y2ms 's (Table 8.1) we see that 1£ is a sum of tenns which are either
independent of time. or vary as exp(int), exp(-inl). exp(2int). exp(-2int).
Y;o(a,p) Y;,(a,fIj Y;,_,(a,p) and recalling that T = 27f/n. we see that when
ao,o = Y2~O(O.O) • aO,2 = yi,o(O,O) °0,-2:: Y2.,O(O,O) (8.115) t=nT • n=O,I,2 ... (8.120.)
where (a, fJ) gives the orientation in spherical coordinates of the spinning axis
%s in the Xp. yp. %p (principal axis) coordinate system. The result is
'!>(I) = exp ( - i'YHo1,(1 + [( - a)') ,!>(O) (8.12Ob)
just the result we would have if the time-dependent terms were missing. Note
1£timeindep= ~'YfIHo[, -
{1+(I(-'i7)+ [(I(LO-O'LO) (3Ca"a-l)
2 that this result says that at the magic angle the time development over integral
multiples of the rotation period is independent of the orientation of the crystal
+ (}
' _O'TR)(,inaCOS2
\TR 2
P)] (3CO"Bo-I)}
2 .
axes relative to Ho. Thus, if we applied a 7f{l. pulse to a spinning sample. the free
induction decay at times given by (8.120a) would act as though there were no
(8.116) anisotropy to the shift tensor. If one had a powder sample which was not spinning,
(This expression can be checked in the limiting case that 80 = O. hence the spin the spread in precession frequency arising from shift anisotropy. 6.wshif~' would
axis lies along Ho. Then the spinning has no effect and the time-independent cause the free induction to decay in a time 1/6.wshif~' If, however, the sample is
tenn is the entire Hamiltonian. Then 0' = 0 or 0' = rr(2, {J = 0 or 1r(2 correspond spinning we can see from (8.120b) that the signals are rephased at times given
to.flo lying along the three principal directions. The result is that for the three by (8.120a). thereby producing a string of echoes.
cases we get shifts respectively 1(3 - 0'3. 1<1 - 0' 1, or J( 2 - 0'2, where 3, I, and Notice that we have not placed any requirement as yet on how fast the
2 stand for the zp. x p , and Up directions.) spinning must be, in contrast to the previous siwalion concerning narrowing of
The other very interesting point to consider is the time dependence when dipolar line broadening. For the dipolar case we compared the rigid lauice line
one is spinning at the magic angle. Examination of (8.110) shows that the time width ~d with n. What frequency should we compare with w in the present
dependence of 1£ comes from the factors Y2ms (80, 1r(2 - nt). Ulilizing the prop- case? Dimensionally, there is only one relevant frequency. the total excursion
402 403
of the precession frequency over the spinning cycle. Note that all our equations I'AQT- OkHz Proton decoupled p3l spectra at 119.05
.'ig.8.IO.
so far imply there is a single crystal being rotated since there is a well-defined MHz of barium diethyJ phosphate spinning
aL the magic angle. The figure illustrates
orientation of zs in the xp, yp, zp coordinate system. However, for a powder, • how, in the limit of slow spinning, the
spinning sidebands reproduce the shape or the
all possible orientations occur. For the rest of our discussion we will focus on
a powder. Then, the maximum frequency excursion one can have in a rotation Ilonspinning spectrum [8.28]
cycle is less than
1'Ho[(J(zI'Zp - qzpzl') - [(I{:cI':C1' - q:cp:cp)] == .6wshifL (8.121)
where we are defining the zp and xp axes as having the maximum and minimum "OOT - 0.94 kHz
shifts respectively.
So we ask, what happens when we vary the relative sizes of fl and .6wshift.? b
Suppose, first, that the powder sample is not spinning. Then, as we have noted,
following a rr/2 pulse, the transverse magnetization will decay to zero in a time
:::::: l16.wshift. The Fourier transform of this free induction decay would give us
the line shape. If we then applied a 11" pulse at time tll" we could refocus the
magnetization into an echo. If t1l':» 1/.0.wshifL' the echo and the free induction
decay would be well separated in time. Indeed, we could apply a string of such
11" pulses (a Carr-Purcell sequence) pnxlucing a string of echoes. The Fourier
transform of anyone echo will give us the powder line shape in frequency
space. If, instead, we took a Fourier transform of the sIring of echoes, we would
now have introduced a periodicity. The resultant J1ansform differs from that
of a single echo in the same way that a Fourier series differs from a Fourier
integral. The transform of the string of echoes would consist of 5-functions
spaced apart in frequency by the angular frequency l/t rep where t rep is the
time between successive rr pulses. The Fourier transform of a single echo would I'OOT" 2.92 kHz
be the envelope function of the spikes. If now one shortened t rep , one would
get to a point where t rep :::::: lI.6wshifL' Then the echo would not have decayed
completely to zero when the 11" pulse is applied, so that over a single period d
one no longer would have the complete shape of a single free induction decay.
The Fourier transform of the sequence of echoes would still have a center line
plus sidebands spaced in angular frequency by 1/t rep , but the envelope of these
, ,
16 a 0 - a -16
sidebands would not be the powdeI line shape (the rransform of a single free Freq (kHz)
induction decay).
Stejskal et al. [8.27] realized that the same situation would arise from magic
angle spinning. When n« .6wshifL, the NMR signal following an initial 11"/2 Once n
becomes comparable to or larger than .6wshifL> the satellites occur
pulse decays rapidly compared to the period of rotation, hence is essentially outside the envelope of the nonspinning line. As can be seen fIOm (8.117), the
identical to what it would be with n '" O. However, after one full rotation there coefficients A\(P,S) and A2(P,S) ale multiplied by lin, hence become progres-
is an echo, as described by (8.120). Indeed, a train of "spinning echoes" will be sively smaller than the lalger n. This is part of a general theorem discussed by
formed, analogous to a Carr~Purcell train. The Fourier transform of any echo will Andrew et al. [8.22] that the second moment is invariant under rotation. Thus,
give the n '" 0 powder line shape. The Fourier transform of the train of echoes the intensity of lines spaced by multiples of must fall off as l/n 2 in the limit
n
will give o-function spikes spaced apart in angular frequency by n, with an of large n.
intensity envelope versus frequency identical to the nonspinning powder patlern. When there are many spinning sidebands, how can one tell the central line?
This situation is illustrated beautifully by Fig. 8.10 showing the data of Herzfeld The easiest technique is to change n. One line will be undisplaced - the cenrral
et al. for a p13 compound [8.28,29]. line.
404 405
The beauty of the slow spinning regime of Stejskal, Schaefer. and McKay
3(M/-rIS)r/5 M[
is that it gives one simultaneously a high precision determination of the isotropic H(
1 r[5 ) = S - -3- (8.122)
average shift, as well as a picture of the associated powder paHem. It also removes r lS r lS
the difficulty of spinning at a rate faster than ~Wlihirt., a genuine difficulty in the We consider M[ to lie along the k-direction, and consider three cases in which
strong fields of superconducting magnets when one has large shift anisotTOpies. S lies at a distance a respectively along the X-, y-, and z·axes (see Fig.8.ll).
A common use of the slow spinning regime is to slUdy low abundance Application of the formula shows that when S is on the z·axis it experiences a
nuclei (e.g. C 13 ). so that there is no like-nuclei dipolar coupling with which to field
contend, combined with standard decoupling to any high abundance nuclei like 2M,
HI whose mutual spin-flips would otherwise broaden the spinning side bands. HI(O, 0, a) • - 3 - (8.123)
a
In some instances it is desirable to remove sidebands in order to simplify the
whereas when S is on the x· or y-axes
spectra. This can be done by applying 1( pulses synchronously with the rotation
as shown by Dixon [8.30) and by Raleigh et al. [8.31].
HI(a,O,O)' HI(O,a,O)' - M.,I (8.124)
a
If, for some reason, S were to jump rapidly among the three positions, spending
8.10 The Relation of Spin.F1ip Narrowing to equaJ limes in each on the average, it would experience a time-averaged field
Motional Narrowing < H/ > given by
<H[>
The strong spin-spin coupling characteristic of solids was essential for achiev-
ing the high sensitivity of double resonance. Its disadvantage is that it produces
line broadening, and thus may obscure details of the resonance line such as the
existence of anisotropic couplings (chemical shift, Knight shift, or quadrupole (8.125)
interactions). Within the past few years, a varielY of clever techniques have been
introduced which utilize strong rf pulses to eliminate much of the dipolar broad-
, ,..---... , ,
ening. We now lum to the principles on which the ideas are based. The highly
ingenious concepts owe their development to several groups. The pioneering ,, ,,
experimenlS were performed by two groups, one headed by John Waugh [8.4], / ,
Ihe other headed by Peter MallSfitld [8.S]. Subsequently Robert Vaughan [8.6] , s
(,j
and his colleagues have contributed imponantly, as have many of the scientists , x
[8.32,33] who worked with Waugh, Mansfield or Vaughan.
The essence of these schemes is to apply a repetitive sel of rf pulses [8.34-
39) which produce large spin rotations and which, by a process akin to motional
narrowing, cause the dipolar coupling to average to zero. Each cycle consists of
a small number of pulses (8 pulses per cycle is common). To achieve the effect, s
Ihere need to be many cycles within the normal dephasing time of Ihe rigid lattice ,
line width. Since the rf pulses must produce large spin rotations (90" pulses are I "F---- (b)
typically used), and since the rotations should occur within a small fraction of
a cycle of pulses, one needs H ,'s which are large compared to the rigid lauice s
Ijne breadth.
\, ,I
, ,,
We have remarked on the similarity between the multiple pulse schemes and
motional narrowing. We stan by making the analogy explicit. For convenience
we shall name the multiple pulse schemes "spin-flip narrowing". Ag. 8.lla-e. The magnetic field or [ III S for lhree localions
Consider two spins, I and S. Let 1" [5 be the vector from spin I to S. The of S all the nme distance, a, from ,pin I. The dashed line
magnetic field H[ at spin S due to I is then IF---- (c) indicates II. magnetic line of force
406
407
This result is the essence of motional narrowing. Though for our case we picked a quick rotation of both spins to the side position of Fig. 8.12b. Wait another
only discrete locations for spin S, the same result is found for continuously time 'T.
variable sites for which the strength of the interaction goes with angular position What is the average magnetic energy < Em > over the interval 3'T? Using
81 s of rlS with respect to the z-axis as 3cos 2 81S - I. This averages to zero (8.126) and (8.127) we get
over a sphere.
Suppose now we consider a variation in the above picture. Suppose we - .!...(-2MI MS'T MIMs'T ll'hMST)_o
< E m> - 3T a3 + a3 + a3 - . (8.128)
position S at (0,0, a), Le. on the z-axis, and consider the orientation of both M I
and Ms (Fig. 8.12). That is, over such a cycle in which, at stated times, we apply selected 7fn.
In tenns of the orientation of the vector r I S with respect to MI' it is useful rotations to both spins, we can make the magnetic energy average to zero. Thus
to introduce some names. Referring to Fig. 8.11, we denote the (c) configuration by spin-flips we can cause the dipolar coupling to vanish just as we could for
as the on-axis position of S, we denote the position of S in (a) and (b) as the side position jumps in Fig. 8.11. This is the principle of spin-flip line narrowing. We
position. Note that for the on-axis position HIS is parallel to MI, whereas for make the dipolar energy vanish by flipping the spins among selected on-axis and
the side position HIS is anti-parallel to MI and half of its on-axis magnitude. side positions, spending twice as long in the side positions as in the on-axis ones.
Referring now to Fig. 8.12, we see that in (a) and (b) S occupies a side All of the complicated pulse cycles are based on exactly this principle.
position, whereas in (c) it occupies an on-axis position. Moreover, since we
have taken M , and Ms parallel in all parts, for configurations (a) and (b) the
magnetic energy E llIag = -Ms' H,s is: 8.11 The Formal Description of Spin-Flip Narrowing
(8.126)
From the previous section we see that it is possible to cause the dipolar in-
whereas for (c) teraction 10 average 10 zero if we cause the spins to flip between the "on-axis"
-2NhMs and "side" arrangements. For the conventional motional narrowing, narrowing
Emag)e = -Ms' HIS = a3 (8.127) occurs when the correlation time Tc and the rigid lauice line breadth (in fre-
Suppose, then, we began with the configuration of Fig. 8.12c, the on-axis quency) DwRL satisfy
arrangement with MI and Ms parallel. After time 'T let us quickly rotate M I TcDwRL $ I (8.129)
and Ms by 7fn. about the y-axis to the side position of Fig. 8.12a. After a 'T give
Withoul jumping, a set of spins precessing in phase initially get out of step in a
lime ..... (l/OwRL)' The condition expresses the faci that for motional narrowing
S
(,)
, to occur, the jumping must occur before the dephasing can take place. In our
~x
example of motional narrowing, the correlation time would be on the order of
, the mean time spent in one of the three configurations of Fig.8.11. If T is the
time in anyone orientation, roughly
Tc :::::: 3T . (8.130)
(
In a similar way we expect that spin-flip narrowing will worle only if the
(b) (c)
S spins are flipped among the needed configurations before dephasing has occurred.
S
Thus we get a condition on T:
(8.131)
" "
The better the inequality is satisfied, the longer the spins will precess in phase.
( We expect, therefore, that we shall wish to flip the spins again and again
through lhe configuration of Fig. 8.12. This we can do by applying a given cycle
FIg.8.I2. The two spill!l Ai, and M. are oriented parallel to one another. In (a), (b), lind
of spin-flips repetitively. The basic cycle will bring Ihe spins back to their starting
(c) their magnetic moments are parallelrespcctive!.y to the Z-, Ih and z-directiolls point at the end of each cycle.
408 409
A fonnal description of what takes place begins with a specification of the ¢(tt) = ei..,lI,tw l o .p(ti) (8.131)
Hamiltonian. We shall use it to compute the development of the wave function
in time. We write lhe Hamiltonian as where, for example, "'{Htt w is chosen as 1ffl giving
(8.134)
Table 8.2. Varialion of the wave function with time for. three-pulse sequence
Time
1t3 = 2: Bij(rjj)(Ii .Ij -
ij
3I::iI::j) (8.135)
I; (just before PI)
It (just after ~)
,,= "('.)
.p=P,¢(I,,)
'Ho includes the chemical and Knight shifts qUi."H~ is the secular part of t. + T,- (just before 1'\) '" =exp( -i1ti•• 1"1 j")~ ¢(t.. )
the dipolar coupling, the coefficients Bij(f"ij) including the distance and angular =UI.,(TdPt.p(I.)
factors [see (8.54a»). t. + Tt (just .ner P,) ¢ = ~Ui .. (TdP, tP(t.)
The objecl ofspin-jlip narrowing is to cause ~ to vanish while maintaining I. + Tl + Ti" (just before 1\) tP = Ui•• (T1)1'\ Ul •• (Tl) PI ,,(t.)
1to nonzero. I.. +TI +T1+T3- ¢(I,,+d = UihtCr~)/"JUi ... (T1)P:!Ulnl(Tl)Pl\/.,(t,,)
To illustrate the principles of spin-flip narrowing, it is useful to idealize the Dust before (II + l)th cycle] == UT ¢(! .. ) defining UT
situation. Correclions to the idealization are important as a practical matter. We
return to them in Sec!. 8.13.
The idealization is to consider that 11rf is zero except for very short times We can thus write
during which it is so large that 'Hint can be neglected in comparison. This ap-
proximation enables us to say that 'Hrr produces spin rotations at the time a pulse (8.139)
is applied, but that between pulses the wave function ¢(t) develops under the
where UT is independent of n. Since UT is a product of unitary operators, it is
action of the time independent Hamiltonian 'Hint. itself unitary. After N pulses, we have
Thus between pulses the wave function at lime t can be related to its value
at an earlier lime t I by (2.49) ,p(tN) = u!f,p(to) (8.140)
¢(t) = exp{ - (ilh)'Hint(t - tt»)¢(t t) so that the problem can be considered solved if the effect of UT can be deduced.
= Uint(t - tlhb(t) (8.136) Let us examine one cycle, and introduce the unitaty operators p j - I which
are the inverses of the Pi'S. For a unitary operator
which defines Uillt. The effect of pulse of amplitude HI and duration t w along
the a-axis (a = x, y, z) at time tt is to produce a transfonnation by the unitary Pi-I = P/ (8.141)
operator Pj for the ith pulse, where the" stands for complex conjugate.
410
411
Then we write UT as
(8.149)
UT :::: Uill~(T3)P.JUin~(T2)P2Uin~(Tt )Pt
:::: P3 P2 PdP\-t p 2- t P3- 1Uinl(T3)P3P2 PIHP.-l p 2- t UinL(T2)P2Ptl which leaves the problem lhe .fame.
X [P1-tUinl(T)Pd (8.142) Thus, if R were a spin rotation operator giving a +1r{2 rotarion of the spins
about some axis, R1tR- 1 musr transfonn the coordinates of Ihe Hamiltonian
Now we showed in our example of Sect. 8.10 that a complele cycle of spin- corresponding to the same rotation.
flips should get us back to the starting point so that we can repetitively flip the If 1t:: -..,hHoIt and the spin is in a Slate corresponding to the spin-up state,
spins among "on-axis" and "side" configurations. Therefore the cycle PI, P2, PJ a transformation, R, which rotates Ihe spin function into the up spin state along
should get us back where we started. Hence the +y-direclion requires rotating 1t Ihe same way, which is done by replacing
(8.143) I: by I" in 1t.
We can therefore interpret
giving us
p- l 1tP:: (P- I )1t(p- I )-1
UT :::: (P1- 1P2- 1P;'Uinl(T3)P:! P2P j][P1- 1P2-IUinL(T2)P2Pd (8.150)
X (PI-IUinl(TI)Pd (8.144a) as the 7t corresponding to the rQlation P- I • We can consider the following two
descriptions of the effect of 1t~ following 3. 1fn. rotation:
:: Uinl(TJ)[PI- 1P2- 1Uinl(T2)P2PI][PI-IUinL(TI)Pl] (8.144b)
i) Use ~ untransfonned acting on a t/J which is rotated +wn.
The meaning of the individual tenus can be made evident by several trans-
formations.
Erst, consider a unitary operator P and a Hamiltonian 1t, and the corre· ii) Leave the spin function alone bUI rolate the spin coordinates in ~ by
sponding U: -wn., the inverse of lhe spin rotation in (i). Equation (8.144) corresponds to (ii).
P-IU(t - to)P:: p-l e -(i/l)1{(t- 1o) P . (8.145)
By expanding the exponential, insening p-l PC:::: 1) between factors, and re- If a rolalion is made up of several rolations in succession, the inverse consists
grouping, we find that of lhe inverse rotations perfonned in the opposite order. Thus, if the spin is flipped
by P2P\, the inverse transfonnmion Q is (P2PI)-I, so the 1t transformed by
p-I U(t - to)P = e.p[ - (Uh)(P-'1iPXt - to») (8.146) the inverse is
so that the effeci of p- 1 UP is 10 cause U to develop in time under a transfonned
(8.151)
Hamiltonian.
If we were to evaluate the expression in the exponent of Ihe transfonned The prescription for finding the transfonned 1t conesponding to the ith
Uin~(T2)' interval Ti of an n pulse sequence is to take the spin rotations PI, P2 Pi,
Pit P2-I1tinL(T2)P2Pj (8.147) which preceded the interval and transform the coordinates in the Hamiltonian
by applying rhe inverse rotations in the reverse sequence (e.g., first p;-t, then
we could first do the trunsfonnation p2-t'HinlP2, then sandwich the result be- Pj-=.ll , ••••... , lastly PI-I).
tween p t- 1 and PI and evaluate that, using the appropriate expressions for ex- We now define the three rransfonned Hamiltonians 11.", 'HB, and 'He as
ponential operatOrs. 111is order of application of the opemtors is the rever.fe of
the order in time of application of Ihe spin rQlation operators. Why is that? 1t" :::: PI-I1tintPI
Consider a SchrOdinger equation 1tB :: P l- l P2-I1tinLP2PI (8.152)
_':.iN=1i'" (8.148) 11.e :::: 1tint
i at 'f'
We can transfonn this equation with a unitary operator R, which is independent On expanding the exponentials we get
of time, to the problem
412 413
UT =ex p ( -*1iCT3)eXp( -*1iBT2)eXp(-*1i"'11) ('H~)B : 'L Bij(Ii ·Ij - 3lzi l zj )
i<j
i (~>c • L: B(I; . I j - 3I,;I,j) (8.158)
= 1-
1i (1iCTJ+1iB1'2+1-£",Tt) i<j
(8.159)
+-l(i)2 2 2 +'H. 2 7"22 + 'H.",7"I)
- (1-£C7"3 B
22 + ... (8.153)
2 I, These results could be achieved if p,-I were equivalem to a 1r/2 rotation
If the 7"'S are short, about the x-axis which would transform Izi imo [yi and if P 1- 1p-I were a 1r12
rotation about the y-axis which would transform Izi and Izi. 2
(8.154) These pulses would also transform 'Ho
(1io )", : -h(wo -w)III-llwo 'Lazzi1yi
where by 1(I'HA,B,clhl)1 we mean a quantity of the magnitude of typical matrix
elements of the transformed Hamiltonians. Under these circumstances, UT is well
(8.160)
approximated by keeping only the leading two teons on the right of (8.153). The
condition on 7"i is similar to the requirement on the correlation time 7" for there
to be motional narrowing.
Introducing the period of a cycle. t c = Tl + 7"2 +"f'J, we get
Setting w : wo for simplicity. we would then get
(T'J
UT = I - -i 1ic- +'H.B-+'H.A-
7"2 7") )
tc (8.155.) liwo""
II tc te te Ho : -3 ~ azzi(Izi + I y ; + lzi)
i __ ,
= I -1i'Hinttc (8.155b)
(8.161)
';:! ex p ( -*'H.inttc) (8.155c)
where (b) defines 'H.in~' the average 1iinh and (c) serves to remind us that over a (8.162)
cycle the system develops to a good approximation as though 1iint were replaced
by its average 1iint' defined above. Equation (8.162) shows that the chemical shift and Knight shift are reduced
Now we expand. 1iin~ into its elements by v'3 (i.e., multiplied by 1/Jj) from their value without averaging. We have
succeeded in making the dipolar coupling vanish without eliminating the chemical
'H.in~ : -h(wo - w)Iz - liwQ 'Laui1zi + 'H~ shift. To the eXlent that the term in (w - wo) is included (it represems frequency
offset or field inhomogeneity). it is also reduced in the same proportion.
" 110 + 11~ (8.156) According to (8.139) and (8.155c), ¢(to) and f/J(t) are given by
The trick then is to choose the pulse cycle (PI> P2, P3, etc.) so that we eliminate
the dipolar coupling but maintain Ihe chemical shift and Knight shift information:
,,(t) . ul( ,,(to)
: exp( - (i/li)1tinl(t - to»)¢(to) (8.163)
(8.157.)
provided
-h(wQ - w)lz - hWQ 'Lazzi1zii-Q (8.157b)
t = to + Nt c (8.164)
Thus if How abour ¢(t) at other times? Consider a time t in the Nih interval such rhar
(~)A : 'L Bij(li .Ij - 3111iIlIj) t:tO+Ntc+tl (8.165)
i<j
414 415
The time t) could fall into anyone of the time intervals T), T2 • ••• Tj which make add either a steady-state or a pulsed spectrometer to observe the character of
up the basic cycle t c . Over anyone complete cycle there is no large change in ¢ the resulting resonance. In practice. the amplifiers of most steady-state apparatus
if (8.152) is true. However, each large pulse produces a sudden big spin change would be blocked by a large rf pulse at their input, with a recovery from blocking
_ typically a 1fo, rotation. These large pulses might, for example. progressively which would be much slower than the short time T between spin-flipping pulses.
cycle the spin along the X-. y-. and z-axis in the rotating frame. If one always Consequently. to see spin-flip narrowing one goes to NMR apparatus designed
observes during the ith interval. however. the effect of the big pulses in successive to handle intense rf pulses-one uses a pulse rig.
cycles will always have returned the nuclear magnetization to the same direction With coherent pulse apparatus. one picks out a particular component (Mz
in the rotating frame. or My) of the magnetization in the rotating frame. If one is adjusted to detect
Thus we can write M z • then one finds in general that it is largest in one panicular interval Ti of
the cycle. perhaps zero in others. One could start with the system in thermal
,,(t) = e<p( - (VIi)1{;n'(' - '0»)"«0) (8.166)
equilibrium. apply a 7(0, pulse about the x-axis to rotate the magnetization to the
when t and to are both within the same subinterval Tj of tc' % - Y plane. and then start the spin-flip cycles. For simplicity of the discussion
If they are within different intervals Tj and Tj. with we assume we can apply 7ro, rOlations about the %-. y-. and z-axes in the spin-flip
cycle. though in practice z-axis rotations are not used. A possible set of pulses
Ntc<t-to«N+l)t c (8.167)
to initiate and carry on spin-flip narrowing is shown in Fig. 8.13. In Fig.8.13a
we should properly go back to expressions like (8.136) and (8.137). 1f we consider we see M along the direction of the slatic magnetic field (the +z-direction). AI
thm the individual T'S are so short that (8.154) is true, then 'Hint does not produce t = 0, Po rolates M to the +y-direction. This is called the preparing pulse. At
much of a change during t e , but the pulses i + 1, i +2, i + 3, ...• j - I, j do. Thus, time TO we initiate the spin-flip pulses Pl. P2, P3 which successively rotate M
defining by -rr/2 about the z-axis, +rr/2 about the y-axis. and lastly +2rr/3 about the
-i + j + k axis. The rotation P3 is deduced as being the inverse of P2P) and
P == PjPj-I ...• Pj+2Pj+l (8.168) can also be thought of simply as PI-I P2- 1 mathematically. or physically as the
there are two expressions which are nearly equal: successive steps -7I:n. about the y-axis followed by +71:/2 about the z-axis. Note
between t and to as though 1{inl acted the whole time, and was followed by the i
i .......... ----...., I'J
(d)
,,
rotations Pj+ I, ....., Pj-I. Pj in quick succession.
--- ' -, P,
i
i, ,,
,
(l')
\
,, ,
i
8.12 Observation of the Spin-Flip Narrowing ,
How can one experimentally observe the effect of spin-flip narrowing? Ordinary Fig.8.13a·e. The erred of ~he various pulses 8S llCCll in the rota~ing frame. (a) Thc initial
motional narrowing can be seen either by steady-state or pulsed apparatus as very thermal equilibrium configuration wi~h lhe ml<gnctizalion M along the :-direction (lhe
direction oflhe static field). (b) The rotation Po produced by the preparing I)ulse puls M
narrow lines in the former case. and as long Bloch decays or slowly decaying along the y-axis. (c) The first pulse P of the spin-flip cycle rolales M lo lhe +z-direetion.
echoes in the laller. (d) The second pulse ~ rotates AI to the +:-direction. (e) The third pulse 1'\ must be
To produce spin-flip narrowing large rf pulses must be applied to the sample. the inverse of ~ Pt. H is a rolation of 21f/3 aboUl the (-i +; + k) axis. (Noll' lhis is not
the same as a -1f/2 rotation aboul the +z-direction)
Assuming one has the equipment to apply such pulses. one could then in principle
416 417
that ahhough P J looks superficially like a rOlation of -1(/2 about the x-axis. it ,p(t) = e -(1M"; .. ' Po,p(O-) (8.173)
is not. as can be verified if one considers rotating a three-dimensional object in
PI, P2. P3 instead of just a vector M pointing initially along the y-axis. This
If we ask for t/J(O later in the cycle. we use (8.170)
distinction is imponant because lld is a second rank tensor, thus it should be ,p(t) = Pe-(I/.),,;o,'Po,p(O-) (8.174)
thought of as an ellipsoid rather than as an arrow.
If the apparatus were adjusted to measure M z • and if it were tuned exactly where P = PI or PIP2 depending on whether t lies in the interval Pt-to-P2 or
to resonance. including any possible chemical and Knight shifts. there would P2- to-PJ.
only be a nuclear resonance signal between pulses PI and P2. If there were a We now need to express the fact that the initial wave function 1/J(0-) cor·
resonance offset (as with a second class of nuclei chemically shifted from the responds to the initial thermal equilibrium of the sample. To do so we express
first), M z would gradually change in the PI-to-P2 interval for successive pulses. y,(0-) in terms of the laboratory frame 4»(O-} instead of the rotating frame. In
undergoing an oscillation (see Fig.8.14) as we show below. It would appear general the rotating frame wave function t/J and laboratory function ifJ are related
during the PJ-to-Pt window as well as time went on. by
Let us tum to a mathematical formulation of the calculation of the observed 1/J(t) =e-ilo.ltl·ifJ(t) (8.1750)
NMR signal. For concreteness, we compute the time development of M z in the
rotating frame < Mz(t) >. We will use the language of wave functions rather so that at t "" 0- we get
than density matrix. Given the wave function !/J(t), we have (8.175b)
< M,(t) > = < ,p(t), M,,p(t) > Taking I n) to be the eigenstates of the secular Hamiltonian 1(. in the laboratory
= '''<,p('),1,,p(r» (8.171) frame. we then have
We assume that we have a pulse sequence similar to that of Figs. 8.13 and
(8.176)
8.14. At t = 0- we let the wave function be t/J(O-); then the preparation pulse
gives us Thus "
,p(0+) = Po,p(O-) (8.172)
<Mz(t» = <t/J(t).Mzt/J(t»
The wave function then develops under 'Hin~. If we ask for it in the interval PJ
to PI. there will be an integral number of cycles PI followed by P2 followed
='Yh E cnc~(m I Po-le(i/.)'Hiall P- t IzP
",m
by PJ, so that. using (8.166).
xe-i/.)'H;a,IPo!n) . (8.177)
We now make a statistical average over the coefficients Cnc:" utilizing the fact
that the states are occupied according to a Boltzmann distribution of the Hamil-
tonian 1(.. which describes the system before application of the pulses:
_._ e- E.. lkT (n le-'HlkT I m)
(,) cnc m = Onm Z = Tr{e 'HlkT} (8.178)
P, P, p, p,
Substituting (8.178) into (8.177) we get
(8.179)
In the high temperature approximation we can write
(II)
e-"/kT (I -7{JkT)
/'0 1', 1'1 I'j P, 1'1 1', P, 1'1 (8.180)
Flj;.8.14a,b. Grllph of ~he OllCilloseope signal of the resonance venus time for ~he pulse
Z "" (21+I)NG
sequence when (a) apP"ra~us is se~ exactly a.~ re&Onan<:e, or (b) II chemical shifl. causes
~he M z signal to change with time appearing in ~he PJ·to-p. inter....11llI it diminishes in
where there are No spins of spin I. The leading term vanishes, so that the
~he p'-to-f\ window nonvanishing term gives a I/kT contribution.
.,8 419
If we then wrile
::: NO'Y2h2I(l + l)Ho
3kT
[.!.. + ~
3 3'0'
(woont)]
J3
1£ ::: ~'YhHoIz (8.181)
ttansfer the PO-I from the lefl to the right in the ttace, using 1 2
:::Mo [ 3+3cOS ~ (WOU"')] (8.189)
Th{ABC}: T,{BCA}
and if we pick Po 10 make
(8.182)
where Mo is the thermal equilibrium magnetization.
We see, therefore, that the chemical shifl-Knighl shift tl zz shows up as an
PolzPO- 1 = I z (8.183) oscillatory contribution to the < Mz(t) > ,reduced in frequency by 1/../3 from
the normal frequency shift W1)Ou one would see if one could resolve il in a
which rotates the spin from the z·axis to the x-axis. we get steady-state resonance.
2 H {- ~ } Jf more Ihan one kind of chemical or Knight shift is present, the resultant
<M (t»:::'...,2h °Tr e(i/1I)'Him,tp- I I Pe-(i/Il)ni·'(I < Mz(t) > 's add, Ihe respective strengths being in proportion to the number of
z (2I + I)No kT z z·
(8.184) nuclei possessing that chemical shifl or Knight shift through the Mo faclOr.
A useful exercise for the reader is to derive (8.184) using the density matrix The form of (8.189) is that of a conStant term plus an oscillatory term. By
language. For simplicity we take P ::: I (Le., an integral number of pulse cycles taking a Fourier transform one can deduce a "reduced chemical shift frequency"
PI. P2, P3 pass between t ::: 0 and the time of the measurement). (WOtl u/../3). II is possible to utilize Fourier transform NMR attachments therefore
Now, to decompose a transient signal when there exists man: Ihan one Knight shift-
chemical shift to unscramble the discrete frequencies.
1£int = - m;o 2;:Uni(lzi + Izi + IIJi)
•
(8.185) 8,13 Real Pulses and Sequences
Our discussion up to now has been directed 10wardS explaining the principles
where [z'i = (1/J3)(lzi + IIJi + IziJ. Suppose all spins have the same (Tn' Then
of spin-flip narrowing. Trnnslating these notions into prnclice. we encounter cer-
we get
tain problems. We discuss these briefly, Ihen list a few of the pulse sequences
1£inl::: - hwo~zIzl (8.186) which have been developed to cope with them. An excellent discussion of these
problems has been given by Haeberlen in his NATO Summer School lectures. I
Now in terms of the prime coordinates we have, in an obvious notation, Vaughan and colleagues [8.38,8.39] give a thorough theoretical discussion in-
cluding explicit fonnulas of use. See also the references in the Selected Bibliog-
I z ::: Iz'cos(x, x') + 1IJlcos(x. y') + Iz'cos(x, z') raphy.
Izl ::: Izcos(x, Xl) + IIJcos(y, x') + lzcos(z, x') (8.187)
8.13.1 Avoiding a z-Axis Rotation
so that
exp (
-iwouutIz
J3
')I zexp
(iwouutIzl)
J3
Pulse rigs apply HI'S in the x - y plane. By choosing the rf phase, a single one
of these can produce rolations about any arbittary axis lying in the z - y plane,
but not about the z-axis. To produce a z-axis rotation with a single pulse would
::: cos(x, x') [IzICOS (WOjizt) + III,sin (wojizt) ] require adding a coil wilh an axis parallel to Ihe z-axis plus the electronics to
energize it. In order to avoid so doing, one can go to a four-pulse technique. That
+ cos(x, y') [-l z 'Sin (w0./izt) + Iy'COS( W0.,ij-zt) ]
in a sense is equivalent to composing a z-axis rotation from successive x-axis
and y-axis rolat.ions. Ordinarily, however, the two pulses are spaced apan in time
+ cos(x, zl)Iz' (8.188) so that the basic timing intervals can be maintained more simply.
Using the fact that Tr{Iz ' III'} = 0, and Tr{I;,} ::: Tr{~I} we get, after evaluating
1 The 1974 NATO Advanced Study Inslilule held in Leuvcn, Belgium fcalur«! a
the traces, number of userul lectures on NMR in solids./fa,,/)#!rl"/1 aoo Van /f«~ in p"rticulllr lreat
spin-nip nflrrowing in [8.401.
420
421
To write sequences, it is convenient to introduce a notation based on the fact effect can be eliminated to first order by having l[1>/dt during turn-on and dtPldt
that1r(l pulses are nonnally used. For example, consider the following expression during turn-off equal and opposite. Making l[tPldt zero is not practical, but the
which defines one four·pulse scheme. balancing during turn-on and tum·off is.
(8.190) Among the various pulse sequences used or referred to in the literature,
X(-r,X, T, Y,2T, Y,"',X,T) .
we shall list three. They are known as WAHUHA, HW·8, and REV-8. These
The first "X" represents the preparing pulse, a 1fn. pulse about the +x-axis. The mysterious symbols are made up from initials of their inventors plus numbers
parentheses enclose the pulses of a single cycle, which in the example indicated which tell how many pulses there are per cycle.
is considered to stan with a time interval T prior to application of a 1((1. pulse The WAHUHA is named after J.S. Waugh, L.M. Huber, and U. Haeberlen
about the -x·axis (or a -1fn. pulse about the +x·axis). This sequence could also [8.34J. \( is
be written as
X, T(T,X, T. Y,2T, Y,T,X, T)
X,T(X,T,Y,2T,Y,T,X,2T) ,
This sequence can be sampled at the times corresponding to just before the
which simply chooses to call the staning point in the cycle as the time just prior parenthesis by measuring M lI • The name WAHUHA refers to the contents of
to the first -:I: pulse in the chain. This notation makes explicit the presence of the parentheses. The preparation pulse can be chosen to give convenient sets of
two intervals of 2T and twO of T. The first notation is useful in proving certain timing. The panicular sequences shown have the property that the X,T, at the
theorems which utilize the time symmetry of a single pulse cycle. start are in fact like the last pulse and interval of the cycle; hence no special
timing intervals are needed at the stan. That is, the preparing pulse in this case
8.13.2 Nonidealily or Pulses is pan of the cycle.
For the preparing pulse shown, the magnetization is initially turned to the
Pulse Duration. We have utilized instantaneous rotations in the discussion up y--direction. Thus, if one detectS My at the inverval between the X and -X
to now, which would require an infinite HI of zero duration. Actual pulses have pulses, one will get a signal.
nonzero duration. One's first guess might be that there would be major problems We can write the HW-8 [8.35] as
as a result of the nonzero duration. Theoretical investigation of this point reveals, (T,X,2T,X, T, Y,2T, Y, T, Y,2T, Y,T,X, 2T,X, T)
however, that it is not serious [8.34,35,38,39]. lbe point is that slow rotations
can still maintain the property of averaging out the dipolar coupling. They are in There are a number of equivalent variations.
this sense equivalent in motional narrowing to allowing (3 cos 20 - I) to average The MREV-8 [8.38] sequence is
to zero over a conrjmwus variation in 8 as opposed to the discrete variation of X, T(T, X, T, Y,2T, Y, T,X,2T,X, T, Y,2T, Y, T,X, T)
Fig. 8.1. The important experimental consequence is that much lower If power
is needed that one would at first suppose. This can be viewed as pair of pulses X, Y separated by T followed at 2T later
by their inverse y, X.
Inhomogeneous HI. As we discussed in Sect. 8.3 it is always difficult to make MREV demonstrate the highest resolution obtained with their MREV-8.
HI homogeneous over a sample, panicularly when one is attempting to keep
the rf coil small to cut down on the power requirements (for a given Q and
Hj, the power needed is proportional to coil volume). This problem is a bad 8.14 Analysis of and More Uses for Pulse Sequence
one in that it causes the rotation angles to vary throughout the sample. It can be
removed to first order, however, by using the trick of Mejboom and Gill or of
pairing positive and negative rotations about any given axis, thereby achieving a The importance and utility of Illultipulse techniques goes far beyond spin· flip
refocusing effect. line narrowing. We turn now to several examples, as well as to illustrations of
some calculations of the effects of the pulse techniques.
Phase Modulation. Turning on an HI which points along a given direction Before beginning our discussion, it is useful to introduce a notation and 10
in the rotating frame sounds easy, but in practice is hard to achieve. Usually the demonstrate some useful relationships. We shall be following the effect of pulses
phase of the HI swings during the buildup and during the decay. Such swings on individual components of spins such as I:r:, I y • or I: and on dipolar tenns.
can be labeled as modulations of the phase angle tP of the rf. They are equivalent For the dipolar tenn we introduce the notation 'Hon(a = x, y. z) defined as
to being off-resonance L\w = dtPldt during the time that dtPldt=l-O [8.41). Their
423
1i&& = L Bik(/j ·lk - 3I&jI&k) (8.191) PI = X then P 1- J = X) and then using the left-handed rotations or by instead
i>k pretending that one is calculating
Explicit calculation then shows that
~4~J ~.I~
1iu +1iIlIl +1iu =O . (8.192)
using right-handed rotations. For a pair of pulses PI followed by P2. then the
In general. the effect of pulses Pi will be rotations. To discuss the effects lefl-handed rotations
of nonzero pulse lenglh:i. one must treat the Pi 's as time-dependent operators
PI-tP2-t1&P2PI (8.I99b)
such as
would be replaced by the right-handed rotations Pt.l'2. etc.
(8.193)
p) P21&P2-J p)-t (8.199<:)
where t is the time variable during the pulse. For our present purposes. however.
we shall Ireat the pulses as negligible duration (so-called 6-function pulses) and Using either approach, one can work out a table for (8.198) which is lhen useful
moreover treat them as all being 1tll pulses. for all applications in which we list 1a across the top and in each row list first
We then have to evaluate expressions such as the operator PI. then the effect of Pt on 1&. The result in the table is the same
whether we treat PI as a left-handed rotation and use (8.198), or a right-handed
P1- 1I: P, (8.194a) rotation and use (8.199a).
=
i>k
L Bii(lj ·lk - 3IlI jIlJk)
Z y .- , (8.201)
i>k
where the bar indicates a negative. Suppose then that in the sense of (8.198) one
= 1iYIJ (8.196) has the operator HYII acted on by PI = X. Then the table shows us that
Thus. if
Pi] I&PI = kl{J k= ±1 then (8.197) (8.202)
Let us now illustrate this with a calculation of the effect of a very useful
Pi 1 H&&PJ = Hpp pulse sequence utilized by Warren et al. [8.42] to refocus dipolar dephasing. They
whether k = +1 or - 1. employ it as pan of their method of generating multiple quantum coherence of
The effect of a (Iefl-handed) rotation PJ in the expression a particular order, a topic we discuss in the next chapter. In discussing magic
echoes, we showed that one could refocus the dephasing arising from dipolar
Pit I&PJ (8.198) coupling by creating a dipolar coupling of negative sign. In particular we showed
can be worked out once one knows PI either by first computing P 1- 1 (e.g. if how we could create a dipolar Hamiltonian 1id given by
424
425
1{d = -~1{n (8.203) have omitted the "T' and the "1{" since only the subscript and sign is needed
to specify .the entry. On the third iine we list P2, then the product P2 PI, which
We now show how Warrell et al. created such an 1{d by pulses.
we simplify to the extent possible. The last line labeled "sum" contains the sum
The basic approach is founded on (8.192)
of the time intervals, and the sums of the products of the operators with the
1{z", + 1i yy + 1i u = 0 time intervals for which each operator acts. The average values are then just the
weighted sums of the operators divided by the duration. Using (8.204) we see
Thus
that the pulse sequence does produce an average Hamiltonian
-11.",,,, = 1i yy + 71.:u (8.204)
Therefore, if we create a Hamiltonian which spends equal time as 71. yy and 1{u, (8.206)
its average value will be }inn. Of course, the pulse sequence must be repeated continuously for as long as one
Since without pulses 1id = 1i u , we simply need pulses to switch 1i u to wants to satisfy (8.206).
71.'11'11' Clearly, all we need is X's or X's for the Pi'S. The simplest thing one can A pulse sequence which gives the same average dipolar Hamiltonian but
do is to switch 1i u to 71.'11'11 with PI = X, then switch it back, P2 = X, so that eliminates the chemical shift tenn and has some other advantages is given by
Warren et a!. [8.42]:
(8.205,)
X,~X,~X,T,X,T,X,T,X,~X,T,X,T
for a sequence
We evaluate it in Table 8.4 utilizing (8.201).
(X,T,X,T) (8.205b)
To follow the action, we make a table. We need to follow two tenus in the Table 8.4. Sequence (X, T, x, T, X, T, X, T, X, T, X, T, x, T, x, T)
Hamiltonian: the dipolar coupling and the chemical shift. Ir we simply wish to
create a negative 1id for all spins without regard to their chemical shifts, we need Duration P; P, ... Pi I, 11..
a pulse sequence which will make all chemical shifts zero. We will find that our X X
simple (X, T, X, T) sequence has problems with chemical shifts. We construct a 1i yy
426 427
Table 8.5. Sequeoce (X,,., Y ,2T,Y, ",X, 2,.) So if we knew what QIII was. we could immediately evaluate (8.212). Since
Durll~ion p. J~ ... P; I, M.. fJ will be x, y, or z (or %. y, Z). we can handle any eventuality if we know the
results of Qlz• QI", and QI~. A procedure which will provide us with these
X
, three results is to replace the single column I~ with three columns Iz,!",!z. A
2
y
j7
XY
XYY = X
•
y
..'"
yy
portion of the table would then look like Table 8.6.
T.ilble 8.6. SBml)lc ~Ilble to help ill eVlllua~jng the effcc~ or QP•• I
X XYYX= I
2 , Line f. I, f,
"
Sum 6 2(r. + J, + Iz ) 2(1l:... +1l,,+1l.. ) I PJ ... p.. _ 1
Average ~(l.:. + J, + J.) f(ll zz + 11" + 11,.) 2 I~II I" I,
3 PI ... P.. (= Q)
4 I",,, I., I.
S P, "'P"P"+ J
For multiple quantum excitation, Warren et al. [8.42] show that it is useful 6
to generate an effective operator proportional to 'H"" -'Hu . In panicular, they
find
!('HYfI -'Hzz ) = i(2'H fIfI + 1l u ) (8.201.) To compute line 6. we first find the effect of P n + 1 on I z • I y • and I z . There will
is generated by the eight-pulse sequence be three results. We look for them on line 2. Thus. suppose
After a bit of experience in calculating tables, one soon discovers a problem. Then
Suppose one has just calculated the effect of the operator Pj ... P,d/p = la'
(8.208) So we would enter I rr on line 6 under the I z column.
acting to the right. Suppose it produces a result I a in the I~ column. We next This approach requires evaluating three columns. but it makes the step of
have to calculate the effect of the operator going from one line to the next much simpler - one has only to evaluate two
operator steps per column. hence six operator sfeps per line.
(8.209)
Now it would be easy if the operator we had to calculate was Pn+l Q since then
(8.210)
This situation would let us take the result of Q. which we already have. and
apply Pn+l to it. However. what we must calculate is instead
QP..+1
which has the PII +l operate before Q operates. Therefore, we must go through
all n + I operator calculations. a tedious process for long pulse squences. Now
suppose
Pn+lI~ = IfJ (8.211)
Then
(8.212)
428
429
9. Multiple Quantum Coherence
9.1 Introduction
In Chapter 5 we introduced the density matrix as a powerful tool for the analysis
of magnetic resonance experiments, In analyzing the case of spin !'we saw
that the diagonal elements of e were connected with the magnetization parallel
to the static field, and the off-diagonal elements were related 10 the transverse
components through the equations
(5.257.)
11<
(M,(O) ~ 2[,++(0 - e--(O] (5.257b)
431
detect elements such as in (9.1), we naturally wonder if there is any experimental In fact, for an N -spin system with total z-component of spin I::T given by
method for probing the existence of such matrix elements. A closely related N
question is whether there is any way to excite such matrilt elements if they are
~-E~ ~
initially zero. k:::t
We are already in a position to say that in cenain cases we can eltcite
the eigenvalue M of ['IT is a good quantum number. Thus, if we have a set of
these elemenls because indeed we have analyzed in detail an eltperiment which
quantum numbers M and a, where a are all the other quantum numbers needed
generates them. Referring to our analysis in Sec!. 7.26 of the pulse sequence so~ ,
(9.2) I,TIMa) = MIMa) (9.7b)
we see that in (7.338) we produced tenns of {l which, eltpressing {l in operator we define matrix elements
form, are proponional to
(M'a'Ie(OIMa) (9.8)
1z 5" (9.3)
as n-quantum matrix elements where
These gave nonvanishing elements of {l
n _IM-M'I (9.9)
(7.3410)
We follow the convention of calling n the order of the matrix element, and
(7.341b) say that the existence of a nonvanishing matrix element (9.8) describes n-Qrdtr
cohertnct.
where in each matrix element the upper signs go together, and the lower signs It is immediately evident that for a system of N spins of spin I. M can range
go together. from N I to - N I, so that the largest n one can have is 2 N I. Thus, for protons,
The matrix elements (7.341 a) join states of energy difference, i1E, given a spin system involving an H2 group (as in a CH2 fragment) should pennit
by generation of two-quantum matrix elements, a system involving three protons
i1E::: ± 1i.(WOf +wos) (9.4) (as in a CH3 group) should pennil generalion of both two- and Ihree-quantum
and the matrix elements of (7.34lb) join states differing in energy by matrix elements.
The generation of a given order of coherence thus enables one to determine
i1E::: ± 1i.(wOf - WOS) (9.S) what spin groupings exist if one can detect the existence of orders different from
If the two spins were identical, so that WOf ~ Wos, we would refer to these one. NOle, however, that a pair of CH2 groups contains four protons, so one
energy differences as two quantum and zero quantum and we would say that the may ask, when should one consider a CH2 fragment to be a pair of protons, and
pulse sequence (9.2) has generated 0- and 2-quantum matrix elements of (l in when should it be considered to be part of a larger group?
addition to the I-quantum matrix which we observe directly. We thus see that there are exciting possibilities to use multiple quantum
A next question is, clearly, can we detect the zero- or two-quantum matrix phenomena to classify spin groups, but that there are also imponant things we
elements given the fact that (7.248) shows that resonance directly sees only the need to understand, such as, how do we create and how do we detect the various
one-quantum matrix elements? Before proceeding with that question, we discuss orders of coherence?
a bit more the concept of 0-, 1-,2-, and in general n-quantum matrix elements. Multiple quantum phenomena are encountered in all branches of spec-
In some cases, mf or ms are not individual constants of the motion. For troscopy. Bodenllausen [9.1), in his excellent review of the field, includes a
example, consider two identical spins, with a chemical shift difference and a discussion of the manifestations of multiple quantum phenomena in cw magnetic
spin-spin coupling involving tems such as 1· S. If the spin-spin coupling is resonance speclroscoPY. It was the advent of pulse methods in multiple quantum
comparable in magnitude to the chemical shift difference, ml or ms are not speClroscopy which caused the technique 10 become of central importance as an
constants of the motion. It is still true, however, that the operator, F::, for the almost routine speclroscopic tool.
total z-component of angular momentum In two fundamental papers, Hatallaka et aJ. [9.2,3] showed how to generate
and to detect multiple quantum coherence. They sludied the Al 27 resonance
(9.6) of A1 20 3, which can produce multiple quantum transitions since the spin of
commutes with the Zeeman energy and the secular part of the spin-spin coupling, Al27 is 5/2. Aue et al. [9.4]. in their famous initial paper on two-dimensional
and thus its eigenvalue, M. is a good quantum number. Fourier lransform NMR, show that, using two-dimensional Fourier transform
432 433
spectroscopy, one can observe both zero- and double-quantum speCtra. They v = Vt cos Wit + V2 cos W2t (9.13)
describe several ways of producing the desired matrix elements.
would produce a current
Since the publication of the basic papers from the laboratories of Hashi and
of Emst, the field of multiquantum coherence has exploded_ 1n addition to the j(O := A(V? cos 2 Wit + vl cos2 w;et + 2Vt V2 cos WI t cos w;et) (9.14)
review by Bodenhausen. there is an extensive review by Weitekamp [9.5]. There Utilizing the fact Ihal
are useful reviews by Emid [9.6J and by Mu/Wwitz and Pines [9.7]. As in aU
aspects of pulsed NMR, the book by Ernst el al. [9.8) gives a thorough and cos WI t cos W2t := !{ cos (WI + W2)t + cos (Wt - W2)t] (9.15)
detailed account of the principles and practice.
Our goal is to explain the physical principles. We turn next to showing that
we get
it is the nonlinearity in the equation of motion which makes possible excitation of
multiple quantum coherence. We initially deal with cases in which we are tuned
i = ~ { V12 [1 + cos (2w 10] + vlp + cos (2wzO]
to discrete transitions, showing how we can then progressively pump higher and
+ 2VI V2[ cos (WI + wz)t + cos (WI - wz)tl } (9.16)
higher order coherences.
Nexi we deal with the much more common experimental case of applying
pulses of sufficiently greal amplitude to excite a number of transitions, discussing
Thus the current i contains the frequencies 2w), 2w 2 , + W2, WI - W2, and 0w,
(which is WI - WI and w;e - W2).
as well how to detect the fact that we have generated multiple quantum coherence. A similar thing happens for the densily matrix equation
We then lum 10 melhods of singling out in the detection coherence of a desired
df! i
order, and lastly to the problem of how one can selectively excite a panicular d' • ,,(eli -1/e) . (9.17)
order of coherence.
Since this equaJion involves a product of f! with 1l, if both f! and 1i are time
dependem on the right hand side, they couple to tenns on the left hand side
which oscillate at the sum or difference frequencies. We want to show that Ihis
9.2 The Feasibility of Generating Multiple Quantum is just what we need to take a {} which has an n-quamum coherence and pump
Coherence - Frequency Selective Pumping it either up 10 an (n + I)-quantum coherence or down to an (n - I)-quantum
coherence by means of a time-dependent 11. which oscillates at the frequency
Consider for the moment a two-spin system. We have seen (Sect. 7.5) that we can difference between Ihe nand n + I (or n - I) quantum coherence.
produce two-quantum elements of f! by applying a driving signal at WI = wOJ, To make those ideas more concrete, let us consider a system with a set
followed by one at Ws := Wos. Among the mauix elements produced is of nondegenerate energy levels, a static Hamiltonian 1iO' and a time-dependent
drive tenn V(O given by
(- - 1,,(')1 ++) • (9.10)
iwt
V(t)/h:= Fe + F·e- iwt (9.18a)
which oscillates in lime as
where the· refers to the complex conjugate.
exp [ - i(wOI + WOS)tJ (9.11) For example, if we have a group of identical spins as in (9.7) acted on by
We now proceed to show that this result is closely related to the idea of the a rotating magnetic field, we get
beating of the frequencies of two oscillators. Vet) = -,!tlll(J'Z:T cos wt - lyT sin wt)
(e e-
Beats are produced by putting input signals into a nonlinear device so that
the output signal contains tenns proportional to the product of the input signals.
Thus if we had a device whose electrical propenies are characterized by the
= -,"/till, [[Z'f
iwt
+2
iWI
) + ~IYT(eiwt _ e- iwt )]
property that the current i through the device is proportional to the square of the := _ ,!i.ll, (ueiwt + l-e- iwt )
2 T T (9.1gb)
voltage V applied across the device
i.AV' (9.12) so that, using the usual notation WI =..,1l I,
application of a voltage (9.18<:)
434 435
Note that a rotating field applied in the opposite sense is obtained by treating W Substituting (9.18) for V/IJ in (9.24), we thus get
as negative. . dl
Let the eigenstates of 'Ho be denoted by quantum numbers, J, k, 1 each of exp(-iwj/t)....::l!..dI
t
"" i L:, ejk {exp [i(w - Wjk)tjFkl + exp [ - i(w + Wjk)t)FkI}
which stands for some sel of quantum numbers such as M, 0' of (9.7b). Then we
define - i L:
, {Fjk exp [i(w - Wk/)t]
(9.19)
+Flkexp[ -i(w+wk/)l]}ekl (9.25)
Then, denoting the interaction representation by a prime, we have
The first two tenns on the right give the sum and difference of wand Wjh
ee:O = e -(i/A)11ot {!' (l)eCi/A)11o t or (9.20) the third and fourth lenns on the right have the sum and difference frequencies
of W with Wk/. This is the same situation we gOt for our example of the nonlinear
(9.21.) currenT/voltage relationship (9.12) and (9.16) in which the product of V1(t) with
V2(t) gave the sum and difference frequencies. Here what is happening is that
we consider the elements of e' on the right hand (either e"k or ekl) to be the
(9.21b)
n..quantum coherence which we wish to pump by means of'v up to the (n+ 1)-
quantum coherence or down to the (n - I)-quantum coherence represented by ejf
The density matrix then obeys the matrix equation
on the left hand side. The point is that since ell [the (n + 1)- or (n - I)-quantum
d i coherence] is to be produced, its time derivative must be different from zero.
dt {ljl = h ~ [ejk('Hoh:1 - ('HO)jk{!kEl Let us suppose, then, that at t "" 0 the matrix elements ejl and ek/' and thus
ejl and e'kl' are zero, but that we have previously excited the mauix element ejko
+ .!. L [ej' V,,(t) - Vj,(t)e"J (9.22) for a particular value of k, labeled ko. So, we have a nonzero ~jko' hence f!..krJ'
h , Let us suppose funher that V(t) possesses a nonzero mauix element betkn
states ko and 1. Then, we have
If we had only the tenns involving 'Ho, (!jk would oscillate at exp(-iwjkt). We
can eliminate them from the right by substituting (9.2Ia) into (9.22). We also
utilize the fact that
exp(-iWjft) dt ""
de'·
iejko {exp [i(w - Wjko)t]Fkof
of time (Le. that dej/dt "" 0) if V were zero. Consequently, w~ expect that Then the condilioll that F pump ejko up to Iljl is that
mauix elements of e' to change slowly compared to the frequenCies Wjl' Wjk. F kol i- 0 and (9.29.)
Wk/' Therefore, to a good approximation the left hand side of (9.24) oscillates at
Wjf. This requires that the right hand side also oscillates at the same frequency (9.29b)
if (dtl;/dt) is to be something other than zero.
Wj -WI ""Wj -wko +w (9.29c)
436 437
pump-
This equation may be interpreted as saying that the beat between the Energy Fig.9.J
ce (freque ncy Wj - wig)
ing term F (frequency w), and the n-quantum coheren ...., 3 3 1--1 Fig. 9.2
coheren ce. [If
must add to give the frequency, Wj - WI, of the (n + l)-quan tum
W is negative, the left hand side is an (n - I)-quan
tum coheren ce.]
can also be written as
Equatio n (9.29c)
(9.29<1)
2 '·-1 1-.)
W.&:o -WI::W .
match
The meaning of this is that to excite ejl from ejko (or ejko)' W must
the resonance frequency of the transitio n ko to 1, and V must have a matrix .... , 2 I•• )
which joins ko to 1.
ce
If we stan with a diagonal {J, we can first produce a one-quantum coheren
with a 1((2. pulse. The discussion above shows that we can pump this up to a two--
quantum coherence. That in turn can be pumped to a three-quantum
coherence, ....,
can apply a rf which excites one transition labeled 1, '1 , 3 wilh ener&ies E, " E, E,,..ex""'---· -' h h
etc. Note that we are here assuming we . Threeel
hFig. 9.1.
lergy
.
levels
( <=eU3$ "-'I, l<.'2
""3 -l<.'2 '# "'2 _ "'I'
""3 In til'" notallon of 9.19). Note that l<.'2 > l<.'2 > "-'I and we assume
L_
on the right hand side of (9.25). Thus,
at a time since we kept only one term
an HI sufficie ntly weak to pick out the lhe lhree levels labeled
if there are splittings, we are using Fig.9.2. The energy levels of a pair of spin-t nudf!i. We focus on
than the spacing between spectra l 1, 2, 3
individual lines. It must therefore be smaller
cy selecliv e exciuu{ on. A strong
lines. This mode of pumping is called frequen
it in lhe
HI covering all the splittings is called nonselective excitation. We treat '
d 1!31 . it I d .
~~I
l I
-*eil<.'2 3tv23e; 1 (9.35)
next section.
levels
--;It "" -he "-':)1 V32e21 and ""
438 439
(9.39) One can represent the basic equations (9.28) and (9.29), describing selec-
tive el(citation of (n ± 1)-quanlUm coherence from a state containing n-quantum.
If we satisfy (9.39) with an equality, then we gel
coherence graphically. as shown in Fig.9.3. We use a solid line between two
tP£'~lI (WI)'2 £'31 - O ' dd'2£'21+(WI)2,=O
t _ (9.40)
energy levels to denote the matril( element of V joining those states. Dashed
-;ur + "2 t2 2 f'21 lines between states represent matril( elements of f' (or t/) between those states.
Then, the requirement that to generate a "if" matril( element on the left of (9.28)
which we recognize as hannonic oscillator equations. Writing down the general
demands a jkg and a kgl matril( element on the right simply means that the two
solutions, then evaluating the constants of integration with the help of (9.37), we
dashed lines plus the solid line fonn a closed loop, as in Fig. 9.3a. Then one also
gel automatically satisfies the energy conservation requirements (9.29c) or (9.29d).
, ,
£'31(t)=e31(O)cos (W"). , (0) Sin. (W")
T +1£'21 T Note that although V is in resonance with the 2-3 transition, we have
assumed it is not in resonance with the 1-4 transition since we did nOI connect
,nd those states with a solid line. In Fig.9.3b we show that if we tuned V(w) to be
e~!l(t) = e21 (0) COS (wt) + ie~l1(O) sin (w~ t) (9.41) resonant at the 1-4 transition, it cannot pump f'21 into f'31 since the lines do nol
fonn a closed figure. However, looking at Fig.9.3c we see that V"I will pump
lllUs, if we turn on Hi. e~lI will transfer its amplitude to l?31' which will then f'21 into ~24. the zero-quantum transition.
transfer it back 10 e~l later tInd so on. Note also that these figures show (Fig. 9.3a) that V23 can pump l?JI> the two-
Note that if t is applied for a lime t w generating a 11" pulse (WI t w = 11") so quantum coherence. into l?21, a one-quantum coherence. Also (Fig.9.3c) VI4 can
that pump the zero-quantum coherence into the one-quantum coherence l?21.
We began Sect. 9.1 with some general remarks about the density matril( of
(9.42)
a system in thennal equilibrium, asking whether or not one could experimentally
it'
1
Vn W \ 911.W)I
2
1
Thus. the full coherence has gone from one-quanlum coherence to two-quantum 2 \ I"
coherence at time two \'{"'11 I , \
'''\ I
1 ,
These same equations show that if we start with pure two-quantum co-
=
herence [£131 (0) f- 0, e:u (0) 0). we can transfer it to one--quantum coherence
911 \
1 \
1
I 1
\1 A V41
,,,""" /< ,
f'JI(tw)=O • (9.44)
Now, "
~21 = (+ -1ll'1 + +) (9.45) 1 \ VII
gives the $-spin transverse magnetization. Thus. a 27r pulse applied to the I- FIg. 9.3a-c The four energy levels of a pair of spin-t nudei. (a) The density matrix d-
spins reverses the sign of the S-spin transverse magnetization. This result is a ements e~1 and 1<'31, which oscillate at l.o.I~1 and ""':II respectively, Rrc CQupled together
i
manifestation of the spinor character of spin systems. It was employed by Stoll
by the lime-dependent perturbation V oscilla.ting at fre<luency .... provided (i) the matrix
clement V23 is nonvanishing and (ii) "" + W11 = lo.IJl. The solid tine (V13 ) ,md the two
et al. [9.9, to] and Wolf! and Mehring [9.11) to demonstrate the spinor nature dashed lines ("11, "31) form a closed loop. (b) H "21, (!:II, and V do NJI form a dosed
i
of a spin panicle. Mehring el al. [9.12J have utilized this fact in an ingenious loop, as in (b) where w is tuned to the 4--1 transition making V41 the driving term, V
Joel! not couple "21 to "31. Thus, V CQuid not pump the one-quantum coherence ~I' (c)
version of spin echo double resonance for electron spin echoes when the electron However, V4 \ can pllmp hi into 124 since the two dashed lines and the solid line form a
is coupled to a proton giving a resolved hyperfine splitting of the proton. closed Iigure
440 441
detennine whether all eij (i • j) were zero for a system in thennal equilibrium. 1ieff = --yllhoFz +haIlzl2z --yh(HIIlz +H2I 2z)
The fact that we can pump n up or down gives us hope. Thus if we want 10
inspect a panicular eij. we look for a sequence of transitions to pump which = --yhhOFz + hahrh: _ -yh( HI; H2 )Fz
eventually enable us to pump the ij coherence (if it is nonzero) into a pair of
levels kl whose coherence, t!kl, is directly observable. Alas. a simple example
will suffice to show that the usual excitations will not do the job in all cases.
--yl{ HI; H2 )(lb - hz) (9.49)
Consider two identical spin i
nuclei (e.g. a pair of protons) with identical The tenn Itz - hz now connects singlet to triplet states, as can be seen by
chemical shifts. Their energy levels are the singlet and triplet states (Appendix evaluating matrix elements between Ill) and Joo) using the spin-up and spin·
H). Designating Ihe total spin by F, we have down functions it and fJ fO express the IFM F) Slates:
F= II + h Fz = liz + 12z (9.46) III) = ,,(1),,(2) .nd (9.50.)
=
For the two spins, we have F = 0, M F 0 for the singlet, F =
I, M F = I, 0,
I
-I for the triplet. 1(0) = j2[,,(I)P(2) - P(I),,(2)] (9.50b)
We now demonstrate a type of matrix element of t! which we cannot probe.
Since We find that
Fz == I tz + I2z (9.47) 1
(II Ill. - 12' 1(0) = - j2 (9.5Oc)
has no matrix elements connecting the singlet 10 the triplet, the only nonvanishing
Yij'S are entirely within the triplet. l11OS, if we wish to inspect a matrix element Thus, if we had initially a nonzero malrix clement (10IU<O)IOO), we could use
the coupling of (9.49) through its matrix element (l111ielfloo) [or (OOI1ielftll)]
(FMFlelF' MF') = (II1e100) (9.48) to produce a matrix element (101 elll), which is one of the standard observable
we could only pump it to stales such as (lOlt!loo) or (I - II£'Joo) in order to matrix elements of e in the triplet (F = 1) state.
generate a closed set of lines. Figure 9.4 illustrates how we could use (Illeloo) Another approach would be to apply a spatially inhomogeneous static field.
to feed (IOleIOO). But (lOlelOO) is not observable either. nor is (I - lIeloo) This would mix the states F = 1, MF = 0, and F = 0, M F = 0. Such an effect
which we could produce as in Fig.9.4b. Thus, we can only succeed in pumping is of course equivalent to having a chemical shift difference between the two
our original unobservable matrix element (9.48) into other nonobservable matrix nuclei. In practice, it would be difficult to make H2 sufficiently different from
elements! HI for there to be an observable effect. A more practical approach would be
In principle we could apply a spalially inhomogeneous ahemating field of to utilize lhe coupling to a third nucleus nearby, provided there is one which is
strength HI at nucleus I and H2 at nucleus 2 so that we had a rotating frame closer to spin 1(2) than to spin 2(1). The third nucleus will cause the singlet and
Hamiltonian triplet Slates of the first nuclei to mix, making the rf singlet to triplet transition no
longer forbick1en, or one could drive the third nucleus with an rf field, to cause
it to excite the singleHriplet transition of the first pair.
101 Ibl The use of frequency selective pulses makes it easy to see how multiple
11'-
lI
I---< 11, -1191001 quantum matrix elements can be generated. It also gives us great control in
generating a particular element of e which we mighr desire. However. it also
II°'I---:<
.
110191001
"':>:>-- 10.01 11,01
V
r--_ \
11019100)
"-
-.;1
\
10,0)
requires that each time we pump n either up or down we be tuned precisely to
the appropriate transition. Thus, referring to Fig.9.3a, to produce the element
1?31 starting from a syStem in thennal equilibrium we might first pump the 1-2
V transition at frequency
1111 --(1119 100 1 11,11---
W=W21 (9.51)
Fig, 9.4. (a) An alternating potential joining the states J10) and lit) can couple (lotf?IOO)
to (t1If?IOO). If we wanted to insp«tan element (tllf?IOO) whidl is itself not directly
to produce en, then pump the 23 transition using
observable, we a~ uJ\$uccessfulsince we can only pump it into an element (tOI(lIOO),
which is also unobservable. (b) If we produced (toteIOO), a furlher IIpplieation of \' tuned W=W32 (9.52)
to the JI- t) -110) trllnsition will produce (t -lleIOO) which is still nOl observable
to produce e13. This would require shifling the frequency W from W21 10 W3to
442
443
something of a nuisance 10 do. In practice, one usually works with large HI's hence that there were no groups of three protons within a given molecule. They
which can flip spins corresponding to many multiplet lines. Such pulses are called thereby showed that the species CCI-I3 was not present.
nonselective. They cover all the lines within an angular frequency width
444 445
n == "(Ho - w (9.54b) M == m, +mS "1' -=
HI m,'+ mS
' (9.57)
representing the possibility of being somewhat off resonance either from the are the same (M' = Ivl), hence are zero-quantum matrix elements.
existence of chemical shifts. or because the experimenter has deliberately chosen For line 2, the terms have
to be off resonance.
Recall that this form of coupling would give exact results for a pair of spins M'=M±1 (9.58)
with dipolar coupling, bUI would not describe twO spins with a pseudo-exchange
hence are one-quantum mauix elements. Line 3 has products of two raising
coupling. See Appendix H. For more than twO spins. the form of the interaction
operalOrs or two lowering operators, hence corresponds to
term ([lS:) is only an approximation even if the coupling is dipolar. However,
though Ihis fonn of coupling gives only approximate results in thai case, it makes M'=M±2 , (9.59)
possible simple explicit calculalions. enabling us to get results whose physical
significance we can explore. which are two-quantum matrix elements. Lastly, line 4 has I z or Sz multiplied
We utilize the spin-operalor method (Sect. 7.26) to follow the density matrix by either a raising or a lowering operator, hence has
in time. DenOling times 0-, 0+. r-, r+, etc. as indicated in Fig. 9.5 to indicate
times just before and just after the pulses at t = 0, T, etc., we take the density M'=M±1 (9.60)
matrix at t = 0- to be again a one-quantum coherence.
e(O-) = (l, + S,) (9.55) Utilizing Table 9.2 we see that at t = 0-, the density mauix consists entirely
of zero-quantum coherence. It is converted to one-quantum coherence by the first
omitting the constant of proportionality. Then the development of e is given in
X(1rI2) pulse. At t = T-, just before the second X(1r!2) pulse, we note £:I(T-)
Table 9.1 up to the time just after the second pulse. We utilize the shorthand
still consists entirely of one-quantum coherence. However, there are now (our
notation terms (lines I through 4) as opposed 10 the single term at t = 0+. Half of these
cos (0,/2) == C; sin (nT) == Sh ' etc. (9.56) terms arise simply because we are off resonance. That is, they come from the
effect of the off·resonance term n(Iz + Sz). The other half arise from the effect
It is useful to express the spin components I z , Ill' etc. in terms of raising
of the spin-spin interaction term alzSz . For lines I and 2, the elements of £:I are
and lowering operators such as proportional to cos (o.T(2). For shon T, those terms survive even if a goes to
I 1 zero. Thus, they would be present even if I and S were nOt coupled together. In
I. = -(P"
2 + r) , I, = 2', (P" - n . the limit of uncoupled spins, line I corresponds to magnetization which is flipped
We have collected the results (or products of twO spin components in Table by the first pulse into the z-y plane, and then is flipped back onto the z-axis by
9.2. Utilizing these results we get the last column of Table 9.1, which expresses the second pulse. Note that as , goes to zero, the combined effect of the two 1r(2
e{,+) in terms of raising and lowering operators. Utilizing them, we see that pulses is to flip the magnetization onto the negative z-direction. They are thus
line 1 of the table conesponds to elements (m,mslelm'/ms) for which equivalent to a single 1r pulse.
447
446
T.b~ 9.2. ProdllCl$ of 5pln oomponcnu cKprcs5Cd in lcrms
We note that while line 3 gives double-quantum coherence, line 4 gives
of r.lslnl .nd lowcr;nll opcr.tors
single-quantum coherence. The distinction between these two lines arises solely
from the precession effect of being off resonance. The lines are identical except
I.S~ .. ~/,(5' +5-)
2 that the tenn cos nT of line 3 is replaced by sin nT for line 4. If one were
I,S, _~/,{S' -5-) exaclly on resonance (n = 0), line 4 would be zero. However, one could men
excite this single-quantum tenn by making the second pulse be Y(1f/2) instead
5,1, _~5,(r +r} of X(7ffl). In the process, me double-quantum coherence would vanish. The
point is mal if one is exactly at resonance, the phase of the second rf pulse will
...
S,I, _~s,(r -r) determine whemer one gets single· or double-quantum coherence. We discuss
these points more in lhe next section.
I 5 "~{/'s'+rs-+rs-+rs')
IS
" .
--~(I"s'-rs'-rs-+rs')
IS .2.(r5'-rS--I·S-+r5')
>' <ti
9.3.3 Evolution, Mixing, nnd Detection of Multiple Quantum Coherence
We have now seen how Ihe first two pulses produce zero-, one- and two-quantum
coherences. We now wish 10 see how these develop in time. Moreover, since
neither zero- nor double-quantum coherences are direclly observable, we want
10 see how me third pulse converts them to something we can measure. The
conversion is conventionally called "mixing". Since we observe nuclei by single-
Line 2 for a = 0 corresponds 10 two uncoupled spins which are flipped inlO quantum coherence. we therefore wish to see how the third pulse converts zero-
the x-y plane by the first pulse and left Ihere by me second pulse (because mey and double-quantum coherence to single-quantum coherence.
point parallel to HI of the second pulse). We merefore conlinue our study of how fJ develops in time under the com-
On the other hand, Jines 3 and 4 are proportional to sin(aT/2). Therefore, if bined effect of me Hamiltonian of (9.54a) and a mird (mixing) rf pulse at time tl
one swilches off Ihe spin-spin coupling (a _ 0), these tenns vanish. They do nOI after the second pulse. Using the spin-operator methods, we construct Table 9.3
therefore arise for uncoupled spins. In order to observe these terms, one mUSI to describe the evolution and Table 9.4 to describe the detection periods. From
choose a time T such Ihal sin (aTn.) is appreciable, hence T'S for which me tables we can follow the effecl of the differenr interactions (spin-spin and
Zeeman) on me time development of the density matrix. We now discuss some
aT/2 ::::: 1rfl, 31fn., elc. (9.61) important features the tables reveal.
3
If a represenlS dipolar coupling as in solids, a ::::: I/r where r is the distance First, we back up a bil 10 express the lime development of fJ. During Ihe
between spins I and S. Thus, by selecting T, one can pick out the values.of.a time interval between pulses, the Hamiltonian (9.54a) is static, hence the lime
or of Ihe distances T which will contribute. For example, the proton pairs 10 development of the density matrix obeys an equation similar to (5.85)
a CH2 group have a characteristic H-H distance. The existence of such a pair ("1,(1)1"') = ("Ie(O)I,i)e,p(i(w,,, - w,,)) (9.62)
will give rise to a characteristic splitting of the proton NMR line into a pair
of lines whose separation is proportional to the coupling. This splitting, first where
observed by Pake [9.14] is often referred to as a Pake doub/~t. Our a.naly~is
(9.63)
shows that the existence of the coupling, hence of the doublet, Will also gIVe nse
to the possibilily of generating double-quantum coherence. The pulse spa~ing r
to produce the maximum coherence depends on Ihe strenglh of the couphng, a, W mrms = -n(m, + mS) + wnJmS = -{}M + am/ms (9.64)
hence on the magnitude of the splitting. Note that one can distinguish coupling
between protons in the same molecule from coupling of protons in diffe~nt where
molecules either by setting T to correspond to the desired distance, or by.studYlOg M=mJ +ms (9.65)
the effect on the Iwo-quantum signal strength of diluting the molecules 10 a non-
proton-containing matrix, as in the experiments of Wang et al. Extensiv.e SlUdies gives the total z-componenr of spin angular momentum. Utilizing (9.62) we see
of this sort have been carried out by BQllnl et al. [9.15], demonstrating large then that the time dependence of fJ between pulses is given by
clusters of protons in a given molecule.
44. 449
Tablt: '.J. Time devdopmml of g durilll tM r:¥OIulion and mWlII period Table 9.3 (ronlinued)
, • , • ,
Number of
quanta in
column 1
Column no.;
Operalor;
I
Column la~l; g(r')
)
01,(/,+1,)
,,(HI,) g(r+ft) •
Num~r of quanta
X(II/2) represented
Line number by column 1
• -(I, S,+ I,S,)Sb2S~ -(I, S,+I,S,)C:Sb 2S~ - (/, S,+ I,S,)C~C;So2S~ (I, S, + I, S,)ChC:So2S~ 'Q
• + (/:S~+I. S:)2S:Sg 2S;
-{/,S~+/~ S,)ShC;So2S~ + (/,S~+ l~s,.)ShC:So2S~ 'Q
OQ
,, +(I,+S,)CoC; + (/,+ S,)~CoC;
'Q • -{/,S,+ I, S,)C~ 2S;SDC~ - (I, S,+ I,S,)c:ch2S:S o C; (I;S~ + I. S:)cb2S;Ch2S;soc: _ (S. + 1.)CbS;Chs;5 0C:
'Q
.."" - (I~ S,+ I, S~)ClDCO 2S~ -(/.S, + I, S~)C;ClQCo2S;
'Q """ (I,S,+ I, S,)ChC;So 2S; (I,S,+ I,S,lC;chc~So2S; -(I. S:+ I:S.)Cb2S;ChC;So2S~.. -(l~+S.)cbs;chc:sos;
'Q
.. ""
(S~ + 1~)Chs:sDS:
- (I.S: + I:S.)2S:ChSg2S;
{S~+I.)C;Chs;sos:
""
" (S,I,+ I,S,l2S;chs:sos:
450 451
exp(i[n(M - M ' ) + a(m~mS - mrms)]t) Thus, we know that the state in which all the spins point up (with MF ==
= exp (W(M - M')t)exp (ia(m~mS - mrms)t) (9.66) N/2) and the state in which all spins point down (M = - N /2) are both eigen-
where the first factor arises from the Zeeman offset. and the second from the spin- Slates. Using a notation Iml.m2.m3• ...• mN) for the m-values of spins 1, 2.
spin coupling. Recalling thai n = 1M - Mil gives the order, n. of the coherence 3, etc,. we know that the state IN/2) in which all the spins point up (i.e. with
we see immediately that the time dependence of the frequency offset uniquely M = N!2) is given in terms of individual spin m values by
labels the order of coherence of the element_ This fact is evident in Table 9.3. (9.720)
The time dependence of the Zeeman offset is given by the terms involving andn
t 1. They are to be found in column 5. and are re-expressed in column 7. Since and the state with all spins point down is
M - M' = 0 for zero-quantum terms. (M - M')ntl = O. hence the Zeeman 1- N/2) =I-!. -!..... -!) . (9.72b)
offset term is merely a conslant independent of n or 1I. This prediction agrees
with the table where the zero-quantum terms (line I) have no terms depending on It is easy to show by explicit evaluation using (9.70) that these two states have
ntl' The one-quantum terms (lines 2. 3. 4. 5. 8, 9,10. II) all involve either Ch Zeeman energies which are equal in magnitude but opposite in sign, whereas the
or Sh. The two.-quanlUm terms (lines 6 and 7) involve (Ch)2 - (Sh)2 = C~o dipolar energies are equal. Therefore. the energy difference between the states.
which gives the time dependence of the N -quantum coherence. is independent
or 2ChSh = S~n' hence are at frequency 2n.
of the strength of the dipolar coupling.
A second important fealUre of the table can be seen by comparing column
We turn now to the detection method period described by Tables 9.3 and
3 with column I. showing the effect of the spin·spin coupling (column 2) on
9.4. This requires converting the various orders of quantum coherence back to
~. Examination of line 6, the matrix elements corresponding to two-quantum
the observable one-quantum coherence, the step called mixing. Since a 1r(2 pulse
coherence, shows that the spin·spin coupling has no effect on ~. To get this
initially convened one-quantum to n-quantum coherence, it is natural to ask if
result using the operator method, one gets eig1l1 terms. They can, however, be
~ would reverse the process. Indeed, in Table 9.3 we analyze the effect of an
added together. Using relationships such as sin 2 +cos 2 = 1 and I~[11 = iI~/2,
X(1f!2) pulse, which is just the inverse of the original X(7f/2) pulse.
they can be reduced to a single final term identical to the starting term. One
The result, in column 7 of Table 9.3. is labeled in column 8 according to
naturally wonders if there is some deeper significance - indeed there is.
the orders of the resulting quantum coherence. For example. line 3 represents
The effect of the spin-spin coupling is given by the second factor of (9.66)
zero-order quantum coherence. II would be undetectable.
exp(ia(m~ms - mrms)l) • (9.67) In Table 9.4 we keep only the coherences which are single quantum during
which. for the two-quantum term, has the detection interval, but label the rows by the order of the quantum coherence
mr = ms = i and m~ = ms = -i (9.68)
during the evolution period.
In column 3 there are tenns such as (l~S~ + I~Sz) which eventually give
or vice versa. But no signal because when we form Tr {g(J+ + S+)} the [z and Sz cause the trace
452 453
For the one-quantum coherence the conversion of the initial z magnetization to one-quantum coherence for the
{F+>' = 2e- iS7t2 cos [a(t} - r - t2)/2] cos (ntd sin (f}r) (9.76) preparation period, the fact that zero-, one-. and two-quantum coherences have
been brought into existence for time interval tl' and the conversion back to
which, in complex fonn, gives single-quantum coherence for observation.
(F+)I = ~[e-i(n-aj2)12(ei(!Haj2)tl +e-iC l?-aj2)1 1) We have seen that the pulse sequence X(-rr/2) ... r ... XC1r/2) will produce
zenr, one-, and two-quantum coherence. As we remarked earlier. if we study
x (eiCS7+aj2)T +e- i(S7-a/2)T) column 10 of Table 9.1, we note that the zero- and two-quantum coherence terms
+ e-i(J?+a/2)t2(ei(l?-aj2)t1 + e- i(l?+aj2)t 1 ) (lines I and 3) are proportional 10 cos f}r, whereas the one-quantum terms (lines
x (e i(S7+a/2)T +e-iCS7-a/2)T)]
2 and 4) are proportional to sin nr. Therefore. ifn = 0 (exact resonance) we
would not produce one-quantum coherence. The question, how, when we are
For two-quantum coherence we get tuned exactly to resonance. can we produce one-quantum coherence with a pair of
spins. turns into the question. how do we produce odd-order quantum coherence
(F+)2 = _2ie- int2 sin (at2/2) cOS(2f}tl) cos (f}r) sin (ar(2) for an arbitrary number of spins. when tuned exactly to resonance? We turn now
= ~ [e-icn+a/2)t2 _ e- i(l?-a j 2)t2] to a simple approach.
8 The disappearance of the one-quantum coherence for a pair of spins can
x (e2int1 + e-2int1 )(ei(S7+a/2)T _ eicn-aj2)T be traced back to column 5 of Table 9.1 where we see the effect of the spin-
+e-iCS7-a/2)T _ e-icn+a/2)T) spin coupling alone [i.e. before the resonance offset term f}r(1z + Sz) acts]. It
(9.77)
produces (ly + Sy)C; on line I and -(lz:Sz + IzSz:)C:; on line 3. These tenos,
Referring back to (9.62), we recognize thac the elements of the density matrix if acted on by X(1r/2), produce (!z + Sz)C; (a zero-quantum coherence) and
will have time variations -(!z:Sy + I y S z )2S:; (two-quantum coherence).
e+ i(l?±a/2)t To get the one-quantum terms we allowed f}T(lz + Sz) 10 turn I z into -Iy •
(9.78.)
I y into I z , then we applied an X(1r!2) pulse. Therefore, if we have f} = 0, so that
for the states M - M' = +1, will have I z and I y are not rotated by 1r/2, we should get one-quantum terms by rotating
e- i(l?±aj2)t the pulse from being an X(1r!2) to being a Y(1r/2) pulse. Indeed, we see that a
(9.78b)
y(7I'"!2) pulse would leave (Iy + Sy)C; alone as a one-quantum coherence. but
for M - M/ = -1, will have tum -(IzSz + I z S z )2S; into (IzSz + I z S z )2S:;, a one-quantum term.
e 2int (9.78<:) In general, of course, f} =f O. Then the points we have just discussed show
that production of a particular order of quantum coherence depends on the amount
forM-M/=2,and one is off resonance. That is frequently a disadvantageous situation. It can be
e~2int remedied by making f} effectively zero by means of a spin echo, then choosing
(9.78<1)
the phase of the second pulse to get the desired order of coherence. Thus. we
for M - M/ = -2. We know from the Hermitian property of f! that whenever can use a pulse sequence
we have generated an M - M/ = p matrix element, we have also generated an
M - M' = -p element. We note next that all three {F+)fl'S have the factors X(,/2) . .. T/2 ... X(,) ... T/2 . .. [Y('/2) or X(,/2)]
exp [ - i(f} + a!2)t2] and exp [ - i(f} - a(2)t2]. This corresponds to the fact that to produce the multiple-quantum coherence. One can also then eliminate the
during t2 the spins are precessing in a given sense. offset effects during the detection period by introducing a 1r pulse Cfor example
We can identify the role of the preparation interval by looking at the r- at t2 = T12) to give an echo (at t2 = T).
dependent terms. Note that these all involve single-quantum terms since their
time dependence involves the frequencies (f} ± 012). We note that the time de-
9.3.4 Three or More Spins
pendence tl tags the evolution period, and the order of the coherence is imme-
diately obvious from the f} dependence which is exp(±Oif}tl) for zero order, We have seen that a pulse sequence X('rr/2) ... r ... X(1r/2) will produce zero-.
exp(± Iiiltd for single order, and exp (± 2intl) for double-order quantum co- one-, and two-quantum coherence from a pair of spins if T is sufficiently long.
herence. Thus, if one studies how (p+) varies with tl. one can isolate the zero-, What happens if there are a larger number of coupled spins? Will this sequence
one-. and two-quantum coherences. The expressions for (F+)p show clearly produce higher orders, or must one add more pulses? We shall show that the
sequence still works.
454 455
1
i
We approach this problem two ways, first by considering three spin nuclei We have written (9.86) in a manner intended to help indicate how, using the spin
with a coupling of the form haikljrlh. which we can solve exactly, then we operator formalism. each term arises.
will look at the general case which we will treat by an approximation method. We note first of all, that only spin I has either x- or y-components. Secondly,
Let us stan then with a Hamiltonian for three coupted spins which general- we note that only coupling constants 0t2 and 031 involving spin I occur. This is
izes that of (9.54a): because terms such as 023Ir2Ir3 commute with Iyl' Thus. if we had four spins,
we would get from I yl terms involving 012, 013. and al4 but no terms involving
1t:::: -hillzT + h(aI2I%1Ir 2 + a23Ir2Ir3 + a3lIr3Irl) 023, 02~, a3-l.
where We note that we can easily find the contributions from I y 2 and I y 3 from
IaT :::: Icrl + I cr2 + In 3 a:::: :to y, Z • (9.79) (9.86) by cyclic permutation of the indices. Expressing Izl and Iyl in terms of
raising and lowering operators shows us that all terms of (9.86) connect states
(Note, we shall assume below that the coupling between any pair, such as al2.
can also be wOllen as a2l.) As can be seen from the case of two spins, it is the
M and M' which satisfy
spin-spin coupling terms which lead to multiple quantum coherence. Therefore M::::M'±1 (9.87)
we need to focus on the second term of (9.79). Then, as with (9.55), we take
e(O-), just before the first pulse, as Thus all are one-quantum operators.
The frequency off-set term in 1i, -hillrT' will cause a precession of the
,,(0-): I'T (9.80)
x- and y-eomponents of spin, as can be seen from column 7 of Table 9.1. It
so that at t :::: 0+, just after the first X(.,./2) pulse, will leave the Irk (k:::: 2,3) terms alone. but will conven 1%1 or Iyl inlo linear
(9.81) combinations of IZI and II/I so that we should replace Izl by
e(O+) :::: IlIT
Then we wish to calculate 1;1:) cos ilt - I yl sin m (9.88)
exp (-io 12/r I I r 20 exp (-ia23 Ir2Iz3t) exp (-ia31 Ir3/%1 t)/1/1 +cos (a31 T/2)2 sin (OI2T(2)[ cos (at)I zl I r2 - sin (ilt)III1 I z 2]
x exp (ia31 Ir3Id)exp (i023Ir2Ir3t) exp (iaI2I%1/r20 (9.84) +2 sin (03 I T/2) cos (012T/2)[COS (ilt)I>:\I z2 - sin (nt)Iyl l z 3]
In (9.84) the exponentials involving I r2Iz3 commute with Illb hence anni- -2 sin (031 T/2)2 sin (012T/2)[ cos (ilt)II/I Ir2Ir3 + sin (nt)Iz1Ir2Ir31
hilate one another so that (9.34) reduces to + tennsobtained by permuting the indices 1,2,3 (9.90)
exp (-iaI2/r 1I r20 exp (-i03 I Iz3Ir 1t)Iyl This Hamiltonian is now acted on by the X(rr/2) pulse. In order to see what
X exp (i031 Ir3Irl 0 exp (i012Iz I Iz2t) (9.85) spin coherences are thereby produced, we need only examine what happens to
the spin factors. TIle result is a table of operators before and after, and the order
We can now apply either of the (I,3) operators to get a result similar to column
of quantum coherences which can arise from each term. Denoting the values
5 of Table 9.1. Then we apply the (1,2) operators 10 that result. TIle final result
is M-M'::p
,
Ix, .. !(I,'+/,)
" ajkT/2 ';;t 1r/2 (9.94)
l 1.,1,2 ,x,/~2" 2.. (I,' + I n(l; -/,-) :1:2,0
If, however, two spins are close, the third far away, two of the three cou-
• "
-1"/~l· -~/,,(1" -/,-) pling constants are small, so that while two-quantum coherence can be produced
,
1~1 "l
" quickly, three-quantum coherence is produced slowly.
lx' 1<) lx' I,l • 2..
.;
(I" + /,-)(11' -11> :1:2,0
We can also analyze the two-pulse method for nuclei whose spin is not
-I"/~I" -~/,,(/1'-ln restricted to I = ~,and for arbitrary fonns of spin-spin coupling, such as dipolar
6
,
1,,/<)
" coupling instead of coupling such as aIzSz , by using a power series expansion.
I" /,//'1 -I" 1~2 I~I· -~ I" (l i -1 2- )(I{ -Ii) :1:2,0
Suppose then we have a Hamiltonian in the rotating frome
8 I", 1,1/'1 Ix,IY1IY1·!(/,' +1'-)(/2' -/ 2 )(/1' -Ill
8
±l, ± 1
1-l = -hfllzT + L 1-ljk == 1-l: + 1-l88 (9.95.)
j>k
where the fl represents being off resonance and 1-l88 and 1-ljk'S are the secular
1 + It p= +2 0'
part of the spin-spin couplings. That is
4i II 2
[1-ljk' IzTJ =0 , [1-l 88 ,I:Tl =0 (9.95b)
.!-I+ r p=o 0' For example, for dipolar coupling
4i 1 2
1 1-l j k = B j k(3I:jlzk - Ij' h) (9.96.)
4i II Ii p= -2 (9.91)
= B j k(2Iz jlzk - IZjIzk ~ IyjIyd (9.96b)
Once one has made such a table, one can work out the result in one's head
As in (9.7b), we then have quantum numbers Al and 0: where
since an Izk or an Iyk gives ± I, so a product of two such operators gives p
values of 1 + I = 2, 1 - 1 = 0, -I + 1 = 0, -I - I = -2. IzTIMo:) = MIMo:) and 1-l ss IMo:) = E(> IMo:) (9.97)
We see that the three spins give us p values ranging from +3 to -3, as
Then, the time development up to just before the second X(1r:I2) pulse is given
we proved earlier. Now, however, we can see how long T must be to produce
by
Ihe coherence. Note that if a12 ~ GI3, that the three-quantum coherence will
develop in a time when sin (aI2T/2) ~ 1 or UI2'/2 ~ 7r/2. If al3 «: al2, then
we must wait until a13T/2 ~ 1r/2. In the process, as time develops, sin (aI2T/2)
{?(r-) = exp ( -* 1-lr) X(1rI2)I:TX(rr/2) exp (* 1-lr)
will have been through several oscillations. Thus, we must wait for the weakest
=ex p ( -k1-lT) IyT exp (*1-lr) (9.98)
coupling.
Suppose, however, we had three spins in a line so that spins 1 and 3, on
opposite ends of the chain, are weakly coupled. Then we expect = exp (ifhIz) exp ( -*1-l88 T)IYT exp (* 1-l88 T) exp (-iflr I:)
(9.92) The spin-spin exponential can be expanded using the theorem (see for example
[9.16])
Then our expression (9.90) would imply GI3 would control how long it would
take to generote a three-quantum coherence. Thai conclusion would, however, eABe- A = 1 +[A,BJ+~[A,[A,BlJ+ (9.99)
be wrong since, had we analyzed the time dependence of I z 2 instead of I zl ' we
to give
would have gouen a three-quantum coherence dependent on
458 459
iI7r1 Equation (9.107) is, however, the condition for one-quantum coherence. In a
e(r-) = e+ • { fYT - *r['Hss, IyT)
similar way, one can show that every tenn in the series (9.101) generates one-
+ (.!.)2
tl 2!
..!..-[1i,u, [1i [yT]J + ... }e
88
• -inrI, (9.100)
quantum coherence, a fact to which we will return.
Returning to the two-spin Cllse, (9.100 and 103), we can now apply the
Zeeman operators exp(inTI.:T) to get for the density matrix at time T-
Consider now the first commutator
[H.., IyTI = L [Hjk,!,,] (9.101) e(T-) "" [yT cos DT + [zT sin DT - ~ fl I2[Izl [.:2 cos DT
,
i>k
- [yI I t 2 sin Dr+I.: I [z2 cos Dr-IdIY2 sin Dr] (9.108)
If 1 is different from both j and k. the commutator vanishes. Thus, we have Then, the X(1r!2) pulse will produce
typical lenns such as
e(T+) "" - I.:T cos DT + IzT sin DT
(9.102) 3.
- 2'flI2[(IxIIy2 + I y 1Iz 2) cos DT
which are just whm one would have for .a pair of coupled spins. It is straightfor-
ward to evaluate the commutation for a specific choice of H 12. Thus, for dipolar + (Itl I y2 + I y1 I z 2) sin Dr] ... (9.109)
coupling, we get Expressing the [z and I y operators in tenns of raising and lowering operalOrs,
[1i12,(IyJ + [y2)] = BI2 [2Izl I z2 - 1:0:1/2:2 - [yl[y2,(1y1 + [y2)] (or using Table 9.2) we have
= B I 2[2(Izl,!yl)Iz2 - Uzl,Iy d I :t2 e(T+) "" - ItT cos DT + Ix'!' sin DT
+2(Iz2.Iy2)Iz1 - (lzz,!y2)Iz d 3i T{ + +
= -3iUz1 l z2 + 1:1 [z2)BI2
~ 2'Bt22i (II I 2 - II Ii) cos DT
which, using (9.97), which says 'H S8 is diagonal in AI, gives [Hjk, [Hln!' [ynn ' (9.112)
Iy'f "" ~i([,t - IT) so that (9.106) ['H 12' ['H 12,Uy l + I y 2)ll (9.114)
or three-spin terms such as
M""M'±l (9.107)
(9.115)
'60 461
In this mannerlwe see that each higher order or T in the series (9.IOO) adds one
X(It Ii !jJX
more possible spin to the couplings.
We can now see another userul point by looking in detail at (9.103). AI· = (1z1 - iI: I )(lz2 + il:2)(I",3 - il:])
though lhe top line involves products of three spin operators, the commulation .= 1",1/",2/",3 - i/",t 1",21:3 + i/", I 1:2 /",] + 1"'11:21:]
I
step reduces the prodUCI to two operators (e.g. ZI with [22)' In a similar way, - iI: 11",2 1 z3 - 1:1 1z2 1:] + J:t /:21~3 - il:II:21:3 (9.118)
one finds that the three·spin tenn (9.115) involves three spin operators, such as
[zll1/2Iz]. In general each higher tenn in the series involves PrOOtiCts of the When we express [zl> 1",2, and /",3 with raising and lowering operators, we see
previous tenn Wilh a pair of spin operators, bUI lhen the commutator reduces the explicitly how the three-quanlUm coherence arises. TIle lenn (9.ll7b) will give
num~r by one. Thus, each successive lenn has one more spin operator in the XIi h:IJ:X : (lzt - iI,t)ly21yJ
product. Since it takes three spin products to get a three-quantum coherence, the
: [.1;\ I Y2I y] - il: 1l y2 / Y:1 (9.119)
T 2 lenn is the first one in the series which can give three-quanlUm coherence.
The T] tenn is the first one which could give four-quantum coherence. Note that TIle tenn l"'t1Y21Y3 contains three-quantum coherence.
we say "could", implying that il does nO! necessarily do so. Thus, if one has From our analysis of the power series expansion, using (9.9a), our con.
two spin-j nuclei, the highest-order coherence one can produce is second, hence c1usion is. then. that the pair or pulses X(1:(l) ... T •.• X(7(fl) will generate all
the T 2 , T , and higher tenns cannOI produce more lhan second-order coherence orders of quamum coherence permitted. While it is a question of proper choice
(double-quantum coherence) in this case. To find what coherences are aClually of T to achieve lhe desired orders, in general, all allowed orders will be presem.
realized, one mUSI look in detail at the lenns or use the general rules about the We thus have IwO problems yet to consider. (I) How can we detecl a particular
maximum coherence 2NI of a group or N nuclei of spin I. desired order? (2) Is it possible to generate a particular order without generating
Suppose in (9.115) one has a three-spin prodUCI involving coordinates of other orders? We tum to these IOpics in the next two sections.
spins 1,2, and 3. Expressing the 11:1: and lyA: 's in tenns of raising and lowering
operators, one might ask, will there be tenns in ()(T-) Uust berore the second. 9.3.5 Selecting the Signal of a Particular Order of Coherence
X(1fn) pulse] such as
We turn now to a discussion of how [0 select the signal arising from a panicu-
ItriI: ? (9.116) lar order of coherence. We have already mentioned one approach at lhe end of
The answer must be "no" since this is a three-quantum tenn whereas we have said Sect. 9.3.3, where we noted that if the signal is Fourier analyzed with respect to
each tenn in the infinite series prior to the second X(7fn) pulse has one-quanmm the evolution time t], the coherence or order p gives rise to lines at frequency
coherence. Permissible terms then would be of the general fonn pil, where n is lhe amoum one is off resonance. Thus. introducing a deliber-
ate resonance offset by an amount n
large compared to the spin-spin splillings
IiliIt or (9.117a) produces groups of lines in which the spectra of the different orders are well
separaled. This approach, and a variant or it (lime proportional phase incremen-
It 1:2[,] (9.l17b) tation, TPPI) in which the offset is actually zero but is made 10 appear nonzero,
has been widely employed. especially by the Pines group [9.5,7J. We discuss
These may be thought of as grouping a raising (or lowering) operalor (e.g.
TPPI at the end of this section.
Ii) with either zero-quantum operators I:k
or with paired raising and lowering
AnO!her powerful method was imroduced by WotOIUi and Ernst [9.I7J. We
operators (e.g. Ii It)· turn to il now. Their method is based on a recognition that the phase of the
When these are operated on by the Zeeman operators. they simply multiply
the r:or I; by a phase exp(inT) or exp(-Wr) respeclively, or leave the
I:k's alone. Thus. the second X(7rn) pulse still acts either on expressions such
multiple quantum coherence depends on the phase of the exciting pulse in a
manner which depends on the order of the multiple quantum coherence. Their
concept also underlies the scheme for production of a selected order of coherence
as (9.117a or b).
which we take up in the next section.
Now
The basic Wokaun-Emst scheme may be described as follows. Let us denote
xIi X: X(l~t + i/y 1)X: 1",1 - il:1 by Xf>' a 7(/2 pulse applied about an x'-axis in the rolating frame where the Xl.
axis lies in the x-y plane. making an angle ¢ with the x-direction according to
X/ 1X: /"'1 + il,l so lhe equations
x' "" x cos ¢ - y sin ¢ (9.120.)
462
1
I;,:I = I;,: cos ¢ - I y sin ¢
= eil.t/J I;,:e-il.t/J
(9. 120b)
(9.12Oc)
I Then, utilizing the fact thm
X = eiJrTff/2 (9.125')
Note Ihal Xl is rotmed from x in a left-handed sense about I z by an amount 1J. we have
Wokaun and Ernst apply a sequence
Xt/J = exp [i!(l;,:T cos ¢ - I lIT sin 1»] (9. 125b)
Xt/J ... r ... X",,,.tl ... X ... t2 (9.121) Defining
recording the complex signal 5(T, 1>,tl, t2) as a function of t2, for fixed values R=ei/.Tt/J (9.125c)
of the parameters T and tl, and for a succession of appropriately chosen phases
1>. Let us write the signal for ¢ = 0 as a sum of contributions G p from the various we gel
orders of coherence p (p = - N 10 +N)
X", = exp(iRI;,:TR- I 1r!2)
+N = ReilrT1r/2 R- 1
5(.=0)= L G" (9.122)
p=-N = eil.Tt/J Xe-il.Tt/J (9.125d)
and then consider 5(1J) to be known as a continuous function of 1J. The essence We now wish to express the ¢ dependence of the signal, 5. Abbreviating
of the Wokaun-Ernst method is then that the Gp's can be found as transfonns of 5(T, 1>,tt> i2) as 5(t/J), we have
5(.)
,. S(t/J) = Tr { If [e -(i/h)'H"J2 X e -(if h)'H"tl J£:I(T+, t/J) [inverse]} (9.126)
G =
P
-!...
h
J 5W,e-; pod. . (9.123)
where (I(T+,t/J), defined as
o
Since in practice Ihere are 2N + I values of p, measurement of 5(1)) for 2N + 1 f!{,,+, t/J) == Xt/Je-(i!h)'HT X",IzTX;le(i/h)'HT X;I (9.127)
values of 1> should suffice, as we show below. is the density matrix jusl after the second 1r/2 pulse, for the case that both
To understand the Wokaun-EmSI theorem, we need to
11' pulses have phase shifts t/J. The tP dependence of S(t/J) therefore arises in
1. examine Ihe general form of the contributions of a given order of coher- £:I(T+, ¢). Expressing Xt/J utilizing (9.125d) and utilizing the faci Ihal I zT ' and
ence, thus exp (iIzT¢), commutes with H, we get
2. examine the effect of a phase shifl on the differenl orders of coherence, f!{T+, t/J) = eil,Tt/J Xe-(i/h)'HT X IzX-leCi/h)'HT X-le-i/,Tt/J
,nd
3. show how to ulilize Ihis knowledge to pick OUI the contributions G p of = eil.Tt/J £:I(T +, O)e-il.Tt/! (9.128)
the particular order of coherence. We recalllhm (1(,,+,0) is the density matrix just after the second X pulse,
We consider a system characterized by a Hamiltonian in the rotating frame hence it contains all the multiple quantum coherence. In general, for a given
spin system, £:1(,,+,0) consists of a number of terms representing not only the
1{ = - E hDkI:k + E 1{jk == 'Hz + 'H u (9.124a) different possible orders of coherence, bUI also the various ways of generating a
k k>j
given order. Thus, for a three-spin system, of spins !'
one can generate p = 0,
where for the spin-spin interactions 'Hjk we keep only the secular terms so Ihal ± 1, ± 2, ± 3. Consider p = I. It could arise from a variety of operators such as
['Hjk> IzT] =0 (9. 124b) It It12: It h.hz ItIiIi (9.129)
Then, we have the quantum numbers NI, a such that plus all the operators one can get by permuting the labels 1,2, and 3. Utilizing
a symbol f3 to designate all such various ways of generating a given order, we
(9.1240)
can clearly break £:I(r+ ,0) into a set of components Yp{J of given p
464 465
Now, a given order, p, is distinguished by having p more raising operators by p, and in which the coefficients of the hannonics are the Gp's. Therefore, if
than lowering operators. (If p is negative, there are more lowering operators than one knew S(tjJ), one could deduce the Gp's by taking the Fourier transfonn of
raising operators.) 5(<1).
We can then easily evaluate g(r+,4J) as Thus, suppose we have deduced S(f). (We can do this by measuring at
e< r +, f) "" eiT'TtI> g( r+, O)e -it ,T.;P = l.:e
i ' ,1'.;P9p13e -i/,T.;P (9.131) a number of discrete values fi and using a smooth interpolation procedure to
estimate S for values between adjacent pairs of values fi') Then consider I, the
P.P transfonn of S,
But,
h 2.
ei/'T h±eif,.;p "" ±e±i.;p eif'T Izke-i/'T "" Izk
Taking expressions such as (9.129), inserting factors exp (-ilzTf) exp (ilzTf)
(9.132)
1= J5(<I)e-;P'Od<l (9.137)
between each pair of spin operators, and using (9.132), one gets results such as
o
Substituting (9.1 36a) for S(f) we get
e i/'T4> It h ze- iT""4> "" It h ze i4> 2.
(9.133)
I = L:G p Je;(p-p')Od</
p 0
That is, the effect of the f rotation for any operator with p "" 1 is simply to
= 21rG I + '" G p [e 21l"i(p-p') - I]
multiply that operator by exp (if). In a similar way, the rotation operators acting p ~ itp' p)
p f. pi
on a p "" 2 operator merely multiply it by exp (2if). In general, for the operators
= 21rGpl (9.138)
9p/3'
0'
eif,T4>Yp/3e-i/.T4> "" 9p/3e iP4> (9.134) 2.
11<,+.<1)=
p=-N
t (L:9pp)e;po
/3
(9.135) there is an upper limit to the quantum order [the quantity N of (9.135)]. In
Olhers (for example an effectively infinite solid of dipolar coupled nuclei, such
as CaF2) N has no limit. Suppose we had the fonner case. Then, there are 2N + 1
where N is the maximum allowed value of p for the particular spin system. values of p, consequently 2N + I Gp's. Clearly 2N + 1 measurements of S(f)
Substituting this expression into (9.126) we get the signal as a function of f to should suffice to deduce the Gp's.
be Suppose, then, we pick the 2N + 1 values of f, which we label fk' as
2.
5(<1) = pEN e;poT{r:t [e-O/")"" Xe-Olh)'Ht, ] (1gpp) [iowse]} <I, = '2N + 1 (9.140.)
N k=0,1.2..... 2N (9.140b)
"" L G p e iP 4> where (9. 136a)
p=-N Note if we chose k "" 2N + I, we would have a value of f "" 21r, which produces
the same shift as ¢ "" 0, a value we have already obtained from k "" O. Then
G p"" Tr{ r.t [e-(i/h)"Ht 2xe-(i/h)1"f.t l ] (~yp.B) [inverse]} . (9. 136b) 5(<1,) ,,5, = L:Gpexp(ip<I,)
p
+N
The individual tenns 9p/3 depend on r, on the ilk'S and the strength of the spin-
spin couplings as illustrated in the tables we have worked out for g(r+) for the
= L: Gpexp [ipk2p/(2N + 1)] (9.141)
p",,-N
two- and three-spin cases. Examination of (9.136a) shows that it says that S(tjJ)
can be represented by a Fourier series in tjJ in which the harmonics are labeled This is a finite series. Consider then the transforms
466 467
I:S(1,)e>P (-ip' 1,) = L: G e>p [i(p - p')1d 5(0)+5('11") = L:Gp(eipO+eip1r)
k k,p " p
= L: G p L: "p [i(p - p')kh/(2N + I)J (9.142) = L: GpO + I) + L: Gp(l - I)
p , p even p odd
= L: I' = L: I' - L: I' Clearly for p = 0 this will give 4Go. For p '" I this will give
1.=0 1,,=0 k=2N+ 1
G 1(1 +ei1r/2 +ebr +e i31r / 2) (9.150)
= ) I f2N+l
=
(
L:I' (1_1"'+1)= -1-1 if 1. 1 (9.144) The four complex numbers may be thought of as four vectors in the complex
1,,=0 plane (1. i, -1. -i) which. when added tail of one to head of the prior one,
Examination of (9.144) shows that as long as p 1= pi the exponent in f is form the sides of a square, giving a zero result. Thus there is no signal from Gt,
neither zero nor a multiple of 21l', hence f =F 1. But first-order coherence. Likewise. there is no signal from p = 2 or p = 3. But for
(9.145) p=4weget
f2N+l = e i (p-pl)211" = 1
G4(eiO +ei4..-/2 +eiS..-/2 +e il 2..-/2)
hence the sum over k in (9.144) vanishes.
If p = pi, f = 1, and the sum over k is trivially 2N + 1. Thus (9.151)
That is, all four terms exp (i4¢) are equal 10 I.
( - -I - )"'5(")
(9.146)
G I = L. '1'1. e -;P'¢' with Thus we will record Go and G4, but not Gj, G2. or G3. In a similar way
p 2N+l k one can show that we will not record Gs, GG, G7. but will record Ga. In fact.
as we see from (9.151), we will get a nonzero result whenever all the tenus are
1k=k2,/(2N+I) , k=0,1, .. .,2N of a given order are 1. Since for a phase shift increment of 8
Therefore, measurement of S(¢k) for these 2N + t values of <p will enable one L Gp[l + eip8 + (e ipO )2 + ... J (9.152)
to decompose S into the contributions. Gp , from the various orders of coher- p
ence. Since Gp is a complex number, one must record signals S(r/J) as complex we will get signals when
numbers, Le. in quadrature, to evaluate the fannula of (9.146).
It is sometimes useful to carry out a simpler version of adding signals. One e ip8 = 1 or (9.153a)
picks a set of phases (e.g. O. rr(2. '11". 3'11"/2) and then a~ds the signals together.
Sometimes one adds signals of one phase but subtracts signals of another. To see pIJ=kh , k=0,1,2,3 (9.153b)
how this works it is useful to keep in mind a graphical picture. From (9.136a) Thus, if we get a signal for p = 2tr/O, we will also get a signal for 2p, 3p, etc.
we have One sometimes speaks of a sequence as being "pk selective", since if it gives
+N order p it will also give k times any integral value of k.
5(1) = L: Gpo;P. (9.147) Note that as long as we add all the signals, we get Go. the zero-quantum
p=-N signal. To eliminate it we add half the signals and subtract the other half. Some
hence each order can be represented in the complex plane by vectors exp (ip¢). cases given by Wokaun and Ernst are
Consider then an example with ¢ = 0 or '11". Then
468 469
add q, =0, 1f gives p= O. 2. 4. 6, 8 and longer times one finds one exciles most of the low-p elements of e, as well
add q, = O. subtract; = 1f, gives p = 1,3.5,7,9 as matrix elements of progressively larger p. Thus, the stalislical order is being
add q, = O. rn. 11", 311"n gives p = 0, 4, 8, ... shared by more and more elements of e. As a result, individual elements gel much
add q, = 0, 1f, subtract; = -rr/2. 3rrn.., gives p = 2,6, small~r. One lends to lose in two ways. (I) During preparation one spreads the
add q, = 0, 2-rrf3, 4rrl3, subtract if> = -rr13 + (0, 2rrf3,4rr(3), gives p = 3, 9. order among many different elements of f!, and (2) the detection pulse feeds back
These simple sequences are particularly useful under circumstances where many elements (Malf!IM'o') in which though M - M' = I) may be the same,
the allowed values of p are limited. Thus, if p = 9 can be ruled out, the last there is desu1Jctive interference from the range of values of the other quantum
sequence gives one just p = 3. numbers, a.
We remarked mat one method of displaying spectra by quantum order is It is thus clear that 10 get strong signals one should first of all restrict the
to tune off resonance by an amounl n.
Then lines associated with order p can elements of f! which one excites to just the desired ones. For example. if one is
be identified by Fourier transforming the signal with respect 10 Ihe evolution interesled in the six-quantum spectrum, one would like to excile only the six-
time tl, since they occur at a frequency offset of pil. As we remarked, there are quantum coherences. Then. having excited the desired coherences, one would
disadvantages with working off resonance. A solution is to use a technique called like to read them out at a later time in a manner which avoids any further loss
time proportional phase incrementation (TPPI). TIle idea is to work at = 0, n of coherence. We turn now to these mailers. Many of the ideas involved here are
exact resonance, but in collecting dala for the variable t 1 (the evolution lime) to from the studies of Pines and his students. The two review articles by Weilekamp
introduce a phase shift which depends on tl. If the phase and tl obey a linear [9.5] and by Mwwwilz and Pines [9.7b] are especially useful for study, as are
relationship the basic papers by Warren el a!. [9.18.19) and Yen and Pines [9.20).
(9.154)
9.4.1 Genef'::.ting a Desired Order of Coherence
Then as tl is advanced the extra phase is just as though one were off resonance
by an amount Llw given by Can one generate one and only one desired order of coherence? We have seen
that it is possible to delcct an order 1J and integral multiples of p by cycling
Llw= A (9.155) the phase of the exciting pulses relalive to the detection pulses. We will now
Then, when one does the Fourier transform of the signal with respect to tl, show how to use phase cycling to generate coherence of a given p and multiples
an effective frequency offset obeying (9.155) is introduced. This approach is thereof.
unaffected by inseniOIl of 'II" pulses to refocus magnetic field inhomogeneities FIrsl. however, it is useful to introduce anOlher way of looking at generation
or real frequency offsets aI any part of Ihe cycle (Le. al me midpoint of the of multiple quantum coherence. Utilizing two pulses. we have seen how the firsl
preparation cycle, the evolution cycle, or the detection cycle). 1fn. pulse convens the density malrix from its initial value, I:, to I z , producing
one-quantum coherence. As time develops, spin-spin correlations develop in lhe
density matrix as is evident either from the spin operalor fonnalism [e.g., (9.90)]
or from lhe perturbation expansion (9.100). However, the density matrix still
9.4 High Orders of Coherence possesses only one..quantum coherence. The second 1fn pulse, by acting on one-
We have seen some simple ways of producing and delecting multiple quantum quantum terms like I:II:2It which express spin-spin correlations (in this case
coherence. For many purposes lhese memods suffice. For example, if one wished of three spins) produces multiple quantum coherence through operators such as
to distinguish protons in a CH3 group from protons in a CH 2 group. one need It Ii It· We can express mis fonnally by writing the equation for e(T+), where
only to be able to produce and detect two- or Ihree-quantum coherence. What T+ is the time just after the second 11"/2 pulse. For convenience we will take the
happens if one is concerned with much higher orders? II is easy to see that second pulse to be an X, though the argument we present could be done, a bit
there is a fundamental problem. Each quantum order contains "information" less elegantly, with an X pulse. Since X undoes an X pulse. we have
- that is. it represents order in the statistical mechanical sense. But we are X = X-I so (9.156)
confident mal the tOlal slatistical mechanical order will never exceed its initial
value representing the sample's initial magnetization. If one pumps only low-
order quantum coherences, the statistical order is shared by only a modest number
e(T+) = X-I ex p ( -*1iT)X I:X- (*1iT)X
I
eX P (9.157)
of (MaleIM'aI)'s. However, if one sets the time, T, of the preparation to longer Suppose for concreteness we are exactly at resonance, with 1{ being only the
spin-spin coupling. We take
470
471
1t = 1t zt (9.158) coherence?" In fact, as we shall see. we can find a way to generate coherence
where 1t zz could be the usual secular dipolar coupting, such as of order pk. where k = 0,1 •.... By special means, one can eliminate k = O.
Then. the means to distinguish among the other k values is by the duration of
1t zz = I:: Bjk(l~jI~k + IyjIyk - 2IzjIzk ) (9.159) the e?tcitation period.
j> k How then do we limit generation to an order pk? We will find that there
Then are two parts to the task. The first part is concerned with finding a way to limit
-~1tzzr)x _~(X-l1tHX)T]
the order, and the second part. with making the amount generated large. In order
X-I ex p ( = exp [ to keep the discussion sufficiently general, we will therefore not as yet specify
the Hamiltonian. However. we suppose that as with 1t yy it can generate various
= ex p ( ~~1tyyr) (9.160) orders of coherence, p. That is
so that 'Ii = I: 'lip (9.166)
472 473
'H<p = L 'Hpeip<p (9.170) (9.176.)
p
the complex numbers of (9.175) fonn a closed polygon of exterior angle P<P'
This expression is reminiscent of (9.136) and the sum in (9.175) vanishes unless
S(,,) = L: Gpe;P. eip,p = 1 (9.176b)
p
Therefore, if we pick a p, which we call ]Jo, that we wish to generate, we
The way we got selectivity in signal detection was to form a sum of signals for a
just need to select <P to satisfy (9.I76b). Therefore if
sequence of phase shifts. How can we use this idea for Hamillonians? We have
already found out how when we studied spin-flip line narrowing. What we must VOr/> = 2'11" or <P = 2'11"lpo (9.177a)
do is employ the conditions for a time average Hamiltonian. That is, we switch
which from (9.175) gives
among the various values of the Hamiltonian, fonning a cycle of desired values,
then repeat the cycle again and again for some desired total time. n = ]JO (9.177b)
The development of the system for one cycle in which the Hamiltonian
we will have a non vanishing 'Hay' We will also have a nonvanishing 'H. for p's
takes on n values
which are integral multiples of po. Thus, we will select values p
'HI, 1=0 to n-l (9.171)
each of duration Tk> is given by the operator
l p=pok
¢ = 2'11"/vo
k=0,1,2 using
(9.178)
exp ( -*'Hn- 1'1",,_1 ) .•• exp ( -* 'H2T2) ex p ( -*'HOTO) For all other values of p, (9.175) and thus 'Hay vanish.
The theorem we have proven holds true for any Hamillonian provided we
';i exp (-2.t, L'HITI)
1
(9.172)
satisfy the condition for average Hamiltonian theory:
(9.179)
where the equality holds provided the Tk are sufficiently short. Then the average The sons of correction tenns can be looked at using the Magnus expansion (see
Hamiltonian Appendix K, and references therein). Reference [9.20] discusses correction tenns
1 in great detail.
'Hav = t L 'HITt where (9.173a)
o 1 If we now look at the sequence of a pair of '11"/2 pulses which we have been
considering, we have for 'Ho
10 = L:TI (9.173b) 'Ho = 'H yy (9.180)
1
So, if we picked a set of n phases <Po, ... , <Pn-l each acling for the same duration From (9.165) we see it contains only the tenns p=-O and ±2. Now, we know
that the two-pulse sequence can generate all values of p, but those larger than
Llr, we would have
two require the T 2 or higher tenns in the expansion (9.164) of the exponential.
(9.174.) These are the tenns we do not wish to have since we require in (9.174) that the
average Hamiltonian be valid.
The question is, then, what can we do to create an 'Ho which gives high
= ~ L: H
p
p ("f:1e;P11»
1=0
(9.174b)
orders of p without violating the conditions for average Hamiltonian theory to
hold? The solution was found by Warrell et at (9.19,20]. We tum now to an
explanation of their approach.
where in the last step we took <PI = l<p. For a given p, the sum over I is nothing The Hamiltonian 'H. yy resulted from the action of the pulses X and X(=-
bUl X-l) acting on 'H u . The crucial steps are (9.160). In panicular,
We have seen in the previous section [note in particular (9.150)] that if Of course, we start with 'H u - that is what nature gives us. We get 'H yy by
474 475
applying a transformation (physically realizable!) with the operators X-I and
X. Let us then look for some other operator - call it R - to use instead of X. of producing -?-lzz (actually -'Hzz(l). The first method was used to produce
That is. ?-lo is to be produced by applying R to HH' Thus magic echoes in Sect. 8.8. It involved producing a strong HI in the x-direction
in the rotating frame which one phase shifted by 11" at periodic intervals. Several
?-lo:: R-1HHR so that (9.182) other'methods involved trains of X and X pulses, as for example shown in Table
8.4.
exp [ - (Vli)'ltot] = WI [exp [ - (i/li)'It"tJ] R (9.183) Another useful sequence produces
What properties should R possess? The trouble with X is that i( produced solely 1{n = k(1{yy - 7-{zz) • (9.187)
z.ero- and two-quantum coherences. What we need is an 1-lo that is rich in the
values of 1) which we desire to excite. That is, in (9.169a) we want large Hp's which, for dipolar coupling given by (9.159), is
for (he values of p we desire.
It is clear that no simple rotation operator like X or Y will do, since it
Hll :: L Bjk(IzjIzk - IyjIyk)
j>k
simply rotates the basic dipolar coupling. However, we remember that actually 1
(9.184)
::"2 L Bjk(It rt + Ii Ii:) (9.188)
j>k
where H I is the strength of the rf field, and II' the pulse duration. This term is This sequence arises from the sequence of (8.207). which we rewrite as
nothing but an approximation to the time development operator for the system
while HI is on, an approximation bec<luse we neglected the dipolar coupling (M. X, d,x, Ll,X. d, X, Ll,X. Ll',x, Ll.X. d, X, iJj2) (9.189.)
while Ht was on. More generally, we should think that where
R:: exp [- (i/h)Hntp] (9.185) d::2Ll+t p (9. 189b)
where 1-l n is the Hamiltonian which acts during the time t r . Now, ?-In can be and t p is the pulse duration [9.19). The fact that Ll' oF 2Ll is to correct for finite
a real Hamiltonian, such as is the case when R:: X, or it can be some effective pulse length. Note that since (9.188) is a two-quantum operator. we can easily
Hamiltonian created by the experimenter by manipulating a real Hamiltonian create its negative by phase shifting the pulses by 11"(2, i.e. by replacing X and
with pulses. X by Y and Y. In practice the schemes utilized are all based on sequences of
The game, then, is to try to guess an H n which does what one wants, then, 1r(l pulses.
if it does, figure out how to create it. Note that in order to create both R and its We now want to show that if we have such a dipolar Hn we can generate an
inverse R- 1 , we must be able to create both Hl/ and -Hn. There are not many Ho which is rich in large quantum orders HJI' Referring to (9.182) and (9.183),
choices. We have already excluded Hn's which are proportional to I z1" IYT' or we have
I~1" All that is left is
11.0 :: R-1HzzR
Hn :: 'H zz (9.186.)
:: exp [+ (i/h)HnT]Hzz exp [ - (i/h)7tnT] (9.190)
'Hn:: 'H yy (9.186b) where T is the length of time we let the system evolve under Hn or under
-1{Il· Note that in the case 'H1l:: 'H zz , we produce a negative (-Hn) of the
?-In = 'HH 01 (9.186c) form -H u {2. which requires that we let it act for a time T' which is twice T.
However, if we set T':: 2T, (9.190) still holds.
?-In:: ?-l~I~1 (9.186<1) We now expand the products using the relationship [9.16}
where z' is an axis other than x. y, or z. We could also take linear combinations e"Be-"=B+[A,Bj+[A,[A,B])+... (9.191)
of the operators. or take these operators multiplied by a constant ?-l~~ alone
is not useful since in (9.183) R would commute with the exponential, leaving Substituting this into (9.190) gives
?-lo :: ?-lH' However. ?-lu, ?-lyy, or ?-l~I~1 all will do the job. Before showing iT 1 T2 I
this, we recall that we get ?-lu; or ?-lyy simply by sandwiching ?-l~~ between two ?-lo = H zz + "h['H n , 1i~~] + '2 2T h2 [HII, [Hn.Hzzll + . (9.192)
1r(l pulses (Y and Y for ?-lu, X and X for H yy ). We have seen several ways
Evaluation of the commutators is straightforward but tedious. The general charac-
476
477
ter of the result is clear. however. Successive terms involve progressively larger We may describe the total preparation period to generate coherence Po (and
numbers of spin products, hence include terms which correspond to progressively integral multiples thereof) as follows:
higher orders of quantum coherence. That is, they contain t~rms of progressively I. .We apply 1in for time T. 1iu for time L\r'. then -Hn for time T (or
larger p. If T is very shon. only the first few terms. con~bute. and ?~ly small 7"). This takes time LlT.
values of p result. But if 11lniT ~ I. many terms WIll be Included, gIVIng large 2. With the phase of 1in shifted by 2Jr/Po. we repeat step l.
p's though each individual term is still of order 11l.uJ
• We now have the operator llo. ll
If we define a time L\T' for which zz alone
acts, and a time L\T for the time in the cycle for which we have a fixed phase.
3.
4.
With the phase 1i n shifted by 2(21fIPo) relative to step I, repeat step 1.
Continue the phase advances until ¢ = (Po -l)(2'11"IPo) relative to step one.
The total time to complete (1)-(4) is PoLlT, and constitutes one cycle. The
¢,.
(9.193) average Hamiltonian produced by this cycle is
LlT=T+T+LlT'
LlT' I
But the development of f! over the total time LlT is given by 'H" = Ll L 'H p (9.200.)
T p
R- 1 exp ( -*ll U L\T
1
)R = ex p ( -in-l1iunLlT') where the prime on the summation indicates including only those p's for which
Note that we multiply the summation by L\T'/L\T since during each interval
Suppose then that we wish to select I' = Po. From (9.178) we see we must picle
LlT. 1i1/> is acting only for the time LlT'. The rest of the time (T+ 7") is spent
a phase shift 4J for (9.175) given by
generating Rand n-
t.
~=h (9.195) The basic cycle is then repeated for as many times, m, as is necessary
Po for the amplitude of the desired coherence to grow sufficiently. Thus the total
For one cycle. ¢t will then, according to (9.175) and (9.195). need to take on preparation time 'p is mLlT. To give an idea of what is involved, we may quote
the different values the numbers reponed by Warrell el al. [9.18] to produce a selected fourth-order
coherence for protons of benzene molecules oriented in a liquid crystal. They
~l =0, (2./Po), 2(2'!l>o), 3(2P//>o) ... 0>0 -1)(2·/Po) (9.196)
use the sequence (9. I 89a) to produce
This cycle lasts for a time t c given by the number of phases n limes the duration
1io= n- l 1iu n (9.201)
of each phase
with a 2.4 lis duration of X (or X)
tc =nLl, =PoL\T (9.197)
,=3.0ps , 2T'=8.8p.s • T=2ms , L\T'=35jts (9.202)
However, in order for the average Hamiltonian theory to hold. we must be
able to approximate The entire cycle was repeated four times.
To get rid of zero-order terms, one must reverse the sign of HI/>
for half the
exp ( -i1itfl,o_tLl,') ex p ( -i1ioLh') by values of ¢ in the cycle. For a Hamiltonian such as lhat of (9.201) this can be
achieved by choice of the phase ¢, since adding 'Ir/l to ¢ changes lhe sign of
'HI>.
eXP[-*(1iI/>,O_1 + + 1iO)LlT
1
] (9.198) We now have a situation in which many more than two pulses are used to
generate the lJo quantum coherence. We may then introduce the unitary operator
This condition is satisfied if
U(T p ) given by
l1ioiLlT' <: I (9.199)
U(rp ) = "p [(Vh)'H"Tp ) (9.203)
Thus. it is not the long time L\T which comes into satisfying the Magnus expan-
to describe the development of f! during the preparation interval of duration 'p :
sion. but the (shorter) time LlT'. Since 1io is independent of L\r'. we can set
T and T' (if T' =f:. T) to make 'Ho what we desire. then set Ll,' to satisfy the (9.204)
Magnus expansion.
479
478
laJ u v 1t
where u is taken to be the Hamiltonian in the absence of pulses. It turns OUt
to be conveniem to Ihink of V as ending in an X pulse (Fig.9.6b). We then
'. t,
'. t,
redefine V 10 make Ihal explicit, replacing V(Tm ) with XV(T m ) so Ihal
Ibl
u v x
(Iy ) ::: Tr{ IyXV(Tm)exp ( -*1t u t l ) U(Tp)I,U-I(Tp)
" t,
'. L.Ib x ex p ( *1t 1
u t l ) V-I(Tm)X- }
Fig. 9.6a,b. More general preparation lind dc~eelion pulses represented by unilary opera-
to", U lind V, preparation and mixing times Tp llnd T m , evolution and detection times
t 1 and t2 respectively. In (b) we introduce explicitly Illi X(Tr/2) pulllC just before the
'" Tr{ (X- 1
IyX)V(Trn)exp ( -*1i u t l ) U(Tp)I:U-I(r p )
detection period so that V is defined II.S all the mixing pulses which precede the X{'If/2)
pulse x ex p (*1tu 1
t 1) V- (Tm)}
We then let the system evolve freely for a time tl. To observe the effects of what
happened during fl. we need to mix down to one-quantum coherence. Previously
::: Tr{ V-I I , vex p ( -*1tu t l )UIzU- 1 exp (*1t u t l ) }. (9.207)
we did this with a third 7r/2 pulse. As we shall see, we can do better by applying To understand the coment of this equation, we evaluate the trace using the eigen-
a fancy sel of pulses, closely related to those which prepared the multiquantum states of the system. Recall that we have M, the eigenvalue of I: T , as one quan-
coherence. Let us call the lolal mixing time T m . Then we represent mixing by a tum number, and a family of others related 10 the dipolar energy abbreviated by
unitary operator V(Tm ), such thai ~. We thus denOle states by IMa). It is sometimes convenient to use a single
e{Tp + tl + Tm ) = V(Tm)e(T p + tl)V- 1(Tm ) (9.205) symbol r or .! to stand for the pair. We define W r then by
The preparation, evolution, mixing, and detection periods may then be thought E,. = EM~ '" hw r and (9.208a)
of as in Fig. 9.6.
We now necd to investigatc the optimum way to carry out the mixing so then (9.208b)
that we do not lose Ihe advamages of confining the coherence we have generated
to a small number of orders, p. (l.) • L (rJV- 1I, VI.)e-;··I, (.IVI,V-'lr)e-C"I,)
r,'
(9.209)
9.4.2 Mixing to Dctcct High Ordcrs of Coherence r,'
Our task now is to find an appropriate set of one or more mixing pulses to Let us first look at a simple limiting case: what we get if both U and V
provide large detected signals when we have generated large-order coherences. leave (! alone. This would be the case if there were no spin-spin coupling (isolated
The Pines group has done extensive work on this problem [9.20]. Specifically, spins) and no rf pulses Other than the final X pulse. Then U :::: 1, V ::: I. Since
given a preparation specified by U(Tp), whal should we conSlJUCt for V(Tm)? We the states Ir), Is) are eigenstates of I the matrix elements are diagonal, with
of course evemually detect the signal as a funclion of t2, but in order 10 study the result W r • vanishes, giving "
the nature of V we shall focus on t2 '" O. Thus. if we were going 10 delect (Iy)
at t2 =0
(I.). L I(Mall,IMa)[' (9.210,)
MOt
(Iy ) ::: Tr {Iye{T p + tl + Tm ))
matrix I:T .
(9.206) This signal is, of course, the maximum we can get since it corresponds to
x eXP(*1tutl)V-1(Tm)}
rotating the total magnetiuttion perpendicular to the static field. There is no loss
in order in that process, hence this is the maximum signal we could get. It is
480 4.,
conceivable that during the development period, tl> there is some loss of order. says that both V-I and U "rotate" I z • Unless V-I and U produce the same
Let us therefore next examine the general result (9.209) for the case that we "rotation" the signal is diminished from its maximum value. In general. the
eliminate such losses by making t I = O. Then, going back to the last line of bigger the relative "rotation", the smaller the net result. In particular, suppose
(9.207), we get (lJ/')'I=O as U generates panicularly large p's, but Y- t does not. Then U will put the order
of the system into large p, but Y will not be able to bring it back down to zero
(1:\1)(1 =0 = Tr (Iz{VU IzU- 1V-I)} (9.211)
order to pennit the X pulse to conven it to transverse magnetization.
Comparing this result with (9.21Ob), which gives us the maximum possible The Pines group has made extensive use of sequences satisfying (9.212)
(Iy ), we see that if we make V=U- I
VU = I Le. (9.212.)
to generate and then observe large-order coherences.
It is the author's opinion that the Warren. Weitekamp, Pines melh<Xi of
v=U- 1 (9.212b)
selective excitation of muhiple quantum coherence can be thought of as a kind
we will guarantee the maximum signal. Physically this condition states that what of graduation exercise in spin Hamiltonian manipulation. It requires that one
U does. V undoes. This statement is clearly a general kind of condition analogous be completely at home with the idea that pulses manipulate the Hamiltonian,
to fonning the conventional echo or the magic echo. We can compare this resull with average Hamiltonian theory, possess knowledge of the specific kinds of
with what we would get if we used U to create the coherence. but only a single Hamiltonian one can generate and how to generate them. It illustrates then how
7I:n. pulse for the mixing. Since as we see in (9.207) we have already explicitly once one has these details in mind, one can advance to a higher level using
included a final pulse X, this situation corresponds to having V = I: the ideas as building blocks, much as in electronics one thinks of oscillators.
amplifiers, mixers, pulses and so on, without then being overwhelmed by the
(I,)" =0 =T'(I,(UI,U-')j (9.213) details of each circuit.
How does this compare with the result when Y = U-I? U is a unitary transfor-
mation. This expression reminds us of the dot product between a vector (1z) and
the same vector rotated (by U). It should be less than Tr {I;} or for that matter
than Tr{(Ulz U- 1)2}. Indeed
T,{(UI,U-')'j =T,{UI,U-'UI,U-'j
=T,{UI;U-'j =T'{I;U-'Uj =T'{I;j (9.214)
Consider then Tr {lIz - (U I z U-I)]2}. Since Ihis is the trace of the square
of an Hennitian operator, it must be positive. Thus
oST, {(l, - (U I,U-'))'} = T, {I; j + Tt «U I,U-')'}
- T, {I,UI,U-'} - T, {UI,U-' I,}
(9.215)
Pennuting the order of the operators of the last trace, and utilizing (9.214), we
thus get that
(9.216)
In general, as in the Schwarz inequality, we expect the inequality to hold unless
U is the identity operator. Thus. the signal of (9.213) will be less than the
maximum.
We can think of the signal for tl = 0 along these lines: the expression
(9.217)
482 483
10. Electric Quadrupole Effects
10.1 Introduction
So far we have considered only the magnetic interactions of the nucleus with
its surroundings. To be sure, by implicalion we have considered Ihe effect of
Ihe nuclear charge, since it detennines the electron orbits and where the nucleus
sits in a molecule. However, we have not considered any electrical effects on
Ihe energy required to reorient the nucleus. That such effects do exist can be
seen by considering a nonspherical nucleus. Suppose it is somewhat elongated
and is acted on by the charges shown in Fig. to. I. We see that Fig.1O.1b will
correspond to a lower energy, since it has put Ihe tips of the positive nuclear
charge closer to the negative external charges. There is, therefore, an electrostatic
energy that varies with the nuclear orientation. Of course,1 turning the nucleus
end for end does not affect the electrostatic energy. Consequently, for spin ~
nuclei the electrostatic energy does not split the ml degeneracy.
-q -q
+q +q +q +q
-q -q
(a) (b)
Fig.IO.I. (a) A cigar-shaped nucleus in the field of four charges, +q on the z-axis; -q on
the y-axis. The configuration of (b) is energetically more favorable because it puts the
positive charge of the ends of the cigar closer to the negative charges -q
1 See references to "Quadrupole Effects" in the Bibliography and the articles by Cohen
and Rei! and by Das and lfalin under "Books, Monographs, or Review Articles" in the
Bibliography.
485
•
10.2 Quadrupole Hamiltonian - Part 1 (We note that sometimes Poisson's equation applies instead. Some care must then
be exercised because we are, of course, interested only in the orientation depen-
To develop a more quantitative theory, we begin with a description in tenns of dent part of the potential, and must therefore subtract the spherically symmetric
the classical charge density of the nucleus, (!. We shall obtain a quantum me- parts.) If one has a nucleus at a site of cubic symmelry,
chanical answer by replacing the classical (! by its quantum mechanical operator. Vu = Vyy = V.;z (cubic symmetry) (10.8)
Classically, the interaction energy E of a charge distribution of density (! with a
potential V due to external sources is which, combined with (10.7), makes all three derivatives zero. The quadrupole
coupling then vanishes. This situation arises, for example, with Na 23 in Na
E = j g(r)V(r)dr (10.1) metal. The face-centered cubic crystal stnlclure puts each nucleus al a site of
cubic symmetry.
We expand V(r) in a Taylor's series about the origin: It is convenient to consider the quantities QO'(j defined by the equation
V(r)=V(O)+LX aO
w) \ a'v
+21 LXaX/3a a
) + ... (10.2) Qa/3 = j(3xcrx/3 - ocrpr )gdr
2 (10.9)
0' Xa r=O . 0',/3 X(J/ X/3 r=O
In tenns of the QcrP's, we have
where Xo' (a=1,2,3) stands for x, y, or z, respectively. Defining
- aw)
2
~ j x a x/3gdr = !(Qa/3 + j 8cr /3r gdr) (10.10)
V."
Xa r=O Va /3 = aa'v
a )
X(J/ :r:/3 r=O
(10.3)
As we shall see, the introduction of the QO'p'S amounts to our subtracting from
we have the left side of (10.10) a term that does not depend on the orienllltion of the
(10.5)
" 2
VO'/3 =0 if a f. {J where 1,p(0)1 is the electronic probability density at the nucleus. The orientation inde-
pendent term, A.E, of (IO.lt) becomes
Moreover, V must satisfy Laplace's equation:
(10.6)
LJ.E = ~ EV.... j r20dT =_4;e J,p(O)J2 j r 2qdT
\7'V = 0
"
This equation, evaluated at the origin, gives us This LJ.E wi\.l...bC"different for two nuclei of the same c1l11rge but different charge distri-
bution~'iliOi:opcs), or for two nuclei of the same mass and charge but different nuclear
(10.7) ~a.tes'(iSomers).In an electronic transition between an 8 and a p-slate, LJ.E will make a
LVcrO' = 0 contribution that will in general be different for different isotopes or isomers. Effects also
" show up in nuclear transition [10.21.
486 487
To obtain a quantum mechanical expression for the quadrupole coupling, we These can be shown to obey the equation
simply replace the classical e by its quantum mechanical operator e(op), given
by (lm"IQ~;)IIm'") = C(lml~(loIp + Iplo ) - 'opI'IIm') (10.17)
e(OP)(r) '" 1:: qAo e(r - rAo) , (10.13) where' C is a constant, different for each set of the quantum numbers I and '1'
•
where the sum runs over the nuclear particles, 1,2, ... k . .. N, of charge qAo'
In order to justify (10.17), we need to digress to discuss the Clebsch-Gordan
coefficients, the so-called irreducible tensor operators TLM' and the Wigner-
Since the neutrons have zero charge, and the protons a charge e, we can simply Eckart theorem.
sum over the protons:
e(Op)(r) '" e 1:: S(r - rk) (10.14)
protons 10.3 Clebsch-Gordan Coefficients, Irreducible Tensor
By substituting (10.14) into the classical expression for QOlf3' we obtain the Operators, and the Wigncr-Eckart Theorem
quadrupole operator Q~;>:
The Wigner-Eckart theorem is one of the most useful theorems in quantum
Q~1) '" JO%et%f3 - Detpr 2 )e(oP)('r)dT mechanics. In order to state it, we must introduce the Clebsch-Gordan coefficients
C{LJ'J;MMJ,MJ), and the irreducible tensor operators TLM' We shall first
'" e 1:: J(3:t a zfJ - 6a fJ r2 )S(r - rk)dT state the Wigner-Eckart theorem and then define the Clebsch-Gordan coefficienls.
protOll!l Next we shall discuss the irreducible tensor operators, and lasdy we shall indicate
'" e 1:: (3ZakZ,Bk - Set,Br:). (10.15) the derivation of the Wigner-Eckan theorem.
protonll We consider a sel.0f wave func.lions characterized by quantum numbers J
We have, then, a quadrupole term for the Hamiltonian 1tQ' given by and J' for the total angular momentum, M J or My for the z-eomponenl of
angular momentum, and as many other quantum numbers '1 or 1'/ as are needed
to specify the state. We are then concerned with calculating the matrix elements
'liQ = (51 '" Q(oP)
L..- Vop .p
(10 16)
.
o,p of the operators TLM' using these functions as the basis functions. The Wigner-
Eckart theorem states that all such matrix elements are related to the appropriate
The expressions of (10.15) and (10.16) look exceedingly me~sy to handle Clebsch-Gordan coefficients through a set of quantities (J'111 T L II J'1'/) that de-
because they involve all the nuclear panicles. They appear to reqUlr: us t~ m:at pend on J. J', '1. '1' and L but which are independent of M J' M J" and M. Stated
the nucleus as a many-panicle system, a complication we have aVOIded In diS- mathematically, the Wigner-Eckart theorem is
cussing the magnetic couplings. Actually a similar problem i.s involv~ in ~h
magnetic dipole and electric quadrupole cases, but we have Simply aVOIded diS- (JMJ"ITLMIJ'MJ") = C(J'LJ; MJ'M MJ)(J" II T L II 1'"') . (10.18)
cussion in the magnetic case. Lei us now define the Clebsch-Gordan coefficienls. They are encounlered
The quadrupole interaction represenled by (10.15) enables us to treat prob- when one discusses the addition of two angular momenta to form a resultant.
lems of much greater complexity than those we encounter in a discussion of We therefore consider a system made up of two pans. Let us describe one part
resonance phenomena. When performing resonances, we are in general con- of the system by the quantum numbers L and A!, (Q describe the tOial angular
cerned only with Ihe ground stale of a nucleus, or perhaps with an excited state momentum of that part and its z-component. Let us use the quantum numbers
when the excited state is sufficiently long-lived. The eigenslates of the nucleus J' and MJI correspondingly for the second part of the system. For the system
are characterized by the tolal angular momentum I of each state, 21 + I values as a whole we introduce quantum numbers J and M J . We have, then. wave
of a component of angular momentum, and a set of other quantum num~rs '1, functions 'l/JLM and ,pJlM ,,10 describe the two parts, and tPJMJ for the whole
which we shall not bother to specify. Since we shall be concerned only With the system. The function tPJA:J can be expressed as a linear combination of product
spatial reorientation of the nucleus for a given nuclear energy state, we shall be functions of Ihe two parts, since such products form a complete set:
concerned only with matrix elements diagonal in both I and '1' Thus we shall
need only matrix elements of the quadrupole operator, such as tPJMJ '" L C(J'LJ;MJlMMJ),pJ'MJ,tPLM (10.19)
J'MJ,;LM
(Im'1IQ::)IIm''1)
488 489
The coefficients C(J'LJ; M)'MM J ) are called the Clebsch·Gordtln coefficicms. We shall wish to compute matrix elements of the TLM'S. We are familiar
Certain of their properties are very well known. Forexample, C(J' LJ; M)'M M J ) with the fact that it is possible to derive expressions for the matrix elements
vanishes unless MJ "" M + M)'. A second property, often called the triangle of angular momentum from the commutation rules among the components. It is
rule, is that C(J'LJ;M;MMJ) vanishes unless J equals one of the values possible to compute the matrix elements of the TLM'S by means of (10.21) in a
J' + L. J' + L - I, .. . IJ - Ll, a fact widely used in atomic physics. similar manner. Let us illustrate.
Let us now define the irreducible tensor operators Tu". Suppose we have We have in mind a set of commuting operators J 2 , J., plus others, with
a system whose angular momentum operators have components Jz, J". and J:. eigenvalues J, MJI and '1. We use ." to stand for all other quantum numbers
We define the raising and lowering operators J+ and J- as usual by the relations needed. We wish to compute matrix elements such as
J+ == Jz +iJlI J- == Jz - iJ" (10,20) (10.25)
One can construct functions G of the operators of the system and examine the By means of the commutation rule
commutators such as [J+, G], [J- ,G], and [Jl' G]. It is often possible to define
a family of 2L + I operators (L is an integer) labeled by an integer M(M ::: [Jz,Tu·,]::: MTUf (10.26)
L, L - 1, _.. - L) which we shall term irreducible tensor operators T,.M, which we have
obey the commutation rules
(10.27.)
[J±. T LM ) = JL(L+ I) - M(M ± I)TLM ± 1 (10.21)
B",
Another example of a Tl/Ir can be constructed for an atom with spin and orbital ::: (MJ - M)')(JMJ1JITLMIJ'M)',,')
angular momentum operators sand " respectively, and total angular momentum
J. Then we define the operators
1+:::/z+ilu r:::lz-il u (10.23)
One can then verify the operators T tM defined by
1 1
Til ::: - ..j2'+ TJO::: I: TI_l ::: ,jir (10.24) Equation (10.27b) shows that
obey (10.21). (Actually the operators of (10.24) form components of an irre- (JMJl1ITUI1J'M)'11')=0 unless MJ-M),:::A{ (10.28)
ducible tensor TIA'f with respect to the operators 1+, 1-, and Il as well as J+ I
J-, and Jz.) We may write the Tl/Ifs of (10.22) as TI/",(J), to signify that they In a similar way we may find conditions on the matrix elements of the other
are functions of the components Jz , J", and J. of J. The Tl/If'S of (10.24) are terms of (10.21). Thus
in a similar manner signified as T1J\f(l). (J MJ.I(J±. Tl.AII1J' Ml"')
It is helpful to have a more physic"l feeling for the definition of the operators
TLAf by the commutation rules of (10.21). We realize that angular momentum
= JL(L + I) - M(M ± 1)(JMJ.ITl.Af ± i1J' M J ,"') (10.29)
operators can be used 10 generate rOiations, as discussed in Chapter 2. It is nOi
surprising, therefore, that (10.21) can be shown to guarantee that TLM trans- (J MJ"IJ±Tu"IJ'M J 'l1')
forms under rotations of the coordinate axes into linear combinations TUf ', in
= (JMJ.IJ±IJ MJ 'f I.)(JMJ 'f I"ITLMIJ' Ml"")
exactly the same way that the spherical harmonics YL/lf trnnsform into linear
combinations of YUI ' 's. This theorem is shown in Chapter 5 of Rose's excellent = J J(J + I) - (MJ 'f I)MJ(JMJ 'f I.ITLMIJ' Ml"') (10.30)
book [10.3]. By combining (10.29) and (10.30), we obtain the other recursion relations:
490 491
trix elements of Tr,M 's are related. The relationship is called the \Vig"er-Eckart
../J(J + I) (MJ" I)M;{JMJ" hlITwIJ' M J,"') theorem:
- (J MJ'1ITLA1IJ' MJ' ± 1'1')';JI(Jf + I) MJ'(MJ' ± 1) (10.37)
=../L(L+1) M(M±I)(JMJ '/ITuf±dJ'MJ , " ) . (10.31)
where the notation (JI'/II TL II J',./) stands for a quanlity that is a constant for a
We note thal the only nonvanishing ICnTlS must satisfy (10.27b). However, if any given J,£,J',I'/,I/ independent of M;. MJ'. and M.
one term in (10.31) satisfies this relation, all do. Equation (lO.27b) and (10.31) As we can see specifically from (10.22) and (10.24), for a given L and AI
constitute a set of recursion relations relating matrix elements for TLM to one there may be a variety of functions, all of which are Tr,M'S. The Clebsch-Gordan
another and to those of TL/l1" 111cse equations lurn OullO be sufficient to enable coefficient is the same for all functions TUf that have the same Land M, but
one to solve for all TLM matrix elements for given J, )', lJ,1/ in ternlS of any the constant (JI'/II T/, II J',/) will depend on what variable is used to construct
onc matrix element. the TLAI·S.
A further insight inlQ the significance of the recursion relations is shown To illustrate this point further, let us consider a particle with spin 8 and
by returning to the Clebsch-Gordan coefficients. In so doing, we shall sketch the orbital angular momentum I and position r. The total angular momentum J is
proof of the Wigner-Eckart theorem. given by
As is shown by Rose, the C's obey recursion relations identical to those
J::: 8+1 where (10.38)
of the TL/lf'S. We shall derive one: the selection rule on M. M;. and MJ'.
Consider the operator
J: ::= L: + J~ where
(10.32) 1. = ~ (y~ z~)
I Bz
- By
J:oJi;M J = M;oJi;MJ
1 =
11
~(z~
i ax
-%~)
8z (10.39)
L:'PLM = MtPr,M
(10.33)
J~¢J'MJ' = MJ'¢J'M J,
TIlen. using (10.19). consider the following matrix element of the operator J:: I, = ~(%~ -y~)
1 fJy ax
(tPLMtPJ' M J,' J:tJiJ MJ) = MJ(!/;LMtPJ' AlJ,' 1Ji; (HJ) We shall now list two T2M 's: one a function of the angular momentum J; the
=MJC(J'LJ;MJ'MMJ ) (10.34) other, of the coordinate r. One can verify that the functions of Table 10.1, which
we shall call Tu,r(J) and T 2M(r), indeed obey the commutation rules of (10.21)
where we have let J: operate to the righl. But. writing J: as L:+J~ and operating with respect to J+, J-, and J:.
on the functions to the left. we get We have used the notation T2M (r) as shorthand for a TU,f constructed from
(wur¢;'M J,' Jzwn.r) = (M + AJJ')C(J'LJ; M;,M M;) (10.35) the components z, y, and z of r. There is an obvious similarity between T2 Af(J)
and T2M(r): Replacement of J+ by (z+iy), J- by (x-iy), Jz by z will conven
By equating (10.34) and (10.35), we find TU1(J) into T2M(r). This similarity is a direct consequence of the similarity of
(10.36) the commutation relalions of components of J and r with J7.' J II , and J z :
(M + MJ' - MJ)CU ' LJ; MJ'M MJ) = a
This equation is quite analogous to (1O.27b). provided we replace Table 10.1
1
(IMjIJIT/,flIIJ J\!!J"/) by
T" J+'
C(J'LJ;MJ,MMJ) . 1'21 -(J.;+ +J+;.)
T" # ( 3 ) : ' _ ;')
One can proceed in a similar manner to compute matrix elements of the rais- T'_I J.;- +;- J.
ing and lowering operators, 10 get equations similar to (10.31). In fact the 1'2_2 J-'
C(J'LJ;M;,/",fM;)'s obey recursion relalions identical to those of the
(J M;I'/ITLAI IJ' M;,,/)'s. As a result, one can say that Ihe C's and the ma-
493
492
[Jz.y] = iz (10.40.) Iz = L, lzk + 3z k ' etc., for Iv and I, , (10.43)
[Jz • J y ] = iJ, I etc. (1O.40b)
where l z l.: and Szl.: are the x-components of the orbital and spin angular momenta
where (10.403) can be verified by means of (10.38) and (10.39). It is clear Ibat of the kth nucleon; and by recalling that
any function G(x. y. z) of x. Y. z, constructed from a function G(Jz , J y, J,) of
J z , J1/' J, by direct substitution of x for J z , and SO on, will obey the same
[tzkoYk) =iZk (SzkoYk]=O , etc. • (10.44)
commutation rules with respect 10 J z • Jv' and Jz. Thus, if a function of J ZI we see that
JVI J, is known to be a TuJ, the same will be true of the function fonned
[Iz.Yk)=izk , eiC. (10.45)
by replacing J z • J 1I • J, by X. Y, z. respectively. The only caution we note in
procedures such as this is that we must remember that the components of some The lenos 3xokXPk -ooprl are linear combinations of T 2M (rk)'S such as found
operators do not commute among themselves; so that for example in T21 (J) we in the right-hand column of Table 10.1.
have the symmetrized product J+ J, + J,J+, not 2J+ J,. The method of direct Equation (10.41) applies in a somewhat more general fonn not only to
replacement will work for olher variables as long as they obey commutation TLM'S but also 10 functions that are linear combinations of TLA's, all of the
relations such as those of (10.40). For an excellent review anicle including tables same L. Thus consider such a function F(P), which is a function of the operators
of T/,M's of various Land M. sec [10.4]. p'
Returning now to (10.37), let us consider two Tu",s, one a function of F(p) = LClft",TU"(P) . (10.46)
variables q and the other a function of variables p. Then (10.37) tells us thai M
1
(J MJ11ITLM(q)IJ MJ"/) Let us define a function C(q) of lhe operators q, using the same coefficients aM:
, 1 (hili TL<q) II J',/) C(q) " L"MTLA/(q) (10.47)
= (JMJ"ITLM(P)IJ MJ"I ) (J" II TL(P) II J'"') (10.41) M
Since the factor (Jf/ll TL(q) lJ J'11')/(J'111 TL(P) II J',./) is a conSlant (that is, in- Then one can easily verify, using (10.41,46,47), that
dependent of M, MJ. and MJ'), we see that we can compute all lhe matrix (JMJ"IC(q)/J'M J ."') = (JUJ"IF(P)IJ'Up ,i)
elements of TL}.f(q) of fixed J, J'. fl. and rl from knowledge of the constant
(J" II TL(q) II J',J')
and of the matrix clements (J M J '1ITLA1 (P)lJ' M Jl 11').
x (J" II TL(P) II J''I') (10.48)
One word of caution is necessary. It may be thai (10.41) is nOI meaningful.
since for some operators p. the matrix element (JMJ'1ITLi\,,(p)IJ'MJI '1') van- We may apply this theorem to show thai
ishes even though the matrix element (J MJ'1ITLM(Q)IJ' M J ,'1') does not. An
example of such a case is when TLM(P) is made up of componenlS of J. Then
(Im'1le L (3xokX,8k - oo,8r~)Jlm''1)
k(proloOS)
all matrix elements in which J' .. J vanish. Of course (J 11 T2(J) II JI) vanishes
(l.T~ + TpT.)
100. so that (10.41) becomes indetenninanl. = (lrnql3 2 - '.pT'ITrn'q)C (10.49)
where C is a constant,3 the same for all m. m', 0', and p. We can express C in
tenns of the matrix element for which m = m' = I, 0' = f3 = z as follows:
10.4 Quadrupole Hamiltonian - Part 2
(lT"I< L (3,1- '·1 ITT") = C(TT"13T; _ T'ITTq)
k(protons)
We now apply the Wigner-Eckarl theorem to evaluate the matrix elements of
=CT(2T-l) . (10.50)
Q~;>. Now
Since the quantum number '1 is llssumed to be associated wilh a variable that
Q~1) = e L (3XO'kX,8k - oal3r~) . (10.42)
commlltes wilh 12 and I,. we can omit it in evaluating the right-hand side of
k(pl"Otons)
(10.50). We shall also define a symbol eQ:
By recalling that I z • IIJI and I, are the operators of the total angular momentum
of the nucleus 3 Do not confuse C with th~ symbot for the CJebsch-Gordan coefficients.
494
495
oQ· (II.,I' L: (3zl- rIIII") (10.51) symbols, '1 and q. called the asymmetry parameter and the field gradient. by the
k(protons) equalions
Q is called the Quadrupole momellt of the nucleus. We have. by combining eq'" V"
(10.50) and (10.51),
'1'", Vu - VIIy (10.57)
'Q V"
C = 1(2l- I) (10.52)
The case of axial symmetry. often a good approximation. is handled by taking
The fact thai we are concerned with matrix elements internal 10 one sel of the axis to be the z-direction. giving '} '" O.
quantum numbers I, 1/ enables us to use (10.49) and (10.52) to replace in Q:;> Since we have seen that the raising and lowering operators often provide
panicularly convenient selection rules. it is useful to write (10.53) in tenns of
the Hamihonian. All matrix elements diagonal in I and 11 are just what we should
calculate by adding an effective quadrupolarcontribution 1-tQ to the Hamiltonian: I+, I-. and I, for an arbitrary (that is. nonprincipal) set of axes. By defining
Vo ",Vu
'Q 1) L..
'liQ' 61(2l " V.pl,U.lp
J
+ lpl.) - 5.pl2 ) .
(10.53)
.,/J
V± 1 '" V,z ± iV,~ (10.58)
II is interesting (0 note lhat of the nine components of Q~o;>. only one V±2 '" levu - VlIlI}±iVzlI
nuclear constant, eQ, is needed. The reason is as follows: The faci Ihal the we find by straightforward algebraic manipulation that
nucleus is in a state of definite angular momentum is equivalent to the classical
statement Ihal the charge has cylindrical symmeuy. Taking z as the symmetry 'Q(22 --
1iQ'" 41(21 _ I} Vo(3!., - I ) + V+1(l I, + I,I )
axis, the energy change on reorientation depends. then, only on the difference
between the charge distribution parallel and transverse to z: + V_I (.P" I" + I"I+) + V+2(r)2 + V_2([+)2) (10.59)
/ z2(}dT and / z2(}dT . Equation (10.59) gives a form of the quadrupole coupling that is particularly
useful when considering relaxation for which the principal axes are not fixed in
This gives us the critical quantity space but rather are functions of time. An anempt to use principal axes would
then be exceedingly cumbersome. We shall not attempt to describe nuclear re-
/(z2 _ z2)(}dT "'! /(2z 2 _ x 2 _ y2){ldT laxation by the quadrupolac coupling, although it is a very imponant mechanism
in insulating crystals, often dominant at room temperature.
'" ! /(3z2 _ r 2)gdT (10.54)
6I(;~- I) [Vu(3I; - 2 2 symmetry such that Vu '" Vyy for a set of principal axes). Let us then consider
1iQ '" 1 ) + VIIII (3l; - 1 ) a magnetic field applied along the z'-axis where in general the z- and z'-axes
differ. Then we have, for our Hamiltonian.
+ Vu (3I"2 - l 2 )] (10.55)
This expression can be rewritten, using Laplace's equation Eo VCfO = U, to give
1i '" -'YnliHoI", + 4It22;~ I) (31; - 2
1 ) (10.60)
496 497
z' z J<·ig.l0.Z. Axesz', z', and z
- 311 --r-
8
_1/ 2
------
'I,
111-% ------
-r-
(al
TH,
(b)
perturbation theory, Defining the x'-axis to lie in the plane containing z' and z,
we have (see Fig. to.2) Fig. JO.3. (a) Effect ora quadrupotfl coupling in first or<kr. The shiflli of all1evels for J ::: ~
ha"fl the same magnituck (b) Spectral absorption corresponding to the energy tevels ~f
I:: '" I::, cos 8+lz ' sin 8 . (10.61) (a). The central line i$ unaffected by lhe quadrUI)()le coupling in first order
!
A set of levels is shown for 1 '" in Fig. 10.4.
The effect of the quadrupole coupling is shown in Fig. 10.3 for the case of a spin We note that there is a degeneracy of ± m conesponding 10 the f::.ct that
I '" ~. It is helpful to note, since turning the nucleus end for end does not affecl the electrostatic energy. If an
alternating magnetic field is applied with a nonvanishing component perpendic~
I: [3m' - [([ + 1» = T< (3[; - [') = 0 , (10.64) ular 10 the z·axis it produces nonvanishing matrix elements of ILlml '" 1. It
m can therefore produce resonant transitions between the quadrupole levels. It is
lhat the quadnlpole coupling does nOI shift the center of gravity of the energy in
first order. Moreover, lhe shifts of +m and -m are identical. With these points
in mind we realize that the energy levels must look as shown.
±t
One interesting result is that for a half-integral spin the m '" ±! levels
are shifted the same amount and the t.ransition frequency between them is un- ±t
affected in first order by the quadrupole coupling. The! !O -! transitions are Ftg.IO.4. Enflrgy levels of a quadrupole coupling when the
quite insensitive to effects such as crystalline strains which may tend to shift ±l Zeeman coupling is negligible
498 499
customary to speak then of "pure quadrupole resonance". although the transition In general for a or p = z. y, z we have
is still induced by magnetic dipole coupling to Ihe allernating field.
An imponant observation related to (10.66) is that when I is and so !' i, Valli = -
p rJ
, (3X.Xp
- - -Ii •
r a"
)2 (lO.68b)
on (in general when 1= n +~, where n is an integer) Ihe energy levels are all
doubly degenerate in Ihe absence of a magnetic field but that for integral spin Equation (10.67) and (10.68) emphasize through the I/r 3 dependence that charges
the degeneracy may be completely removed, as with the m = 0 state. This result close to the nucleus have the most imponant effect. We may suppose that the
is an example of an imponant theorem. due to Kramers. and applies to both electrons belonging to the atom containing the nucleus would make major contri-
eleclTon and nuclear magnetic resonance. Kramers' theorem stales: bUlions to Vu . Such is indeed Ihe case. However, if we have a closed shell. the
For a system of angular momentum 1= n +!.
where n is 0, 1.2. and so electronic charge is spherically symmetric. and there is no quadrupole coupling
on. the degeneracy of any state can never be completely lifted by electric fields. (see. however. some funher remarks on closed shells below). The case of an
A corollary is that when a system is composed of an odd number of spin, incomplele shell is readily illustrated by an example of a single p-electron in an
particles. electric fields can never completely lift the degeneracy. orbit zf(r).
The degeneracy is commonly called the Kramcrs' degeneracy. Proof of its We wish, then. to compute the quadrupole operator for Ihis example. Since
existence actually depends on the propenies of the system under time reversal. the electronic molion is mpid. we shall average the expression of (10.68) over
the electronic orbit. This procedure is equivalent to saying that. of the total
Hamiltonian describing bolh the electron orbil and the nuclear spin. we shall
10.6 Computation of Field Gradients compute only matrix. elemetlls that are diagonal in the electron orbilal quantum
numbers. and that we shall negieci the perturbation of the electron orbit by the
nucleus.
We have seen that the quadrupole coupling depends on Ihe second derivatives We have. then,
Va ,8 of the potential, which reduce for the case of axes x. y. z. which are principal
_ /.,•• <3 CO,20 - I)
axes to V.u and V:n : - VylI . The potential V arises from external charges of either
other nuclei or electrons. It is a straightforward matter of taking derivatives of
Vu - -e 'Pe ,. 3 tPcdTc
the potential 10 show that a charge e at a point z. y. z produces a Vu at the 2 (3cos 28 - I) r 2 f2(r)
= -e cos 8 3 sin8r2drdfd8
origin of / r 4x
Vu = e
(3z 2 _ r 2 )
r
5 (10.67) : -, I~ C') (10.69)
r2 = x2 + y2 + z2 where we have designated the electronic charge as -e and where. as usual,
In terms of spherical coordinates we ha ve (see Fig. 10.5) (117- 3) is the average of 1/,.3 for the p-orbit. We note that large-Z atoms. for
which (l/r 3) is very large, will have large field gradients. This trend is shown by
11" (3 cos 2 8 - 1) Table 10.2., which lists typical values of e 2 qQ of halogen nuclei in covalently
Vu· =e 3 (10.68.)
r bonded crystals.
It is interesting to note that the values of e2 qQ of Table 10.2 put the f~
, r quency of pure quadrupole transitions for covalently bonded halogens at much
higher frequencies Ihan that of Iheir Zeeman transitions "InBo for typical labo-
, ratory magnetic fields.
o
Table 10.2. Typicl\l vl\luC$ or e2'lQ for hlllogen nuclei in covl\lently bonded crystals
Nucleus e 2qQ 11oo1llzi Q [10 24 cm 2 1
/' y
""- ,/
~
--------_':#
80
SO.
-7.97 X 10- 2
0.30
2000 -0.59
x Fig. 10.5. Location or a chiLrge ... in terma or the spherical coordinates r,',~
500 501
When the electronic wave function contains a mixture of s- and p-states
(a "hybridized bond") the s-part conlributes nothing to the quadrupole coupling.
11. Electron Spin Resonance
A similar situation arises when a halogen atom is in a state corresponding to a
mixture of a pure covalent bond (that is, a p-state) and an ionic bond (closed
shell). The quadrupole coupling of the ionic bonding vanishes. One can therefore
utilize quadrupole couplings to study bond hybridization, degree of covalency,
double bonding, and so on.
The fact that the closed shell electrons are very close to the nucleus makes
it important to consider their distortion from spherical symmetry. For example,
a charge e will produce fields that will disturb the closed shell electrons. This 11.1 Introduction
effect has been studied extensively by various workers. It leads 10 a correction
to the gradient V:O: due to e alone. The actual field gradient V:: is in fact given
by
So far we have confined our attcntion to nuclear magnctic resonance, although
many of the basic principles apply to clectron spin resonance. We have also
(10.70)
considered questions concerning the electrons, such as the quenching of orbital
The quantity ,.(r) is called the Sternheimer antisliielding factor, after one of the angular momentum and the magnetic coupling of the nuclear spin to that of the
workers who has made some of the mOst important contributions to understanding electron. In this chapter we shall add a few more concepts that are imponant to
of the phenomenon (10.5). The fact that it is a funclion of the distance r from the study of electron spin resonance l but which are not encountered in the study
the charge e to the nucleus is emphasized by writing ,.(r). In general 1'(r) < I of nuclear resonance.
as long as e is well inside the closed shell charge distribution. Once r is well Probably the major difference between electron and nuclear magnetic reso-
outside, i becomes independent of r. We shall denote this value by 1'00< Some nance is the fact that the nuclear properties such as spin, magnetic moment, and
theoretical values for I - 'Yoo are shown in Table 10.3. quadrupole moment are to a very high degree of approximation unaffected by the
surroundings, whereas for electronic systems, the relatively much greater physi-
Table 10.3. Theoretical values of 1 - 700 cal size and the much smaller energy to excited states make the system strongly
Ion J - 700 dependent on the surroundings. An atom, when plnccd in a crystal, may have
angulnr momentum, magnetic moment, and quadrupole moment valucs entirely
CI 48
Cu+ 10 different from those of its free atom. It is as though in nuclenr resonance we
Rb+ 51 had to compute 'YII' I. and Q for each material in which the nucleus was to be
ea+ 99 studied.
The fact that the state of an atom in a solid or liquid is very different from
that when it is free means that we cannot predict the properties or even the
As we can see, the correction is enonnous, amplifying the direct effect V:o:
existence of a resonance from the free atom electronic angular momentum and
by one or two orders of magnitude.
magnetic moment.
The existence of the Stemheimer effect greatly complicates the determina-
For example, a sodium atom has zero orbital magnetic moment and angular
tion of nuclear quadrupole moments. It is difficult to know how accurate the
theoretical ,.'s are. However, examination of (10.68) reminds us of the magnetic
i
momentum, but it has a spin of and a corresponding spin magnetic moment.
The magnetic properties can be studied by the method of atomic beams. In
dipole coupling between a nucleus and an electron spin. The radial and angular
sodium metal, the valence electrons fonn a conduction band, with substantial
terms are the same as the A and B tenns of the dipolar coupling. Since the
pairing of spins. However, there is a weak electronic spin magnetization whose
nuclear and electronic magnetic moments are known, it is therefore possible to
spin resonance has been studied. In sodium chloride, the sodium gives up its
use measured hyperfine couplings to get the average of (3 cos 2 8 _1)/r 3 . Since
outennost electron to complete the unfilled p-shell of the chlorines. The result
we are not using closed shells (the hyperfine coupling of a closed shell vanishes,
is a zero spin magnetization and no electron spin resonance. Even if one has
since the electron spin is zero), the Sternheimer factor is only a small correction.
atoms whose bonding is covalent, as in molecular hydrogen, there is usually
This technique has been applied to the alomic beam experiments of halogens and
is the basis of the most reliable experimental measurements of nuclear quadrupole
moments. I Seoe ~r~rencdl to -Electron Spin Il.eson"nce- listed in the Bibliography.
502 503
no net spin magnetiuuion because the electron spins pair off into a spin singlet. Often the electric field in an atom points nldially outward and is a function
There are exceptions, of course, such as the oxygen molecule. As we remarked in of r only, so that
connection wilh chemical shifts, the orbital angular momentum is often quenched, r
so that there is no first·onler orbital contribution to a resonance. E(r) = -E(r)
r
We see Ihal mOSt insulators will nOt exhibit a resonance, unless one takes
special pains to unpair the spins. Some atoms, such as those in the iron group or Then Ex p becomes (l/r)E(r)r x p '" (1IIr)E(r)L. This circumstance
rare earths, have incomplete inner shells. Even when ionized, they still possess a leads to the well-known fonn of the spin-orbit coupling, utilizing the spin-
net moment Thus neutral copper has a configuration (3d)104s. Cu++ has (3d)9, orbit coupling constant ,),:
which is paramagnelic. In an ionic substance such as CuS0-t· 5H 20 (copper 'Hso",,),L·S . (11.2)
sulfate), the copper atoms are paramagnetic, and a resonance results.
We may list several classes of substances or circumstances in which one For free atoms that obey Russell-Saunders coupling, the spin-orbil coupling
may expect to find resonances, although in individual cases the general rules may gives rise to the spliuing of states of given Land S and their classification
break down: according to the total angular moment J '" L + 5, L + 5 - I, ... , IL - 51.
4. The coupling of the electron spin and orbital magnetic moments to an
I. Materials containing atoms of the transition elements with incomplete inner externally applied magnetic field.
shells; as, for example, the iron group or rare earths. 5. The magnetic coupling of the nuclear spin to the electronic spin and orbital
2. Ordinary metals, the conduction electrons. moments.
3. Ferro-- and ferrimagnets. 6. The coupling of the nuclear electrical quadrupole moment to the electronic
4. Imperfections in insulators, which may trap electrons or holes. For exam- charge.
ple, the F-center (electron trapped at the site of a missing halogen ion in
an alkali halide) or donor and acceptor sites of semiconductors. Let us tum now to an example that will illustrate the role of some of the
more important tenns. We begin in the next section with a discussion of the
Tre'Hment of all these situations on a unified basis is SO hopelessly general role of the crystalline fields and spin-orbit coupling. In the Section 11.3 we shall
that none of the interesling features emerge. The approximations important in one consider the coupling to the nuclear magnetic moment.
problem may nOt be at all justified in another. For example, if one is dealing with
the resonance due to Cu++, one knows already a great deal about the electronic
wave function, since it will be closely related to that of a free Cu++ ion. One
can therefore start by considering states of a free copper ion. On the other hand, 11.2 Example of Spin-Orbit Coupling and Crystalline Fields
there is no equivalent to the "free ion" if one is dealing with an F-center. We
could not, therefore, define a set of "free ion" states. For our example we shall consider the case of an atom at the origin of a set of
What we shall do is list some of the more important interactions and then coordinates. possessing a single p-electron, acted on by four charges equidistant
consider several examples that represent rather different physical situations but from the origin, two of the charges being positive, two negative, their magnitudes
which involve the major phenomena. all being the same. The details of the arrangement (see Fig. 11.1) are seen to be
The principal teons in the electron Hamiltonian will consist of: identical to those thai we discussed earlier when we considered the phenomenon
I. The electron kinetic energy. of chemical shifts.
2. The electron potential energy. Often it is convenient to divide this into a Neglecting nuclear coupling, we have for the Hamiltonian of the electron
"free ion" potential energy plus one due to the crystalline surroundings, the of charge q (q negative):
so-called crystalline potential. Such a decomposition makes sense provided
there is such a thing as a "free ion", but, as remarked above, it would not z
have meaning for an F-center. y
3. The spin-orbit coupling. An electron moving in an electric field E expe·
riences a coupling of the spin to the orbital motion 'HSO:
eh
+ lo'ig.ll.l. ArrangcmCllt or two positive lind
?iSO' - -2 5
2m c 2
· (E x p) (11.1) + • two ncglltive charges, all equidistant rrom the
origin
S04 505
1 { =1- (
2m
p--A
c
<)' +Vo+V,+.\L·S+2/3H·S (11.3) Tabte I Lt. Spin-orbit coupling consbnts per electron for sevenol .toms
Alom Coupling constant
where A is the vector potential associated with the applied static magnelic field (em-II
H, Vo is the potential of the "free alom", VI is me potential due to the fOUf
charges, and 2f3H· S represents the coupling of the electron spin moment to the B 10
external field. We are here using fit the Bohr magnetoll, to express the electron C 28
F 211
magnetic moment. It is related to Ie. the electron gyromagnctic ratio, and J.f.e • CI 440
the spin magnetic moment, by the equation Br 1842
-I(
2m
p--A
c
q p2
<)2 =---(p.A+A'p)+--A'
2m
q2
2
2me 2mc
( 11.5)
iron group.
Let us first consider the case that VI is much larger than A. At first. we
consider the effect of VI only. For the situation of Fig. 11.1, the effect of VI will
II is convenient to specify the vector p<>(cnlial as be to lift the orbital degeneracy. The resultant energy levels, shown in Fig. 11.2,
are all twofold degenerate because of the electron spin. We denote the wave
A = ~Hx r ( 11.6)
functions as xf(r)u m • and so on. where the function Urn is a spin function. If
which gives us there were no spin·orbit coupling. the spin would be quantized independently of
1( <)2 = 2m
2m p- ~A
p2
-
qh q2
2mc H . L + 8mc2 H (x
2,2 12
+y ) (11.1)
the orbital state so that the urn's would be the usual eigenfunctions of S:/, where
z' is the magnetic field direction.
where, as usual, L = (I/i)r x 'V, and where the axes x' and y' are perpendicular
to the field direclion Zl. (We distinguish between the field direction Zl and the yj(r)u.
crystalline axis z). The teml proportion to H2 gives the usual diamagnetism. It zj(r)u",
!Urns out to be unimportant, compared to the tenn H· L, in influencing the elec- Fig.ll.l. Energy levels of the three p-sta~('$ under Lhe
xj(r)u", poLential of Fig. I l.l
tron spin resonance. By utilizing the fact that f3 = eliflmc, we have. accordingly,
as the Hamiltonian
p' Let us consider, then, the effect of the two remaining terms:
1{=-+,8H·L+Vo+VI +AL·S+2/1H·S . (11.8)
2m pH·L+AL·S . (11.9)
We shall consider that the principal energy tenns are the kinetic energy and the
"free atom" potential Vo. We shall treat the remaining tenns by a perturbation We examine the sons of matrix elements these terms possess. "There are two
method. For our example we shall think of the three degenerate I>-states xf('·), SOrts: those that connect the same orbital state, and those that connect different
yf(1'), and zf(x), which are solutions of the free atom potential Vo. We shall orbital states. The former are clearly the more important, if they exisl, since the
assume that the coupling to other free atom states is relatively unimportant, orbital splillings are so large. We have, then. matrix elements such as
so that the effect of the remaining tenns in the Hamiltonian can be found by
(11.10)
considering only the submatrix of the Hamiltonian involving these three orbital
states.
For practical laboratory fields the tenns H· L and H· S are only about (11.11)
I cm- t whereas VI may be a substantial part of an electron volt (that is,
506 507
where ciT stands for an integral over spatial coordinates. and dTs. over spin
variables. We see that both integrals of (11.10) and (11.11) involve J¢:m,>.L· 5t/Jzm ciT dTs = >.fJ E E
J xf(,)L,xf(,')dT (11.12)
. X
w=y,z q,q'=z,y,z
(m'IS,' Im)l(xIL,' Iw)(wIL,lx) + (xIL,lw)(wIL,' Ix)JH,
Recalling our earlier discussion about [he quenching of orbital angular momen- E E • (11.15)
z w
tum, we realize Ihat the integral of (11.12) vanishes. Therefore the only nonva· For computing mauix elements internal to the ground orbital Slate, (11.15)
nishing malrix elements of tenus pH· Land >.L· 5 connect states differing in is equivalent to our replacing the lerms plI· Land >.L· 5 by an effective teon
the orbital energy. They have. therefore, no effect in first order. in the Hamiltonian (1t'eff):
We have discussed this very problem in the absence of spin. noting that there
was no first-order tenn pH· L. since the states xf(r) and so fonh correspond ?t,rr ~ L S,H,'),P L [(xIL,lwXwILtlz) +(xILt1w)(wIL,lz)]
to no net circulation of the electron. A similar remark applies to Ihe spin-orbit q,q' w E z Ell'
coupling. The spin is coupled 10 states in which the electron has no preferenlial
circulation. The average magnetic field due to orbital motion seen by Ihe spin ~ PL S,a,q'H,' (11.16)
vanishes. ,.q'
We know, however. from our discussion of chemical shifts that the teon Since the manix elements thai make up aqq' rransfonu under coordinate rotations
pH· L will induce some orbital circulation. The spin will not. therefore, expe- like L q and Lq'. lhe aqq"s are components of a second-rank tensor. Examination
rience a suiclly zero field due to orbital motion. of (11.16) shows that is a symmetric tensor (a qq , = aq'q).
We can think of solving exactly for the wave function under the influence of For our particular case we can compute the malrix elemenlS from the oper-
the applied field and of computing the matrix elements of >.L· 5. using the exact ators
az - ay
wave function. As a practical matter we use penurba[ion theory to compute [he
effect of pH· L on the wave function, keeping only the first tenn.' We have, L. ~ ~(y~ ,~)
I
then. for the modified wave function tPzm
they do not involve the applied field H. By neglecting them, we are finding the I
field dependent coupling energy. (The second-order teons in spin-orbit coupling
L,zf(') ~
,
-7yf(,-)
produce no splitting when the spin is l) Thus the matrix elements of L z vanish, those of L y connect only to the state
z/(1'), and those of L: connect only to yf('·). This gives us contribulions only
2 We have two terms in the Ilamiitonian, {JH· Land >.L· S, neither of which gives
from teons with q :::; q'. Thai is, the x-. y-, z-axcs arc principal axes of the leosor
a fil'!ll-order contribUlion. [n ~on(1 and higher ordel'!l, both terms perturb ll,e wave
funClion. Bocause of the simi[arily belween lhe present problem and lhe ehemical shift, a'l'l" Specifically we have, using (11.16) and (11.18),
fil'!lt of all we trellllhe effoct of the term {JJI' L on the Wllve function. It may seem more
sensible to take the much larger AL· S term first. As we shall see, however, our fllll\1
answer involves I'll l\PI>ro:o;iml\tion lhat is proportional to the product {J11' Land AL· S 1t'err:::;211 (E :E SyHy + E >. E SzHz) (11.19)
(it is the interplllY between the two energics th,,~ givcs the effe<:l). For this ptlrpOlle i~ is
z z z y
imm<\tcrial which interllction we usc to perturb the wave function. By combining this result with the Zeeman teon, 211H· 5, we obtain a spin
Hamiltonian for the ground orbital state
508
509
H = {3(gzzH z S z + gyyHySy + gz: HzSz) (11.20) Equations (11.29) and (11.30) emphasize the fact that for a given orientation of
H, the splitting of the spin states is directly proportional to the magnitude of
where H. Frequently one talks about the g-shift, a tenn that refers to the difference
9u = 2 , 9yy = 2 ( 1 - E
z
AEz) between 9 and the free spin value of 2. From (11.21) and (11.30), recognizing
that both E z - E z and E y - E z are positive, we see that positive values of
A make 9 less than or equal to 2, whereas negative A'S make 9 greater than
We may employ the dyadic notation g, defined by or equal to 2. We associate positive A'S with atomic shells less than half-full,
9 = igzzi + i9 yyi + kgzzk (11.22) negative A'S with those more than half-full. Another tenninology is to remark
that electron resonances give positive A and hole resonances give negative A'S.
10 write the interaction as We shall return to this point in Sect. 11.4, where we shall find that a great deal
'Ii=pH· Ii ·5 (11.23) of caution must be exercised in this simple interpretation as a general rule.
The size of the g-shift clearly increases with the nuclear charge, as we noted
in place of in Table 11.1. Its magnitude also depends on the magnitude of the splitting to the
'Ii = 2pH· 5 (11.24) excited states to which the orbital angular momentum couples. Using an energy
of about 1.3eV (IO,OOOcm~l) and a A of lOOcm- 1 , we see 2 - 9 ~ 0.02, a
Comparison of (11.23) and (11.24) shows that the combined effect of the spin~ readily observable effect.
orbit coupling and orbital Zeeman energy is as though the real field H were We note that the g-shift arises because of the interplay between the spin-
replaced by an effective field Herr, given by orbit and orbital Zeeman interactions. It is analogous to the chemical shift that
arises from the interplay between nuclear spin-electron orbit coupling and the
H - H·
err- 2
9 -'H 9zz + 'H 9yy + kH 9zz
-t Z 2 J y 2 z 2
(11.25) electron orbital Zeeman interaction. In both cases we say that the spin (electron
or nuclear) experiences both the applied magnetic field and a sort of induced
with the resonance given by magnetic field. All such phenomena involving the interplay of two interactions
'Ii=2PH,rr. 5 . (11.26) can be viewed also as an application of a generalized form of second-order
perturbation theory. This is in fact Ramsey's method for deriving the chemical
Since gzz, gyy, and gzz are in general different, the effective field differs from
shift fonnulas. We shall illustrate by computing the g-shift.
the actual field in both magnitude and direction. If we denote by z" the direction
The problem is treated in general in Appendix D. There it is shown that the
of the effective field, it is clear that a coordinate transfonnation will put (11.26)
perturbation effectively adds a tenn, 1t new , to the Hamiltonian, which has matrix
into the fonn elements between states 10) and 10/); for our example these elements have the
'Ii = 2PH,rrS." (11.27) same orbital pan xf(") blll may differ in spin function.
Defining a perturbation leml 'H pert by
where Herr is the magnitude of Helf' The resonant frequency wo therefore sat-
isfies the condition 'H perl = AL· 5+ {3H· L (11.31)
flWO = 2{3He ff we find a matrix element (OI1t nc wIO') between 10) and 10'), as outlined in Ap-
~-,------:--:---:--;: pendix D, given by
= {3 H;giz + H~9~y + H;g;z
(OI1t
new
10') = L (OIHperdn)(nl1t per d<Y) (11.32)
= {3H 0~9iz + 0~9~y + o~g~z (11.28) /I Eo E tl
where 01, 02 and 03 are the cosines of the angle between H and the x-, y-, and By substituting from (11.31), we get
z-axes. Often one writes (11.28) as
(OI'li",wIO') = L [(01.\£' 5In)("IPH· £10') + (OIPH· £1")("1.1£·510')
tlWo = g{3H (11.29) n Eo En Eo E/I
where the "g-factor" is defined by the equation
(OIPH· £In)(nIPH· £10') (01.\£·51")("1.1£· 510')]
+ + .
222222 (11.30) Eo E/I Eo - En
g= gzzoJ + 9yy Cf 2 + 9z:03
(11.33)
511
510
The first two tenns on the right give the g-shift we have calculated. The last two I<lg. J1.3. Ener&)' I~yel$ for L ;:;; I, 5 ;:;; 1 re-
tenns shift the twO spin states equally. They do oot, therefore, either produce a sulting from a spin orbit coupling >.L· ~
splitting of the doubly degenerate ground state or contribute to the g-shift. (If
the spin were grealer than 4, however, such a tenn could give a splitting of the
ground spin state even when H:: O.) The last two teons are just what we should
have had if either perturbation were present by itself. Our previous calculation We consider next the effect of Fl' Here il becomes convenient to assume
of the g-shift did not give them because it treated the effect of one tenn of the a specific fonn. Assuming that the potential arises from charges external to the
perturbalion (PH· L) on the other (AL . 5). The method could be extended to atom, Ihe potential in the region of the atom can be expressed as a sum of the
find aU the tenns included in (11.33), but we see that direct application of (11.32) fonn
gives us a systematic method of getting all tenns. On the other hand, the physical FI ;:;; L Clm ,.Il'jm (11.39)
principles of the first calculation are somewhat more apparent. I,m
In the example we have discussed so far, we have considered the crystalline
where the YIIlI'S are spherical hannonics and the Clm'S are conslants. If the
potential VI to be much larger than the spin-orbit coupling constant A. As a result,
potential is due 10 the charges of Fig. 11.1, it vanishes on the z·axis and changes
the orbital angular momentum is largely quenched, and the g-value is very close sign if we replace z by y and y by -z (a coordinate rotation). It is a maximum
to the spin-only value of 2. This situation corresponds to the iron group atoms
on Ihe x-axis and a minimum on the y-axis for a given distance from the origin.
as well as to many electron and hole centers. We tum now to the opposite case,
The lowest I in the series of (11.39) is clearly I:: 2. Of the five 1:: 2 functions,
one of strong spin-orbit coupling and relatively much weaker crystalline fields,
.:z:y, xz, yz, 3z 2 - r 2, .:z:2 - y2, only the last is needed. We have, therefore, as an
as encountered in the case of rare earth atoms. approximation, insofar as teons for 1> 2 are not required,
If the spin-orbit coupling is dominant, the situation is in first approximation
similar to that of a free atom. In fact the Hamiltonian (11.40)
p' where A is a constant. (We shall see shortly that no higher terms are needed for
'Ii = - + Vo +AL ·5+ pH·L +2fJH'5+ VI (11.34)
2m an exact treatment.) We have, then, to consider the effect of VI on the states of
is identical to that of a free atom except for the tenn VI. As a first approximation Fig. 11.3. Two sorts of matrix elements will be important: those entirely within
we consider just the effect of the tenn >.L· 5 on the state formed from spin a given J, such as (JM; IVtlJMJ), and those connecting the different J states.
functions and the three p-states xf(r), yJ(r), and ZJ(I·). The sum of the angular The fonner will be the more important because they connect degenerate states.
momenta Land S is the total angular momentum J: We can compute the matrix elements internal to a given J by means of the
Wigner-Eckan theorem, for we notice that, with respect to L z , L y , and L:
J::L+5 . (11.35) (hence with respect to J z , JIJ , and J:), F I is a linear combination of Tu,,'s.
By squaring, J, we have That is, the commutation relations of VI with J:, JT. ± iJy show VJ 10 be a linear
combination of T 2M 'so (In fact Vt is proportional to T22 + T2-2.) Thus we have
~L·5= ~(J' - L' - 5') (11.36)
2 (JMJlVtlJM~) =A(JMJlz' - y'IJM'/)
with eigenvalues Esc given by :: CJ(JMJIJ; - J~IJMJ) (11.41)
~
£So = 2:[J(J + 1) - L(L + I) - 5(S + I») (11.37) This is equivalent to our replacing VI by the operator 1ft.
Since we are concerned with states all of which are characterized by an orbital Hl :: CJ(J; - J;) (11.42)
quantum number L and spin quantum number 5, the possible values of J are as long as we only compute matrix elements diagonal in J.
£+5, £+5-1, ... IL- 51. For our example, L:: 1, 5::::t!, so that J:: ~ or As an allernative, had the potential VI been B(3.: 2 - r 2), we should have
!. !
The J :: ~ and states are therefore split apart by an energy spacing used
512 513
We have still to consider the effect of Ihe magnetic field terms. Again lei thai a quadrupole coupling cannot split a pair of spin i levels). For J :: 3, we
us consider only the malrix elements dia{:onal in J. This means that we wish to have ~
have 1 .
IJJ) = J2(%+'Y) ( 11.52)
(JMJIPH· L +2pH· SIJMj) = pH· (JMJIL +2SIJMj) (11.44)
We can apply the Wigner-Eckart theorem to the malrix element, which, with where we have denoted Ihe spin function S~:: +~ by ul/2' Therefore we find
respect 10 J, is a linear combination of TtM's. We can therefore write
(J MJIL + 2SIJMj) = 9J(J M J IJIJMj) (11.45) ~/2 = BJ
'3 I 2 + y2)f2(r)(3z 2 - r 2)dr
2"(x (11.53)
where 9; is a conSlanl for a given J, independent of M; or M J. We recognize Th~ angular ~nions of the integral can be carried through, using spherical co-
this problem to be the same as that of compuling Ihe Zeeman effect of free atoms, ordinates, to gIVe
and the constant 9; is therefore the familiar Lande 9-factor: ~ 2B-
J(J + 1) + S(S + 1) - L(L + I)
'""3/2:: --r'
15 (11.53,)
g; = 1 + 2J(J + I) (11.46)
where r 2 is the average value of r 2 for the p-states.
Equation (11.45) is equivalenl (as far as malrix elements diagonal in J are The ~p.in. Hamiltonians of (11.48) and (11.49) do nOI include any malrix
concerned) to our having a term 'Hz replacing lhe two magnetic terms, where elements JO,"lng states of J = iwilh those of J = !.
If we wish to include
'Hz is given by such effects, we can actually apply the Wigner-Eckart theorem in an ailemative
manner. All matrix elements of VI are between states of L = I. Thai is, we are
"/iz=gJPH·J . (11.47)
concerned. only with (LMr,lVdLMJ). However, the commutators of L z , L y •
By combining (11.42) and (11.47) with (11.36), we obtain an effective spin and ~, with VI show that VI is a linear combination of T 2M 's. Therefore all
Hamiltonian 1te fr' which describes our problem accurately within a given J: matnx elements are of the form
1tcfr=Cj(3J;-J 2)+9;pH.J (11.49) These matrix elements are equivalent to those we should have had by replacing
VI by an equivalent Hamiltonian 1t1:
The tWO tenns on the right of (11.48) and (11.49) lift the (2J + I)-fold degen-
eracy of each J Slate. We shall not discuss the details of handling this problem 1t1 = GL(L; - L;) or 1t) = C'L(3L; - L 2) (11.55)
other than 10 remark Ihal clearly it is formally equivalent to the solution of the In terms of 1tI, the effective Hamiltonian 1te rr. which will give all matrix ele-
problem of a nucleus posscssing a quadrupole moment actcd on by an electric ments fonned from the three p-states xf(r), yf(r), and zf(r), is
field gradient and a static magnctic field.
So far we have not computed the constants C; or Cj. We lum now 10 Ihat 1te rr=>.L'S+C L(L;-L;)+PH.(25+L) or
task, illustrating the method by computing C J. 1teff :: >.(L· S) + Ci,<3L; - £2) + pH· (25 + L) . (11.56)
We have. using (11.43)
B(J M;13z 2 - r21JMj) = Cj(JMJ I3J; - J2lJMj) (11.50) Equation (11.56) reduces 10 (11.48) or (11.49) for matrix elements diagonal in
J.
By choosing /If; = Mj = J, we have In the absence of an external magnelic field, the energy levels of Ihe Hamil-
tonian of (11.56) would remain at leasl doubly degenerate, according to Kramer's
B(JJI3z' - r'IJJ) = Cj[3J' - J(J + 1)1 (11.50.)
=
theore~, ~Iated i~ Section 10.5 since £ I, S =!.
It IS Interesung to note that the derivation of (11.45)
c'J = J(2JB _I)(JJ I
3,' -r 'I JJ) (11.51)
(JMJIL + 2SIJ Mj) = YJ(JMJIJIJMj) (11.45)
Now for J = 4, all matrix elements of 3J; - J2 vanish (analogous to the faci using the Wigner-Eckart theorem. is quite analogous to the stalement about the
514 515
nuclear magnetic moment i-L and spin I connecting states of different crystalline energy such as x f(r) and y f(7'). We shall
(11.57) neglect the elements that are off-diagonal. We have, then, only matrix elements
such as
which, stated more precisely, is
(IMI"IIM') = 1t.(IMIIIIM') (11.58)
(~msmJIHJSrlxmsml) = JtP~,u;"sxf(r)11.rSr
, ,
x tPm,u m ' xf(r)drdrsdrJ (11.62)
We expressed the potential of Fig. 11.1 by means of only one tenn in the
expansion of the potential, that with I = 2. We might have supposed also that where ms and mr stand for eigenvalues of Sz and J z ; tPml and ums are nu-
tenns 1 = 4, 6, and so on, would have been needed. (The odd 1's are not needed clear and electron spin functions; and where dr and drs stand for integration,
for this example, owing to the inversion symmetry of the charges.) If we had respectively, over electron spatial and spin coordinates, and drl over nuclear spin
included an 1=4 tenn, we should have then needed to compute matrix elements coordinate. It is, as usual, convenient to leave the specification of the quantization
such as of both electron and nuclear spins until later, since the appropriate quantum states
will depend on other parts of the Hamiltonian. We therefore omit the electron and
(LMLlrIYtmILM~) (11.59) nuclear spin functions and integrations, computing only J xf(7')Hrsrxf(7')dr.
This integral leaves the nuclear and electron spin coordinates as operators. We
with 1 = 4. However, with respect 10 L1;' L y , and L z , rlYtmis a Tim' Therefore
shall therefore denote it by 11.rs:
we can apply the Wigner-Eckart theorem 10 evaluate it. Recognizing that such an
integral is closely related to the combination of angular momentum (by means of
the Clebsch-Gordan coefficients), we note that it will vanish unless L and I can
HJS = Jxf(r)11.J Srxf(")dr (11.63)
couple to fonn an angular momentum L (the triangle rule). For the case L = 1, By substituting (11.61), we obtain
I = 4, we can see that L and I could combine 10 give angular momenta of 5,
4, or 3, so that the integral must vanish. In fact, for L = I, only I = 2 gives
nonvanishing matrix elements. We need not, therefore, bother with 1 = 4, 6, and
HJS =iein1l2 J ,.13 [3(I ·2r)(s.r)
r
~I'slx2f2(7')dr (11.64)
so on, in the expansion of the potential. The tenns such as JzxSyY will contribute nothing, since for them the integrand
is an odd function of x or y. The other tenns can be expressed as the product of
angular and radial integrals, giving
11.3 Hyperfine Structure
(11.65)
We have not as yet considered the magnetic coupling of the electron to the nearby where, as usual, (1/1- 3 ) denotes the value of I/r 3 averaged over the state xf(r).
nuclei. The basic fonn of the interaction has been discussed in Chapter 4. We If instead of a p-state we had an s-state. or more generally a wave function
distinguish between s-states and non-s-states. If; containing some s-state, we could compute a corresponding tenn Hrs arising
s-states : from the {j-function coupling:
81f ,2
Hrs r = 3ieinll I· 5{j(r) (11.60) HJS = J1f;*(r)H rsr 1/;(r)dr = 8 1f ")'<)")'011.2 11/;(0)1 2 I· S
3
(11.66)
The most general interaction is. of course, the sum of the couplings to the inter-
Non-s-states :
actions of (11.60) and (J 1.61):
li2
H = iein [3(l.r)(5.r) -1.5] (11.61)
rSr r3 r2
hrs = Jh b(I')1 2 {8; i<)inrt2I.5~(r)
The effect of the hyperfine coupling can be illustrated by considering the
example of the preceding section in which the orbital angular momentum was + 101"h'[3(I'T)(S'T)] -I'S}dT (11.67)
rJ r2
quenched (corresponding to a crystalline potential VI much larger than the spin-
orbit coupling A). In the Hamiltonian matrix there will then be elements of It will be linear in the spin variables Jz , Jy , I z and Sz, Sy, and Sz, being of the
the hyperfine coupling diagonal in the electron orbital energy as well as those general fonn
516 517
11.IS: L: ACiOlSoJol (11.68) We see that I z commutes with (11.73). We therefore take the states to be eigen-
Ci,Ci':;e,y,z functions of I z with eigenvalues mI. The first-order energy is therefore
but with the A's symmetric (that is, A Cia , = Aa'a)' One can therefore always find E.= fJguHmS - 'YnhHmI + AzmSmI (11.74)
the principal axes such that A aa , is diagonal, with values Act and the hyperfine
coupling given as Since I 2 , 52, and I z and 5 z all commute with the Hamiltonian of (11.73) we
can take the eigenfunctions to be a product of a nuclear spin and electron spin
(11.69) function:
5'8 519
r--~-j I- 2,.H,-j
The Hamiltonian becomes
I-----;:;~ _I I
1t = 2{JH:,5,' -inuR:,I" + A z l z 5 z + A y I y S lI + A:I,S:
_I
(ll.80)
We shall continue to assume 2{JHo ;» A z • A y • and A, so that, to a good approx-
imation, the Hamiltonian commutes with S". We now seek to find those other
l,.Ho A. ~
2A' parts of the Hamiltonian that will be diagonal in a representation in which 5:, is
J<'ig.lI.5. NudcllT ~nanc' when the IlU- Fig. 11.6. Nuelcar resonance for all elec- diagonal. To do this, we express the spin components Iz.Iy • I" and 5 z , Sy, and
clear Zeeman energy islllrger than IA. msl t
tronic spin of when the hypcrfine cou- S: in terms of the primed axes. (Actually it would be sufficient to express 5 z • 51/
drawn on the llSSumption of an eleetron pling is larger than lht nuclear Zeeman
and S, in terms of 5 z " 5y" and 5". However. the transformation of the nuclear
spin of ! coupling
axes enables us to see more readily that the nuclear quantization direction is not
parallel to that of the electron). Noting that y = y'. we have
Sz: = Sz:1 cos 8+S" sin 8
w""
A.lmsl
h
± H
"'In 0
{I 1.79) Sy = Sy'
5 z = 5,. cos 8 - 5z:' sin 8
For an electron with spin i.
the result is shown in Fig. 11.6.
I z = Iz:' cos 8 + I" sin 8
(11.81)
An examination of Figs. 11.4. 11.5. and 11.6 shows that the electron reso-
nance will enable one 10 measure the hyperfine coupling tensor and the nuclear I y = Iy'
spin, but will nOI by itself enable one to measure the nuclear moment. When I, = I,. cos 8 - I z ' sin 8
combined with the results of a nuclear resonance, even the nuclear magnetic
By substituting these expressions into (11.80). we find
moment can be found. This latter feature is of particular importance in the slUdy
of color centers because it enables one to identify the nuclear species that gives 1t = 2{JHo5" - inhHoI,. + I z .5z ,(A z cos2 8 + A, sin2 8)
rise to the hyperfine spliuing, since the nuclear i'S are known from other exper- + Iy'5y'AI/ + I,.5,.(A z sin 2 8 + A, cos2 8)
iments. The nuclear resonance, on the other hand. enables one to measure the
+ (1z:.5,. + I,.5z ' )(A z - A,) sin 8 cos 8 (11.82)
spin of the electron system.
We have so far restricted ourselves to an orientation of the static field along To first order. those terms in lhe hyperfine coupling involving 5 z • or 5y'
one of the principal axes of the g- and hyperfine tensors. Interesting new effects possess zero diagonal elements in the 5" scheme of quantization and therefore
arise when the field lies along other directions: The axis of quantization of may be omitted. In second order. they will contribute energy shifts of the order
the nucleus becomes different from that of the eleclrOn and in fact depends on of A 2/2{JHo. where by A 2 we mean the square of a matrix element of order A z •
whether the electron is oriented parallel or antiparallel to the field. For the sake A y , or A,. and the 2{JHo comes in because the "excited state" differs in electron
of simplicity we shall illustrate these effects for the case where the electronic spin orientation in the static field. The reduced Hamiltonian 'Hred' which results
g-factor is isotropic. We shall take the field direction i to lie in the x-z plane. from dropping all lerms involving Sz or 5y • is then
where x, y, and z are the principal axes of the hyperfine coupling. The orientation
'Hre<J = 2{JHoS" - 1'nliH01" + {Iz,(A z sin 2 8 + A z cos 2 8)
is illustrated in Fig. 11.7.
+ I'l'.(A z - A,) sin 8 cos 8j5z' (11.83)
, Of course 5 z' commutes with 'Hred' However, I" does not As far as the nucleus
is concerned. 'Hrcd corresponds to a nucleus coupled to a magnetic field with
components Hz •• H y" and H" of
}f
1 .
H'l" = -~ (A, - A z ) Sin 8 cos 85,.
1'n,l
x'
hyperfinc tensor A, relative lo lhe direction of the stalic
field " H z '-H
-O-
2
(AzSin 8+A:cOS
~
2
1'n/l
8)s :'
520 521
These expressions involve the operator 5 z" by which we mean, of course, that the 1-'lg.It.8. The axes (:1:',::2) are rolaloo an angle q, aboul the 111 :<
effective field depends on the electron quantum state. By denoting the eigenvalue 112 uis ~Ialive lo lhe aXelI (ZI, zd
of 52' by rns. we have
522
523
clear echo case? For a pair of coupled nuclei of spins I and S respectively, the
(11.91) quantization of one nucleus is to an excellent approximation independent of the
orientation of the other nucleus. Consider, then, application of a 1r/2 ... r ... 1r
puls~ sequence to the 5-spins. If the I-spin has a particular orientalion, it main-
This refinement will slightly modify the treatment of (I 1.81-84). tains this orientation despite the S-spin pulses. Therefore, whether the I - S
The other refinement worth mentioning is to include the fact that the nucleus coupling enhances or diminishes the 5-spin precession rate does not matter.
exerts a magnetic field on the electron. II therefore should playa role in the clec- The dephasing produced thereby between the first and second pulse is refocused
(ron spin quantization. Thus, if the nucleus is flipped, the electron quantization between the second pulse and the echo.
direction changes slighdy. This change is ordinarily small and can frequently be Now consider the electron case. Suppose the electron (5) is initially pointing
neglected. However, there are some situations in which it is important. Waugh up. Let the nucleus (I) be pointing up along the effective field. Thm is, the
and Stichter [11.2] have called this slight change in electron spin quantization nucleus is in a stale u+ of (11.90) and the total system is in the state 1+ +).
"wobble," They point out thai it gives the electron a degree of freedom much The 1r/2 pulse now tilts the electron. As a result, the nucleus suddenly finds the
less costly in energy than flipping between the stales 0' and p. They point out effective ficld acting on it tilted. The nucleus thus begins to precess about the
that it may playa role in nuclear relaxation by paramagnetic ions at very low new field direction. Consequently the field it produces on the electron changes in
temperatures since wobble can be excited by the small thermal energies, but time. Clearly the magnetic field acting on the electron during the interval 0 to T
electron spin flips between states 0' and (J cannot. will, in general, differ from Ihal between T and 2T. Therefore. the echo will not
refocus the spin-spin coupling completcly. If, however, T corresponds exactly to
a 2'/T precession of the nucleus, then the spin-spin coupling will have the s:lme
time average during the two intervals, hence the echo will be perfec!. Thus,
11.4 Electron Spin Echoes we expect Ihe echo envelope to oscillate with T such that maxima in the echo
envelope correspond to r's that are integral multiples of the nuclear precession
The use of pulse techniques, in particular of spin echoes and stimulated echoes, period. We will see in our fonnal treatment of the problem that the description
has become important in electron spin resonance. We turn now to some special we have given is exact. In that connection, we can note that 10 think about lhe
features that arise for this case which are not encountered for spin echoes in quantum treatment of the effect of a '/T/2 pulse on the electron system, the general
nuclear magnetic resonance. As we have seen, if we have two coupled nuclei concepts developed in Sect. 7.24 are helpful. In particular, they help us deal with
belonging to two different species (e.g. 13C and IH), the spin-spin splitting of the problem that there is not just one nuclear precession frequency. Rather, there
one species produced by the other does not affect the size of the echo. Indeed, it are twO, depending on whether the electron is in its spin-up or spin-down state.
acts just like a static magnetic field inhomogeneity. Therefore, although it affects We begin by expressing the reduced Hamiltonian (11.83) ill different nota-
the free induction decay, it is refocused perfectly at time 21" by application of the tion. First we define the frequencies a, b, We, and W n :
second pulse of the echo sequence. In order to observe the spin-spin coupling
on an echo, one must do a double resonance experiment (Sects.7.19-21). rl2a :: Ax sin:! () + A z sin:! ()
For an electron (spin 5) coupled to a nucleus (spin I) the situation is dif- h2b :: (A'l; - Az)sin () cos ()
ferent owing to the fact, which we have just discussed, that in many cases the W
n
::inHO (11.92)
nuclear quantization direction depends on the electron spin orientation. As we hWe :: 2(JHo
shall see, the consequence of this coupling is that the amplitude of the elec-
tron spin echo is modulated as a function of the time 1" between pulses by the Then
1- S spin-spin coupling. The modulation frequencies are the splitting frequen-
'Hu'<J :: h[we5 z' - Will..., + (a1z1 + b1'1:' )25z/] (11.93)
cies of the nuclear magnetic resonance. These are the frequencies one observes
in ENDOR. Therefore, the electron echo envelope in effect reveals the ENDOR We now !Tansfonn this Hamiltonian to the electron spin rotating frame at fre-
frequencies. It is this special feature which has made spin echoes such an impor- quency W
tant technique for electron spin resonance. This phenomenon was first reported
De:: We - W (11.94)
by Rowan et al. [11.3] and subsequently elucidated in greater generality by Mims
[11.4 1. using the operator exp(iw5z d) 10 !Tansfonn the wave function
Before providing a formal description, let us examine physically what is t/J1::e-iwlS.,t/J with (11.95)
happening. In particular, why is the electron echo case different from the nu-
525
524
1i~ed '" fl[neS~, -wnl~, + (al~, + bl:t:,)2Sz'] (11.96) (11.102)
To save in writing symbols, we now drop the primes from Xl and z' in (11.96), The directions with respect to the z-axis are shown in Fig. 11.9. The angles 6+
understanding that from now on the new unprimed axes are really the old primed and ~_ shown in the figure obey the equations
axes.
To solve for the eigenstates we note first that Sz commutes with 1i~ed' b b
tao 8+ '" - - tao 6_ '" - - (lU03)
Therefore, the eigenstates, 1/J1 , of 1i:.w can be labeled by ms, the eigenvalue of Wn-a wn+a
Sz. Thus We then define the unit vectors k+ and k_ as
1fred ¢'(ms) "'1i(OemS -wnl: + (al: + bIz)2ms]¢'(ms) (11.97a) (11.104)
which give for the components of I along w+ and w_
= ~il.ms + 'li/(msl] .p'(ms) • (l1.97b)
1z+ :::: k+ . I , 1z _ '" k_ . I (11.105)
where 1i/(ms) is an operator in the components of I, given by
These enable us to write
1i/(ms) '" -1i(Wn1: - (a1: + b1:t:)2ms] (11.98)
1i,(!) ::::-flW+·1=-/lW+1:+
For ms :::: +! we have (11.106)
'li/m = -"(w" - a)I. - bI.] (11.99) 1i,( - i):::: -llw_· I'" -llw_lz _
and for ms '" -! Then if we define the eigenfunctions 1.1+. u_, v+. v_ by the equations
z Then, returning to (11.97) to express 1i~ed' we find that when 1t~ed operates on
z
any of the four functions
z+ 1++) =:(tu+ 1+ -) =: (tIl_
(11.109)
1- +) "pu+ l- -) " pv_
that they are eigenfunctions of 1t~ed with eigenvalues
Wn +a I I I W+ h(ilo - w+)/2 , Ii(nc +w+)12 ,
(J 1.110)
6_ I Wn - (J I
h(-ilo -w_)/2 h( - ilo +w_)/2
L_.J.:__
-b- x
I 1__ -"---_8+
-L-
_
x
respectively. The four functions of (11.109) therefore solve the reduced Hamil~
tonian.
FIg. 11.9. The two angular momentum vectors '"'+ and '"'_ in the :t: - : plane, defining Utilizing the definitions of 6+ and 6_ given by (11.102) and shown in
the angles 8+ and 8_ made by '"'+ and '"'_ respectively with respect to the ~-axis. Note
the sign conventioll (indicated by the arrows) for positive 8+ and 8_. The axes:+ and Fig. 11.9. we can calculate the angle 60 between z+ and z_. shown in Fig. 11.10.
,_ lie along '"'+ and '"'_ respectively
(11.111)
526 527
•
[
z J.'ig.1I.10. The angle 80 between the angular momentum where for simplicity we do not write out the contents of the inverse brackets and
vectors \<>+ and ",_. With 0+ and (L both positive in the where the subscripts I and 2 on the brackets serve to distinguish the contents of
sense of thO) arrows, (}o = 8+ + 9_
the brackets.
E.xplicit evaluation of X S1(1T)T(t - r)XS(rr) gives us
X
s 1
(1t)T(T)X 5(11")
XS(1rI2) ... T ... XsC1r) (11.114) Moreover, we get a more compact notation if we define the unitary operators
R(Sz) and R(-Sz) by
Then defining the lime development operation T(t), in the electron spin
rotating frame we have R(S,) " exp{ -i[(aI, + bI.)2S, -woI,]r} (11.121)
,ad
T(l) =ex p( -*1{~edt)
R(-S,) " exp{i[(aI. + bI.)2S, +w"I,)r} (11.122)
= «p{ - i [D,S, + (aI, + bI.)2S, - woI,] I) (11.115)
where R(-Sz} is obtained by replacing 5 z of R(S:) by -5:. Then we get
We take the density matrix e(O-) prior to the first pulse as ,q(O-) = Sz. Then
the signal at time t (we take t to be after the 11" pulse), (S+(t», is given by
(S+(I)) = Te{S+ e(l)) We now evaluate the (face using a complete set of basis function. For the com-
= Te{ S+ [T(I - r)X s (·)T(r)Xs(·/2)] , S,[ 1,' 1 plete set we pick initially the two electron States a and fJ with quantum number
ms and the two nuclear states u+ and u_. which we denote by the quantum num-
= Te{ S+[Xs(.)Xs'(.)T(1 - r)Xs(.)T(r)],S,[ I,'} ber p.. which can be +! or -!.
Thus, using the fact that (msllIS_lmsp./) = 0
(11.116) unless p. = l, and so on,
528
529
(S+(2T) = L: (msIIIS_lmSII)(msPIR( -S:)R(S:)lmSIt') Then, for any function j expressible in powers of the components of I,
"'S,I'
...5·.. ',1''' R1/2fR_I/2=exp(iw+Tlz)jexp(-iw+rlz)=j , (11.132)
X (msJ/ lS'y Im~Jt') (lnsp'IR-1 (Sz)R- 1(-S:)lm~plI) since. every component of I is rotated onto itself for this value of W+ T. Therefore
(11.123)
R-l/2Rl/2R=:/2Ri/~ =R_ 1/ 2R=:/2 =1 (11.133)
In order to have the sums be diagonal sums, we get that m~ = ms and p" = JJ.
Now the matrix element of S_ requires that Thus
I I
ms=mS-I. (11.124) (S+(2T» = 2iT<d1j = 2i(2[+ I) . (11.134.)
Therefore, we get contributions only from But this is, apan from the usual sign change in a 1f!2, 11" echo, just the signal
ms = -t ' ms = +! (11.125)
we would have immedialely after applying an XS(1I:!2) pulse:
(S+(O+» = T<{s+S,) = iT< { S~) = i(2[ + I)Trs{ S~}
,. ,
and therefore (mspIS-lm's II) = 6", ",' -I' Then we get for Ihe matrix elemen!
. S(S+I) i
of Sy =1(2[+1) 3 (2S+I)=2(2[+I). (l1.I34b)
(ms/IS,lms/) = WS,I-!) = ;i WS+I-!) = ;i (11.126) Therefore, if W+ T = n2:lf we get an echo corresponding to the full magnetization.
A similar argument holds for the condition
so that
(S+(2T» = ;i L (!,'IR(-S.)R(S.)IV)
Il,Il'
W_T = n211:
Now, lcuing (he operator S: of R(Sz) and R(-S:) operate, we replace S: R_l/2RI/2R=~/2Ri/~ = exp(iw_ TI z_ )R 1/ 2 exp( - iw_ TI z_ )n-;/~
by ms (and -Sz by -ms) in the matrix elements. Then, defining the operators
T
= exp(iw_ I z _) exp [ - iw_ T( RI/2Iz_Ri/~)]
R- 1/2 =exp{i(alz+bl.l"+wnI:)T} , (11.136)
(11.128)
R 1/ 2 = exp{i( - aI: - bI", +wnlz)T} , Now RI/2Iz_R~/~ provides a rotation of the veclor k_Iz _ about the vector
we get w+. When this rotation is an imegral number of cycles
(S+(2T» = ;i L ("IR_1/2RI/2I''')(I''IR=:/2R~AI")
/,,/,'
R 1/ 2 exp( - iw_ rI z _) Ri/~ = exp( - iw_ T I z _) (11.137)
otherwise
= ~iTrJ{R_I/2RI/2R=:/2RI>~} , (11.129)
Rl/2R=;/2Ri/~"4 Ri/~ (11.138)
where TrJ means a trace over the nuclear spin eigenstates only. We can express
and the producl
the operators R 1/ 2 and R_ 1/ 2 in a ronn that gives them a simple physical
meaning now by llIilizing 01.101 and 105) R_1/2R1j2R=:/2Ri/2"41 (11.139)
R t / 2 =exp(iw+Tlz+) , R_ 1/ 2 =exp(iw_Tlz _) (11.130) Clearly these equations express mathematically our initial physical description
Therefore, R 1/ 2 generates rotations at angular velocity w+ about the z+-axis, of why lhe spin echo envelope is modulated as T is varied, returning to its full
and R_ 1/ 2 gener.ates rotations about the z-axis at angular velocity W_, amplitude when either
Suppose now that T corresponds to an integral number of 211" rotations about w+T=n211" or w_T=n211' , n=I,2,3 .... (11.140)
the z+-axis:
We now wish to evaluate the lTace explicitly. Denoting the eigenvalues of
w+T=n211" . (11.131) 1z + by p and of 1z _ by q so that
530 531
(11.141)
If 00 = 0, so that {here is no change in the quantization direction, (1 t.147) reduces
we get 10
TrdR_l/2Rl/2R=~/2Ri/2} = 2 ,
TTl {R_l/2RI/2n-~ :/2Rl>~}
which 'agrees with (11.134). That is, it is independent of ,. Thus, the oscillation
= I:: (q IR_ I / 2 Iqll ) (0.1 IR 1/ 2 1qll) (u"in: :/210'1/1) (alII IR~/~ 10") comes about because of the change in quantization direction of the nucleus when
,"
,,11:,,11' (11.142) the electron spin flips.
Note also that if either
= L e+ iUW _ T
(aIR 1/ 2 Iqll ) e-iUItW_T (qIlIR1/2 Iu)
cos(w+,) = I or cos(w_,) = 1
",17"
= E eiW_T(<T-u") (0"IR 1/ 2 Iqll ) (a IR1/ Iu)
Il
2
(11.143)
Tr I {R-l/2RI/2R=:/2R0~}
(T ,U"
4
= 2 [ cos «()0/2) + sin 4 (00/2) + 2 sin2«()0f2) COS 2«()012)]
=
2 2
,,',," = 2 [ cos «()o2) + Sin «()o2)t
1',1" "II ,I''''
=2 .
x (fJ. "IR-II
1/2 I' "')( II- "'I (1 )
Thus, we get the full signal, as we showed with (II 133).
and relabeling 0'" as (/ and using the faci that the states III) are eigenstates
of I~+, we get
11.5 V, Center
TTl {R_l/2R'/2R=:/2Rl/~}
= L ei~.LT(O'-"") (ulp)eiw+TI-' (11.10-')(/7'1//) e-iW+T/1' (/l'la)
A particularly interesting example of the application of the ideas of the previous
",u '
1'",1 sections is the discovery and identification of the so-called Vk center by Kiinzig
(11.144) and CaStner (11.5]. The detailcd analysis of the spectra will enable us to discuss
the g-shifl more thoroughly, including the effects of having several electrons and
more than one force center.
Since 17, ql, J~, and fJ.' can each take on two values (+! or -!) there are 16 Kdnzig and Castner's first work in electron spin resonance was followed
by a set of beautiful experiments by De/becq [11.6], who combined optical tech-
lenns. The factors (ulp.) are given by expressions such as, if
niques with electron resonance to detennine (I) the optical absorptions associated
17=! fl=-! with the center and (2) the energies of the excited stlltes. A full account of all the
(ulr) = Jv~u+d7" (11.145)
work on Vk centers would take us too far afield. We shall remark on the method
of identification of the center, and on certain features associated with the g-shift,
where the integral is easily evaluated using (11.109): which are not found in "one-atom" one-electron centers.
It is helpful to begin by, so to speak, giving the answer. The Vk center is
(0"::: ~I/~ = -!) = sin(Bo/2) (11.146) formed in alkali halides by X-raying crystals al a temperature near that of liquid
nitrogen. In this process, electtons are ejected from the negative halogen ion,
and so on. It is then straightforward to show that
changing it from a closed-shell configuration to one with one electron missing
Tr I {R_I/2Rl/2R=:/2Ri/2} from the p-shelJ. The ejected electron may have a varicty offatcs. We shall simply
assume that not all recombine wilh neutral halogens. As an example, consider
2 2
= 2 [ cos 4(00/2) + Sin 4(00/2)] + 4 sin (00/2) cos (00/2) CI; the neuttal chlorine atom is unstable and pulls togcther with a neighboring
Cl- to form what may be convenicntly called a Cli" molecule. The Cl-CI axis
x [COS(w+ ,) + cos(w_ ,) - cos(w+ ,)cos(w_ ,)] (11.147)
turns out to lie in 110 or equivalent crystal directions, as illustrated in Fig. lUI.
532 533
+ + 1'lg.IUI. A eli" molecule, or V. center. in KC!. The «nl~ may recognized to arise from the coupling of the unpaired electron spin with magnetic
be thought of ILS a hole, denoted by h, trapped on Il pair of Cl-
moments of a pair of chlorine nuclei.
+Y:J + ions
Let us explain. There are two isotopes of chlorine: C135, which is 75 percent
a~undant; and C~37. 25 percent abundant. Both are of spin i.
+ ((/+ bUI have slighLly
different magnelJc moments ("(37/7'35 = 0.83). As we shall show. the seven-line
pa~lem arises from a pair of Cl35 ·s. Let us assume that Ho lies along a principal
+ + axiS of the eli; call it Ihe z-axis. By recalling our discussion of the preceding
section and generalizing it slightly, we see Ihat the electron resonance condition
will be
The electronic sttucture turns out to be very similar 10 (hat of the p-electron
with quenched orbital angular momentum, discussed in Section 11.2, the electron 1
w = h[9uPHO + A,(m) + m2)] (11.148)
with the unpaired spin being in an orbit whose axis is parallel 10 the bond direction
of the eli" molecule. Coupling to excited stales gives a g-shift thai varies as the where m, and m2 are the m-values of the two CI nuclei. Therefore the frequency
magnetic field orienlalion is varied with respect 10 the crystal axes, and the depends on m, and m2 only through their sum ml + m2. The largest m, + m2
coupling of the unpaired spin with the nuclear moments of the two chlorines i ! i
is + ~ = 3. The next largest is + = 2. This value of ml + 1»2 is also found
gives rise to a hyperfine coupling. if mt = ~. m2 = !.
Since we assume the nuclei to be distributed at random
As we have remarked. this center was first discovered by elccuon spin among their m-Slates, the line at m) + m2 = 2 will be twice as strong as thai
resonance. A pattern observed for a case where the stalic magnetic field is along for m) + m2 = 3. We indicate the possible combinations of mt and m2 in Table
the 100 crystal direction is shown in Fig. 11.12. 11.2,
At first sight the spectrum seems too hopelessly complicated to unravel, but The recognition Ihat the seven main lines could arise from a scheme such
fortunately a convenient starting point is the set of seven prominent lines. They as we have described was the first clue as to the nature of Ihe resonance. If
are nearly equally spaced and are of intensities 1: 2: 3: 4: 3: 2: I. These were one accepts this clue, Ihat Ihe electron spends equal time on two CI atoms, one
mUSI next consider what happens for CJ; molecules in which the nuclei are bolh
Cl37,S or one is a 35 and the other is a C1 37 . The probability of finding a
a
~ 0
3
single atom of CI 5 is ~; that of finding a CI37 is t. Therefore the probability
of finding pairs is as follows:
10 a 35 _ CJ35:
20 CJ35 _ CJ37:
~30 CJ37 _ C135:
Cl 37 _ C137:
~"
• H
I II. ,I 1
" 1 II' 'I I The generalization of (11.84) for unlike nuclei may be wriuen as
"
70
w= *
[9uPHO + (A;5 m35 + A;7 m37 )J] (11.149)
100
80
90
,
. , , ...
, , 3.3
, , U, , ..
, kilocauu
,
in an obvious notation. The hyperfine coupling, as we have seen, is proportional
to the nuclear "('s, so that
A
3S
7'35
A37 = "Y37 (11.150)
Fig. 11.12. VI: rellOnance in KCI for ~he static field parallel to a 100 crystallographic axis. The result of having nonequivalent nuclei is that configurations such as (3 I)
t 3 . ~'~
(This figure kindly supplied by Klinzig and Casfnl!r)
and (~.~) no longer gIVe the same frequency. If we call the (i.~) intensity
534
53S
Table 11.2. Combinations (ffil,ffi7) to make the same total valuc m\ + ffi7, and the We can see that in general there will be several hyperfine patterns whose centers
corresponding frequcncy and statistical weight of gravity are displaced because of the anisotropy of the g-factor. Moreover, the
hyperfine splitting is itself strongly anisotropic. A~ being much bigger than either
A:t" or. A y (z being the bond axis). The anisotropy is interpreted as indicating
thal on each atom the individual bond function is a linear combination of an
(~,~) 3A 1
s-function g(r) and a p-function zj(r).3
(~,~)(~,~) 2A 2
The hyperfine coupling becomes, then, using (11.65) and (11.66),
(:I.2' ~1)(.1
2
1)(_.1 :I.)
7'2 2'7
lA 3
(;l _i!)(1 _1)(_1 1)(_i! ;l) OA 4
7' 2 2' 7 2'2 2'2
(-~, ~)( -~, -t)( 1, -~) -IA 3
(-t,~t)(-t,-~) -2A 2
(11.151)
(-~,-~) -3A
where 0 2 is the fraction of p-function, (l/J.J) the average of (l/r 3) for the state
"unity", the line that for like nuclei would have intensity 2 splits into two lines
zj(r), and where a factor of 4
multiplies the expression of (11.67), since the
wave function spreads over two atoms. (We have neglected renonnalization due
of unit intensity when the nuclei differ; that of intensity 3 splits into three lines to overlap of atomic functions.) There is a near cancellation of the two terms
of unit intensity, and soon. The positions are all predictable, using the measured multiplying (II + 12)·5, leading 10 a strong anisotropy.
t
A~5 and (11.91). The intensity of the 35-37 lines is that of the outennost line
37 The fact that the electron ranges over more than one atom presents a new
of the 35-35 spectrum. There is in addition a set of seven lines from the Cl problem in calculating the g-shift. In the example we slUdied, we represented
35
pairs. also having a predicted position, with an intensity ~ that of the Cl pairs. the spin-orbit coupling by a form >'L· 5 appropriate to a free atom. The origin
All these lines are found at the proper positions and with the proper intensity. In about which the angular momentum was measured was, of course, the nuclear
this way many of the lines are accounted for. charge, since it is motion of the spin with respect to this charge that produces
An additional factor in determining the spectrum is the fact that the g-shift the spin-orbit coupling. When there is more than one nucleus, it is not apparent
displaces the position of the center of the hyperfine panems. The principal axes which nucleus to choose as the origin. The dilemma is resolved by using the
of the g-tensor with respect to the molecule are shown in Fig. lI.l3. As we see more basic fonn of the spin-orbit coupling:
in Fig. 11.13. if the magnetic field were perpendicular to the plane of the paper
(pointing in a (001) direction), it would be parallel 10 the y-axis of the center. eli.
'HSO=-2 22S.(Exp) (11.152)
IC it were along the 100 or 010 directions, it would make an angle of 45° with me
respect to the z-axis of the molecule in Fig. 11.13. For any given orientation of In this expression E is the electric field through which the electron is moving, and
the magnetic field with respect to the crystal axes, there are in general several p is the momentum operator of the electron, (M)'V. Since for an isolated atom
classes of Vk centers in tenns of the angles made by Ho with the principal E is directed along Ihe radius vector from the origin, Ex p ex: r x p, the angular
axes of the center. If H 0 is parallel to a 111 direction. there are two classes of momentum. The usual >.L . 5 expression therefore follows from (11.152). Since
l
molecules. If HO is parallel to the 100 direction, of the centers have their bond the electric field E is largest near a nucleus, the principal contribution to 11.50
axes perpendicular to Ho and ~ have a bond axis making a 45° angle with Ho· comes when the electron is near a nucleus.
In addition to the presence of two force centers, we have the complication
that we must deal with more than one electron. In fact the Vk center lacks only
, one electron to fill the valence shells of its two chlorine atoms. In order 10
proceed funher, it is helpful to describe the electronic states. We shall describe
+ + 3 The electronic wavc functions are discusse<1 in the subsequent text. We may remark
Fig.II.B. Principal axes :1:, y, ~ of the g-tensor for the V k that the bond function refcrroo to here is the function ~\ + ~2 mentioned in the latcr
+ + center. The y-axis points out of the paper discussion.
536 537
is from the zi + Z2 to %1 - .1:2, since the electronic dipole matrix element here is
...,, , /' ..... the largest (in fact, corresponding to a dipole moment arm equal to the length of
"-~ .- '
,,~ ~/ the molecule).
We see, therefore, Ihat we muSt generalize our previous discussion to account
for two new features: the lack of a single force center, and the fact that more
than one eleclrOn is involved.
In order to illustrate the first point (more than one force center) without the
complications of the second (more than one electron) let us consider an example
+
Fig. 11.14. Functions (a) %1 %2 and (b) %1 -':2 shown 5Chematiclllly. The function of (a) is in which we have only one electron occupying the VA: center orbitals. Then the
higher in energy, !lincc I) it hll.!l more nodCllllnd 2) the node lies ftt a region of aUractive ground orbital state function !Po is, neglecting overlap,
potential for the e1eclron
1
!Po "" /2(ZI - .1:2) (11.153)
them in a molecular orbital scheme in which the molecular orbitals are made
up as linear combinations of the free-atom p-states. We shall denote by :1:1 an which has quenched orbital angular momentum.
atomic p-function zf(r) centered on 3tom 1 of the center. The atomic functions In our earlier discussion of the g·shifl for a problem with only one force
are thus %1> YI> zl. %2. Y2, and %2. the z-axis lying along the bond. The functions center we saw that the g-shift arose from the interplay between the spin-orbit
%1 + %2 and %1 - %2 are shown schematically in Fig. 11.14. A study of the figure coupling and the slight unquenching of orbital angular momentum produced by
shows that %1 - %2 corresponds 10 a lower energy than does %1 + %2. since it the orbital Zeeman energy. When more than one force center is present there is
has fewer nodes and lends to concenlrate the electronic density between lhe two no unique point about which to measure angular momentum and it is natural to
atoms where it can share their altractive potential. The Slates are in fact referred return to the more basic fonn of the spin-orbit coupling:
to as bonding (ZI - Z2) and antibonding (ZI +Z2). In a similar manner it turns out eli.
that XI + X2 and YI + Y2 are bonding, and XI - X2 and YI - Y2 are amibonding. "S·(Exp)
'liSO=-2
m ,
(11.154)
(The z-states are the so-called q-states and the X or y states are the lI'·states.)
The energy levels of these states are shown schematically in Fig. 11.15. Actually In the absence of an applied magnetic field this expression is correct but in the
the states Xl + X2 and VI + Y2, which are degenerate in the free molecules, are presence of a magnetic field described by a vector potential A we must modify
not degenerate in a crystal, but we neglect that splitting. it to oblain a gauge invariant result:
" A2
1ioz "" -2( p ' A + A· p) + __ (11.156)
2mc 2 2mc
'. + .1:2 (u,,) We can consider that 1iOZ and 'HSO together constitute a penurbation 1iperL :
"y, --• "J
Y2
(11'",) 1iperL "" 1iso + 1ioz (11.157)
,,
,
"y, + "2"1 (r,,)
Fig. IUS. Molecular orbiUlIs ronned rrom ,..sta~ in ..
halogen molecule ion. The allowed optica.ltransitions into
the unfilled <1'" orbit &reshown by dashed line!l. The ll'..hil'l
We are concerned with calculating maaix elements designated by an orbital
quantum number n and a spin quantum number o. In panicular the effect of1ipert
is, according to Appendix D, equivalent to our having an additional interaction
t
" - '2 (u,)
"transitions" an shown by a solid line 1in_ whose matrix elements diagonal in the ground orbital state 10) are
538 539
where
1
(11.158) L(R) " 7(r - R) x 'V (11.164)
1
is the dimensionless angular momentum operator about the arbitrary point R.
where the prime on the summation means omitting n = 0 and where we have Integrals such as the lower one on the right of (11.163) are readily evaluated
neglected the spin contributions 10 the energy denominators. 11te g·shift arises by the methods of (11.17) and (11.18) provided R is chosen as zero so that the
from keeping just those terms of {I 1.158) that are linear in the vector potential angular momentum is measured about the natural origin of the atomic orbitals
and the electron spin. Thus we get (Oo"l'H.1g[OoI) as thac make up the functions In). We shall call this the "natural" gauge.
An even more imponant point is seen by examining the first-order tenn of
(Ocr l1i<191 0a/ ) = 2~~c2 (oaI5.E x~AloO"I) (11.161),
e2h
e2h ~I (OaIS·E x plnull)(nq"lp·A + A'pjOql) 2m'c3 (aISla') (OlE x A(R)IO)
0
Eo-En
0
540
541
use a mixture of gauges: one gauge when in the vicinilY of one nucleus, the other (OlE x A 10) =(volE x Alvo) + (volE x Alvo) . (11.173)
in the vicinily of a second nucleus. Such a trick is aClually possible, provided we
can neglecl cenain overlap integrals. Lei us Slate a theorem; lhen we shall outline Then we introduce two vector polenlials. AI and All, which differ by a gauge
ils proof and then show how Ihe lheorem enables us 10 use such a lechnique of transfonnation:
several "natural" gauges to treat the problem of muiliple force centers. AI = A'I + V? (11.174)
Let us therefore consider a system with two atoms. The ground stale 10)
will be a linear combination defining Ihe funclion ,p. (That (11.174) is simply a gauge transfonnation follows,
of course, from Ihe faci that it satisfies the requirement V x A' = V X A").
100=uo+vo (11.168) We then substitute A' for A in integrals involving u's. and All + V,p for A in
where uo is a linear combination of atomic orbitals on Ihe first atom and Vo is a integrals involving v's. By collecting tenns. we gel
linear combination of alomic orbitals on the second alom. The excited stales In)
are also linear combinalions: (Oul1ll1g l{)q/) = + 2~Z~ (uI5Iu'). {(UoIE x A'luo) + (VOlE x Aillvo)
(11.169) + _1_ E' (OlE x pln)[(unlp· A' + AI . pluo)
We shall neglect all contribulions 10 matrix elements involving a produci of a u 2m n Eo En
and a \/. This approximation is often good, but can lead to errors in some cases. (un Iv' A" + All • vivo)] + c.c.
We have, then, as our theorem Ihat the combined effeci of the spin-orbit + Eo _ Ell + (volE x V,plvo)
and orbital Zeeman coupling is to give a g-shift characterized by
+ _1_ L;' (OlE X vln)(v"lv' "71 + "71' vivo) + c.c.}
(OaIH.ogIOo-') 2m n Eo E"
(11.175)
= 2~zhc3 (qJ5Iq')· {(UoIE x A'luo) + (volE x A"I1l1J) To derive our theorem, (11.170). we must show Ihal the lenns involving
¢ add to zero. By making use of the faci Ihat we are neglecling overlap [enns,
+ _1_ '£' (OlE x pln)[(uII1A' . p + p' A'luo) + (vnIA" . P + p' AI/lvo») + c.c.} our proof is equivalenl to showing thai the quantity (Oul1ll1g(¢)lvou/). defined
2m n Eo-En below. vanishes:
(11.170) eZh {
where A' and A" are any vector potentials Ihal give the static field Ho
(0.I H .o,(1)lvoo') " - 2
mc
2 3('1 5 1")' (OlE x "71I vo)
542 543
nOling that, neglecting overlap, Therefore. when the upper sign applies, the two tenns in the square brackets
,'I. [
(Oo'l'HLlg(t/J)lvoql) = 2m 2 c3 (uISI0"') (DIE x 'V¢lvo)
= WztlE x plxt) Of (z,IE x plx,)J (11.185) But. by the symmetry of the atoms,
Since the two atoms are identical and since E is large only near a nucleus, (unlA1 'p + P ·Alluo) = (V,dA2' P + p' A2lvo) (11.190)
(11.186) so that we have. neglecting the first-order tenns such as given by (11.170),
544 545
(o.I1i",IO.') = 2PA(.ISI.'). [(Z1l
L 1I<1)«II L 1I ZI)
E:"I_z~ - E:l:1-2:2
+ (z1I L 1IYI)(Y1I L ll z
E Z \-Z2 - EY \-1/2
il). Ho (11.191)
@
This is equivalent 10 having
where
546 547
Then the spin-orbit and orbital Zeeman coupling combine to give matrix elements we get
equivalent to OUf adding a tenn 1(. 6 9 to the Hamiltonian, where
(OSMI'Ii"gIOSM') =
(OSMI'Ii"gIOSM') = (OSMI'lisoAIOSM') j"inlosM)
+ :L (OSMI 1isoolnS' M")(nS' MlIl?iozlOSM') '''';11 105M')
nS'M" Eo-En
+ (OSMI1fozlnS' M")(nS' M/lI11'sooIOSM')
+ L:' (/.1'Il!OZ
I) 1/'.)(1'.1'Ii[I) II.')
SOO
resuictiol1 A El - ~ll
Eo - En (11.198)
where we have neglected the spin in the energy denominators, and where we are + L:' (lal1t~~ 11'0')(/'a'l1igil/a')
~LrkLion B
also keeping only the terms that give rise to a g-shift. E,- E"
In the case of the Vir center, we may take the wave functions 105M) 10 be (11.202)
where by restriction A we mean:
a product of one-electron molecular orbital states, properly antisymmetrized. The
calculation proceeds along lines similar to thai of the indiIttt nuclear coupling 110) is occupied in 105M)
in Section 4.9. Lei us denote the state (Zt - %2)/../2 containing electron number 110') is occupied in 105M')
1 with spin-up (m = +!) as Ita) is occupied in neither 105M) nor 105M')
u,,-.z=J,+(1) . (11.199) and by restriction B we mean:
Since the total spin of the VI; center is ~. the Siale 105M) = IOi-l) is then 1/0) is occupied in 105M)
11) I.... 110') is occupied in 105At)
I0'2'2
p
:: /iTf'''p)-I) PU':'_':l,+(I)u':'_':l,_(2) Ita') is occupied in neither 105M) nor 105M')
X It~. +~1,+(3) ... :. +:1,+(11) (11.200) Equation (11.202) will hold for any system in which the wave function can be
ta);:en as a product of individual spin functions.
That is. all the orbitals except u:. +:1,- are occupied by an electron. We wish, of course, to incorporate our earlier theorem, (11.170), to enable
It is convenient to denote functions such as (Zl - Z2)/,fi by a symbol I and us to use "natural" gauges. This can be done readily by noting two things: The
the spin quantum number by (I (since m is also used for the mass of an electron).
In this notation the individual electron orbitals are denoted as 1/0). As discussed first is that, since 1lgi is independent of spin,
in Section 4.9, all matrix elements of (11_198) arise from one-electron operators,
so that the states joined by the operators can at most differ in the occupation of
(l'I'lig~lt.) = (l1'lig~lr) = (/"I'lig~I'.,) (11.203)
one orbital. Thus we find that we can express (05MI1tsoAI05M') in terms of By utilizing this fact, we can see the second point. If we remove the condition that
the one-electron operator 1i~gA: II' a) be unoccupied in 105M) or 105M') from restriction A, and that Wa') be
unoccupied in 105M) or 105M') from restriction B. the extra terms we acquire
(OSMI'lisoAIOSM') = L: (l.I'Ii~gAI/.,) (11.201) will exactly cancel in pairs. We can therefore write
10' ,/0"
(OS M 1'Ii", lOS M')
where 10 goes over all values occupied in 105M) and where la l goes over all
values occupied in 105MI ). We do not include matrix elements (lal1iSOAl/lal)
= L: (l'I'Ii~gAII.')
10';/0"
where zt i- I, since these states imply a change in the occupation of molecular
1
orbitals. That would imply that the ground state possessed orbital degeneracy, a I) 11'.)(/'.I".A
(I. 11i(oz ) II.') + (/.11i(seo
'~SQQ
I) III a l )(I' a' I1t(l)
oz II.')
circumstance we do not wish to consider. + L:
/(1;10";1' E, - E" (11.204)
The second-order terms are handled in a similar manner. The sum over
electrons can be converted to a sum over orbitals occupied in the ground state where now we require only that I/a) be occupied in 105M) and 1/0') be occupied
and sum over n to a sum over orbitals not occupied in the ground state. Therefore in 105M').
548 549
Consider now all the teoos of fixed lu and lu': where the components Sq are. for example,
N
(I) , ",(lql1ig~II'q)(I'ql1i~\5,IIu') Sz = L Sz; etc. (11.210)
(lql1isOAllq) + L, E E,
I' I I i=l
For the VA: center, however, rather than employ the Wigner-Eckart theorem. we
+ (lql1i~&ll'ql)(ra'l1{gill<"') (11.205) shall simply evaluate (11.208). Symmetry tells us that the principal axes of the g-
E,-E" tensor are the x-. y-. z-axes of the molecule, where the z-axis lies along the bond.
This is identical in (ann 10 (11.161). II can therefore be convened 10 the expres- Suppose, therefore, that the static field lies along the x-axis and that M is taken as
sion involving the mixed gauge. Let us therefore define the eigenvalue of Sz. Since this is a principal axis. the only nonvanishing matrix
elements have M = M'. Of course we can verify this by explicit evaluation of
(11.208). S is, of course, 4.
Let us compute (ll.208) for M M' The = = !.
simplest way to discuss the matrix elements is, then, in terms of a diagram of
and
the states. Since we are computing a diagonal term in (OSMI1t.dgIOSM'). the
(lu!U(1)lt u') = 2:2nc3 (uIStlu')· [(udA'l' PI + PI . A'duII ) states labeled 110') and 110") of (11.208) must be identical (we must relurn the
electron to the state from which it was vinually excited). Setting 0' = 0 and
+(v/IA'i,pt +Pl,A'{lv/,» (11.206)
neglecting the terms involving (luJV(l)llu), we have
In tenns of these definitions we can rewrite the expression of (11.205) as
(OSMI1i"gIOSM)
(I) , L:'(lqIU(I)lI'q)(l'ql1i~2ollq') = I: [(lUIU(I)II'O)(l'UI1i~~llu)+c.c·l
(lqlV Ilq) + E E (11.211)
" I - I' Ef Ef'
Ir )occupied
(lql1i(I) II'q')(YqW(I)IIu') I" )unoccupied
+ SOO (11.207)
E,-EI' where 110') occupied or 1/'0') unoccupied refer respectively to whclheror not lin)
Therefore we find is occupied or [1'(7) is unoccupied in the ground Siale. For S =~, M = the !.
(OSMI1i"gIOSM') = I: (lqlV(1)llq') states 11'0') and lIn) can be summarized by a diagram in which solid arrows
designate states occupied by electrons, an arrow pointing up. T. referring to
IUi 'I1'
By means of the Wigner-Eckart theorem, it is possible to show that (11.208) ---t+- y.-y,
implies that all matrix elements (OSMI'HLlgIOSM') can be obtained from a -t-1 Xl - X,
550 551
Fig. 11.20. The double arrow indicates states joined
by the matrix elements of 11.120) for a field in the
:I:-direction, The arrows indicate the electron spin
= (Z1l 1\1)IYI)( -~IS\I)I- D (11.216)
quantization, r signifying spin parallel to the static
field and 1 signifying spin antiparallel. The dashed
We obtain in this manner
arrow 1 state is vacant in the ground state
+i=::!:±:===:
:-t1 Yl + Y 2
X I +Xl
(o~22~1'Ii"gIO~~)
22
= _ 2PH.(-I/2)IS\I)I- 1/2)
E + E ZI Z2 yl - Y1
(11.217)
(Zl - ~['H~go:l:IYI -~) "" 2~;c2 [1 zi(E x p):l:yldr] In a similar manner we get
x (-~IS\I)I- D (11.213)
gyy ::: 2 (I - E ZI +Z1
>.
E:I:\ -:1:2
) 9n : : 2 (11.220)
552 553
shift). The close proximilY of these two states would make l.dyul > 1.d9zzl, 12. Summary
and 9"y, of course, is Slj]] 2, since Ihe peninenl matrix elements vanish.
It is clear that Ihe 12 states are less Ihan half full, yel the predominant
g-shift is that of a "hole". We see, Iherefore, Ihat we must use extreme caution
in characlerizing centers as "eleclron" or "hole" centers simply from the g-shift
data.
We should also comment that we have assumed very simple functions with
no overlap between atoms, to compute the matrix elemenls. In general, we would
need to make correclions both for overlap and for the possibilily thai the funClions We have considered a variety of effects - line widths, chemical shiflS, Knight
:1:), Yl, ZI> and so on are linear combinations of alomic orbilals, as we did in shifts, hyperfine splillings - a bewildering array of seemingly special cases. As
discussing the hyperfine coupling. However, these corrections do nOI aller Ihe we look back, we see some effeclS that occur in first-order penumation theory,
principles, although they do complicate the numerical calculations. others thai require a higher order. Since we have discussed the phenomena one by
one, it is appropriate to summarize by wriling a single Hamiltonian that includes
everything. As we coOlemplale it. we should remind ourselves of the significance
of each tenn. We write below Ihe Hamiltonian describing a nucleus interacling
with an electron in the presence of a magnelk field Ho. We define the vector
potenlials Ao. associated with the field Ho. and An' associated with the field at
.
the electron owing to the nuclear moment (An = P X rlr 3 nonnally). We also
define the quanlity
1f=7~+-Ao
1
,
C
Then we have Ihe following Hamillonian:
(12.1)
2
'l.J=_Ji
I~ 2mv
n2 +V0+ V~rrl\
.1"",,,,,, k;~'lio .leo'_ """.Iial
m~ '.'/' ;~po'.~';,,1
.1...1t...
01,
'h. n.w ._". d... ,~
Io•••ad <>l' .U'....... IOi4t ,h.
oth 1.. "..... .._
We can add to Ihis the coupling of nuclei with each other and the magnetic
coupling of electrons with each other.
554 5S5
Problems
Chapter 2
dF = '!'[1{ F] + of
dth' at
where 8Flat represents an actual derivative of F(O with respect to lime.
2.2 Equation (2.22a) gives an expression for (JJz(t» for a particle of spin
!. Generalize the expression for a spin I.
557
b) Compute the wave function in the laboratory reference system during and b) show that the solution of these equations is
after the pulse. a(t)= ~[a(O)(1 +cos wlt)-c(O)(l -cos WIt) + ihb{O) sin Wit]
c) Compute the (It;r,(t)) during and after the pulse.
i .
1I(t) = 11(0) cos WI t + ../i(a(O) + c(O)] Sill WI t
2.5 A coil of length I, cross-sectional area A. and n turns is wound in the
fann of a solenoid. The axis of the coil is in the x-direction, and a static field
Ho is in the z-direction.
c(t) :::: 4[c(O)( I + cos WI t) - a(O)(1 - cos WI t) + i hb(O) sin WI t]
c) Show that these equations give
a) Assuming the nuclear moments are in thennal equilibrium. what is the
nuclear magnetization per unit volume Mo prochx:ed by Ho. in tenns of "'(t) = tp(O) if WI t ::::: 2:1f
H 0 and the static nuclear susceptibility xo'1 in contrast to a spin ~ particle for which "'(t) = -!/J(O) if Wit = 2:1f.
b) Compute a numerical value of xo for protons in water at room temperature.
using the fonnula 2.8 In Sect.2.9 the existence of spin echoes is derived assuming that the "7rn."
N'"(2h 2 I(1 + 1) pulse gives a :lfn. rotation which is positive in the right-handed sense. This im-
xo = 3kT plies a negative '"(. Carry through the derivation assuming the :lfn. rotation is
positive in the left-handed sense, as with a positive '"(. showing that the echo is
where N is the number of spins per unit volume. (For procons. '"( can
formed along the negative y-axis of Fig. 2.13.
be found from the fact that the resonance occurs at 42MHz for Ho =
104 Gauss.)
2.9 Consider an experiment in which inirially the magnetization is at its thennal
c) The magnetization Mo is turned by a 90 0 pulse. Derive an expression for
equilibrium value. kMo• along a slatic field kHo. At t = 0 the magnetization
amplitude Vo of the voltage induced in the coil by the precessing Mo.
is invened by a 1f pulse. It then grows back towards its thennal equilibrium
d) Make a numerical estimate of Vo for procons, assuming a Io-tum coil 2cm
exponentially according to the equation
long, t em in diameter. and an Ho of S<XX>Gauss.
dM z Mo-M z
2.6 Suppose the coil of Problem 2.5 has inductance Lo and resistance Ro -;u= T1
and is in series resonance with a condenser C. At a time Ti> M t is inspected by applying a 7rn. pulse and observing the
a) Derive an expression for the voltage across the condenser in lenns of the initial amplitude of the free induction signal generated in an rf coil oriented
induced voltage Vo of Problem 2.5. Lo, Ro. C. and Q(= Low/fl). transverse (o Ho.
b) Using the numerical estimate of Problem 2.S(d), and assuming the coil has a) Show that M z • the value of the transverse magnetization immectiately after
a Q of 100, compute the size of the voltage across C. the :lfn. pulse. is
2.7 Rotation behal/ior of a spin J particle: Consider a particle of spin I. with Mz = MoO - 2e- TJ / TI )
eigenstates of It given as "t. "0, U_1 corresponding to the eigenvalues of 1, O. b) Assume now that owing to natural line breadths, the transverse magne-
and -I respectively of It. Consider that it is acted on by a rotating field, HI> tization decays rapidly to zero. Suppose one now waits a time 1'2. long
exactly tuned to resonance as in the discussion of the spinor behavior of a spin compared to the time for transverse magnetization to decay. then again
~ particle in Sect. 2.6. applies a 7l' pulse, and again applies a second 7l'n. pulse at time 1'1 later.
a) Taking the wave function in the rotating frame, '1//. to be Suppose one repeats this pulse sequence many times:
"/(t) = a(t)u1 + b(t)uo + c(t)tt_1
... !I+-1'I_1I +-1'2"'" 11+-1'\"'" II··· etc.
and defining WI as ,HI> show that a. b. and c obey the differential equa- 11" 11"/2 :If 11"/2
tjons Show that the initial value of the free induction decay following the 11"/2 pulse
db .Wl becomes
- = I-(a+ c)
dt J2 M z =Mo[1-2e- T!/'r1 +e-(TI+fl)/'I'I]
5S8 559
2.10 This problem concerns lhe calculation of the time at which a spin echo a) Show thaI the expectation value of the total x-component of spin, obtained
occurs for a (7r(2, 11') echo sequence, when the duration of the pulse is not taken by adding up the contributions of all the spins (Le. integrating over dho),
as zero. Refer to the end of Sec1.2.9 and Figs. 2.15 and 2.16 for the notation. vanishes at t = 2r.
The goal is to derive lhe relationship of (2.73), r' = r + IhH I, in the approxi· b) Show that the time development of expectation value of the total y-
mation that we neglect corrections to r ' involving (I/H t )2. In all that follows, component of spin (Iy,todm after the 90° pulse (but before the 1800
this approximation is assumed. Thus tan (lJH/H 1) = eH/H t etc. pulse) is the same as (Iy,tot(t - 2T). the behavior after the peak of the
spin echo.
a) Prove that the "11'(2 pulse" rotates the magnetization, eM, from its initial
direction along the z-axis to the x'_y plane (where the y and y axes
2.13 From (2.170) we have thaI
coincide), and that it makes an angle LJ.8 = HI/eH with the negative y
b)
axis.
Prove that the projection of eM in the x·y plane makes the same angle.
2 I
xo=--
• wo
J "( ')~.!
00
X w uw
LJ.8. with the y-axis.
c) Show that the effect of the ":If pulse" on the projection of eM in the x-y °
Show that if X~ax is the maximum value of X" in an absorption line.
plane is a reflection in lhe x-z plane, as in lhe "infinite HI" 11" pulse (i.e.
II :If WQ
there is no correction for lhe ":If pulse" similar to the correction for the Xmax = "2 xo LJ.w
",..(2 pulse'').
where LJ.w is a suitably defined line breadth.
d) Deri'e (2.73)
Assuming that the line width of the nuclear resonaoce of protons in water
(2.73) is 0.1 Gauss broad (because of magnet inhomogeneity) and that Ho = 104 Gauss.
compute x'~ax for water. and also compute the maximum fractional change in
e) Show that (2.73) is also valid for an echo produced with a (311"/2. 211")
coil resistance for a coil of Q = 100.
pulse sequence, but that (2.74) becomes
, 2 2.14 1be response of a cenain piece of material to a step of magnetic field of
T =T+3:1ft3r/2 (2.74)
unit height. applied at t = 0, is
2.11 Consider an operator R and its inverse R- t (Le. RR-I = R-tR = I). M 5tep(t) = xo(l - e- t / T )
An example might be a) Compute x'(w) and Xll(w).
R = e i81• b) Show that x' and X" satisfy the Kramers·K.ronig relations.
which is associated with rotation about the z-axis. 2.15 In Appendix F. an expression is developed for the complex magnetization
Let G be some operator function which in general does not commute with
MI;+iM y
R (i.e. RG - GR = 0). Prove that
R-le iG R = ei(R-IGR) "i,c:H"I"M=oT"["2"+c-T
MI;+I.M y =- ,.:(",a"'..,+,.:a,,,,,,)I2,,,J
(I + O"ar)(1 + O"bT) 1
a) Show that in the limit of very slow jumping (rew >- 1) the absorption line
2.12 This problem is concerned with the quantum mechanical treatment of has the shape of two distinct resonances at frequencies
the free induction decay and the spin echo from noninteracting spins subjected
to a distribution of static field strength. Ow
w='Y H o±2'
Consider the time development of the wave function in the rotating frame
for a spin echo in which the HI is applied along the x-axis in the rOlating frame. b) Show that in the limit of very fast jumping (rew« 1) a single resonance
The first pulse is a 1r{l pulse, and the second, applied time T later, is a 1r pulse. results at frequency
Take the distribution of magnetic fields H to be symmetric about wh, so that
w = '1HO
if ho =: H - who lhe distribution function P(hO) of ho is symmetrical about
ho = O.
560 561
2.16 In Appendix F the Bloch equalions are employed to analyze the case the frame which rotates at "(Ho. Draw a picture of M a and Mb in the x·y
of a group of nuclei which spend equal times on the average in either of two plane at a time at which M a makes an angle 9 with the x-direction.
sites at which the resonant frequency differs. b) Draw a picture of the vectors oMa '" G.Mbot and oMb "" -GtMbot
This equal time aspect appears in the statement that the two quantities C 1 . which would occur if G I is now switched on at the time corresponding to
and C2 are equal, and in the use of Mo/2 in the term involving HI. Suppose pan (a).
the equations were to be used to describe a problem in which the nucleus could c) Show by vector addition the new vectors
jump between twO sites which were not equally populated on the average. For
M~ "" M a +oMa and Mf, = Mb+oNh
example site "a" might be at a higher energy than site "b" so that the thermal
equilibrium populations differed.
a} If M a and IIifb are the thermal equilibrium static magnetizations, what
must the relationship be of G 1 to C2?
Chapter 3
b) Set up the appropriate form of (F.13) for this case and derive the expression
for the resulting M~ + iMlJ' i
3.1 A pair of identical spins of II "" 12 = is coupled by their magnetic dipole
moments. Assuming zero external static magnetic field. show that the proper
2.17 A static magnetic field H(r} may be expressed as eigenstates of the spins are the singlet and triplet states, and then find the ener-
gies of the different states.
H = "'''
where, is the magnetic potential which, in free space, satisfies 3.2 Suppose in Problem 3.1 that a static magnetic field Ho is applied par-
allel to the intemuclear axis.
",',,=0 _
a) Find the energy levels and eigenfunctions as a function of Ho.
If the magnetic field of a laboratory magnet is assumed axially symmctric about
b) An altemating magnetic field is applied perpendicular to the internuclear
the z·axis, show that to lowest order the inhomogeneity can be expressed as
axis. Find the allowed transi!ions, their frequencies and relative intensities:
(1) for Ho much less than the dipolar coupling; (2) for Ho much larger
H-HO=Z(O::) than the dipolar coupling.
where H is the magnitude of H, and Ho is the magnitude of H at the origin
and where (8HI8z) is evaluated at the origin. 3.3 Equation (3.43) involves Tr {(1i~, l1-JIJ+}. Prove that this trace vanishes.
2.18 Show, following the methods of Problem 2.17, that the most general 3.4 Consider two identical spin i
nuclei. Let I: =: II: + I2: be the total z·
form of H(x, y, z) based on including the spherical harmonics of 1"" 2 or less in component of angular momentum. Evalume Tr {Ii} by explicit evaluation of
the magnetic potential is the diagonal matrix elements for two schemes of quantization: (a) the mi. m2
scheme and (b) the I, M scheme. Show that the answers agree with each other.
H(x.y,z) =Ho +Ax+By+ Cz
3.5 Consider a nucleus of spin I. Compute Tr {I:I7;.} and Tr {I2 Ii}.
2.19 Utilize the expression for H(x,y,z) of Problem 2.18 and the methods
of Appendix G to get a more general expression for the effect of diffusion on 3.6 Consider a group of N nonintcracting spins of spin I and gyromagnetic
the echo envelope decay than that of (6.15). ratio 1'. The total wave function can be taken as a product of the individual spin
states, and the total energy as the sum of the individual eigencnergies. Evaluate
2.20 Appendix F uses the Bloch equations to discuss the problem of a spin the expression for Xl/(w) of (2.190) to give the absorption. For simplicity, let
which can precess at either of two natural frequencies, depending on whether the Z "" (2I + I)N and exp(-EalkT) "" I, the expressions for the high-temperature
spin is in site "a" or site "b". limit.
a) Suppose that T2 '" 00, H t "" 0 (free precession), that the quantities CI 3.7 The electrostatic exchange coupling between two electrons can be rep·
and C2 are zero, and that at t "" O. M a and Mb lie along the x-axis in resented by adding a term A5,· 52 to the Hamiltonian.
562 563
a) Prove that this lerm commutes with the Zeeman energy. 4.2 Calculate the numerical size of the magnetic field produced by the Of-
b) Prove that addition of such a term to the direct dipole coupling does bital motion of an electron in the n = 2, I = I, m = + I state at the nucleus of a
not affect the second moment computed when assuming dipolar coupling hydrogen atom. (Neglect all effects associated with the electron spin.)
alone.
4.3 Consider the states xf(r), yJ(r), and zJ(1') split by a crysral potential
3.8 Consider three operators, A, B, and C. Prove that A(x 2 - y2). Let the syslem be slarted al time t "" 0 in a state
n{ABC} =n{CAB} =n(BCA} ,,(0) = (x+iy) !(,)
,fi
3.9 Consider a nucleus with spin ~ whose Hamiltonian is 1i = Hz + 1tQ. Compute the expectation value of the z-component of angular momentum, show-
where ing that it oscillates in time between the values + I and -1 at the angular fre-
quency of oscillation LJlh.
Hz = 'YnhHoI~ Note that your result corresponds to the classical picture that the effect of
The fonn of 1tQ is similar to the one that sometimes arises when a nudeus has the crystal field is to cause the plane of the circular orbit to tum in such a way
an electrical quadrupole moment. An alternating field is applied to the system to as to reverse the sense of circular motion periodically.
produce absorption.
4,4 The hyperfine coupling of s-states may be found simply in the approx-
a) Prove that 'Hz and 'HQ commute.
imation that a nucleus is a unifonnly magnetized sphere. In this problem, we
b) Trealing 1iQ as analogous to 1i~ in Seclion 3.3, and assuming A <:: "'(nnBO derive the famous s-state fonnula for that model.
prove that A unifonnly magnetized sphere of magnetization M per unit volume can
(w) = "(nHO be represented by current distribution flowing on the surface, the current being
proportional to M· n. where n is the unit outer nonnal. We shall represent a
and find (Ll.w 2 ). nucleus by such a sphere of radius R. Consider the current to flow in circles
about the z-axis; the surface current density J(8) is then
3.10 Derive (3.61) for the case of a unifonnly TOtating pair of nuclei.
J(U) "" Jo sin 8
Chapter 4
•
4.1 In Section 4.4, gauge transformations are discussed.
a)
b)
Using (4.18), show that (4.19) is true.
Prove that the operator . ""
JoutoC J into paper
rx (~V-~A)
for the angular momentum is gauge invariant.
c) Consider an s-state !/J(r) = u(l') in the absence of a magnetic field. A uni- a) Show that the magnetic field H inside the sphere is unifonn and that
fonn magnetic field with vector potential Ao = iBo xr is applied. Derive outside the sphere it is a pure dipole field.
an expression for the resulting angular-momentum expectation value, and b) Show that the field inside is
evaluate the answer in units of Ii. for the ground Slate of hydrogen, assum- 8.
ing Ho = IO,OOOGauss. H""T Jok
c) Show that the magnetic moment 11 of the sphere is (41l'13)R J Jo.
564 565
d) Show mal 5.2 Consider a system of N spins that interact with one another via a dipole-
-H, = JH,lu 2
(')ldT =
8", 2
3"lu(O)1 " dipole coupling and with an external static field H in the z-direction. Assuming
a density matrix f! given by
where u(r) is a spherically symmclric function that does no! vary too e-'H.jkT
rapidly within R of the origin and where H; is the z·component of field e=
due to the sphere. Z
corresponding to thermal equilibrium (where Z is the partition function), show
4.5 An atom has a single valence electron in an s-Slate and a nucleus of spin I. that the thennal equilibrium expectation value of the total magnetization is
The electron spin·lallice relaxation time is so short thai the nucleus experiences
only the time-average magnetic field of the electron. Derive an expression for (M,) = (M.) = 0 • (M,) = N~'!fi32:~ + 1) H
the resonance frequency of the nucleus when a sialic field Ho is applied. giving
your answer in tenns of the electron susceptibility Xc. Discuss the temperalUre in the high-temperature approximation.
and field dependence: (a) at high temperature (where kT >- "tchHo) and (b) at It is interesting to nOle that these equations are of the fonn M = CHIT
low temperature (where "fcltRO "" kT). Curie's law, and that the constant C does not depend on whether H is large 0;
small compared with the local field due to neighboring dipoles - in contrast to
4.6 In the text the Knight shirl was calculated by firsl-order penurbalion the- one's naIve expectation.
ory. using the fact that the static magnetic field Ho causes a repopulation of the
electrons among their spin states. It is possible to obtain an expression for the
Knight shirt by using second-order perturbation theory and by assuming that the 5.3 .Consider in a metal a system of nuclear spins that interact with a dipolar
applied field varies spatially in such a manner that there is no repopulation. Thus, couphng only. By means of (5,49), prove that the spin-lattice relaxation time in
suppose the applied field is in the z-direction, varying with x as zero static field is one-half its value in a strong field (Tl '" l/aoo).
566
567
5.7 Consider a system with three energy levels 1, 2, and 3. An alternating
interaction V(t) '" V cos wt is applied nearly at resonance with the rransition
t/>= ~c"tlnexp( -~Ent)
between states I and 2. where en are independent of time, and Un are the eigenfunctions of the time-
a) Write down the differential equations for the density matrix analogous to independent Schrooinger equation of energy eigenvalue E".
(5.394) and (5.395). Prove that the expectation value of the energy for such a function is inde~
b) In the limit of negligible saturation, compute (M:.:(t) and show that the pendent of time, Le., that the energy of the system is conserved.
width of the resonance is affected by the relaxation to level 3. (This is the
phenomenon of lifetime broadening due to transilions to a level that is not 6.6 Derive (6.46)
directly involved in the spectral line.) W(w) = i"{2 Hfg(w)
from the fonnulas given in Sect. 2.12, using the high temperature approximation.
Chapter 6
6.7 In Sect. 6.4, formulas are developed which show that for slow changes
of the magnetic field for which the applied field Hi is always much larger than
6.1 Consider (6.3). Let M:.: '" Mo, My::; M~ ::; 0 at t::; O. Show that for times the local field,
less than approximately TI, M~ and My will remain small, so that M will lie
H Hi
along HI, and will decay to zero exponemiaJly in T2. - r- - and Mf=Mi
Or - OJ
6.2 Show that the expression for the average energy given in Sect. 6.3 Show that these results can also be obtained by considering what happens
to the population of the 2I + I Zeeman levels of the individual spins for an as-
E::; LPnEn sembly of N noninteracting spins. (The lack of imeraction corresponds to setting
n
HL = 0.)
can be found as a derivative of the partition function Z with respect to {J
({J == l/kT) thus confinning (6.1). 6.8 Consider an adiabatic demagnetization experiment similar to that of Sect. 6.6.
Show that the sign of the spin temperature in the rotating frame depends on
6.3 Show that the average energy E of a syStem of N nuclei acted on by whether HI is switched on when Ho is above or below resonance.
an applied field H 0 and coupled together through a dipolar Hamiltonian Jid is For the case of negative spin temperature, draw a figure showing M and
-
E::; -C
(H6+ H r)' where
He in the :l;-Z plane for several values of He as one approaches resonance.
8
6.9 Let HL = 3Gauss, Make a graph of MIMo versus ko for three cases
2 I 2 c::; N"{2h 2 I(l + I)
CHL = k(2I + 1)N Tr {1td } and 3k i) HI = 1 Gauss
ii) HI = 3 Gauss
6.4 The entropy (1 is given by iii) HI = 9 Gauss.
7.3 Consider the nuclear Overhauser effect for a pair of spin! nuclei as in (H(mi »eff = i(H,)s + k[(ho)s - (ArnII"Ysh»)
Sect. 7.5. b) Draw a scale picture of (H(mJ»eff for the case A = 1"S li(h o>S, (ho)s =
a) Show that if (WITcP« 1 and (WSTCP« 1 in (7.47), that (7.61) becomes (Hds .for mJ = ! and mi = -!.
c) Show that the energy eigenvalues of (7.190) are given by
.'!.«I,) - Io) = '«([0 - (I,» + ~(50 - (5,» E(mJ, ms') = -1"JhHomJ -,stuns'
dl
aod x {[(ho)s ~(AmdYsh)12 + (Hd~}'/2
'H' = -''/I;IHoIz -1" s li[(h o)sSz +(H,)sSzl + AIzS z (7.190) Consider the case of an X/(rr!2) pulse. Then, the time-dependent expectation
value of J+ is
Since [[z, 'H'] = 0, the eigenstates may be characterized by mI, the eigenvalues
of I:.
570 571
(I+(t)) = T' (r+ e(l» will not produce an echo for a system of two identical spin ! particles coupled
:= Tr {/+ exp(i(fll f: + flsS:: - alz Sz)t)X,e(8) by a Hamiltonian
x X,I exp (-i(n!I: + nss: - aI:Sz)t)} 1t.:::: AlrSr
Show thai the term in (I+(t» arising from Sf in f!(8) can be wriuen as 8.4 In order to grasp the meaning of the various spin-flip narrowing pulse
(1ShHoIZkS) x Tr{I+S.:}. which vanishes. sequences. one can begin by assuming the dipolar broadening is negligible. that
there is negligible magnetic field inhomogeneity. and follow what happens to
7.10 Using the relationships for spin 'i panicles: the magnetization. Employ this method to the three-pulse cycle of Sect. 8.12 to
verify the results of Fig. 8.14.
c iS• S = cos (5:9) + i sin (5:9)
8.5 Employ the method of Problem 8.4 to describe the magnetization vec-
cos (5:9) = cos (On)
tor in the rotating frame through the first cycle for the four-pulse sequence
sin (5::8) = 2S: sin (8/2)
(r.X. T. Y.2T. Y.r.X. r).
Chapter 10
Show thal
e iS• SSze- iS• 8 = Sz cos () - Sy sin 8 10.1 It is stated in Sec\. 10.3 that (lO.27b) and (10.31) provide a set of re-
cursion relations among the elements (J MJI'/ITLM IJ' MJ''I1') for the various
possible values of MJ. M'. and M J ,. and a fixed sel J. L. J'. 1'/. '11'. For Ihe
case that J :::: J' show Ihal it is indeed true that specifying one matrix element
Chapter 8 (for example, that for which MJ :::: M J , :::: J) enables all others 10 be computed
by using Ihe recursion relations.
8.1 Suppose that a spin echo is produced by applying the H I of the first pulse
along the +::z:-axis in the rotating frame, rotating the magnetization to the -y-axis, 10.2 Verify that the functions Tu,.,(J) of Table to.l in Secl.IO.3 satisfy the
and a 7( pulse a time T later with H I along the +y-axis. commutation relations of a T 2A1 with respect to J.
a) Show thaI an echo is formed at 2r along the -y·axis.
10.3 Consider an axially symmetric potential and a weak static field The
b) Show that if HI. when on, is produced by applying a linearly polarized
allernaling field in the sample coil along the laboratory frame x-axis Hamiltonian is then
2
11. e qQ 3 2 2
Hx{t)=2HI coswt • t<T , :::: 41(2/ _ I) ( I: - I ) - 'Yn liH l r ,
572 573
10.5 Show that a charge e located at a point xo. YO. Zo produces a field 11.3 Equation (11.65) says
gradient
Vzz ==
8'V)
--2 of Vzz
_ 3zij - rfi
- e 5
"!-IS = 'Ye'Yn h '( T) 2 z 5 z - I· S}
r 3 '5(3l
8z Z,II,Z=O ro
Show that this follows from (11.64):
10.6 A nucleus of spin ~ experiences an electrical quadrupole coupling
'HIS = 'Ye'Yn h
'J r I [ (I·r)(S·r)
3 3 r2
2 , 1
- 1·5 x f (r)dT
'H = AU; - I;)
a) Show that the energy eigenvalues are 0, ± 2/7 A. 11.4 In Sect. I 1.3 the hyperfine splitting was worked out for the case of an
b) Show that the eigenfunctions are
E-O
_ {~I = (f.)I/'[~'/2 - (W/2~_3/']
~2=(-&)1/2[~_'I,-(~)1/2~3/2]
isotropic electron g-factor. Generalize the result to the case of a Hamiltonian for
a field with x- and z-components only:
1-{ = P(gzzHz + gzzH z ) + A z l z 5 z + A y l y 5 y + A z I,5,
- 'Ynh(Hzlz + H::I,) .
"" = (fa) 1/2 [~'/2 + (¥) 1/2~1/2 + m1/2~_3/21 11.5 Consider matrix elements (O'ILa,lz). where O',er' = x.y.z, which are
frequently encountered in problems for which Ix) is the ground electronic state.
E - 2(7)1/2 A
- { ~4 = (fa) 1/2 [~-'/2 + (¥) I/'~_I/' + m1/2~3/,l Show that there are only two which are nonzero (0' = y, 0'1 = z and 0' '" z,
0" = y) and evaluate the matrix elements for them. These results are useful for
E=_2(7)I/'A
~, = (fa) 1/2 [~'I' - (¥) 1/'~I/' + m1/'~_3/'] ] the following problems.
10.7 Prove that the eigenvalues EI, F:2, and so on of a Hamiltonian 'H =
AU; - /f.) come in pairs ± £1> ± B2, or else are zero. (Hint: Consider the
effect of :n operator R that changes x into y and y into -x).
Chapter 11
11.1 Evaluate the coefficient GJ> defined by (11.41), to obtain an answer anal-
11.7 Consider the atom of Problem I I.6, but now include the spin. Let there be
ogous to (1 1.53a). a spin-orbit interaction ),(r)L· 5 added to the Hamiltonian. Take as basis states
the states IO'ms) where 0' = :t.y, Z and ms = ±~' the spin quantum number.
11.2 In the notalion of Secl.11.2 prove that, with respect to L z , L y • and Lz,
the function x 2 - y2 is a linear combination of TU.f's. a) Show that the matrix elements of the spin-orbit coupling (ul.\(r)L . SI:ts')
which are inlemal to the ground orbital state vanish.
574 575
b) Using penurbation theory, show that the spin-orbit coupling mixes ex· 11011 Consider a single p-electron whose spin and orbit are strongly cou-
cited states into the original ground state wave funclion IX3) and find (he pled so that the J MJ scheme of quantization applies. The nuclear moment and
corrected wave functions Ixs)corr correct to lerms linear in A. electron orbital motion are coupled by a Hamiltonian
cJ Assuming an applied magnetic field along the z-direclion, compute the • •
1{::-2 (PoAn+Anop) (I)
lotal orbital plus spin Zeeman energy by first order penurbation theory, me
i) using functions Ixs) and where An :: 'YnhI X r/r 3 is the vector potential due to the nucleus.
ii) using the corrected functions of (b).
a) Using the Wigner-Eckart theorem, show mat for matrix elements diago-
d) Show that the Zeeman splitting of (c)-(i) is L1E = 2{3H but for (c)-(ii) it nal in the electron quantum number, J, (1) is equivalent to an effective
is of the fonn L1E =: gf3H, and find the expression for 9 correct to terms Hamiltonian
linear in A,
1ie ff:: AJJo]
11.8 An atom containing a single p-electron located at the origin is acted on where AJ is a constant for a given J, independent of M J .
by a set of 6 equal charges, q. The charges on the x- and y-axes are a distance b) Find AJ for me J :: ~ state.
a from the origin, those on the z-axis a distance b. Assume b < a. Neglect spin.
a) Show thai the lowest nonvanishing term in the Hamiltonian which de·
scribes spliuing of the p-states is 1{ = A(3z2 - r 2} and determine the sign
of A.
b) Find the proper eigenstates and energies in terms of A and (r 2 ), the mean
square radius of me p-state orbit.
c) Suppose a magnetic field is applied along me z-direction; compute the
3 x 3 Hamillonian matrix.
d) Find the eigenstates which resolve the degeneracy.
e) For which states is the angular momentum quenched?
11.9 Consider the system of Problem 11.8. Assume that the splitting of the
p-states produced by the crystal field is large compared to kT, but that pH <:: kT.
a) Show that when b> a the crystal splilling of the p-states has the opposite
sign from its value for b < o.
b) Curie's law states that the magnetic susceptibility tensor per atom Xoo'
(0:, a' :: z, y, z) goes inversely with temperature
Cerer'
Xero:':: T
Show that for one sign of the crystal field Curie's law applies, show that
for that sign the Z-, y., and z-axes are the principal axes, and evaluate the cor-
responding C zz , C yy , and Cn.
11.10 Consider an atom with a single p-electron, the orbital angular momentum
being quenched by a crystalline field such as that of Fig. 11.1.
By using second-order perturbation, show that interplay of the spin-orbit
coupling )"L S and the coupling (d2mc)(p· An + An p) between the nuclear
0 0
moment and the electron ocbit give an effective spin-spin coupling between the
nucleus and electron.
576 577
Appendixes
e A + B = e B e A eC / 2 . (A.3)
as follows: by taking the derivative of both sides of (A.6) with respect to ..\, we
find
e~>'A(BA _ AB )eM = dR
d!.
- C= (A.8)
579
since AB - BA =: C commutes with A. t
-
Integrating (A.g) we have -=r (B.4)
h
R(>.) = -C>'+constant . (A.9) we get
We can evaluate the constant by setting>. '" 0 and by noting from (A.7) that
R(O) = B. Thererore
~/I(w) = _w_ J
r:.e-E./kT
2kTZ -co a,b
R(~l.-C~+B (A. 10) x (al".lb)(bl".la)exp {(ice. - E,ltVhjexp (-iwt)dt (8.5)
By substituting (A. 10) into (A.6) and using the fact that C commutes with We can use the fact that the states la) and Ib) are eigenfunctions of the Hamil-
B, we get tonian 1i to express the expression more compactly as
dG
-~CG=
d~
(A. II)
Xl/(w) == ~
2kTZ
J ~(ale-1t/kTe
+~
~,
i7iL / Ait
•
e- i1tt / Alb)
which can be interpreted to give -co a,
J
-~
Tr {e- 7i / kT ei1it / l itze-illt/A J1r }e-ilolldt
B. Some Further Expressions for the Susceptibility then define the operator Pz(t) by
flz(t) == ei1tI / Altze-i1tt/h (B.8)
(This appendix requires familiarity with Chaps. 2, 3, and 5.)
we can also express (8.7) as
Equation (2.190) gives an expression for XII. Another expression is fre-
+~
quently encountered in the literature. It provides an alternative derivation for the
moments of the shape function. It can be obtained from (2.190). XI/(w) == 2k;Z J Tr{Pr(t)JJz}e-ilJ1dt (B.9)
-~
x"() hw
w =--
2kTZ
J~e
+~
t...- -E /kT
•
-~
580 581
We see that, setting t =0,
t I += In general, if 1ip is invariant under a rotation of 180° about the x- or y-
2kTZ T,{",(O)",). 2. J
-~
f(w)dw (B.12) axes (as is usually the case), the second term vanishes. This can be shown by
evaluating the trace, using coordinates x' '" x, y' '" -y, Zl = -z, which differ
only oy a 1800 rotation about x. Then 'H. p = H~ by our postulate, so that
By taking the nth derivative or(B.I1) with respect to t, and evaluating at t = 0,
we find Tr {e i('H p/ll)1 /lye -i('H p/ll)t Jlz}
1 tr' I
(1)"
2kTZ dtn Tr {llz(t)llz} t=o = 211"
+~ "
J w f(w)dJ..J (B.13) = Tr {ei('H~/lI)l(- Jjy')e -i('H~/1t)tJlz' }
-~
= _ Tr {ei('H~/")1Jly'e(i/ll)'H~1 /J z '} (B.t9)
We get, therefore, a compact expression for the nth moment of the shape function
few): _ But the last trace is clearly the same as the first. Therefore the trace is equal
J wnf(w)dw
to ilS own negative and must vanish.
We have, then, that the correlation function Tr {llz(t)llz} is given as
-~
(w") -_ __ (i)-"(tF/dt")Tr (JlzO)llz}lt 0
(B. 14)
+J~ T, (",(O)",) Tr {Ilz(t)pzl = cos WOt Tr {ei{'Hp/l)t lJze-(i/l)'HptJjz) (B.20)
f(w)dw
Since this is the Fourier transform of the shape function f(w), we see that the
-~
transient behavior consists of a term cos wot multiplied by an envelope function.
As an illustration let us derive an expression for the second moment (w 2 ). Taking If we define Il;(t) as
the derivative of Ilz(t) gives
p;(t) '" e(i/l)'Hplllze-(i/l)'Hpi (8.21)
~Tr(ei'Hj/l"
dt 2 ,..z
e-i'Ht/l" J
,..z .G)' T, {.m'I'[1i. [1i. ",)).m'I' ",} we can say that the envelope function is Tr {Il;(t)pz}.
By writing the cos wot as
= - hl2 Tr {ei'Ht/ll (1i'llz)e -i'Ht/lI ('H:,llzl} cos wot = !(eiWOI +e- iWOI ) (8.22)
Therefore (B. 15)
we can say that the two exponentials correspond to lines at +wo and -woo If we
(w') • _2.. T, ((1i, ".)'j (B.16) wish to discuss only the line at +wo, f+(w), we can therefore write
h2 Tr {Il;}
+~
This formalism provides a very simple way of generating expressions for the
higher moments. Note, however, that the odd moments all vanish, since few) is
f+(w) = 4k~Z J Tr {p;(t)/Jz }e+ilotole-ilotldt (8.23)
-~
an even function of w.
So far, apart from assuming the high-temperature approximation, we have
left the specification of the Hamiltonian completely general. We can proceed
(B.24)
further if we assume that it consists of the sum of a Zeeman term Hz and a tenn
'H:p , often a perturbation, which commutes with Hz. A typical H p is the terms
A and B of the dipolar coupling. Then, since H p and 1{z commute, This can be rewritten as
large contribution to «w -
wo)"} were it included! This is a "nonnal modes" problem, with solutions obtained by adding or sub-
By following steps similar to those of (8.15), we now find tracting:
+ho = HI - ho = H2 (C. I)
Then we wish to know the correlation function G(r): D. A Theorem from Perturbation Theory
G(T) = H(t)H(l + T) (C.2)
where the bar indicates an ensemble average. In this appendix we shall derive from perturbation theory a theorem that has wide
If the field is HI at time t = 0, then we can write for a single member of utility in magnetic resonance. It is closely related to second-order perturbation
the ensemble: theory but gives the results in a fonn particularly useful when there is degeneracy.
A typical situation in which the theorem has great use is illustrated by the g-shift
(C.3) calculation of Section 11.2. We may divide (he Hamiltonian into three teoos;
where PI(r) and P2(T) are either zero or one, depending on whether at time 1i = 1£0 + 1£1 + 1i2 where (D. I)
T the field is HI or H2' We now perfonn an ensemble average of (C.3) over
the various histories. This replaces quantities PI (r), P 2(r) by their ensemble p'
averages PI(T) and l>:2(r), which are the probabilities that in an ensemble in 1£o=-+VO+VI
2m
which the field was HI at r = 0, it will be HI or H 2 at lime T. HI = 2pH· S (D.2)
Thus we have
H, = AL·S+pH·L
(CA)
Since 1£0 does not depend on spin, its eigenstates may be taken as products of
This equation, of course, assumes thai at T = 0, H(r) = HI, so that as r --+ 0, an orbital and a spin function. We denote the orbital quantum numbers by 1 and
PI(T) --+ I, n(r) --+0. Equally likely is the situation that the field is H2 at T = 0, the spin quantum numbers by a. Then
which will give a similar equation except thai I and 2 are interchanged.
We shall assume the behavior of PI and 1'2 as a function of r 10 be given Hoi/a) = Ed/a) (D.3)
by a rale equation: The states 110') are degenerate for a given 1 because of the spin quantum numbers.
584 585
The tenn 1{t lifts the spin degeneracy. Since it depends on spin only. it has no (D.7)
matrix elements between different orbital states:
where dr and drs represent integration over spatial and spin variables, respec-
(D.4) tively.. By utilizing (0.6) and the Heonitian property of 5, we have
In general the matrix elements of 1{l between staies 110) and Ila') where a #a'
will be nonzero. Therefore the presence of 1{1 still leaves us a group of subma-
J ¢iO/1{¢I,O',dr drs = J 1/;iO/e -iS1{e iS 1/J/IO/ldr drs (0.8a)
trices (lal1{llla') to diagonalize. For our example, since the spin was !. these = (IO'le-iS1{eiSll'O") (0.8b)
submatrices are only 2 x 2 and are easily handled.
The presence of the term 1{2 spoils things, since 1{2 joins states of different where we have used the notation 110') for matrix elements calculated using the
I. However. as a result of the quenching of the orbital angular momentum. the 1/J's. We may interpret (0.8) as saying that we can look either for transformed
matrix elements of 1{2 that are diagonal in l vanish: functions, rPIO/. or a transfonned Hamiltonian, (0.8b).
If we define ?i' as
(laIH,lla') = 0 (D.5)
(D.9)
We may schematize things as shown in Fig. 0.1, where the Hamiltonian matrix is
illustrated and where we have labeled which tenns 1{1 or 1{2 have nonvanishing we may state as our goal the detennination of a Hermitian operator 5 that
matrix elements. generates a transformed Hamiltonian ?i' such that ?-t has no matrix elements
between states of different I.
I, Il I) Presumably S must be small, since the original Hamihonian 11. has small
~~~ matrix elements off-diagonal in 1. Therefore we may approximate by expanding
", the exponentials in (0.9):
I]
{---1---, ---~---·~---i---
,,
'::
,, ----,---
, ,
o 'HllIO/l'Zt'rU ?i' = e-iS?ie iS
Il ,- __ ---'--- ---,---
~; + ... )?i(I+iS- ~; + ... )
{ 0'H2/l = (l-iS-
, , ,
01l ' Uro
: : I
I, I_::..l:i:_:_::'--i--_-l.i-_--.J-_-_--'--l--_- .
= H + .[H, SJ + ( SHS - T
S' H - -2-
HS')
.- ' " .
Fig.D.1. Hamiltonian matrix. The regions in which nonvanishing clements of?i l or 1i2 = H + i[H, S] - ~[[H, SI, SI (D. 10)
may be found are labeled by the shading. The quantum numbers II, 12 lind b designate
different eigenvalues of 1io
Writing ?i =?io +?il + ?i2. we wish to choose S to eliminate 11.2·
By writing out (0.10), we have
The technique that we shall describe below in essence provides a transfor- 1{1 :: 1io + 1{t + 11.2 + i[?io +?i h Sj + i[1t2' 5j - ![5. [S,?ill (D. II)
mation which reduces the size of the matrix elements of 11.2 joining states of We can eliminate the third teon on the right by choosing
different 1. In the process, new elements are added which are diagonal in 1. In
this way states of different I are. so to speak, uncoupled, and we are once again ?i2 + i[1{o + ?iI, S] = a (D.12)
faced with diagonalizing only the smaller submatrices diagonal in I. Then we have
The basic technique may be thought of fonnally as follows. The set of basis
functions 1/J10/ fonns a complete set, but has the troublesome 11.2 matrix elements 1{1 = 1{o + 1{1 + i[11.2. 5]
between states of different I. We seek a rransfonned set of funclions ¢/O/ given i2 i2
by + 2 [[Ho + HI, S], sl + 2 [[H" 5], S] (D.13)
• _e iS •I• (D.6) If ?i2 were zero, S would vanish. Therefore, we expect 5 will be of order
'1'/0/ - '1'/0/
1{2. and the last teon of order (?i2)J. Neglecting it, and utilizing (0.12), we
where S is a Hennitian operator that reduces the size of the troublesome matrix
have
elements. In tenns of the ¢'s, the Hamiltonian matrix elements are i
1t'=Ho+HI+ [H"SJ . (D.14)
Z
586 587
Equmion (0.12) may be put in matrix fonn to obtain an explicit matrix for S. i
(fO'I1-£'jfO'I) = (lcrl'Ho + 'HI + 2 ['H2, SlIlcr')
Using the facts that 1i1 has no malTix elements between states of different 1 and
l
that 1i2 has none diagonal in 1, we have = E/ona , + (lal 1t llla )
X [Ell ~ EI" + EI ~ EIII ] (0.20) is the partition function. Since the energies E a are energies of the N-panicle
The off-diagonal matrix elements are therefore reduced in the ratio of 1i2 10 the system, they may range from -N'YhHo1 to +N'YhHo1 as a result of the Zeeman
difference between eigenvalues of 'Ho, and the states of different 1 are "uncou- energy alone. Of course, the energy -N'YhHoI would occur only if all N spins
were in the state m = I, and is thus quite unlikely on a statislical basis. However,
pled". The matrix elements diagonal in 1 are modified, too. They become, using
(0.14) aod (0.17), we expect to find typical values of lEal';::' .../N'YhHoI combining the m values
588 589
of the N spins at random. Since N ';;t 1023 in a typical sample, how can we If we define m = },f - mj. we can write
approximate Ea/kT <:: I?
I! is clear that no one spin inleracts with many others. so that in some sense
exp (liwoMlkT) ;;: exp (llwom/kT)exp (tlWomjlkT) (E.8)
we do not really need to consider 1023 spins to get a fair precision in computing Then•. using the fact that (ml>m2•...• mi •... I/J"'jlml.m2 •...• m i....)
;I'. That is, assening that 1023 spins are involved is really a fiction. After all. = (mjlp:",jlmjl>. (E.?) becomes
in an applied field of reasonable strength we can predict the location of the
absorption by considering only one spin. "( ) _ 1rhw '"
X w - kTZ L.J exp (mllwolkT)
We believe that the high temperature approximation will hold if the energy ml,m1, ... mj-l,mj+I, ...
of a sing/I! spin is small compared to kT. We wish now to demonslrate how that
comes aooUi. To do so, we shall consider a simplified case. that of N noninteract- x L exp (-m jhwolkT)J(m j I/t",j Imj)1 2
ing identical spins, and assen that a similar argument should hold for interacting j ,m j ,mj
spins provided that the effective interaction between pairs is still small compared x 6[llwo(mj - mj) - llw] (E.9)
to kT so that no drastic phenomena such as ferromagnetism results.
If the spins are noninteracting, we can choose as exact quantum numbers Here we have used the fact that the sum over the N - I coordinates omitting
the individual spin quantum numbers ml.m2, ... ,mN. The energy Eo. then Tni is independent of which j we omit since the spins are identical. But
becomes Z = L exp (mliwolkT L exp (mj1iwolkT) (E. 10)
N ml,m2, ... ,mj-t,"lj+l. IIIj
592 593
where (F.I)
and the amount is proportional to the small time interval ot. C 1 is a constant
which depends on how often jumps occur. Of course, since MOo and Mb are
(F.2) complex quantities of the fonn M z + iMy , (F.8) is a vector relationship. The
constant C 1 is real.
I
0'=-+170
. h where (F.3) The process of (F.8) will diminish lhe b-magnetization by
T,
(F.9)
ho=Ho-w/-y (FA)
When a jump occurs, the a-spins in one CH3 group leave the b-site, diminishing
We wish to describe nuclei al two siles, a and b. Let us denote the total M~+iMy the b-magnetization, and go 10 an a-site, adding 10 the a-magnetization. The
for nuclei at an a-type site as MOo and M~ + iMy for nuclei at a b-type site as nature of the molecular formula for our example requires also that during the
Mb. MOo and Mb are then complex quantities. We define very same jump a-spins go to a b-site, but in a more general case (e.g., if we
had a CDJ group and a CH3 group on the molecule and were considering only
(F.5)
the proton resonance), the jump of a-sites to b-sites is independent of the jumps
the displacement of the frequency from the resonance condition in H o, and of b-sites to a-sites.
introduce the splitting Ow between the resonant frequencies of the two sites. By the same token a-spins will jump to b-sites adding to Mb an amount
Then, were there no jumping, the nuclei at an a-site would obey one differ-
OMb = C2Mo.Ot (F. 10)
ential equation, those at a b-site would obey a second differential equation.
dM
- l Mll.( M,
- : - - + 1 Llw+- MOo+I,-H\
dt T2
'W)
2
.2 (F.6a)
and diminishing MOo by
oMo. = -C2Mo.Ot (F. I I)
dM.
- - : - - M•
dt
+1
T2
,w)
. ( Llw-- Mb+I,-H,
2
. M,2 (F.6b)
So we can say that as a result of jumps
dMo.
& = CI Mb - C2Mo. and (F.I2a)
where Mo is the thennal equilibrium static magnetizalion of the sum of the two
sitcs_ dM.
The steady-state solution of (F.6), obtained by setting dMo./dt = dMtJdt = 0 dt"" = C2 M a - C 1 Mb (F. I2b)
goes through exactly as in Section 2.8, giving two Lorentzian lines of equal These rates can be added into (F.6a) and (F.6b), respectively.
amplitude, one located al
L1w Ow
=-2 or w =7 H O- 2 'w (F.7a)
dMo.
&=-T2
Mo.. (
+1
'W)
L1w+2" Mo.+CI Mb- C 2Mo.+ 1,H\2
. M, (F. 13.)
the other at
(F.13b)
L1w = Ow
2
or 'w
W=7HO+2" (F.7b)
For our example, whenever the molecule reorients a-spins convert 10 b-spins and
To include the molecular motion, we now assume that the molecule reorients, b-spins convert to a-spins, so that Cl = C2 = C.
carrying spins from an a-site 10 a b-site, and vice versa. We assume thai when the These equations are easy to solve in the steady state, since then they are
reorientation occurs, the actual process of going over the barrier is very rapid, two simultaneous linear algebraic equations for the complex quantities M a and
so rapid in fact that the protons do not change their direction. Thus if we lake a Mb·
lime interval ot sufficiently long that a number of spins al b-siles jump to a-sites, The result for the total complex magnetization, Ma: + iMy is
they will add 10 the a-magnetization an increment oMa of the form
·M M M i7 H \MoT[2+r(O"Oo+O"b)!2)
M z+1 y= 0.+ b= (F.14)
(F.8) (l + O"o.TXI + O"bT) - 1
expressing the faclS that Ihe eXira magnetization oMo. has Ihe orientation of Mb' where
594 595
(F.ISa) R the distance to the nearest neighbors and the ± signs represent the fact that
the field due [0 the neighbor changes sign depending on the orientation of the
neighbor. Consequently the resonance of the nucleus in question is spread above
(F..ISb) and below the value w:::: "(Ho. In a solid, a given nucleus has many neighbors,
some aiding, some opposing the external field. Thus instead of having twO peaks
as in the case we analyzed, we get a smear with a width somewhat broader
1 (F.16) than would result from having only one neighbor. We thus make the rough
-=c
T
identification
6w Z,,(/J
The absorplion signal is given by My. the imagi.nary part. of ~F.14). The 2:::: R3 (F. 17)
results computed by Gillowsky and Soika (F.8] are displayed In Fig. F.3. The
. I bele<! I 2 3 and 4 correspond to various values of rliw. NOIe where Z is a small numerical factor representing the fact Ihat there is more than
vanous curves a ..' . I f
that the theoretical results have the same form as the expenmental ~su IS 0 a single neighbor. The process of self-diffusion enables the neighbor atoms to
Fig. F.3. From the curves we see that the transition from the c~se of '~freq~e~t move, so that one neighbor is replaced by another. In the process, the resonance
jlUnping to the case o!rapidjumping occurs when raw ~ 1. Thl~ re~atlOnshlp IS frequency of the nucleus undcr observation may change depending upon whether
widely encountend and is ojgreat importance!or numerOlls appllcatlons a/mag- a neighbor is replaced by a nucleus whose moment is oriented in the same
netic resonance. It describes morional narrowing. Note though the total pattern direction or in the opposite direction from the magnetic moment it replaced. If
changes character when Tliw :::::: 1. there are also changes ap~nl when or6w:> 1 T m is the mean lime a neighbor sits before jumping, we surmise
, _ 1 Thus in Fig. F.2 the low temperature peaks begin to broaden long
m~w_. , . . h'h T:::: T m (F. 18)
before they collapse and merge into a single peak. LIkewise at 19 tem.pera.
lure Ihe peak conlinues to narrow with increasing temperature long after 11 has At low temperatures, then, we expect a nuclear resonance line width order 6w/2
given by (F.17). As we increase the temperature, we expect T m to become shorter.
become a single peak. .' . .. f
In general, associated wilh motion there IS a sphtbn~ 6w, characlenstlc 0 When
Ihe zero-motion (long T) limit, and a physical process which ca~~s Ihe nucleus (F. 19)
to change resonant frequency, with an associated time T descnbmg how often
on the average the frequency changes. . . we expect the nuclear resonance line width to begin to narrow. Exactly such
For example, in a solid, a typical nucleus expenences magnen~ fields i1H effects are seen, as is discussed in Chapter 5.
of order ± p/W where}J is the nuclear magnetic moment of the neIghbors, and Chemical exchange can be studied by its effect on the nuclear resonance
absorption lines. For example, the spin-spin coupling in a molecule gives rise
to structure in liquids, as illustrated in Fig. 4. I I. In liquid CH3CH20H, the OH
2f6w
protOn splits the CH2 proton resonance structure if the OH proton remains on the
, 10- 1
1
molecule. But if there is rapid chemical exchange of the OH proton with other
protons in the liquid, the structure may be washed oul [F.9].
3 10
4 10'
597
596
possible natural precession frequencies. For the case of diffusion in a magnetic Mz. and My arising from precession. If we transform to the reference frame
field, there is a continuum of magnetic fields. For simplicity, we assume the rocating at the precession frequency, we eliminate the precession and can write
magnetic field, though inhomogeneous, has axial symmetry so that in this frame
H(x,y,z,)=Ho+z az
8H
(G.l) 8~:: '" M
DV2 z (0.3,)
In Appendix F we described the process whereby spins switch between sites
of different precession frequency by the nile term (F.12a) and (F.l2b). When a;:y = DV2My or (O.3b)
diffusion can take place, the possible precession frequencies form a continuum
as expressed by (G. I). 8M+
The usual way to describe diffusion is by means of a diffusion equation. --=DV 2M+ (0.4)
at
The use of a diffusion equation in conjunction with the Bloch equations was
introduced by Torrey [G. I ]. We shall follow a rreatment which is a slight sim- using the definition M+ == Mz. + iMtI •
plification of his. . Since M+ describes a two-dimensional effect, (GA) treats vector effects. It
Suppose we had a homogeneous static field in the z-direction, had no ap- IS analogous to (F. 12), if we bear in mind that M+ is a function of position.
plied alternating field, and by some means had produced a nonuniform M z , as in We now proceed as in Appendix F to recognize that when the static field
Fig. G.Ia. Let us suppose T t is infinite. TIlen the total z-component of magneti- is inhomogeneous, we must now include the effect of the spread in precession
zation cannot change, but as a result of diffusion the region of magnetization will frequency.
spread as in Fig. G.Ib, eventually leading to a uniform M z throughout the sample We therefore add the precession driving terms giving
(Fig. G.lc). (Recall that the symbol M z denotes the magnetization density.) The
process of Fig. G.I is described by the equation
aM+(~ y, %, t) = _ i-yh(x, y, z)M+(x, y, z, t) _ M+(:t. y, %, t)
T,
8Mz = DV 2 M (0.2) (0.5)
at ' where
where D is the diffusion constant h(:t, y, z) = H(:::. y, z) - H o (0.6)
If the magnetic field were uniform, and there were initially also :t- and y-
components to the magnetization density, there would also be diffusional effects and Ho is the spatial average field over the sample. Substituting (G.I) for
for Mz. and My. But since the static field causes a precession, two effects are h(:t, y, z~, w~ ~btain an equation describing free precession in a static magnetic
present: (1) changes in Mz. and My arising from diffusion, and (2) changes in field which IS mhomogeneous but possesses axial symmetry:
8M+
- - '" -i-yz -
(8H) M+ - -M+ + DV 2M+ (0.7)
i)t 8z T2
(.) (b) (,)
Equation (G.7) incI~des the natural T2 effects as well as the T, effect. [The fact
that T} effects are Included can be verified by examining (2.86) with HI set
M, M, M, equal ~ ~ro). Note that since aH/8z is evaluated at the origin (:t = 0, y = 0,
z = 0), 11 IS a constant The only explicit dependence of (G.7) on position in the
sample is therefore in the first term on the right-hand side.
If there were no diffusion (D = 0), (G.7) could be solved simply since over
o L o L o L a layer of constant z the equation describes precession in the rotating frame in a
Fig. C.lao<. A sample in II static field in the z-direction is lnllgnetized in the z-direction. constant field h, together with decay at the rate T2. The result is
The magnelization density M~ is IIllSUmed initially to be nonuniform. Figure G.ta shows
the initial M~ as a function of the :a:-coordinate, the sampJ.e extendins from: = 0 to
M+(r,t) = M(r,O)e-I/T'e- i-r z(81l/8z)t (0.8)
z. = L. As a result of diffusion, the magnetization density at a later time (Fig.G.tb)
has spread throughout the sample, with a diminished peak. Eventually, the M. becomes where M(r,O) is the complex magnetization density at t =
O. Suppose that
unifOnlJly spread as in Fig. G.le. Note that if T 1 is infinite, the lotal z-magnetiution, M(r,O) were uniform, as would be the case if it were prepared by applying
which ill the area under the curve of 11-(.(:), is the same in all three casn
a 7r(l pulse to a sample initially magnetized to thermal equilibrium along the
598 599
static field. (We here neglect the slight variation in thennal equilibrium static
magnetization produced by the small inhomogeneity in a sialic field). Consider -V-T
BH into 'II'"
BH
+ "(Z-T
the magnetization density al th, z, and z + L},z where az is a small distance.
Z- I
Bz Bz
Since these three planes are equally spaced. their precession frequencies are Thus at t = T+
equally spaced. Swting in phase at t = 0, the magnetization at z+ L1z develops a
lead in angle over that at z, that at z -.dz develops an angular lag relative to the M+(r,T +)=Moexp ( -T2T) exp (-V (BH)'
7 Dz T')
"3
magnetization al z. the lead angle and the lag angle growing progressively with
time, bUI always remaining equal. When spins from both z + L1z and z - L1z x exp (iJr) exp (i"(z ~~ , )
diffuse to z, they will do so in equal amounts. Thus they will add in equal
amounts increments which lead as increments which lag, giving no net phase We now use (G.12) to describe the development of M+ in lime following T:
change for the resultant magnetization at z. Therefore, we expect that diffusion
will not affect the phase of the development of the magnetization at z, but will
affect the magnitude. As long as we are not near a boundary, every plane z has
M+(r,t - T) = M+(r,,)exp ( _ (t;2 T»)ex p [ -n("( ~~Y(t ~ ,)3]
( . BH )
planes symmetrically Llz above and below with phase lead and lag, respectively,
from which magnetization diffuses to z, the phase lead or lag depending on x exp -I"(z Dz (t - T) (G. 14)
.az independent of z. Thus we expect that the diffusion-induced decay will be
Substituting (G.~3! we see that at t - , = T (or t = 2T), the complex phase
independent of z. We therefore try a solution factors cancel, glvmg
1I1+(r,t) = Moe-tjT2e-i.,(8ff/8:)IA(t) . (G.9)
When this solution is substituted in (G.7) we get a differential equation for A(t) M+(r,2T) = -Moexp(- 7
~)exp[ -v( ~~J2;3] (G.15)
for the magnetization following an initial 7fn. pulse. This equation describes the H. The Equivalence of Three Quantum Mechanics Problems
development of magnetization with time from its initial value Mo at t = O.
We now need to consider what happens if we apply a 7f pulse at time T
producing a rotation about the y-axis. The magnetization density just prior to the In this appendix, we show the equivalence of three quantum mechanics problems:
pulse is 1. A. pair of identical spin ~ nuclei coupled to a strong static field, interacting
600
601
(H. I) Table H.I Table B.2 Table H.3
+ b/2
2 (H.3) 1/>\,-1 WO
(3) 'Ho = -hwoFz + rlc(3F; - F ) pp WO + a/4 f .. -a/4 f .. 0
It will be convenient for all three cases to express the energy eigenvalues in units T1' {'H.} = 0 , T, {'H,) = 0 , T, {'Hol = 0 . (H.l4)
of frequency. Wi. Thus, adding the diagonal elements must give zero. The energy columns of Tables
(R8) RI-3 all add to zero.
Wi == Eil h . The last Hamiltonian is
So ::= -wo(ml + ms) + amlmS
(H.9)
Will/illS
Ji c ::= -hwoF: + hc(3F; - 1F1 2 )
See Table RI. We note that the states Cli{J and {JOI are degenerate. Thus any linear
603
602
i
for F = 1. Obviously I Fl 2 and F; commute with il so we immediately take the Since (l'J1-£pjl) vanishes between F = 1 and F = 0 stales, for cases (I) and (2)
stales .pFM as eigenstates. getting Table H.4. .. . of coupled pairs of spins we get. using the notation aFAI
For all three Hamillonians. we could induce transitions by applymg a trans-
verse alternating field. aoo = constant . (H.2t)
Hz(t) = Hzo cos wt (H.15) The other aFAI's obey the equations
604 605
Chapte r 11. These features give rise to a resonan ce frequency W which, as shown
in (4.187), goes as
0.1)
to the crys-
where (8, q,) specify the orientation of the static field Ho with respect
(8,,p) are present random ly, giving rise 06
tal axes. For powder sample s, all angles
howeve r, quite charact eristic singulari-
to a spread in resonant frequencies with,
. Such spectra are called powder
ties from which wo • wb. and We can be deduced 0.'
B/oemb ergen and Rowlan d [1.1]' We
pattern s. This problem was first treated by
powder s. In doing so, we
tum here to calcula ting the intensity pallems for such
liquid crystals or some
are omillin g discuss ion of fascinating systems such as 0.2
crystals and
form of oriented polyme rs which are concep tually in between single
l'lSlJ.
random powder s.
We express the random orientation of the crystallites by saying
that the o~o'---~20I-'--,,,:o--,,~-,:;--f-;;:;;--:-::--:~. L~.,-- -2:,j
80 le.O
90 100 160
120
probabi lity dP that Ho lies in any infinitesimal solid angle dfJ is given by
¢lldrgrr ul
pJ
'"
(I.2) F'ig.l.I. Curves of constan t frequene Y .In the..t ane for" general anillOtropic chemic,,'
dP "" dfJ "" sin 8 d8 d.#47r = d(- cos 8)d#41f or Knight shift '1'":
4,
Introdu cing the variable z defined as
l(fI)Ll fi = LlA
z=-oo s8
(1.3.)
~ ~
.
we have that dP is This relation ship shows that if we divide the total frequenc y range Into equal
frequency interval th be
s. e area tween successive curves of constan t fre uenc .
P~portlofnalto (he in~ensity at that frequency range, giving us a simple ~Phrc~:
.
dP = d~dz (l.3b)
4~
rlrz plane corre-- picture 0 where the Intensity is high, where it is low.
The advanta ge of the variable s (q" z) is that equal areas in the
now describ e how to calcula te len). We first treat the simplest
We case
. I '
spond to equal probabilities. axla symme try, then do the general case.
Inttodu cing fJ defined as
Axial symme try means that two of the principal frequencies are
the same
(1.4) (e.g,w(J =w,,~w orw =w t et) ' IS
The ca Icu Iatlon .
. easIest if we pick
fJ == W - Wo
th 8 _ 0 . c' (I C Wb, c..
we have e - aXIs to be the symme try axis (W(I = Wb) since then axial sYrnme try
means that the q,..dependent teons of (1 .5) vams ..
. h gIving
n "" W(I sin 2 8 cos2 ,p +wb sin2 (J sin 2 q, + We cos2 8 2 n= + (we - w(I)z2
2 ,p + [(we - w a ) - (Wb - w a ) sin q,]z2 (1.5) W(J
0.8)
= W a + (wb - w a ) sin
frequency, fJ. Therefo re, cu.rves of constan t frequency in the q,-z
plane are straight Jines at
This express ion can be thought of as detenni ning lines of constan t
in the rlrz plane. Solving for z2 we have ~n~~ti: t~~~' 1.2). The area, L1A', between two curves differing in frequency
2
2 n
-wa - (Wb -wa )sin q, ~~ LlA' = 211"Llz
z= 2 (1.9)
(We - wa) - (wb - W(I) sin 4J
SO that it is straight forward to calcula te the curves of z versus
q, at constant whe~ the 21f ~mes from the total range of ¢J, and where Liz is the vertical
er, then, spaclOg approp nate for the range Lin :
frequency in the ,p-z plane. We illustrate such curves with Fig. 1.1. Consid
two curves at frequencies
finding a nucleus whose
nand
frequen
fJ
cy
+ dfJ.
within
As a
this
result
frequen
of
cy
(1.3),
range
the
is
probabi
just
lity is
lity of
the
also
area
the
Llz = (afla,)
- Ll
fl. (1.10)
L1A in the rlrz plane between these two curves. But the probabi
J. Thus we have Such a strip occurs both for O<8<lI"n and lI"n<8<1I". since n is an even
nonnali zed NMR line intensity 1(fJ)df
606 007
z
1 .0 § §
0.8
O.
0.8
0.7
0.6
~5 y
FIg. L2. Curves of COfl5tan~ r~uenc y
in thf: t/J-z plane for axially !Jymmct-
ric chemical or Knight shifts. " is the
nonnl\li~f:d frequenc y defined by (1.16)
y=
O-Wo
we:-wo
From. (I.13) and (1.14) we
We <Wa • and extends from
see that y is always positive whether
a to I:
We:
(Ll6)
> W a or
8, I (Ll2)
an "" 2J(n wa)(w c wa) ~~;.~~~.g that I(y) of (l.I8) is properly normalized. This line shape is shown in
the line extends from Wa on the low end to We on the high end, with
Ilyl
(Ll4)
Wo $ O$we
If. however. We < Wa (the symmetry axis is at the high frequency end of the
spectrum).
(1.15)
We $ O:S Wa
Fi~ I.J. NMR line shape ror An
argument of
In this case. both (0 - w o) and (we - w o) are negative. so thai the aXially symmet ric chemica l or
the square rool in (1.12) is still positive.
yas
It is useful to introduce a dimensionless, variable y in place of n. We
define
o 0.2 0.4
Y
0.6 0.8 1.0 ~night shift tensor. NMn. inten_
sity 1(1/) versus norm"li~cd fre-
quency II
608 609
,.,
1.0
":1g.l4. Lines of e.onlllant frequene.y tI in one pot-
We adopt the convention that WI! lies between w'" and We.. but allow either Wo > We. , tlon of the </>-z plane for 1=0.75
or Wo < We.' Thus W/I can be either the low or the high end of the frequency 0.'
0.'
spectnlm. With these conventions we have
(1.201»
OSfSI 0.'
Note that f tells what fraction the total frequency interval We. -W/I lies between
Wo and Wb' In other words, Wb divides the interval into twO fractions f. between 02
0.4 yzf~O.lS
w'" and Wb' and 1 - f. between wI! and We.' Then we gct from (1.5)
the curves for 'II"{l ~ ¢ ~ 'II" are mirror images about ¢ = 'II"{l of Ihose for 0 ~ ¢ marked these curves on Fig. 1.4. . a e
~ '11"/2. Therefore, we need only consider the curves for 0 ~ ¢ ~ K/2. To calculate I(y) we now consider two lines in the first octant of the tit-
Since (1.22) involves z2. Ihe curves are an even function of z. Thus, we plane, on~ c~rrespond~ng to frequency n, the other to fl+ L1n. Then, recallin;
need consider only posilive values of z. The net result is that if we calculate I(y) L1y IS gIVen by eight times the area, L1A', between the two lines at positive
any areas for the octant 0 ~ z ~ I. 0 ~ ¢ ~ 'll"n. we must mulliply the net result z.
by a factor of eight to get the full area. To get the axially symmetric case with I(y)Lly = 8LlA' (1.29)
Wa = WI!. we simply set f = 0 which gives then
Foc:using on two such curves, we examine Fig. 1.5. We look first at the small
z2 = Y . (1.25) regIon bounded by lines at y = constant and y + £1y = constant. The area £12 A'
of the small shaded region. essentially a parallelogram, is
Figure 1.4 shows the lines of fixed frequency y for the first octant (0 ~ q, ~ 7rn,
o~ z ::; I) for the case f = 0.75. For some values of frequency y the z(¢) curves £1 2A'-Ll
- z Ll." . (1.30)
go from ¢ = 0 to 7rn. Others intersect the z = 0 axis at smaller values of q,. The Th
. e spacmg Y'IS am·mary. but the spacing £1z must correspond to the frequency
. Ll"·
curve which divides these two classes has a y which has z = 0 when ¢ = 7r/2.
Interval. hence is given by (1.10). Therefore
From (1.22) that is
(1.26.) Ll 2A' = 0% L1y Ll¢ (1.31)
y=f . 8y
which means From (1.22) we have
(1.26b) 0% I I I I
'2 ~.;7;=1~f"'S'Cin"j'=;¢:-';7Y=-=;f"s"'in"'=¢
{l=WI!
Equation (1.22) also shows us that at the origin (¢ = 0, z = 0) oy = 2z I - f sin 2 11 =
(1.27) 1 I
y =0 or
=2.jY :-.;r.I===f~s~in'j"=;¢~.;F.l=;(f'7/"y):=""'·nr.,¢ 0.32)
611
6'0
FIg. LS. Curves or
constant frequency in lhe ~, Since y s;: 1, f < fly, and we can identify
, plane for two nearby frequences nand + .dO
n
p' =I .' = Ily (1.39)
giving
·1' (fly) - I . ,
I( ) _
Y - y/(1
4
f) Jo dX/ 1 -l_/Stnx (1.40.)
'1'
-::-YVm(14~/) Jo dx / 1 l(l - y) . ,
= - (I _ f)y sm x (1.40b)
•
Therefore from (1.29-31) we get For the case of y < I (n < W6), where the maximum value of ~, ~max, is
I(y)m-
8 'mJ.. d. (1.33)
less than 7f12. we can transfonn Ihe inlegnl by a substitution of a new variable
~' defined by
2.,fY 0 ../1 - Isin' ."/1 - (fly) sin' • sin 2 ~' "" Ly sin 2 ~ . (1.41)
In this equation, tPmax is the maximum value of ¢ for the particular frequency,
y. Thus, as can be seen from Fig. 1.4, ~' is chosen so that when ~ = O. ~' "" 0, but when ~ "" ~max, ~' "" rrl2. With Ihis
substitulion, it is slraightforward to show that
(1.34)
I rbm.x drjJ
If y < I, tPmax is easily found by
sin2 <PmKx = 7.
setting z = 0 (1.22):
(1.35)
.,fY 1 "/1-sin'."/I-(ylf)sin'.
If ¢mu = 7rn.
the integral turns out to be a complete elliplic integral of the first
1 J0 ../1 - ysin' .. dO'
'1'
kind. GradshU!yn and Ryzhik [1.2] show thai if (1.42)
= ,jJ ../1 - (y/f)sin' ..
O<p2<q2<1 , O<::Z:~1f(l (1.36)
Apart from the change in the inlegralion limit. Ihe right-hand side differs from the
a dx 1 (~) left solely in imerchanging y and f. AI all places, therefore. for y < I (n < w6).
F(n, k) = J
a
2 2
d.
(1.37) o
- (I - y)f Sin 'I'
4 ./2 diP An altemalive nOlalion commonly encounlered for Ihe complete elliptic
I(y) = - J-r.'==;=e=r.'7.-~~F.
.,fY 0 ../1 - sin' ."/1 - sin'.
I (fly)
(1.38) integnl of the firsl kind is K(m) defined as
612 613
tr/2
Jo Vl-msin
I
ley) -
_
,Jf(l
4
y)
K(Y(l-f))(I y)!
(145)
. 0
the energy levels obey (10.63)
Ii'
2
=_ hH + e qQ (3cos 8-1)(3m2 _1(l+I))
2
£Jfn. .., om 41(21 _ I) 2
and for y >f (n >W/I)
_
ley) - lY(1
4
j)
K(I-y)!) y(1 - f)
(I 45b)
.
with transition f~uencies
which brings out clearly the relationship between the frequency regions on the 2 2
= + e qQ 3cos 8 - 1 3(2m + I) (1.48)
two sides of w".
Note that y is the distance in units of normalized frequency WO 4[(21 I)h 2
=
from the "lower" edge (y 0), whereas 1 - y is the distance from the upper This leads to the angular pattem for anyone transition which is similar to that
edge. A similar relation holds for f and 1 - f. The argumenls m of the two of Fig. 1.3, with a singularity at 8 = rrn.. However, for each transition (m. Tn - 1)
ponions of the spectrum both approach m = I al n
= Wb' As m goes from 0 [0 there is a cOlTesrnding one between levels with the m values reversed [e.g.
1, K(rn) goes from about 1.57079 to 00, blowing up at m = L Thus, the line has in addition to ('2',~) there is (-~,-i)], This occurs at (m' + I,m'), where
a singularity at n
= Wb, but is finite everywhere else. The line shape is shown m l = -(m + 1), leading to replacing the factor (2m + 1) by
in Fig. 1.6. Note from (1.45) at y = 0 and y = I the intensity is
(2m l + 1) = -2(m + I) + 1 = -(2m + 1) (1.49)
[(0)· ~K(O) [(I). ~K(O) (1.46) This reverses the sign of the coefficient of (3 cos 2 (J - 1) in (1.48), so that the
pauem is reversed about w = we (Fig.I.7).
so that
Bloembergen [1.4] noted that if 1 is ~. ~. elC., the term (2m + I) vanishes
[(I) If when m = -!.
so that the (i, -i)
transition has zero first-order quadrupole
[(0) • VI::! (1.470)
splitting. He shows that it is shifted in second order by an amount
1.0
0.'
0.'
Ily)
0'
0.'"---- _ e qQ.
21'1
0 e qQ.
'1'1
"'-"'0
Ag. L6. NMR intcn3ity l(y) versus nor- Fig. L7. Powder line shllpe for 1 =
i1 NMR b<-oadened by II first-order, "xilllly symmetric
o~o--'O:':.':--O;';.':---;;O.•,--JL.,o.'::,--;1.'=-O- maJized frequency y, for I = 0.70 c1edric field gradient, The singula~ities at .... - wo = ±e'qQ/4h have Ixoen made linil.e
, by convolution with a na~row ~adening funo:tion,. This ~a.r: cur~e "Is<;> "rises in the
powder line sh"pe of " ""Ir of dlpolar-c:oupled Idenlical spm 2 nucleI, see (1.51)
614 615
I(ul t
FIg. U. The line shape of the ~ to - NM R produces an infinitesimal change in ¢ - hence as the infinitesimal time goes to
transition resultinr; from the powder average zero, so does the change in ,p). Thus the wave function just after the sudden
of an axially symmetric eledric field r;radi-
ent. The broadeninr; ariseS from treatins the change is identical to its value just before. So if we divide time into intervals
field yadient to lIeCOnd order. A. is defined by t .. t2' ; .. ,tN with time-dependent Hamiltonians in each interval 1i lt 1i2. etc.,
(1.50) we immediately get that at lhe end of the nth interval:
1/l(t) '" exp ( -*1ill l n ) ... exp ( -*1i2l2 )ex p ( -{1illl )¢(O) (1.3)
If one expresses this relationship in tenns of the variable z of 0.3a), and ,,(t) = exp ( - ~/ h(t')dt') ,,(0) (1.5)
utilizes (Db), 0.7), (1.10), and (1.1 t), one obtains the angular pattern of Fig.U.
The case of lhe second-order quadrupole broadening of the (!' -!) transition has However, in general
been treated by various aulhors, and is summarized by Gerstein and Dybowski
[1.51·
[hi, hj] ,,0 , (1.6)
The first NMR paper to analyze a powder paltem line shape was Poke's so one is left with (J.3) as the best one can do.
treatment (1.6) of a pair of spin! nuclei coupled by the dipolar magnetic fields. It is useful to have a more compact way of writing this result. This can
This pattern is the same in fonn as that of the first-order quadrupole splitting for be done by a method due to Dyson {J.I]. He introduced the concept of a time
a spin ~ nucleus except that the singularities occur at ordering operator, which we will denote by To (D for Dyson, T for time). Using
it
3 'Y2h
Iw-wol "'4-;:J (1.51)
TO exp ( -*1i1 tl )ex p ( -*1i2t2) '" exp ( -*1i2t2 )ex p ( -{1i1 tI)
(1.7)
0'
,,(t) = "p ( - *
ht) ,,(0) (1.2)
F I F2 ··• F" '"
N
IT Fk
k"'l
(1.9)
616 617
Then, if later we came to evaluate the product and continued our careless ways,
(J.i9)
we might write out the product as
,p(t) "" exp (-i1l It) exp (-i1l2t2) _.. exp (-i1l n t n)t/1(O) . (J.1I) we get the coefficient of Tn as
If the 1lk's did not commute. we clearly would not have rewritten the correct (A2 + AI)n
result (l.3)! The Dyson operator allows us to deal with this problem because if (J.20)
n!
we wrote (J.ll) as Now, if At and A2 commuted, we could write (l.20) as
-*
n I
A fI - k A k
,p(t) "" exp ( 1ln t n )ex p ( -*1ln - ltn-I) ... -'t (n -
_ '"
k)!k! 2 I
(1.21)
k""l A n2 + n An-I
2
A
1+
n(n-t)An- 22
2 2 Al + ... (J.23)
since no matter what order we use to write out the product of the exponentials.
TO tells us to reorder them to get the order of (l.3). which is just (J.t8). That is
We can derive another useful theorem. Consider a product of two exponen- nIl
tials TO(A, + AI)" = L ( _ k)1 _A,-k A k (J.24)
k""O n . kl
(J.l 5)
We can therefore write
where Al refers to some operator at an earlier time tJ, and A2 an operator at a
later time t2. We can use the Dyson operator to wrile this as To ~ eXP(*1iktk) "" To exp (* ~1iktk) (J.25)
!(A2,AI) "" Toe Al + A\ or (J.16a)
We can now deal with a general time-dependent interaction 1l(t) represented
eAleA1 "" ToeAl+A\ (l.16b) schematically by Fig. J.la. First we approximate it by a series of short steps of
duration Tl,T2 • ... (Fig.J.tb) during which we approximate 1t(t) by 1t" 1t2.
At first sight this result is surprising since (1.16) would nOl be correcl without etc. We now use Ihis in (l.12) to wrile
the To present. However, the theorem is easy 10 prove by considering the related
functions as power series ,p(t) "" Toexp(* ktO 1tk Tk )J/J(O) (J.26)
618 619
~ 10) FI,.J.\. (.) A 1,~dop""O"IIl.mi1'o,,;." 1«.)
V(!rsus I (b) The Hamiltoman of (a) approxi·
"(') = Toexp( -* J o
(a(T)"Ii. + b(T)"Ii,JdT),,(0)
'H Itl
nlllted by a series of lIme-independenl n"mil-
lOllillns of amplitude 1ft (I: = 1 1.0 n) lind
duration Tt = exp ( -* J o
[a(T)"Ii. + b(T)"Ii,ldT) ,,(0)
.'
(bl = exp{ -T.["Ii. (J a(T)dT) + "Ii, (J, b(T)dT») ] },,(O) (1.34)
o 0
If we define the average value of a or b over the lime interval t as
'H It I ,
'. - tIf
aCt) =
I
a(T)dT , - tIf
b(t) = b(T)dT (J.3S)
o o
Sometimes the Hamiilonian has a par1icularly simple rime dependence then we can wrile (1.34) as
620 62'
Then for stroboscopic observation The theoretical formulas we have just given are sometimes referred to as
average HamiflQnian theory. For it to hold true, typical matrix elements of 11.,
,penT) = TDe,p ( -k J. nT
H(T)dT ,p(0)
)
(al1ilo/), must satisfy
o (al~la')T -<: 1 (1.48)
kT
=TDe,+~ t( J H(T)dT)],p(O) One useful way to think of this approximation is in terms of rotations. Rotations
k:::1 (k-\)T in three dimension do not commute, but infinitesimal ones do. Thus, if we think
kT of 'H a and 11.b as two rotations, their total effect is changed if we reverse the
=TD IT e,p(-~ J H(T)dT),p(O) (1.41) order in which we perform the rotations. If, however, the rotations are infinites-
imal, corresponding to satisfying (J,48), the order of performing the rotations is
k:::l (k-I)T
Bol irrelevant.
kT
In Appendix K we take up the corrections to the formula of (J,47).
J 11.(1')dT (1.42)
(k-\)T
K. Correction Terms in Average Hamiltonian Theory-
is independent of k, so The Magnus Expansion
,p(nT) = II [TD e,p(- *J o
H( T)dT ) ] ,p(0) = U"(T),p(O) (1.43)
In the theory of spin-flip line narrowing as well as in multiquantum excitation the
Thus, we need only work out U(T) to find !f(nT). effect of applications of sequences of pulses is analyzed using a time-averaged
Suppose, then, that the time interval T is very short so that U(T) differs Hamiltonian. Specifically, in Sect. 8.11 we analyze a three-pulse sequence which
only slightly from the identity operator. In this case, we can keep as our lowest causes the Hamiltonian to be effectively transformed as described in (8.152):
order approximation the first two terms in the exponential expansion of U(T), 1i A ::: p\-I1iintPt
getting
.T 11. B ::: p\-l P2-111.intP2Pl (8.152)
U(T) '" 1 - TDk J H(T)dT (1.44) 11.c ::: 11.int
o
where 11.in~ is the sum of the Zeeman and dipolar Hamiltonians in lhe rotating
However, thinking of the integral as a sum of infinitesimal tenos, we see that reference frame. Then, over one cycle, the effect of the Hamiltonian is given by
we never have a problem of commutating operators since we have no products the unitary operator UT of (8.153)
of 11.(1') at one time with 11.(1') at another time. Consequently we can omit the
time ordering operation, getting
UT ::: exp ( -*11.C1'3 )ex p ( -*11.B1'2 )ex p ( -*11.A1'l) (8.153)
U(T)::: 1 - *J
. T
o
11.(1')d1' S! exp
(
-* J . T
•
11.(1')d1'
)
(1.45) We expand the exponentials and keep just the leading terms, to write
;
UT :::1--(11.CT3+11.BT2'HA1'l)
Defining the time average of the Hamiltonian 11. as r.
::' ex p ( -*(11.C1'3+11.B1'2+11.A1'd)
11.:::
1 T
T J 11.(1')d1' (1.46)
The approximation is equivalent to writing
o
we can write eCebea = ec+b+ a (KI)
for three nOllcommuting operators a, b, c. Since clearly the expression (KI) is
U(T) = e,p( -*HT) (1.47)
only approximate, one would like to know how big the errors are. This problem
622 623
has been treated by many authors. A pioneering paper in this regard is the ';;t exp (e + b+ a)exp [!([c, bJ + [e, a] + [b, a])J
treatment of Haeberlen and Waugh [K.IJ. A useful mathematical treatment is
'" exp [(c+ b+ a) + ~(Ie, h] + Ie, aJ + [b, aD] (K.6)
that of Wi/cox [K.2]. The basic paper is that of Magnus [K3], who derived the
first terms of a series expansion for a time-dependent Hamiltonian 1{(t) as being where in these expressions there is implicit the idea that the expression is to be
equivalent to a unitary operator Vet) given by thought of as an expansion, and is accurate only up to terms involving products
of pairs of the operators.
U(t) = exp ( -*('Ii + ?i(l) + 'Ii(2»t) (K.2) As we show below, the expressions (K.6) is an example of including just the
first two terms, 1{ and ?t(1), of the Magnus expansion. Deriving the result (K,6)
where was easy. It is obvious now how to get the next term for our simple example of
, (K.I): merely by continuing the expansion of the exponentials. We leave that as
- tIJ
1{ = 1{(tddt l (K.3a) an exercise for the reader.
We now want to show that if we evaluate (K.3), we get the same result as
o
. t t1 (K.6). First let us define times tA, tB, and tc as follows. The time-dependent
'Ii(l) = - 2:. Jo J tit,
0
tit, ['Ii(t,),'Ii(t,)] (K.3b) Hamiltonian exists as 1i A from t = 0 to t = tA, as 1iB from tA to tB, and as
1ic from tB to tc. We further define TA' TB' and TC as the durations of the
three.intervals. Then we have, for one cycle, from (K.3)
It
f-c [i
t3/1
6th 0
J J J
'Ii(') = - , dt, tit, dtt!['Ii(t,), ['Ii(t,),?i(tl)J)
0 0 1i = f-c J dtl1i{tl) = dtl1i(t.) + J dtl1i(td+ J dt l1i(tl)]
0 0 tA to
+ ['Ii(t,), ['Ii(t,),'Ii(t,»)J) (K.3c)
!
= -[1i A TA + 1iBTB + 1icTcl (K.7)
Our goal is to give the reader a simple idea of where these expressions come te
from, and what they say in concrete terms. To do so, let us consider the three This is simply the average Hamiltonian we have used previously.
pulse example.
We turn now to the next term, 1i(l), given by (K.3b):
Let us therefore examine the two sides of (KI), utilizing power series
expansion of the exponentials. Thus
I
ec+b+a = I +(c+b+a)+ 2t(c+b+a)2+ ...
'liP! = -2:1, J, dt, J" dt!l'li(t,),?i(t,)] (K.3b)
o 0
I Now t2 is always later than or equal to t 1. Thus, they could both fall into the
= I + (c + b + a) + 2! (c 2 + b2 + a2 + cb + be + ca + ac + ba + ab) + ... same time interval, such as between 0 and t A , or t2 could be in one interval (e.g.
between tA and tB) and t1 in another (in this case necessarily 0 to tA)' Since
whereas (K.4) 1i(t) is a constant during anyone interval, it is useful to break the integrals into
eCebe
a
= (1 +e+ ;~ + .. .) (I +b+ ~~ + .. .) (I +a+ ~~ + .. .) these time domains. We therefore adopt a noration for specifying the limits of
integration:
! tc t1
= 1 + (c + b + a) + 2! (c 2 + b2 + a2 + 2eb + 2ea + 2ba) + ... (K,5)
Jdt2 Jdtl F(t2, tl)
where in (K,4) and (K,5) we have collected all the terms involving products of 0, o 0
I, or 2 of the operators but have not listed those involving 3 or more. Comparing,
we see that (K.4) contains terms such as bc+eb, whereas (K,5) has only 2cb, i.e.
e only to the left of b. We can therefore write
eCebe tl ';;! {exp(e+b+a)} +!(eb-be+ea-ae+ba-ab)+ ...
(K.8)
'" exp (e + b- a){1 + ~([e, b] + [e, aJ + [b, aD)
624 625
Consider the first integral. Since tl and t2 are both between 0 and tA, An imponam application of these equations is to find pulse sequences which
L.... ,2:SL.... I.... h :S t .... cause 11.(1) to vanish, Here is a simple example. Suppose that one considers a
Jd', J d', [H(.,). H(.,)] : Jo d', J d.,[HA.HA] 0
pulse sequence of equal T'S. Then
. . ,
o 0
H(I): -~ L [H;.H,) (K.16)
:0 (K.9)
2Th j >k
Consider then the second tenn on the right side of (K.8) so that the vanishing of 11.(1) arises from the vanishing of the sum S defined as
to t.... In L....
L
Jd', Jd', [H(.,).H('I)] : Jd', Jd'I[HB.HA] S:
j>k
[H;.H,I . (K.17)
I.... 0 I.... 0
(K.IO) For a three-pulse sequence this can be written
= [1i8' 'HA)T8 TA
We therefore get S: [H,.H,I + [H,.HtI + [H,.HtI (K.18)
. te t2 Now suppose
Jill) : - - ' -
2tch
Jd', Jd'J!H(.,).H(tI)]
0 0 1-£1+1-£2+1-£3=0 (K.19)
as was true for the Hamiltonian of (8.158). Then, since
: _ _i_({'H8, 'H",lT8TA + ['Hc, 1t'",]TCTA + (1{c, 1(8)TC1'"S)
2'C" (K.ll) [H"HtI:O .
so that using (K.ll) the term of (K.2) involving 1-£(1). -i1f!..l)tc/h, is we can write
i'H(1lt c 1 (K.20)
- = - - , (['Hs, 1{",)TATS
" 2"
+ [Hc.HAITcTA + [Hc.HBITcTB) (K.12)
Alas. this does nOI vanish. How can we construct a pulse sequence which van-
ishes? Haeberlen and Waugh point out that one can readily find a six-pulse
Now. identifying a. b. and c as sequence which vanishes.
Introducing the notation
i i i
a= --",11. AT", ' b= --",'HBT8 c = --'HcTc (K.13)
[H;.H,I " i. k (K,21)
we get that the correction tenn of (K.6) " we have
1 S = 6,5 + 6,4 + 6,3 + 6,2 +6, 1
2([e. b] + [c. al + [b. al)
+5,4+5,3+5,2+5,1
1
: --.2 ([1t'C.'HB]rC TB + ['HC.11. A]TCTA + ['HB, 11.",]TSTA), (K. 14) +4,3+4,2+4,1
2"
which agrees with (K.12). +3,2+3,1
It is clear that if one has a succession of intervals j = I to N. of duration + 2,1 (K,22)
Tj and Hamiltonians 11.j, the general form of 'H(I) is Now, previously we had
11. 1 + 11.2 + 'H 3 = 0
-2~h ,2: (K,ll.) (K.23)
'H(I) = ('Hj.'Hk]TjTk where
J>k Suppose we keep this condition, and add to it the condition
N
(K.24)
T: I>; (K.Ilb) 'H4+'HS+11.6=O .
j=1 Then, adding the columns of (K.22) venically, and insening items such as 4,4
in the second column, etc. we get that the second, third, and fifth columns add
to zero, leaving solely contributions from the first and founh columns:
625 627
s =6, 5 + 3,2 =['Ii" 'Ii,] + ['Ii" 'li2] Selected Bibliography
It is now obvious that if we pick
Basic Papers
E.M. Purcell, II.C. Torrey, R.V. Pound: Re!IOnaoce absorplion by nuc:lear magnetic mo--
ment.8 in atolid. Phys. Rev. 69, 37 (1946)
F. Bloch, W.W. Hansen, M. Paek&rd: Nuc:lear inductH>n. Phys. Rev. 69,127 (1946)
F. Bloch, W.W. Hansen, M. Paebro: The nuc:lu.r inductH>n experiment. PhYll. Rev. 70,
474-485 (1946)
F. Bloch: Nuc:lear induction. Phys. Rev. 70, 460-474 (1946)
N. Bloembergen, E.M. Purcell, R.V. Pound: Relaxation effects in Iluc:lear magnetic res0-
nance absorption. Phys. Rev. 73, 679-712 (1948)
628 828
R.V. Pound: Prog.Nucl. Phys. 2, 21-50 (1952) . II .12 Coupling
William Low: ~Paramagnclic Resonance in Solids", in Solid Start PhysICS, Supplement 2,
ed. by F. !kit~, D. Turnbull (Academic, New York .1960). . . 11.5. Gutowsl:y, D.W. McCall, C.P. Slichler: Nuclear magnelic fCIIOnance mulliplels in
J.S. Griffith: TM Thea? ofTr(lllSilUm·Mefalloru (Cambridge Umvennly Press, Cambndge liquids. J. Chem. Phys. 21, 279-292 (1953)
1961) E.L. Hahn, D.E. Maxwell: Spin echo measUT(!menls of nuclear spin coupling in molecules.
N.F. Ramsey: NucltoT MonunlS (Wiley, New York 1953) Phys. Rev. 88, 107G-108-I (1952)
N.F. Ramsey: Moiteular Beams (Clarendon, Odord 1956) N.F. Ramsey, E.M. Purttll: InteractiOn5 beh.:ccn nuclear spins in molecules. Phys. Rev.
A. Abragam: The Principles ofNudear Magnefism (Clarendon, Ox£O«I 1961) 85,1>\3-1>\4 (1952) (Idter)
N. Bloembergen: NuclearMag~f~Re/arJJlUm(W.A. ~njamin, New York 1961) N.F. Ramsey: Electron coupled illleracl;ons between nuclear spins in molecules. I'hy•.
J.D. RoberLs: Nuclear Magntl~ ReSDfl(JllCe (McGraw-lilli, ~ew .York 195~) Rev. 91, 303-307 (1953)
R.T. Schumacher: IntroducliOfl 10 Magntlic RtSOfI(J)lU (Bel\lamm-Cummmgs, Menlo Park, N. Bloembergell, T.J. Rowland: Nudearspin exchange in solids: 1'1103 and 1'1 105 lllll.gnetic
CA 1970) resonance in thalliu.,n and ll~allic oxide. Phys. Rev. 97,1679-1698 (1955)
I{ .A. McLauchll\ll: Magnetic Resonance (Clllrcndon, Oxford 1972) . . M.A. Ruderman, C. I{,ttel: Indirect exchange coupling of nuclear magnetic momenLS by
G.E. Pake, T.L. Estle: TM Physical PrillCiples of Electron Paramllgnttic ResontJflCe (BcnJamm- conduclion eleclrons. Ph)·s. Rev. 96, 99-102 (1954)
Cummings, Menlo Park, CA 1973) .. K. Yosida: ~lagnetic properties of CIl-Mn alloys. Phys. Rev. 106, 893-898 (1957)
A. Abragam, B. Sleane)': Elecfron Paramagntlk Resonat1Ce o/TrarulflQn Ions (Clarendon, H.M. McC~lllnell, A.D. McLean, C.A. Reill)': t\nalysis of spin-spin multiplet.s in nude"r
Odord 1970) . magnetic reI:IOnance speclra. J. Chem. Phys. 23, 1152-1159 (1955)
Maurice Goldman: Spill Tentplral/Ut tlfId Nuelear Magntlk RtuJNUlCe in Solids (Clarendon, II.M. McConnell: Molecular orbilal approximll.lion lo e1eclron coupled inlerAction be-
Oxford 1970) . ' tween nuclear spins. J. Chem. Phy•. 14, 46G-467 (1956)
C. Kittel: lnlroductiOfiIO Solid SIo4 PhyslU, 5th ed. (WIley, New Y~rk 1976) Chap. 16 W.A. Anderson: Nuclear magnelic resonance speclra of some hydrocarbons. I'hy•. !lev.
C.P. Poole, Jr., II.A. Faraeh: TM TMory c{Magntlic RUOIVUtCe (WIley, New York 1972) 102,151-167 (1956)
630 631
C.D. Jeffries: "Dynamic Nuelear Polarization- in Progrusin Cryo~nia (Heywood, London
1961) A.G. Redfield: Nuclear magnetic resonance saturation and rotary saturation in solidll.
Phys. Rev. 98, 1787-1809 (1955)
G.R. Khulsishvili: The Overhauser effed and related phenomena. Soviet Phys. - Usp. 3,
285-319 (1960) A.G. Redfield: Nuclear spin·lallice relaxa~ion time in copper and aluminium. Phys. Rev
101,67·68 (1956) .
R.H. \Vebb: Steady-state nuclear polarizations via electronic transitions. Am. J. Phys.
19,428-444 (1961) J.II. Van Vlcek: The physical meaning of adiabatic magnetic llU!ICep~ibilities. Z. PhYll.
Chern. Neue Folge 16, 358 (1958)
J.II. Van Vleck: The coneept oftemperature in magnetillm. II Nouvo Cimento Suppl 6
Serie X, 1081 (1957) , .,
Quadrupole Effects C.P. Stichter, W.C. 1I01ton: Adiabatic demagnetization in a rotating reference llystem.
PhYll. Rev. 121, 1701-1708 (1961)
R. V. Pound: Nuclear electric quadrupole inleractions in crystals. Phys. Rev. 79, 68:>-702 A.G. Anderson, A.G. Redfield: Nuclear spin-laUice relaxation in metals PhYll. Rev. 116,
( 1950) 583-591 (1959) .
N. B1ocmber8en: "Nuclear Magnetic Resonance in Imperfect Crystals", in Report of the L.C..H~bel, C.P. Slichter: Nuclear llpin relaxatiOll ill Normal and SIII)erconduc~il\g alu-
Bristol Conference on Defects in Crystalline Solids (Physical Society, London 1955) mllllUm. PhYll. Rev. 113, 1504-1519 (1959)
l>p. 1-32 A. Andcrson: Nonresonant nuclear spin absorption in Li Na and AI. PI,Yll Rev liS 863
T.J. Rowland: Nuclear magnetic resonance in copper alloys. Electron distribution around (1959) , , ",
mlute atoms. Phys. Rev. 119, 900-912 (1960) L.C. liebel, Jr.: Spin Temperature and Nuc:lcar Relaxation in Solidll Solid SWlt Physics
W. Kohn, 5.11. Voslco: Theory or nuclear rCllOnance intensity in dilute alloys. Phy!l. Rev. Vol. 15 (Academic, New York 1963) "
119,912-918 (1960)
T.P. Das. M. Pomeralllz: Nuclear quadrupole interaction in pure metals. Phys. Rev. 113,
2070 (1961) Rate Effects
T. Kushida, G. Benedek, N. Bloernbergen: Dependence of pure quadrupole resonance
frequency on pressure and temperalure. Phys. Rev. 1G4, 1364 (1956)
H.S. GutO\':'llky, A. Saika: Dissocia~ion, chemical exchange, and the proton magnetic res-
onance In lOme '"'!uoous electrolytes. J. Chelll. PhYll.l1, 1688-1694 (1953)
J.T. Arnold: MagnetiC re!IOnancea of protollll in ethyl alcohol. PhYll Rev 102 136-150
Chemical Shifts (1956) . . ,
R. Kuoo: Note on the lltochastic thoory or resonance abl'lorption. J. Phys. Soc J 9
W.G. Proctor, f.C. Yu: The dependence of a nuclear magnetic resonance frequency upon 935-944 (195-4) . pn. ,
chemical compound. Phys. Rev. 77 , 717 (1950) U.M. McConnell: Reaction ra~es by nudear magnetic I'ClIOnanee. J. Chem Phys 18 430-
W.C. Dickinson: Dependence of the piS nuclear resonance position on chemical com- 431 (1958) . "
pound. Rev. 77, 736 (1950) S. Meiboom. Z. I,uz, D. Gill: Proton relaxation in water. J. Chelll. Phys.17 1411-1412
11.5. Gutowsky, C.J. 1I0ffmann: Chemical shins in the mllgnetic I'ClIOnance of F1S. Phys. (1957) (letter) ,
Rev. 80, 1I0-lll (1950) (letter)
N.F. Ramsey: Magnetic: shielding of nuclei in molecules. Phys. Rev. 78, 699-703 (1950)
N.F. Ramsey: Chernic:a1 effect.s in nuclear rnasnetic resonance and in diamasnetic sus-
ceptibility. Phy!l. Rev. 86, 243-246 (1952) Cross·Relaxation
A. Saika, C.P. Slichter: A note on the nuorine I'ClIOnance shifts. J. Chern. Phya. 22, 26-28
(1954) N. Dloembergen, S. Shapiro, P.S. Pershan, J.O. Artman: Cr06S-reiaxation in llpin SYlltems
J.A. Pople: The theory of chernielll llhms in nuclear magnetic resonance I - Induced Phys. Rev. 114, 445-459 (1959) .
current denllities. Proc. R. Soc. London A139, 541-549 (1957) P.S. Pershan: Cross relaxation ill tiF. Phys. Rev. 117, 109·116 (1960)
J .A. Pople: The thoory of chemicalshifb in nuclear magnetic resonance II -Interpretation
of proton shiftll. Prot. R. Soc. London A1J9, 550-556 (1957)
II.M. McConnell: Theory of nuclear magnetic shielding in molecules 1- Long range dipolar Electron Spin Resonance in Paramagnetic Systems
shielding or protons. J. Chern. Phys. 17, 226-229 (1957)
R. ~man, G. MUrTay, R. Richards; Cobah nuclear resonance llpect.... Proe. R. Soc. D. Dleaney, K.W.f1. Stevens: Paramagnetic resonance. Rep. Prog. Phys. XVI, 108-159
London 141A, 455 (1957) (1953)
T.G. CBlItner, W. Kanzig: The electronic structure of V-centera. J. Phys Chem Solids
3,178-195 (1957) . . .
Spin Temperature G.D. Wat.kinll: EI.eclro~ ~pill resonallee of Mn++ in alkali chloridell: association with
vacancl,:" and IIl~PUfltll:S. PhYll. Rev. 113, 79-90 (1959)
N. Bloembergen: On the interaction of nuclear apins in a crystalline lattice. Phyaica 15, G.D. Watkms: Mo t.IOn of. Mn++-cation vacancy pairs in NaC1: !ltudy by eloctron spin
386-426 (19-49) resonance an d d ~eIceltiC loss. Phya. Rev. 113,91-97 (1959)
E.M. Purcell, R. V. Pound: A nuclear spin llyalem at nqative lemperatu~. Phy•. Rev. G. ~eher: Ob6ervahon of nuclear masnetic resonances via the electron spin re!IOnance
81,279-280 (1951) (Iella-) Ime. PhYll. Rev. 103, 83-4-835 (1956)
A. Abragarn, W.G. Proctor: Experiment!l on spin temperature. Phys. Rev. 106, 160-161 G. Feher: Electronic structure or F centera in KCI by the electron llpln double resonance
(1957) (letter) technique. PhYll. Rev. 105, 1122-1123 (1957)
A. Abragam, W.G. Proctor: Spin temperatUl"e. PhYll. Rev. 109, 1441-1458 (1958) G. Feher: Electron llpin I'ClIOnance experiments on donofll in llilicon _ Elec~ronic struelure
of donors by the electron nuclear double reSOllallce technique. Phys Rev. 114 1219.
1244 (1959) . ,
632
633
11.11. Woodbur)', C.W. Ludwig: Spin resonance of transition metals in silicon. Phys. Rev. Double Resonance in Liquids
117,102-108 (1960)
A.F. Kip, C. Kittel, R.A. Levy, A.M. Portis: Electronic structure of F centers.: hyperline F. BIocb: Recent developments in nuclear induction. Phys. Rev. 93, 944 (1954)
interactions in eledron spin resonance. Phya. Rev. 91,1066-1011 (1953) V. Royden: Measurement of the spin snd gyromagnetic ratio o( Cl3 by the collapse of
C.J. Delbecq, B. Smaller, P.II. Yuster: Optkal absorptioll ofCh molecule-ions in irradi_ spin-spin splitting. Phys. Rev. 96, 543 (1954)
ated potassium chloride. Phys. Rev. III, 1235-1240 (1958) A.L. Bloom. J.N. Schoolery: Effects of perturbing radio(requency fields on nuclear spin
M. Weger: Passage dfects in paramagnetic resonance experiments. Bell 5yst. Tech. J. 39, coupling. Phys. Rev. 97,1261 (1955)
1013-1112 (1960) (Monograph 366J) W.P. Aue, E. Bartholdi, R.1l. Ernst: Two-dimensional spectro!:lCOPY: application to nu-
George Feher, A.F. Kip: Electron spin resonance absorption in metals I - Experimental. clear magnetic resonance. J. Chern. Phys. 64, 2299 (1976)
Phys. Rev. 98, 337-348 (1955)
C.P. Slichter, D.C. Ailion: Low-field relaxation and the study o( ultrll8low atomic motions
Nuclear Resonance in Paramagnetic and Anliferromagnelic Substances by msgnetic resonsnce. Phy•. Rev. 135, AI099 (1964)
D.C. Ailion, C.P. Slichter: Ob6ervation of ultra-slow translational diffusion in metallic
N. Bloembcrgen: Fine structure or the magnetic resonance line or protons in CUS04 lithium by magnetic resonance. Phys. Rev. 1)7, A235 (1965)
• 511 20. Physica 16, 95 (1950) D.C. Look, I.J. Lowe: Nuclear magne~ic dipole-dipole relaxation along the static and
N.J. Poulis, G.E.G. lIandeman: The temperature dependence of the spontaneous mag- rotating magnetic fields: application to gypsum. J. Chern. Phys44, 2995 (1966)
netization in an sntiferromagnetic single crysLaI. Physica 19, 391 (1953)
R.G. Shulman, V. Jaccarino: Nuclear magnetic resonance in paramagnetic MnF. Phys.
Rev. 108, 1219 (1957) ENDOR
N. Jaccarino, R.C. Shulman: Observation of nuclear magnetic resonance in anti(erromag-
netic MnF, Phys. Rev. 107, 1196 (1957)
G. Benedek, T. Kushida: Nuclear magnetic resonance ill antiferromagnetic MnF2 under G. Feher: Electronic structure o( F cente... in KCI by the electron spin double resonance
hydrostatic p~ure, Phys. Rev. 113, 46 (1960) technique. Phys. Rev. 105, 1122 (1957)
\V. Marshall, R.N. S~uart: Theory of trallsition ion complexes. Phys. Rev. 123, 2048 II. Seidel, H.C. Wolf: In Physic.sc{ColorCefllers, 00. by W. Beall Fowler (Academic, New
(1961) York 1968)
634
635
U. Ilaeberlen: lIi,h Ra<Hution NMR ill Solids: StI«:ti~ A~raging: Supplement 1 to Adl'(:ll!Cts E.O. Stejskal, J. Schaefer, R.A. McKsy: Iligh-resolution, slow-spinning magic-angle carbon_
ill MagMtic RUOMIICe (Academic, New York 1976) 13 NMR. J. ~I&&n. 11.eson. 25, 569 (1917)
W.T. Dixon: Spinnin,-.ideband-free and spinning.sideband-only NMR !lpec:tra in spill-
ni~g samples. J. Chern. Phys. n, 1800 (1982)
Chemically Induced Nuclear Polarization (CIDNP)
G.L. Closs; A mechanism explaining nuclear spin polaritations in radical combinalion
reactions. J. Am. Chern. Soc. 91, ..552 (1969) . . Magnetic Resonance Imaging
G.L. Ooss, A.D. Trifunac: Chemically illd~ed n",?lear sp," pohlrlZllllon Illl a tool for
determination of spin multiplicities of radical-pair precursors. J. Am. Chern. Soc. 91, P.C. wuterbur: Image formation by induced local interactions: Exsmple!l employing nu·
4554 (1969) . . I ' l" clear magnetic resonance. Nature 141,190 (1973)
J.II. Freed, J.B. Pederson: The lheory of chemically induced dynamIc Spill po arlza Ion. P. Mansfield, P.K. Granell: NMR "dilTractiOIl~ in $()lids? J. Phys. C6, L422 (1973)
Adv. Magn. 11.C!lOn. 8, 2 (1976) .. A. I{umar, D. Welli, Il.R. Ernst: NMIl Fourier zeugmalography. J. Magn. Reson. 18, 69
C.L. Closs: Chemically induced nuclear polarlzallon. Adv. Magn. Reson. 7,157 (HI74) (1975)
K.K. King. P.R. Moran: A unified description of NMR iml\ging data-colleclion strategies,
snd reconstruction. Med. Phys. II, 1 (1984)
P. Mansfield, P.G. Morris: NMR IrMgiflll ill DiomediciM, ed. by J.S. Waugh (Acsdemic, New
Composite Pulses York 1982)
M.II. Levitt, R. F'r-eeman: NMR population inversion using a composite pulse. J. Magn.
Reson. 33, 473 (1979)
M.II. Levitt: Composiu pulses. Pros. NMR S~t~: 18, 61. (1~86,> Superoperators
A.J. Shaka, J. Keeler: Broadband spin decouphng III IlIOtroPIC liqUids. Prog. NMR Spec-
tros. 19, 47 (1987) . . h' h i ' I J. Jeenes-: Supero~"'tors in magnetie resonance. Adv. Ma~n. Reson. 10,2 (1982)
M.II. Levitt, 11.. Freeman, T. Frankie!: Broad band decouplml III '15 reso utlon nuc ('ar R. Zwanzig: Ensemble method in theory of irreversibility. J. Chern. Phys.33, 1338 (1960)
msgnetic reso;lsllCe !lpedroscopy. Adv. Magn. Reson. 11, 48 (1983)
636 637
References
Chapter 1
1.1 C.J. Gorter, L.J.F. Broer: Physica 9, 591 (19<12)
1.2 E.M. Pur«ll, II.C. Torrey, R.V. Pound: Phys. Rev. 69, 37 (1946)
1.3 F. BI()(.h, W.W. nansen, M. Pac:brd , Ph)',. Rev. 69, 127 (1946)
Chapter 2
2.1 J.l1. Van Vleck: Phys. Rev. 74, 1I6S (1948)
2.2 M. Tinkham : GfOIq) ThQ)ry fJIIIl QUllIIllUll Mechan ia (McGraw -Hili,
New York 1964)
2.3 H. Rauch, A. 7.eitinger, G. Bl'durek , A. \Vilfing, W. Bauspie$
ll, V. Boose: Ph)'s.
Lett. 54A, 425 (1975)
2.4 S.A. Werner, n. Colella, A.W. Overhlluser, C.F. Eagen: Phys. Rev.
LeU. 35, 1053
(1975)
2.5 II.J. Bernstein: Phys. Rev. Lell. 18, 1102 (1967)
2.6 Y. Aharano v, L. Susskind : Phy,. Rev. ISS, 1231 (1967)
2.7 M.E. Stoll, A.J. Vegll, R.W. Vaughan : Ph)',. Rev. AI6, 1521 (1977)
2.8 M.E. Stoll, E.K. Wolff, M. Mehring: Ph)'s. Rev. A 17, 1561 (1978)
2.9 E.K. Wolff, M. Mehring : Phys. Lett. 70A, 125 (1979)
2.10 E.L. lIahn: Phys. Rev. 80, 580 (1950)
2.11 II.Y. Carr: Current Comme nts 20,24 (1983)
2.12 II.Y. Carr and E.M. Pun:ell: Phys. Rev. 94, 630 (1954)
2.13 I. Solomon : Phys. Rev. 110, 61 (1958)
2.14 J. Spobs: Thesis. University of Illinois (1957) (unpubli sll«!.)
2.15 J.J. SpokM, C.P. Slicht.cr: Phys. Rev. 113, 1462 (1959)
2.16 D.F. Holcomb , R.E. Norberg : PhYlI. Rev. 98,1074 (1955)
2.17 Z. Wang: Private commun ication
2.18 C.J. Gorter: ParanulgM/icRtlaxatiorl (Elsevie r, New York 1947) p.
127
2.19 R. Kubo, K. Tomita: J. PhYll. Soc. Japan 9, 888 (1954)
2.20 E. W. Hobson: TM TMoryo f FlUlClioru ofa Rwl Variable aIId 1M Theory
(Cambri dge UnivcrlIity Press, Cambrid ge 1926) p. 353 IT.
of FOfUids Strits
2.21 P.W. Anderso n: J. Phyll. Soc. Jllpan 9, 316 (1954)
2.22 11.5. Gutowsk y, G.E. Pake: J. Chern. PhYll. 16, 1164 (1948), ibid.
18, 162 (1950)
for example
Chapte r 3
3.1 E.n. Andrew , n.G. Elides: Proc. n. Soc., London A1.18, 537 (1953)
Chaple r 4
-4.1 N.r. namsey : Phys. Rev. 78, 699 (1950); Phys. Rev. 86, 213 (1952)
4.2 W. Lamb: PhYll. Rev. 60, 817 (1941)
639
...
<.3
<.5
J.A. Pople: Proc. R.. Soc., London A1J9, 5-41, 550 (1957)
N.r. Ramsey: Phys. Rev. 86, 243 (195'2)
D. Pines. Solid Stale Physics, Vol. I, ed. by F. Seib, D. Turnbull (Academic, New
6.'
6.7
6.8
C.P. S/ichter, W.C. Holton; Phys. Rev. 122, 1701 (1961)
E.I!. Thrner, A.M. Sachs, E.M. Purcell: Phys. Rev. 76, 465 (A) (1949)
R.V. Pound: Phyt. Rev. 81,156 (1951)
York 1955)
<.6 L. Pauling, E.B. Wilson, Jr.: lnlroductionlO Quan/um Mechanics (McGraw-Hili, New 6.' N.F. Ramsey, R.V. Pound; Phys. Rev. 81, 278 (1951)
Vork 1935) p. 232 6.10 . R.V. Pound, E.M. Purcell: Phys. Rev. 81, 279 (1951)
ILT. Schumacher, C.P. Slichter: Ph)'•. Rev, lOt, 58 (1956); 6.11 J.II. Vsn Vleck: Nuovo Ciment.o Suppl. 6, $erie X, 1081 (1957)
...
<.•
B.R. Whiting, N.S. Van der Ven, R.T. Schumacher: Phys. Rev. B18, 5413 (1978)
Ch. Ryter: Ph)'.. Rev. LeU. 5, 10 (1960)
W. Kohn: Ph)'•. Rev. 96, 590 (l9r.4)
6.12
6.13
6.14
6.15
L.C. Hebel, C.P. Slichler: Phys. Rev. 113, 1504 (1959)
A.G. Redf)Cld: Phys. Rev. Leu. 3, 85 (1958)
A.G. Redfield, A.G. Anderson: Phys. Rev. 116, 583 (1959)
R.W. Morse, II.V. 8ohm: Phya. Rev. 108, 1094 (1957)
".10 T. Kjeldaas, Jr., W.Kohn: Ph)'•. Rev. 101,66 (1956)
S.II. Vosko, J.P. Pcradew, A.l1. MacDonald: Ph)'•. Rev. LeU. 35,1725 (1975) 6.16 L. Cooper: Phys. Today 26, (31 July 1973)
4.11
A.C. Gossard, A.M. Portis: Ph)'s. Rev. Lett. 3, 1M (1959); J. Appl. Ph)'•. Suppl. 6.17 A.G. Anderson: Phys. Rev. liS, 863 (1959)
4.12
31,2055 (1960) 6.18 A. Abragam, W.G. Proctor: Phys. Rev. 106, 160 (1957)
W. Marshall: Ph)'•. Rev. 110, 1280 (1958) 6.19 A.G. Redfield; Phys. Rev. 98, 787 (1955)
U3 6.20
".14 R. Walson, A. Freeman: Ph)',. Rev. 113, 2021 (1961) C.P. Slichter, W.C. lIo/lon: Phys. Rev. 122, 1701 (1961)
N. Bloembergen, T.J. Rowland; Ad.. ~letl.1l. 1, 731 (1953) 6.21 B.N. Provotorov: Soviet Phys. - JETP 14,1126 (1962)
4.15
E.L. Ibhn, D.E. MuW('lI: Phys. Rev. 88, 1070 (1952) 6.22 R.T. Schumacher: Phys. Rev. 112, 837 (1958)
4.16
4.17 H.S. Gutowsky, D.W. "'lcCaIl, C.P. Slkhler: J. Chern. Phys. 11, 219 (1953) 6.23 GP. Slichter, D. Ailion: Phyt. Rev. I3S, A 1099 (1964)
N.F. Ramsey, E.M. Purcell; Phys. Re.... 85, 143-144 (1952) (Idter) 6.24 D. Ailion, C.P. Slichtcr: Phya. Rev. 137, A235 (1965)
4.l8
4.19 N. Bloembergcn, T.J. Rowland: Phys. Re.... 97, 167~1698 (1955) 6.25 D.C. Look, I.J. Lowe: J. Chem. Phya. 44, 2995 (1966)
4.20 M.A. Ruderman, C. KiHei; Phys. Rev. 96, 99 (19M)
1.21 F. Frohlich, F.R.N. Nabarro: Proc. R. Soc. London AI7S, 382 (1940)
4.22 K. Yoshida: Phys. Rev. 106,893 (1957) Chapter 7
4.23 J.II. Van Vleck: Rev. Phys. 34, 681 (1962)
7.1 R.V. Pound: Phys. Rev. 79, 685 (1950)
7.2 T.R. Carver, C.P. Slichter: Phys. Rev. 92, 212 (1953); ibid. 101, 975 (1956)
Chapter 5 7.3 A.W. OV~hauser: Phya ~ev. 91, 476 (1953), ibid. 92, 411 (1953)
7.< C.D. Jeffnes (~); ~fIilJ1UC.NIIC~Or~~at~ (Inlerscience, New York 1003)
5.1 C.J. Gorter; ParQmllfic RtfcuatwII (Elsevier, New York 1947) 7.5 D.F. lIolcomb. ThOlIS, Umvennty of IlhnOls (1954) unpublished
5.2 L.C. Hebel, C.P. Slichter: Phys. Rev. 113, 1504 (1959) 7.' F. Bloch: Phya. Rev. 102, 104 (1956)
5.3 J. Korrillga; Physica 16, 601 (1950) 7.7 1. Solomon: Phys. Rev. 99, 559 (1955)
5.'l D. Pines: Solid Stall Physics, Vol. I, cd. by F. Seit:!:, D. Thmbull (Academie, New 7.8 C.D. Jclfries: Phys. Re~. 106, 164 (1957); ibid. 117, 1056 (1960)
York 1955) 7.' A. Abragam, J. Combnsson, I. Solomon: C.R. Acad. Sci. 247, 2237 (1958)
5.5 R.C. Tolman: T~ Principles of Skllistical Mtchania (Odord University Press, New 7.10 E. Erb, J.L. Montchane, J. Uebersfeld: C.R. Aead. Sei. 246, 2237 (1958)
York 1946) 7.11 G. Fel)Cr: Phys. Rev. 105, 1122 (1957)
5.6 I.J. Lowe, R.E. Norberg: Phys. Rev. 107, 46 (1957) 7.12 II. Seidel, II.C. Wolf: In PJrysiai¥CoIorCtnlus, ed. by W. Beall Fowler (Ae..demic
5.7 C.R. Bruce: Phyt. Rev. 107,43 (1951) New York 1968) ,
5.8 R.R. Ernst, W.A. Anderson; Rev. Sci. Instrum. 37, 93 (1966) 7.13 J.P. Gordon, H.J. Zeiger. C.II. Townes; Phya. Rev. 9S, 2821 (1954)
5.9 R.R. Ernst, G. BodenhauS<ln, A. Wokaun; PrifICiplu o{NUJ;letJr MagMtic ResOlWnU;1I 7.14 Nobel Ltctwres-Physu:s 1963-1970 (Elsevier, Amsterdam 1972)
OM and Two Dimt/lS;o/lS (Clarendon, Oxford 1987) 7.15 N.G. Bll.SOv, A.M. Prokhorov; Sov. Phys. - JETP 27,431 (1954)
5.10 A.G. Redfield: IBM J. Res. Dev. 1, 19 (1957) 7.16 N. Bloclllbergen: Phys. Rev. 104, 324 (1956)
5.11 R.K. Wangsness,". Bloch: Phys. Rev. 89, 728 (1953); 7.17 A. Abrsgam, W.G. Proctor: Phys. Rev. 106, 160 (1957)
F. Bloch: Phys. Rev. IOl, 104 (19S6) 7.18 N. Bloembergen, S. Shapiro, P.S. Pershan, J.O. Artman: Phys. Rev. 114, 445 (1959)
5.12 N. Bloembergen, E.M. Purcell, R.V. Pound: Phys. Rev. 73, 679 (19-48) 7.19 p.s. Pershan: Phys. Rev. 117, 109 (1960)
5.13 E.R. Andrew, R.G. ~: Proc. R. Soc. London A2U, 531 (1953) 7.20 A.G. Anderson: Phys. Rev. IlS, 863 (1959)
5.14 D.F. lIolcomb, R.E. Norberg; Phys. Rev. 98, 10H (1955) 7.21 A.G. RcdflC.ld: Phyt. Rev. 130,589 (1963)
5.15 E.F.W. Seymour; Proc. Phya. Soc., London A66, 85 (1953) 7of2 N.C. Fernehus: Proe. of the XIV Colloque Ampere (1966) p. 497
5.16 J.J. Spokas, C.P. Slichter: Phya. Rev. 11), 1462 (1959) 7.23 R.E. Slusher, E.L. nahn: Phys. Rev. 166, 332 (1968)
7.24 M. Minier: Phys. Rev. 182, 437 (1969)
7.25 ~~37eitekaml)' A. Bielecki, D. Zn, K. Zilm, A. Pines: Phya. Rev. Lelt. SO, 1807
Chapter 6
7.26 D.T. Edmonds: Phys. Rep. 211, 233 (1977)
6.1 II.B.G. CllSimir, F.K. duPri: Physica 5, 507 (1938) 7.27 N. Bloembergen, P. Sorokin: Phys. Rev. 100,865 (1958)
6.2 J.II. Van Vleck: Phys. Rev. 57, 426 (1940) 7.28 S.R. lIarimann, E.L. lIahn: Phys. Rev. 128, 2042 (1962)
6.3 J.II. Van Vleck: J. Chern. Phys. 5, 3'20 (1937) 7.29 ".M. Lurie, GP. Slichler: Phys. Rev. 133, AlI08 (1964)
6.4 L Waller: Z. Phys. 79, 370 (1932) 7.30 c.P. Slichler, W.C. lIolton: Phys. Rev. 122, 1701 (1961)
6.5 A.G. Redfield: Phys. Rev.1l8, 1787 (1955) 7.31 P.R. Spencer, N.D. Schmid, C.P. Slichter: Phys. Rev. 81, 2989 (1970)
7.32 D.V. Lang, P.R. Morl\fl: Phys. Rev. III, 53 (1970)
640
641
7.33 A. Pinell, M.G. Gibby, J.S. Waugh: J. Chem. Phys. 56, 1776 (1972); ibid. 59, 569 7.85 0.8. Twieg: Moo. Phys. 10 610 (1983)
( 1973) 7.86 S.J. Ljunggren: J. Mllgn. R~n. 54, 38 (1983)
7.34 V. Royden: Phys. Rev. 96, 543 (1954)
7.35 A.L. Bloom, J.N. Shoolcry: Phys. Rev. 97,1261 (1955)
7.36 F. Bloch: Phys. Rev. 93, 944 (1954) Chapter 8
7.37 R. Freeman, S.P. Kernpsel1, M.H. Levilt: J. Magn. Reson. 35, 447 (1979)
7.38 W.A. Anderson, F.A. Nelson: J. Chern. Phys.39, 183 (1963) 8.1 £.L. Hahn: Phya. Rev. 80, 580 (1950)
7.39 R. Freeman, W.A. Anderson: J. Chern. Phys. 42, 1199 (1965) 8.2 H.Y. CarT, E.M. Pun:ell: Phys. Rev. 94, 630 (1954)
7.40 R. Ernst: J. Chem. Phys. 45, 3845 (1966) 8.3 S. Meiboom, D. Gill: Rev. Sci. Instrum. 29, 6881 (1958)
7.41 J.B. Grut.ner, R.E. Santini: J. Magn. Rcson. 19, 173 (1975) 8.' E.D. Ostroff, J.s. Waugh: Phys. Rev. Lett 16, ioo7 (1966)
7.42 V.J. Basus, P.O. Ellis, II.D.W. lIi11, J.S.WlIugh: J. Magn. Rcsoll. 35,19 (1979) 8.' P. Mansfield, D. Ware: Phys. Rev. Leu. 21,133 (1966)
1.43 M.II. l.evitt, R. forcernall: J. Magll. Rcsoll. 33, 413 (1979) 8.8 t.M. Stacey, R.W. Vaughan, D.O. Ellernan: Phya. Rev. Lett. 26, 1153 (1971)
7.44 M.II. Levitt: Prog. Nud. Magn. Reson. Spcctrosc. 18, 16 (1986) 8.7 J.G. Powles, P. Mansfield: PhYI. Lett. 2, 58 (1962)
7.'l5 A.J. Shalea, J. Keeler: Prog. Nuc!. Magn. Reson. Sped~. 1.9, 47 (1987) U I. Solom~n: Phya. Rev. 110, 61 (1958)
7.46 R. Freeman: Bull. Magn. Reson. 8, 120 (1986) 8.9
~9:'(?9a;6)' K.R. Jeffrey, M. Bloom, M.I. Valie, T.P. lIiMB: Chem. Phya Lett. 42,
7.47 D.E. Kaplan, E.L. lIahn: J. Phys. Radium 19,821 (1958)
7.48 M. Emahwiller, E.L. lIahn, D. Kaplan: Phys. Rev. liS, 414 (1960) 8.10 P.K. Wang: Thesis, University of Illinois (1984) unpublished
1.49 J .B. Boyce: Thesis, University of Illinois (1972) 8.11 P.K. Wang, C.P. Sliehter: Bull. Magn. Reson. I, 3 (1986)
7.50 D.V. Lang, J.B. Boyce, D.C. 1,0, C.P. Slichter: Phys. Rev. Lett. 29, 716 (1972) 8.12 J. Jcener, P. 8roenert: Phya. Rev. 157, 232 (1967)
7.51 D.V. Lang, D.C. 1,0, J.B. Boyce, C.P. Slichter: Phys. Rev. 09, 3071 (1914) 8.13 n96t.;)' W.1. GoJdburl: Phys. Rev. Leu. H, 2SS (1963); Phya Rev. 140, AIUl
7.52 C.D. Makowb, C.P. Stichter, J.II. Sinfelt: Phys. Rev. Lett. 49, 379 (1982); Phya.
Rev. 831, 5663 (19M) 8.14 D. Ba.rnaal, I.J. Lowe: Phys. Rev. Leu. 11, 258 (1963)
7.53 P.K. Wang: Thesis, University oflllillOis (1984) unpublished 8.15 W.-K. Rhim, H. Kcssemeier: Phys. Rev. OJ, 3655 (1971)
7.54 P.K. Wang, C.P. Slidlter, J.II. Sinfelt: Phys. Rev. Lett. 53, 82 (1984) 8.16 H. Kcsse~ier, W.:K. Rhim: Phys. Rev. B5, 761 (1972)
7.55 J. Jeener: Lecturu, Ampere Intemationa.l Summer School, Basko Po~u, YugOll1avia 8.17 W.-K. Rhlm, A. Pllles, J.S. Waugh: Phya. Rev. Lett 2S 218 (1970)· Ph R
BJ,684 (1971) " , ya. ev.
(1971) unpublished
7.56 R. Freeman, G.A. Morris: Bull. Magn. Rcson. 1, 5 (1979) 8.18 I. Solomon: Phya. Rev. Lelt. 2, 301 (1959)
7.57 W.P. Aue, E. Bartholdi, R.R. Ernst: J. Chem. Phys.~, 2229 (1976) 8.19 K. Takegoshi, C.A. McDo_II: Chern. PhYI. Lett. 116, 100 (1985)
7.58 A. Bax: Bull. Magn. Reson. 78,167 (1985) 8.20 U.S. GUlowsky, G.E. Pake: J. Chem. Phy•. 16, 1164 (1948)' ibid. 18 162 (1950)
7.59 D.L. Thmer: Pros. Nud. Magn. Reson. Spcctrosc. 17, 281 (1985) 8.21 I. Lowe: Phys. Rev. Lett. 2, 285 (1959) "
7.60 L. Muller, A. Kumar, R.R. Ernst: J. Chern. Phys. 63, 5490 (1915) 8.22 g:i9)ndrew, A. Bradbury, R.G. Eades: Nature 182, 1659 (1958); ibid. 183, 1802
7.61 L. Muller, A. Kumar, R.R. Erllllt: J. Magn. Res<m. 25, 383 (1977)
7.62 D.T. Pegs, M.R. Bendall, D.O. Doddrell: J. Magn. Re8on. 44, 238 (1981); ibid. 45, 8.23 :ri~~:~~:I~~)farMomell/lUnillQlUlflIumMtduvUcs (Princeton University Press,
8 (1981)
1.63 G. Feher: Phya. Rev. 103, 500 (1956) 8.24 II. Kcssemeier, R.E. Norberg: Phys. Rev. 155, 321 (1967)
7.64 G. Feher, E.A. Gere: Phya. Rev. 103, 501 (1956) 8.25 M.M. Ma~cq, J.S. Waugh: J. Chem. Phys. 70, 3300 (1919)
7.65 E.B. Baker: J. Chem. Phyll. 37, 911 (1962) 8.26 ~.M. Po~tlll: Phys. Rev. 91,1971 (1953)
7.66 K.G.R. Pllchler, P.L. Wessels: J. Magn. Reson. 12,337 (1973) 8.27 .0. Stejskal, J. Schaefer, R.A. McKay: J. Magn. Reson. 25, 569 (1917)
7.67 A.A. Maudsley, R.R. Ernst: Chem. Phys. Lett. SO, 368 (1977) 8.28~ Herdeld, A. RoufOSS(!, R.A. lIaberkorn, R.G. Griffin, M.J. Glimcher' Phi los.
7.68 G. Bodenhausell, R. Freeman: J. Magn. Rcson. 28, 463 (1977) al18. R. Soc. London, Series B 2&9, 459 (1980) .
7.69 A.A. fo,iaudsley, I,. Muller, R.R. Ernst: J. Magn. Rcson. 28, 463 (1977) 8.29 J. He~~eld, A.E. Berger: J. Chern. PhYI. 7J, 6021 (1980)
7.70 G.A. Morris, 11.. Freeman: J. Am. Chern. Soc. 101, 760 (1979) 8.39 \V.T. Dlx?n: J. Chem. Phya. 77, 1800 (1982)
7.71 O.W. Sorensen, G.W. Eich, M.JI. Levitt, G. Bodenhauscn, R.R. Erno,st: Prog. Nuel. 8.31 (~:8·.R) ',Ielg,h, E.T. Olejniuak, S. Vega, R.G. Griffin: J. Am. Chern. Soc. 106 8302
Magn. Reson. Spcetrose. 16, 163 (1983) ; : fag? Reson. 72, 238 (1987) ,
7.72 F.J.M. Van den Ven, C.W. IIilbcrs: J. Magn. Reson. 54, 512 (1983) 8.32 M. Mehrmg: /flgh ResoluJion NMR S~c//"O$co'"' in Solids NMR· B . p. . I
7.73 P.K. Wang, C.P. Slicht.er: Bull. Magn. Reson. 8, 3 (1986) Pro V I II ed b ' Y J , . l\Sle rlllClp es lin d
197:)' o. , . y P. DIehl, E. Fluck, R. Kosfcld (Springer, Berlin, Heidelberg
7.74 E.I,. lIahn, D.E. Maxwell: I>h)'s. Hev. 88, 1070 (1952)
7.75 11.5. Gutowsky, D.W. McCall, C.P. Sliehler: J. Chem. PhYll. 21, 279 (1953) 8.33 U. Haeberle.n: IJigh ~esolu/ion NMR in Solids: Selec/ive Averaging; Supplement No.
7.76 K. Wuthrich: NMR of Proteins and Nucleic Acids (Wiley-lntcrscience, New York 1986) lo Advancts If! Magnellc Resonance (Academic, New Yorl.: 1976)
7.77 P. Mansfield: NMR J!IUI8ing in Biomedicine (Academic, New York 1982) 8.34 J.S. Waugh, L.M. Huber, U. HaeberJen: Phys. Rev. Lett. 20,180 (1968)
7.78 R.R. Ernst, G. Bodenhausen, A. Wokaun: Principles of Nuclear r.hgnetic l1.eso- 8.35 U. Ha.eberlen, J.S. Waugh: Phys. Rev. 175,453 (1968)
nance in One and Two Dimensions (Clarendon, Oxford 1987) 8.36 P. Mansfield: J. Phya. C4, 1444 (1971)
7.79 P.C. Lalilerbur: Nature 242, 190 (1973) 8.37 ri9~)"sJ1eld, M.J. Orchard, D.C. Stalker, K.II.B. Richards: Phys. Rev. B7, 90
7.80 P. Mansfield, P.K. Crannell: J. Ph)'s. C6, L422 (1973)
7.81 A. Kumar, D. Welti, R.R. Ernst: J. Magn. Reson. 18, 69 (1975) 8.38 W.K. Rh~m, D.O. Ellernan, R.W. Vaughan: J. Chern. Phys. 59 3740 (1973)
7.82 P. Mansfield: J. Phys. C 10, L55 (1971) 8.39 ~9~'(~~~)' D.O. ElJeman, L.B. Seh~iber, R.W. Vaughan: i.
Chern. Phys. 60,
7.83 K.K. King, P.R. Monln: Med. Phys. II, I (1984)
7.84 T.R. Brown, B.M. Kincaid, K. Ugurbil: Proc. Natl. Acad. Sci. USA 79, 3523 (1982) 8.40 L. Van Gerven (ed.): Nl«.'.ftar MagntlU; RUOfICIlICt in Solids (Plenum, New York 1977)
642 643
8.41 P.R. Moran: J. Phys. Chern. Solids 30,297 (1969) Appendix
8.42 W.S. Warren, S. Sinton, D.P. Weitckarnp, A. Pines: Phy•. Rev. Lett. 4], 1791
(1979) F.l H.S. Gulowsk)', C. Holm: J. Chern. Phy•. 25, 1228 (1956)
F.2 H.S. Gulowsky, D.W. McCall, C.P. Slichter: J. Chern. Phy•. ll, 279 (1953)
F.3 .B.L. Hahn, D.E. Maxwell: Phy•. Rev. 88, 1070 (UI52)
F.4 J.lI. Van Vied:: Ned. Tijdschr. Natuurkd. 27,1 (1961)
Chapter 9 F.5 H.M. McConnell: J. Chern. Phy•. 28, 430 (1958)
F.6 D.H. Archer: Thesis, lIarvard University (1953)
9.1 G. Bodenhausoen: Pr<'>fl;. Nucl. MaSn. Reson. Sp«trosc. 14, 137 (1981) F.7 P.W. Anderson: J. Phy.. Soc. Jpn. 9, 316 (1954)
9.2 II. lIataoab, T. Terao, T. lIa.shi: J. Phys. Soc. Japan 39, 835 (1975) F.8 H.S. Gulowsky, A. Saib.: J. Chern. Phys. 21, 1588 (19S3)
9.3 II. Hatanab, T. lIashi: J. Ph)'•. Soc. Japan 39, 1139 (1975) F.9 J.T. Arnold: Phys. Rev.IOl, 136 (1956)
9.4 W.P. Aue, E. Biutholdi, R.R. Ernst: J. Chern. Ph)'•. 601, 2229 (1976)
9.5 D. Weitebmp: Adv. Magn. R.eson 11, 111 (1983) G.I H.C. TOrTey: Phys. Rev. 104, 563 (1956)
9.6 S. Emid: Bull Magn. Reson. 4, 99 (1982) G.2 H.Y. Carr, E.M. Pu«:ell: Phy•. Rev. 94, 630 (J954)
9.7 M. Munowib, A. Pines: Science 233, 525 (1986); Adv. Chern. Phy•. LXVI, I (1987)
9.8 R.R. Ernst, G. Bodenhausen, A. Wobun: Principiu ojNlU:lnuMug~/U:Ruonanceill I.l N. Bloembe:r'!en, T.J. Rowland: Acta Metall.l, 731 (1953)
One and Two DilMnsiotu (Clarendon, Oxford 1987) 1.2 1.5. Gradshte)'n, I.M. R)'llhik: Tabu 0/ in/llrols, Sums, &riu, and Products, 4lh ed.
9.9 M.E. Sloll, A.J. Vega, R.W. Vaughan: Phys. Rev. A16, 1521 (1977) (Academic, New York 1980) pp. 177, 904
9.10 M.E. Sloll, E.K. Wolff, M. Mehring: Phy!!. Rev. A17, 1561 (1978) 1.3 B.T. Feld and W.E. Lamb: Phys. Rev. 67, 15 (1945)
9.11 E.K. Wolff, M. Mehring: Phy•. Lett. 70A, 125 (1979) 1.4 N. Bloembe:rgtln: In Dt!ICts in errS/alUM Solids (Phy.ical Society, London 1954) pp.
9.12 M. Mehring, P. HOfer, A. Grupp: Phy!!. Rev. A3l, 3523 (1988) 1-32
9.13 P.K. Wang, C.P. Slichter, J.II. Sinrelt: Phys. Rev. Lett. 53, 82 (1984) 1.5 D. Gerstein, C. Dybowski: Tnlflsien/ TechniqulS in NMR oj Solids (Academic, New
9.11 G.E. Pake: J. Chern. Phy•. 16, 327 (19<18) York 1985)
9.15 J. Baum, M. Munowitll, A.N. Garroway, and A. Pines: J. Chelll. Phy•. 83, 2015 1.6 G.E. Pake: J. Chern. Phy•. 16,327 (1948)
(1985)
9.16 E. Menba.::her: Quan/um MecJw.nics, 2nd cd. (Wiley, New York 1970) p. 167 J.l F.J. Dyson: Phys. Rev. 75, 486 (1949)
9.17 A. Wokaun, R.n. Ernst: Chcm. Phy•. Lelt. 51, 407 (1977)
9.18 W.S. Warren, S. Sinlon, D.P. Weitelcamp, A. Pines: Phy•. Rev. Lett. 43, 1791 K.I U. Haebe:rlen, J.S. Waugh: Phy•. Rev. 125,453 (1968)
(1979) K.2 R.M. Wilcox: J. Math. Phy•. 8, 962 (1967)
9.19 W.S. Warren, D.P. Weitcbmp, A. Pines: J. Chern. PhYll. 73, 2084 (1980) K.3 W. Magnw;: Commun. Pure Appl. Malh. 7, 649 (1954)
9.20 Y.-S. Yen, A. Pines: J. Chern. Phy•. 73, 3579 (1983) K.4 P. Mansfield: J. Phya. e4, 1444 (1911)
K.5 D.P. Burum, M. Linder, R.R. Ernst: J. Ma&". Raon. 44, 173 (1981)
Chapter 10
10.1 J.H. Smith, E.M. Purcell, N.F. Ramsey: PhYll. Rev. 108, 120 (1957)
10.2 L.R. Walker, G.K. Wertheim, V. Jaccarino: Phy•. Rev. Lett 6, 98 (1961)
10.3 M.E. R.o5e: Efmwllaryl'Moryoj ....nguuuMf)tMn1IInc (Wile:y, New York 1957)
10.4 E. Ambler, J.C. Eisenstein, J.P. Schooley: J. Math. PhYll. J, 118, 760 (1962)
10.5 R.M. Stemheimer: Phys. Rev. 84, 244 (1951); ibid. 86, 316 (1952)j ibid. 95, 736
(1954)
Chapter 11
11.1 A.R. Edmonds: AflgulorMomefl/umifl Quof1lumMechaflit;s (PrinCtllOn Univtlrsily Press,
Princeton, NJ 1957)
11.2 J.S. Waugh, C.P. Slichter: Phy•. Rev. B37, 4337 (1988)
11.3 I,.G. Rowan, E.L. Hahn, W.B. Mims: Phys. Rev. 137, 61 (1965)
IIA W.B. Mim.s: Phys. Rev. B5, 2409 (1972)
11.5 T.G. ClISlncr, W. Kin~i8: J. Phy•. Chem. Solids. J, 178 (1957)
11.6 CoJ. Delbceq, B. Smaller, 1'.11. Vusler: Ph)'s. Rev. 111, 1235 (1958)
644 645
Author Index
647
Hill, H.D.W. 310 McCall,D.W. 127,132, Raleigh, D.P. 406 Sorokin, P. 275,276,281, Valie, M.I. 372
Hobson, E.w. 56 141,351,593 Ramsey, N.F. 92,107, 284,289,290,292,294 Van cler Yen, F.J.M. 345,
Weitekamp, D.P. 274,
HOfer, P. 440 McConnell, H.M. 593 132,229,486,511 428,434,463,471,476,
Spencer, P.R. 287 354,356
Holcomb,D.F. 42,44, McDowell, C.A. 392 Rauch, H. 32 477,483
Spokas, J. 42,215 Viln Gerven, L. 421
215,258 McK ..y, R.A. 4(H,406 Redfield, A.G. 34,199, Welti, D. 362,363
Stacey, L.M. 367,406 Van Heeke, P. 421
Holm, C. 592,593 Mehring, M. 33,440 200,201,204-206, Werner, SA. 32
Stalker, D.C. 406 Van Vied, J.H. 14,79,
Holton, W.C. 220,235 Meiboom, S. 367,369, 219-221,230,232, Wertheim, G.K. 487
Slejskal, E.O. 404,406 133,142,219,221,230,
Huber, L.M. 422,423 371,422 234-236,241,242,244, Wessels, P.L. 331,338,
S~ernheimer, R.M. 502 593
Merzbacher, E. 459 274,276,279,281,289 3S1
Sloll, M.E. 33,440 Vaugh..n, R.W. 406,422 Wilcox, R.M. 624
Jaccarino, V. 487 Mims, W.B. 524 Reif, F. 485 SU8IIkind, L. 32 Vega, A.J. 33 Wilting, A. 32
Jeener, J. 319,331,335, Minier, M. 274 Rhim, W.-K. 389,391, Vega, S. 406 Wilson, E.B., Jr. 117
351,380,381,383,384 Monteh ..ne, J.L. 264 392,406,422
Jeffre~ K.R. 372 Moran, P.R. 287,364,422 Richards, K.II.B. 406 Takegoshi, K. 392 Wobun, A. 179,357,434,
Jeffries, C.D. 264 Rose, M.E. 490,492 Terao, T. 433 463,464,469,473
Morris, G.A. 319,325, Walker, L.R. 487
Roufosse, A. <105 Theiss, G. 398 Wolf, H.C. 268
331 Waller, I. 219
Kii.nzig, W. 533,534 Tinkham, M. 32 Wolff, E.K. 33,440
Morse, R.W. 231 Rowall, L.G. 524 Wang, P.-K. 318,345,378,
Kaplan, D.E. 311 Tolman, R...C. 157 Wuthrich, K. 351
Keeler, J. 311 Muller, L. 322,323,331 Rowland, T.J. 127,133, 382,444,448
Munowilt, M. 434,448, 141,143,606,615 Tomila, K. 55
Kempsell,S.P. 307,310 Torrey, H.C. 9,598 Wllng, Z. 45 Yen, W. 398
463,471 Royden, V. 297,307, Wangsnesa, R. 199
Kessemeier, II. 389,391, Townes, C.H. 269 Yen, Y.-S. 471,475,480
309 Ware, D. 367,406
398,399 Thmer, D.L. 319,325 Yosida, K. 133,142
Nabarro, F.R.N. 133,1<13 Ruderm..n, M. 133,143 Warren, W.S. 428,471,
Kincaid, B.M. 36-1 Thrner, E.H. 229 Yuster, P. 633
Nelson, F.A. 310 Ryter, Ch. 122-121 476,477,479,483
King, K.K. 364 Twieg, D.B. 364
Norberg, R.E. 42,179, Ryzhik,I.M. 612 Watson, R. 126 Zax, D. 274,301
Kittel, C. 133,143
Kjeldus, T. Jr. 123,124 183,215,398,399 Waugh, J.S. 247,293,310, Zeiger, H.J. 269
Knight, W. 113 Uebersfeld, J. 264 367,389,391,392,399, Zeilinser, A. 32
Olejniczak, E.T. 406 Sachs, A.G. 229 Ugurbil, K. 364 422,423,624,624,627 Zilm, K. 274
Kohn, W. 123,124
Orcharo, M.J. 406 Saika, A. 596
Korringa, J. 156,157
Ostroff, E.D. 367 Santini, R.E. 310
Kubo, R. 55
OverhaU&er, A.W. 32, Sehaefer, J. 404,406
Kumar, A. 322,323,362,
247,249,250,254,256, Schmid, N.D. 287
363
257,293-295 Sehoolery, J.N. 297,307,
Lamb, W. 103,615 309,310
Lang, D.V. 287 Puhler, K.G.R. 331,338, Schooley, J.F. 494
Lau~erbur, P.C. 357,358, 351 Schreiber,I,.B. 406,422
361 Pa<:kard, M. 9 Schrieffer, J.R. 231
Lee, M. 384,386,388,389 Pake, G.E. 63,80,392, Schumacher, R.T.
Levitt, M.II. 307,310, 448,616 122-124,240,241
311,345,350,354 Pauling, L. 117 Seidel, II. 268
Linder, M. 628 Pegs, D.T. 331 Seymour, E.F.W. 215
Ljunggren,S.J. 364 Pennington, C.II. 301 Shaka, A.J. 311
Look, D.C. 245 Persh,,", P. 272-274 Sh..piro, S. 272
Lowe,I.J. 179,183,245, Pines, A. 247,293,389, Sin felt, J.II. 312,313,318,
387,392,398 391,392,434,448,463, 444,448
Lurie, F. 275,281,283, 471,475,477,479,480, Sinton, S. 428,471,479
284,286,287,289,293 '83 Slichler, C.P. 42,122-124,
Pines, D. 115,124,157 127,141,155,215,220,
Magnus, W. 475,478, Pople, J.A. 105 230,235,245,249,2&0,
624,625 Por~is, A.M. 126,399 257,275,281,283,284,
Makowka, C.D. 312,313 Pound, R.V. 9,199,206, 286,287,289,293,312,
Mansfield, P. 357,358, 229-231,247-250,295 313,318,345,351,378,
361, 362, 36~367, 371, Powles, J.G. 371,372 378,382,444,448,524,
372,406,628 Proctor, W.G. 231,271 593
Maricq, M.M. 399 Prokhorov, A.M. 269,270 Slusher, R.E. 274,287
Marshall, W. 126 Provolorov, B.N. 239,240 Smith, J.II. 486
Maudsley, A.A. 331,335, Purcell, E.M. 9,40,401, Solomon, I. 42,258,262,
338 132,199,206,230,231, 264,372,390
Maxwell, D.E. 127,132, 367-369,392,404,486, Sorensen,O.W. 345,350,
351,593 661 3&<
648
..9
Subject Index
I
- effect on powder pllUerns 60Slf - general pN)pertietl 192ff
Antiferromagnets Correlation time 193
- NMR in 124f COSY 3S0ff
Antishielding factor 502 Cross-polarizlIlion (CP)
Asymmetry parameter, defined 497 - Harlmann-lIl1hll method 277lf
Average lJamiltonian - Pines-Gibby-Wllugh melhod 293fT
calculated for various pulse sequences Cross-relaxation
4231f and etllablishment of spin-lemperature
correctiol15 to the theory of 623lf 1<7
introduced .ol ... lf bet.... een ulllike nuclei 271ff
Crystal fields
- role in chemica.! shins 90,101
Benune, structure $tOOied by second - role in electron spin resonance 50Sff
moments 80ff Current density j(r), quantum
Bloch equatiol'l$ mechanical expression for 93ff
applied to diffusion 597lf Curie's law, derived 163
applied to rate processes S92IT
failure under some conditions 220 Deeoupling of spins 303ff'
introduced 33 Density malrix, su also Redfield theory
modified for relaxations in the of density matrix
ins~antaneous field 217 definition 159
solution in low 111 3511' general treatment lS7ff'
Bloch functions, deAned 117 interaction reprcsentlltion 163lf
Bloch-Wangsncss-Redfield theory 19911' perturbation theory 164lf
Blocmbergen-Sorokin experiment product operator method applied
- described 27511' 344ff
- explained using spin temperature response to a 6-function 174lf
28911' in the rotating refcn:-nce frame 16Str
Boltzmilnn ratio Bil 255 for thermal equilibrium 162
Bond length, determined by magnetic time dependence using eigenfunctions
resonance 82 161
651
- time-depend('nt equation 160 - c1asskal theory II Inductance, effect of X on 37 Magnetic dipolar terms A, B,C, D, E, P,
- time development for l¥;'O spins 325ff - effect of all<'mating fields, c1l1l1Sical Interaction f'epre:lllentation 163f defined 66
of a two-I('...:I system 186ff 20ff Irreducible tensor operators TI. M Magnetic field gradient
- used to derive transition probability - expeo:::tation value 17 - applied to electron spin resonance defined 3SSf
513f - effect wll('n diffU!lion is present 367,
'OOff - quantum m«hllnical 13
Detailed bahlllce 148 - - in a rotating reference - construction from YI.M'S 493 369,597ff
Diffusion frame 29ff defined -190 Magnetic field at nucleus
effed of Carr-Pun:el1 pulse sequenc(' Ethyl alcohol 88 7iM'S 490 - from elcctron orbit (c1assicl\l) 89
367Fr Exponential operators '11M 's ,193 - from electron Sl)]n 108ff
eFred on resonance 367, 597ff dcfined 25 use in electric quadrupole problems - quantum mC(;hanical 95
Dirac equlltion, derivation of .-state expansion for cABe- A 459 4S9f Magnetic resonance imaging (Mltl) 357ff
coupling III Fr a theorem I)f"oved 579,580 Isomer shift 487 Magu('tic susceptibility, X = X' - iX"
Double resonance 247ff use in solving time-independent w;\\"e - atomic theory of X' and X" 59lf
classification of three kinds 247,248 equation 26 J-ooopling 131f complex x, introduced 37
- Cf"O!lSo-relaxation family 271tr Jcener-Brockaert pulse sequence 380f - effect on coil impedance 37
- Pound·Overh ..user family 2Hff Fcrmi function, defined 118 further qmmtum mechanical
- spin-coherence family 2!l5ff Ferromagnets, NMIt in 124lf Kllight shift expressiolls for SSOff
S-flip only 296ff Fourier transform (IT) NMR - experimental faclll 113 mci\Sllremellt of sl)in contribution of
spin-echo variety (SEDOR) 311 - derived 179ff formuh. 122 conduction ele<:trons 122ff
- ('xperimental advantages 183,184 - general theory 114f - orbital in terms of current density
Effective magnetic field, Ifelf , defined 21 J-resolved 21) 323 - relation to magnetization dellsity 116 '0<
I~leclric ficld gradient line shapes in 20 32-1-5 I<orringa relation 156f - relation of X" to I(w) 71
- asymmetry parameter, defined 495 two-dimellsionlll (2D), introduced fo,lagnetization, re[,..lion of rotatillg to
calculated 5001f 319ff Larmor frequcncy 13 linear polarizations 36
- introduced 486 I-rce induction d«ay, defined 23 Laser 269f fo,lIIgnus ('xpansion 623f
- from valence eledrons 501 {(...,), defined 71 Line shapes Maser 269f
Eledric quadrul)()le (Ch..p.IO) 485ff - effect of crystalline anisotropy 605ff Master equation 149
- computation 500ff Gau!e transfonnation, - thf!Ory of dipolar effed on 127ff Matrix el('menu of many-e\.ectron product
examples in strong and weak magnetic - defined 92 I.inear respolllle thf!Ory functions 117
fields 497tr - invariance under 92f applied to pulse NMR IHff Meiboom-Gill method 369ff
II..miltonian g-factor - Ill! basis for Fourier transform NMR Mixer 184f
- for g('neral a.xes 496 - defined 510 179ff Moments
- general treatment 486ff,49'11f expression for 510 - general thf!Ory 51 ff - effect of molccuh.r rotation 84
- for principal aX('5 496 Lande 514 - introduced 51 - further expressions for 582
using raising al\d lowering operators theory 505ff Local field - method of 71
497 - for V k center 5:13ff in double resonance 282 - nth moment, (loin), delined 71
Sternhdmer IIntishielding factor 502 g-shift su g-factor - in laboratory frame 222 - lIC<:ond moment formulas 79,80
Elcctrk quadrul)()le mom(,llt, defined 496 Gyromagnetic rl\tio "t, defined 2 - in robting frllme 235 - ~eroth, firsl, and lICCond calculated
Electron-nuclear double I"l'lIOnance Lorentz local field 125,182 73ff
(ENDOR) 2661f 11, Lorenlzian line, d('fined 39 - (6...,") delined 72
Electron spi n - defined 20 Motional narrowing
- coupling to nucleus l08lf - effect of finite 1/ 1 on time of the ecllO Magic angle - analyzed using the Bloch equations
- role in chemiclll shift 143tr 49 - in the rotating frame (~Goldb"rg 592f
- rol(' in Knight shift 113lf effect of inhomogeneity on pl.llscs 422 ('xperiment) 384ff - delillci1 213
Electron apill echoes 524ff relaled to power level 38,39 - spinning 392ff - simple derivation 213f
Electron spin I"l'lIOnl\nce (ESn) (Chap. II) rell\tion to linearly polarized fl('ld 36 Magic «hoes 388ff Multiple quantum coheN!:nce (01all.9)
503ff rotating 20 Magnetic dipoh.. r broa<kning of resonance 4311f
Electrons Ilartmann-irahn condition 277 lines 65lf bll8ic cxplan"tioll 43-llf
electric field gradient produc<,d by 500 I-jebel-Slichter equation for 1'1 150 Magnetic dipolar coupling reduced by evolution, mixing, and detC(;tion
intNl\ction of electron orbit with Iligh-temperllture approximation spin-flip method ,106ft' 449-155
nudei 89 - generl\1 discussion 589f Magnetic dipolar interaction betwccn frequency selective pumping 434ff
interaction of electron slHn with nuclei - introduced and defined 63 I)airs of nuclei 66 - generating a desired order 471ff
108ff 1I)·periine structure (in ESR) Magnetic dipoh.. r order - nOllsel«tive ~xcitation ...ol4lf
magnetk interaction with nuclei - general thf!Ory 516f - produced by Jeencr-Broekaert sequence - observin! a desired order ..63ff
(Chap.4) 87ff - in the v'" ~nter 533f 380ff - preparation 445-449
Equl'Ition of motion of a spin - produced by adiabatic demagnetization - three or more spins ..55f
- classical, ill a rotating reference frame Impedance, eff«t of X on 38 227,2-45 m(..)
12 Indirect nude"r coupling 131f Magnetic dipolar refocusing dil>olar - defined 51
coupling by «hoes 371ff - derived by quantum lllC."Chani<:s 174ff
652 653
Signal·to-noise ratio ~ ~ used to analyze Bloembcrgen- Time-ordering operator To (of Dyson),
Nuclear Overhauser efTed (NOE) 257fT defined 617
and Carr-Purcell pulse sequences Sorokin experiments
Nuclear polarization see Overhauser Transition probability
367f - used in 1'1 theory of solids 146ff
effect, Solid effed Sternheimer antishielding factor 502
and double resonance 270 - formula and conditions of validity 61
_ and phase coherent detection 42 - introduced Sf
Operators set Exponential operators, see Spin-lattice relaxation
Irreducible tensor operators, Raising
and lowering operators
Single crystal spectra 127ff
Slow motion, theory of effect on Tl~
TI
T"
- defined 243
7.
- role of the lattice in thermal processes
655
654
Springer Series in Solid-State Sciences
Editors: M. Cardona P. Fulde K. von Klitzing H.·J. Queisser